Working memory and neuroplasticity in older people: a behavioural and neurofunctional approach

ii DIREITOS DE AUTOR E CONDIÇÕES DE UTILIZAÇÃO DO TRABALHO POR TERCEIROS Este é um trabalho académico que pode ser utilizado por terceiros desde que respeitadas as regras e boas práticas internacionalmente aceites, no que concerne aos direitos de autor e direitos conexos. Assim, o presente trabalho pode ser utilizado nos termos previstos na licença abaixo indicada. Caso o utilizador necessite de permissão para poder fazer um uso do trabalho em condições não previstas no licenciamento indicado, deverá contactar o autor, através do RepositóriUM da Universidade do Minho. Licença concedida aos utilizadores deste trabalho Atribuição-NãoComercial CC BY-NC https://creativecommons.org/licenses/by-nc/4.0/ iii Agradecimentos À Centelha Divina e aos meus Mentores. Aos participantes, que fizeram deste trabalho mais que uma tese de doutoramento, uma coletânea de memórias, experiências e aprendizagens. Às orientadoras: Professora Adriana, pela oportunidade e confiança que depositou em mim e pela inspiração que me fez cruzar o Atlântico para viver essa experiência engrandecedora. Professora Sandra, por me guiar no aprendizado da neuromodulação e pelo incentivo para fazer o estágio no exterior. Aos meus tutores não oficiais: Célia, pela ajuda nas análises estatísticas e incansáveis revisões dos artigos. Diego, Anabela, Jorge Leite, Rosana e Orquídea, pela partilha de conhecimentos. Aos professores do Departamento de Psicologia Básica: em especial, ao Professor Ferreira-Alves, pela empatia e apoio. Ao Professor Armando, pelo ensino da análise de dados. Ao Professor Óscar, pelas ricas discussões nas reuniões laboratoriais, no primeiro ano do curso. Ao Professor Pedro, pela participação no meu Comitê de Acompanhamento. Ao Professor Felipe Fregni e sua equipa, por me permitir uma rica experiência em seu laboratório. À Carina, pela ajuda com a triagem de artigos na meta-análise. À Sílvia, Catarina e Tâmara por colaborar na recolha de dados. Aos meus colegas do Laboratório de Neurociência Psicológica: em especial, à Diana, pela dedicação e colaboração nos meus trabalhos e presença amiga. À Carla, Alberto, Guida, Sofia, Joana, Sara, e Ella pela ajuda e por serem tão acolhedores. À Tatiana, que me emprestou o carro para fazer as viagens durante as recolhas de dados. Ao Alberto Crego, pela ajuda na tarefa do EEG. A todos os meus amigos: Sobretudo, ao Leandro e a Luciana, sem vocês eu não teria começado essa jornada! À Fabiana, Renata e Sonaira, pela presença carinhosa. À Patrícia, Isabella e ao Felipe que, mesmo de longe, estão sempre perto. Ao Diogo, pelo apoio e por me ajudar a revigorar nas caminhadas no Gerês. À minha família, pelo amor incondicional e por entender minha ausência. Ao CIPSI e todos os/as funcionários/as da Escola de Psicologia. À Associação Gerações, Câmara Municipal de Famalicão, Associação Bomfim e Santa Casa da Misericórdia de Barcelos, pela colaboração nesse projeto. À Fundação BIAL e à FCT, pelo apoio financeiro que tornou possível concretizar esse projeto.1 Àqueles cujos nomes não foram citados, mas que fizeram a diferença neste percurso. “Quem tem amigos tem tudo!” Gratidão!!! 1 This work was funded by the Portuguese Foundation for Science and Technology (FCT, Portugal), with the doctoral Grant reference SFRH/BD/80965/2011. The research project was funded by the grants of Bial Foundation (#286/16) and FCT (NORTE-01-0145-FEDER032152, POCI-01-0145-FEDER-028682). iv STATEMENT OF INTEGRITY I hereby declare having conducted this academic work with integrity. I confirm that I have not used plagiarism or any form of undue use of information or falsification of results along the process leading to its elaboration. I further declare that I have fully acknowledged the Code of Ethical Conduct of the University of Minho. University of Minho, June 27th, 2019. v Working memory and neuroplasticity in older people: A behavioural and neurofunctional approach Working memory training (WMT) has been used to promote neuroplasticity in older people and tDCS has been proposed to boost WMT effects. Nevertheless, there is not robust evidence of WMT effectiveness and the few studies assessing the combination of tDCS with WMT used cognitive tasks as endpoints. However, the use of different markers, as the event-related potentials (ERPs), can be useful to better understand the combined or individual effects of these interventions. Thus, the studies presented in this dissertation aimed to assess WMT effects, as well as, the add-on effects of tDCS. Given the need to use different endpoints to measures WMT-induced neuroplasticity, an additional aim was to assess if the ERPs can be used as indexes of fluid intelligence (Gf), a commonly assessed constructed to infer generalization of WMT. In the first study, we presented a meta-analysis on the effects of WMT in healthy elderly. Small significant and long-lasting gains were observed in working memory (WM), but not in short-term memory (STM) and Gf tasks. Type of training tasks, the adopted outcome measures, the training duration, and the total number of training hours moderated WMT effects. In the second study, we performed a systematic review on the uses of tDCS to boost WM in healthy older adults. The studies suggest that tDCS may modulate WM in this population, improving the accuracy and shortening the reaction time. In the third study, we performed a randomized double-blind controlled experiment to evaluate the effects of 5-day WMT coupled with tDCS in healthy older adults. Fifty-four participants were assigned to one of three groups: 1) WMT (dual n -back task)+active tDCS (atDCS); 2) WMT+sham tDCS (stDCS); or 3) sham task + sham tDCS. During the training, both groups that performed the dual n -back task (WMT+atDCS; WMT+stDCS) improved throughout sessions, with no significant differences between them. However, the “WMT+atDCS” was the only group that presented gains in Gf and verbal STM after training (i.e., next day after the intervention) and at follow-up (i.e., 15 days follow-up). Finally, in the fourth study, we explored whether ERP components (i.e., P2, P3b and the LPC - late positive complex) are associated with Gf in the elderly. Fifty-seven participants performed a continuous performance task and a visual oddball paradigm while EEG was recorded. They were divided into high-performance (HP) and lowperformance (LP) groups according to their performance in the Raven’s Advanced Progressive Matrices test (RAPM). HP group presented significant higher LPC amplitudes in the CPT and shorter P3b latencies in the oddball task when compared to the LP group. Keywords: fluid intelligence; late positive complex; older adults; P3b; tDCS; working memory training. vi Memória operatória e neuroplasticidade em adultos em idade avançada: Uma abordagem comportamental e neurofuncional O treino da memória de trabalho (WMT) tem sido usado para promover neuroplasticidade em idosos e a ETCC tem sido adotada para potencializar seus efeitos. No entanto, não há evidências robustas da eficácia do WMT e os poucos estudos que avaliaram a combinação da ETCC com o WMT usaram tarefas cognitivas como medidas. O uso de diferentes marcadores, como os potenciais evocados (ERPs), pode ser útil para entender melhor os efeitos dessas intervenções. Assim, os estudos desta dissertação objetivaram avaliar os efeitos do WMT, bem como, os efeitos adicionais da ETCC. Dada a necessidade de usar diferentes parâmetros para mensurar a neuroplasticidade induzida pelo WMT, um objetivo adicional foi avaliar se os ERPs procedem como índices de inteligência fluida (Gf), um construto comumente avaliado para inferir generalização do WMT. No primeiro estudo, apresentamos uma meta-análise sobre os efeitos do WMT em idosos. Foram observados pequenos e duradouros ganhos na memória de trabalho (WM), mas não na memória a curto prazo (STM) e Gf. O tipo de tarefas treinadas, as medidas adotadas e a duração/número total de horas de treino moderaram os efeitos. No segundo estudo, realizamos uma revisão sistemática sobre o uso de ETCC para melhorar a WM em idosos saudáveis. Os estudos sugerem que a ETCC pode modular a WM nessa população, aumentando a precisão e reduzindo o tempo de reação. No terceiro estudo, foi realizado um experimento aleatório duplo-cego para avaliar os efeitos do WMT associados à ETCC. Cinquenta e quatro idosos foram designados para um de três grupos: 1) WMT (tarefa dual nback ) + ETCC ativa (aETCC); 2) WMT + sham ETCC (sETCC); ou 3) tarefa placebo + sham ETCC. Durante o treino, os dois grupos que realizaram a tarefa dual n-back (WMT+aETCC; WMT+sETCC) melhoram ao longo das sessões, sem diferenças significativas entre eles. No entanto, o "WMT + aETCC" foi o único grupo que apresentou ganhos na Gf e STM após o treino e no seguimento de 15 dias. Por fim, no quarto estudo, exploramos se os componentes dos ERPs (P2, P3b e LPC – late positive complex ) estão associados à Gf. Cinquenta e sete idosos realizaram uma tarefa de desempenho contínuo (CPT) e um oddbball visual enquanto o EEG era gravado. Participantes foram divididos em grupos de alto desempenho (HP) e baixo desempenho (LP) de acordo com seu desempenho na Matrizes Progressivas Avançadas de Raven (RAPM). O grupo HP apresentou amplitudes superiores no LPC evocado pela tarefa CPT e latências mais curtas na P3b evocada pela oddball quando comparado com o grupo LP. Palavras-chave: ETCC; idosos; inteligência fluída; LPC; P3b; treino da memória operatória. vii LIST OF ABBREVIATIONS ANCOVA - Analysis of Covariance Amp - Amplitude Aospan - Automated Operation Span Task atDCS - Active tDCS AUC - Area under the Curve AV - Auditory-verbal BDNF - Brain-derived Neurotrophic Factor BF - Bayes Factors BWD - Backward CACT - Computer-assisted Cognitive Training CBT - Corsi Block-Tapping Testing CERAD - Consortium to Establish a Registry for Alzheimer's Disease CFQ - Cognitive Failures Questionnaire CI – Confidence Interval CONSORT - Consolidated Standards of Reporting Trials CPT - Continuous Performance Task CWMS - Categorization Working Memory Span CWIT - Color-word Interference Task CVLT - California Verbal Learning Test; DFT - Design Fluency Test DKEFS - Delis-Kaplan Executive Function System DLPFC - Dorsolateral Prefrontal Cortex DS - Digit Span DSC - Digit Symbol-Coding EE - Estimate Error EEG - Electroencephalography eKFA - electronic Questionnaire for Cognitive Failures in Everyday Life EPS - Everyday Problem Solving xiv Chapter VII: Conclusions Conclusions……………………………..…………………………………………………………………….…….………… 206 Future directions…………………………………………………………………………………………….…….………… 207 Concluding remarks……………………………………………………………………………………….…….………… 208 References…………………………………………………………………………………………….…….………………… 208 ANNEXES Annex A - Informed Consent Form……………………………………………………………………………………….. 211 Annex B – Data Collection Instruments…………………………………………………………………………………. 214 Annex C – Ethics Committee Approval………………………………………………..………………………………… 223 xv LIST OF FIGURES CHAPTER I: INTRODUCTION Figure 1 . A. Distribution of the world’s population by age and sex in 2017. B. Life expectancy at birth (in years) by region: estimates 1975-2015/ projections: 2015-2050. C. Percentage of population in broad age groups in the world and by region, 2017. D. Portuguese population by age group. ………………………………………………………………………………………………………… 02 Figure 2. Number of transcranial direct current stimulation (tDCS) articles published per year (2009-2018)…………………………………………………………………………………………………………. 08 CHAPTER II: WORKING MEMORY TRAINING GAINS IN HEALTHY OLDER ADULTS Figure 1 . Schematic representation of the main findings of the current meta-analysis……………. 56 App. Figure 1 . PRISMA flow diagram…………………………………………………………………………… 70 App. Figure 2 . Posttest Forest plots. …………………………………………………………………………. 71 App. Figure 3 . Risk of bias summary graph. …………………………………………………………………. 74 Appendix A - Supplementary data Figure A . Trim-and-fill Plots by Group (measure). …………………………………………………………… 93 CHAPTER III: EFFECTS OF tDCS ON WORKING MEMORY IN HEALTHY OLDER ADULT Figure 1 . Flowchart for identifying eligible studies 101 CHAPTER IV: TRANSFER EFFECTS OF WORKING MEMORY TRAINING COUPLED WITH tDCS IN OLDER ADULTS Figure 1. CONSORT (Consolidated S127tandards of Reporting Trials) diagram……………………. 126 Figure 2. Schematic representation of the sessions.………………………….….……………………….. 127 Figure 3. Fitted data representation of group x session interaction for each outcome. …………… 140 Figure 4. Fitted values (Group x Testing Session x Predictor) for each predictor in RAPM set 1 scores. .………………………….….…..……………………………………………………………………………. 141 xvi Appendix C – Supplementary data (study IV) Supplementary Figure S1 . Dual n-back maximum level (raw data). .………………………….….….. 158 Supplementary Figure S2. Dual n-back maximum level (fitted data). .………………………………… 158 CHAPTER V: TRANSFER EFFECTS OF WORKING MEMORY TRAINING COUPLED WITH TDCS IN OLDER ADULTS Figure 1. Schematic illustration of the EEG tasks.………………………….….…..……………………….. 166 Figure 2. Electrode positions.………………………….….…..………………………………………………… 168 Figure 3. Raw mean scores in the RAPM and mean RT and D-prime for the CPT for each group…………………………………………………………………………………………………………………… 170 Figure 4. ERP waveforms (FZ, Cz and Pz electrodes) comparing LP and HP groups during CPT and Oddball performance. Topographic plot of the ERP waveforms for both tasks in Fz (Top) and Pz (Bottom)……………………………………………………………. .……………………………………… 171 Figure 5. Bar graph representing LPC, P3b and P2 amplitudes and local peak latencies for match/non-match and standard/deviant conditions, as well as deviant-standard difference waveform..………………………………………………………………………………………………….………… 173 Appendix D - Supplementary material (study V) Figure S2. Scatter Plots showing the relationship between LPC amplitude of match stimulis and RAPM (set 2). …………………………….………………………………………………..………………….. 192 Figure S2. Receiver operating characteristic (ROC) curve for predicted scores of RAPM (set II)………………………………….…………………………………………………………………………………….. 193 xvii LIST OF TABLES CHAPTER I: Introduction Table 1. ERP Components of Interest for the Current Thesis………………………………………… 11 CHAPTER II: REVIEWING WORKING MEMORY TRAINING GAINS IN HEALTHY OLDER ADULTS: A META-ANALYTIC REVIEW OF TRANSFER FOR COGNITIVE OUTCOMES Table 1. Effects of Working Memory Training Compared with Control Group by Construct…. 46 Table 2. Moderator Effects………………………………………………………………………………….. 47 App. Table 1 Main Findings of Previous Reviews of Working Memory Training Including Older Adults………………………………………………………………………………………………………………… 57 App. Table 2. Inclusion and Exclusion Criteria……………………………………..……………………… 58 App. Table 3. Characteristics of the Included Studies in Alphabetical Order.………………………. 59 App. Table 4 . Description of the Trained Tasks……………………………………………………………. 61 App. Table 5. Description of the Control Tasks…………………………………………………………….. 62 App. Table 6 . Outcome Constructs and Categories Used in the Analysis…………………………… 63 App. Table 7. Description of the Tasks Used to Assess Near and Far Transfer…………………….. 64 App. Table 8. Influential Studies in Each Group (Divided by Measure)………………………………. 67 App. Table 9. Sensitivity Analysis to Assess Publication Bias and “Small-Studies Effects”……… 68 App. Table 10. Moderation Analysis of Control Group Type…………………………………………… 69 Appendix A - Supplementary data Table A . Combinations of the Descriptors Used in the Literature Search………………………… 89 Table B . References of the Included Studies…………………………………………………………….. 89 IntroductionChapter II xviii Table C. Sensitive Analysis - Posttest……………………………………………………………………… 90 Table D. Sensitive Analysis for Control Groups………………………………………………………….. 92 CHAPTER III: EFFECTS OF TDCS ON WORKING MEMORY IN HEALTHY OLDER ADULTS Table 1. Characteristics of the Studies Included in the Review………………………………………. 104 CHAPTER IV: TRANSFER EFFECTS OF WORKING MEMORY TRAINING COUPLED WITH tDCS IN OLDER ADULTS Table 1. Generalized Multilevel Models Results for each moment per Group……………………. 136 Table 2. Generalized Multilevel models results for between group analysis per moment………. 138 Appendix C. Supplementary material (Study V) Supplementary Table S1. Characterization of the sample……………………………………………… 152 Supplementary Table S2. Descriptive Statistics for the Outcome Measures by Group and Time-point (Pretest, Posttest, Follow-up)……………………………………………………………………. 153 Supplementary Table S3. Hegde’s g Corrected by Baseline for Posttest and Follow-up ………… 154 Supplementary Table S4. Pearson Correlation Coefficients of Transfer Measures Between Pretest and Posttest or Follow-up……………………………..………………………………………………. 155 Supplementary Table S5. Results of Mixed Model Analysis of individual differences………………………………………………………………………………..………………………… 156 Supplementary Table S6. Results of Mixed Model Analysis of Near Transfer Predicting Far Transfer Gains………………………………………………………...…………………………………………… 157 CHAPTER V: LATE ENDOGENOUS ERPs AS MARKERS FOR FLUID INTELLIGENCE IN OLDER ADULTS xix Table 1. Sample Demographic Characteristics…………………………………………………….……… 164 Table 2. Behavioral data for HP and LP in CPT task (mean ± SD) …………………………………… 170 Table 3. Correlations between ERPs, D-prime and RAPM scores …………………………….……….. 174 Table S1. Results of t-test Analysis of Group Difference (LP versus HP) in Amplitude (amp) and Latency (lat)………………………………………………………………………………………………….. 190 Table S2. Results of Bayesian Independent Samples t-tests Analysis of Group Difference (LP versus HP) in Amplitude (amp) and Latency (lat)…………………………………………………………. 191 Table S3. Results of Bayesian Independent Samples t-tests Analysis of Group Difference (LP versus HP) in Response time (RT) and D-prime for CPT task…………………………………………. 191 CHAPTER I INTRODUCTION Introduction Chapter II 1 Introduction Ageing Population The rapid growth in elderly population ageing is a worldwide phenomenon. In 2017, the number of people aged 60 or more was about 962 million, representing 13% of the world population. However, the estimate is that this number will surpass 2 thousand million in 2050, and it will be three times more by the year 2100 (UN, Department of Economic and Social Affairs, & Population Divison, 2017). Figure 1A shows the distribution of world population by age and sex, while Figure 1B shows the life expectancy by region. As can be seen in Figure 1C, Europe has the world’s largest proportion of old people, representing 25% of the total European population. In Portugal, this scenario is not different (see Figure 1D). In 2017, more than 20% of the Portuguese population was aged 65 or over (FFMS, 2018). According to the National Statistical Institute of Portugal, the estimate is that the ageing population will double by the year 2080 (INE, 2017). Challenges of populational ageing and actions to overcome them Typical ageing presents a pattern of a general cognitive alterations that may change the daily life of older adults. Whereas cognitive impairment is evident in the domains of WM, Gf, episodic memory, spatial ability, and processing speed, other domains including language abilities and implicit memory seem to be preserved (Bopp & Verhaeghen, 2005; Deary et al., 2009; Park et al., 2002; Salthouse, 1991, 2018; Verhaeghen & Salthouse, 1997). These cognitive changes are concomitant with alterations in brain volume and size, white matter integrity and myelin, dendritic shape and connection; and blood flow. The first age-related brain changes are more evident in the posterior, frontal, and parietal brain areas. Specifically, prefrontal cortex and the striatum are the areas most deteriorated. Degradation in hippocampus is also reported (for a review, see Humayun & Yao, 2019) Ageing is associated with an increased risk of dementia (Chaves, Santos, Alves, & Salgado Filho, 2015; Pusswald et al., 2015). More specifically, the dementia incidence rate doubles its value every five years after the age of 65 (Corrada, Brookmeyer, Paganini-Hill, Berlau, & Kawas, 2010; Jorm & Jolley, 1998). In Portugal, the prevalence rate of cognitive impairment in people aged between 55 and 79 is Chapter II Introduction 2 12.3%, whilst the cognitive impairment associated with dementia has a prevalence estimated to be 2.7% (Nunes et al., 2010). Figure 1. A. Distribution of the world’s population by age and sex in 2017. B. Life expectancy at birth (in years) by region: estimates 1975-2015/ projections: 2015-2050. C. Percentage of population in broad age groups in the world and by region, 2017. Reprinted from World Population Prospects: the 2017 Revision (pages 2, 8, and 10 ), by UN, Department of Economic and Social Affairs, Population Division, 2017, New York: UN. ©(2019) UN. Used with the permission of the UN. D. Portuguese population by age group. Red= people aged less than 15; Dark blue= people aged between 15 and 64. Light blue= older people (65 or over). Adapted from Fundação Francisco Manuel dos Santos, Retrato de Portugal PORDATA, 2018 edition, page 9 (based on data from INE, PORDATA). Retrieved from https://www.pordata.pt/ebooks/PT2018v20180713/mobile/index.html. Adapted with permission. Introduction Chapter II 3 As a consequence, the growth in elderly population is posing various challenges to all sectors of the society with macro and individual implications in many aspects. For example, from a macro perspective, it brings social and financial issues associated with the demographic changes and the imbalance between taxes and pensions; together with an increased financial burden to the government. There is also a great load on health-care systems and caregivers due to the need for specialized goods and services. From an individual perspective, aging may be associated with dependence, isolation, abuse, reduction in physical capability, and change in many aspects of the family structure (Araújo, Paúl, & Martins, 2011; Gil, Kislaya, et al., 2015; Gil, Santos, et al., 2015; Valtorta, Kanaan, Gilbody, Ronzi, & Hanratty, 2016). Faced with these challenges, a growing interest in preventive actions to help people to live longer with less disability and less functional limitations has been emerging. As a matter of fact, the World Health Organization pointed out that the concept of ageing needs to change so that old age would not be synonymous of dependency (World Health Organization, 2015). Accordingly, initiatives to promote successful ageing emerge in the field of cognitive enhancement, more specifically using cognitive training (Melby-Lervåg, Redick, & Hulme, 2016) or brain stimulation (Hanley & Tales, 2019). On top of that, it is also of paramount importance the study of markers of optimal ageing covering both neurofunctional and behavioural outcomes. These markers would be essential to monitor and evaluate the efficacy of enhancement therapies (Belleville & Bherer, 2012). Moreover, the characterization of typical ageing is fundamental to determine thresholds of misfunctioning in the population that will be important to characterize non-healthy ageing. Otherwise, the identification of abnormal ageing would not be possible (Salthouse, 2018). In this regard, the next sections will discuss about the cognitive training (focusing especially on the working memory training) and the tDCS, techniques which could be used in the promotion of healthy ageing, as well as about factors that may moderate the effects of these interventions. We will also present one section about the neurophysiological signatures of aging, since it could be used as markers of brain functioning in healthy older people. Base of cognitive training Cognitive training is a technique that consists of the practice of structured tasks aiming the improvement or maintenance of cognitive functions (Bahar-Fuchs, Martyr, Goh, Sabates, & Clare, 2019), Chapter II Introduction 10 2012; Pinal, Zurrón, Díaz, & Sauseng, 2015). In this dissertation, we will use the ERP approach, which consists of a time series of scalp-recorded voltage changes time-locked to a given event (i.e., the presentation of a stimulus) (Kappenman & Luck, 2012; Luck, 2014). More specifically, it is the sum of postsynaptic potentials occurring at the same time in similarly oriented cortical pyramidal cells in response to an internal or external event (Luck, 2005). The waveforms are a continuous series of positive and negatives peaks, varying in polarity, amplitude, and duration. The division of the waveform in discrete sources of voltages reflecting neurocognitive processes originates the ERP components (Kappenman & Luck, 2012). Each ERP component is associated with particular cognitive processes (for a review, see Luck, 2014). The commonly analysed parameters are latency and amplitude. Most of the time latency is measured as peak latency, which refers to the time spent between the event onset and the maximum amplitude point within a time window, being related to the timing necessary for the execution of a given cognitive process. Amplitude refers to the difference in voltage between the mean voltage of the baseline period and the largest peak of the ERP waveform within a time window (Polich, 2007). Amplitude is related to cognitive processing demands and efficiency. The most commonly used task to elicit ERPs is the standard oddball task. For the purpose of illustration, in a visual oddball task, a set of two different figures is shown to the participant, one is the target and is less frequently presented (deviant stimulus), while the other figure (standard stimulus) is considered the non-target. Participants have to respond (e.g., mentally counting or pressing a button) whenever they are presented with the target stimulus (for an example of study using a visual oddball task, see Crego et al., 2012). During the analysis, the average of the trials, for standard and deviant stimuli, is extracted for each participant and component, and a grand average may also be extracted across participants. ERPs is normally classified into two groups: the early components (named sensory or exogenous) peaking around the first 100 milliseconds after stimulus and the later components (termed cognitive or endogenous) which reflects stimulus evaluation related to the processing of information (Sur & Sinha, 2009). Age-related changes in late endogenous ERP components are well reported in the literature. For example, aging is related to an attenuated and delayed P3b (Dinteren, Arns, Jongsma, & Kessels, 2014; Falkenstein, Gajewski, & Getzmann, 2014; Lubitz, Niedeggen, & Feser, 2017; Pinal, Zurrón, & Díaz, Introduction Chapter II 11 2015; Schapkin, Gajewski, & Freude, 2014) and abnormalities in this component were observed in mild cognitive impairment and pathological aging (Gu et al., 2018; Lai, Lin, Liou, & Liu, 2010; Olichney et al., 2002, 2008; Waninger et al., 2018; Zurrón et al., 2018). Late positive complex (LPC) differences were also described when comparing older with younger adults (Getzmann, Hanenberg, Lewald, Falkenstein, & Wascher, 2015; ko et al., 2014; Wolk et al., 2009), and healthy older adults and those with cognitive impairment (Waninger et al., 2018) or dementia (Lubitz et al., 2017). Finally, age-related differences in P2 were reported (Bourisly & Shuaib, 2018; Lubitz et al., 2017; Riis et al., 2009; Schapkin et al., 2014; Wolk et al., 2009), as well as differences between normative and unhealthy aging (Waninger et al., 2018). In light of these results, as well as by a thorough visual inspection of grand average difference waveforms, the components identified for the analyses reported on the current dissertation were the waveforms of positive polarity named P2, P3b, and LPC. Their main characteristics are displayed in Table 1. Table 1 ERP Components of Interest for the Current Thesis Component Peak latency (ms) Location of maximum effect Cognitive process Age-related alteration in amplitude and latency P2 100-250 ms Anterior and central Stimulus evaluation and context updating Larger latency Mixed results for amplitude P3b 350-600 ms Centro-parietal, with maximum amplitude over the midline Context updating, attentional resources, detection and rating of a stimulus Lower amplitude and higher latency LPC 500-800 ms Centro-posterior Recognition memory, categorical response, memory match, decision accuracy, and maintenance of a visual working memory representation Reduced LPC Topography involving additional anterior regions Note . LPC = Late Positive Complex Moderators of WMT Finally, one point that should be considered in WMT practice is the variables that may interfere with the effects. The literature in WMT has pointed out some factors as following: age (Borella, Carbone, Chapter II Introduction 12 Pastore, De Beni, & Carretti, 2017; Borella et al., 2014; Zinke et al., 2014); education (Borella et al., 2017); general cognitive ability (Borella et al., 2017); baseline performance (Zinke, Zeintl, Eschen, Herzog, & Kliegel, 2011; Zinke et al., 2014); and training duration (Bürki, Ludwig, Chicherio, & de Ribaupierre, 2014; Lilienthal, Tamez, Shelton, Myerson, & Hale, 2013; Stepankova et al., 2014). Regarding tDCS studies, most of them do not consider interindividual factors (e.g., baseline neuronal state, anatomy, age, brain lesions) in the analysis (for a review see Li, Uehara, & Hanakawa, 2015). However, some studies reported the moderator effects of age (Fujiyama et al., 2014; Heise et al., 2014), education (Berryhill & Jones, 2012), and baseline performance (Katz et al., 2017) in the neuromodulation effects. Therefore, it is important to understand not only if the intervention is effective but also for whom and in which conditions it works best. Research aims Taking into account the points outlined in the previous sections, the studies presented in this dissertation aimed to primarily assess the transfer effects of WMT, as well as the add-on effects of tDCS in this intervention, considering the variables that may moderate the effects. Given the need to use different endpoints to measures WMT-induced neuroplasticity, an additional aim of this research was to assess if the ERPs can be used as indexes of Gf, a commonly assessed constructed to infer generalization of WMT. In order to reach the above-mentioned goals, we performed four studies reported in this dissertation as follow: The first study (Chapter II) aimed to systematically review the literature on the transfer effects of WMT in healthy older adults and to perform a meta-analysis using a robust multilevel meta-analysis technique. This method allowed us to deal with the presence of multiple outcomes in a more sophisticated way, overcoming the limitation of previous reviews in the area. Results are presented and discussed, as well as the factors that moderated the WMT effects. In the second study (Chapter III), we reviewed studies using tDCS associated with working memory performance in healthy older adults. Only four studies met our inclusion criteria (Berryhill & Jones, 2012; Jones et al., 2015; Park et al., 2014; Seo et al., 2011, demonstrating the incipient interest in this area and the importance of further studies in this field. In this chapter, we presented the included studies, discussing the major findings, and providing recommendations for future studies. Introduction Chapter II 13 The third study (chapter IV) reports a double blind (with assessor and participant blinded), randomized, placebo-controlled experiment to assess the short-term (i.e., next-day after the intervention) and long-term (i.e., 15 days follow-up) transfer effects of 5-day WMT coupled with tDCS in healthy older adults. Participants were randomly assigned to one of the three groups: 1) WMT (adaptive dual n -back task)+active tDCS (2 mA; 20min); 2) WMT+sham tDCS; 3) sham task (visual target detection task)+sham tDCS. Moderator analyses were also included to verify factors influencing the transfer effects. In the fourth study † (Chapter V), we described the electrophysiological correlates of Gf performance in older people. The motivation to perform this study was that, in the future, we will verify the effects of WMT associated with tDCS having an EEG measure to complement the behavioral analysis. As we have used a Gf task to assess generalization of WMT, we want to ensure the association between the ERPs with this construct. This association is well described in young population (Amin, Malik, Kamel, Chooi, & Hussain, 2015; Bazana & Stelmack, 2002; Beauchamp & Stelmack, 2006; De Pascalis, Varriale, & Matteoli, 2008; Duan, Shi, Sun, Zhang, & Wu, 2009; N. Jaušovec & Jaušovec, 2001; Schlottfeldt, Mansur-Alves, Flores-Mendoza, & Tierra-Criollo, 2018; Wronka, Kaiser, & Coenen, 2013; Zhang et al., 2007; Zhang, Shi, Luo, Zhao, & Yang, 2006), but not yet in the elderly. Therefore, it is important to verify this relationship in this phase so that the ERPs, can be used to better understand the combined or individual effects of tDCS and WMT. In the final sections (Chapters VI and VII), we present our final considerations and conclusions having a summary of the main findings reported in this dissertation, their implications, and plans for future studies, as well as, a reflection on the challenges we have faced in this journey. References Aichele, S., Rabbitt, P., & Ghisletta, P. (2015). Life span decrements in fluid intelligence and processing speed predict mortality risk. Psychology and Aging, 30 (3), 598-612. doi: 10.1037/pag0000035 Alonzo, A., Brassil, J., Taylor, J. L., Martin, D., & Loo, C. K. (2012). Daily transcranial direct current stimulation (tDCS) leads to greater increases in cortical excitability than second daily transcranial direct current stimulation. Brain Stimulation , 5 (3), 208–213. doi: 10.1016/j.brs.2011.04.006 Amin, H. U., Malik, A. S., Kamel, N., Chooi, W. T., & Hussain, M. (2015). P300 correlates with learning † Participants of study 4 were the same participants of study 3, with three more participants included. Chapter II Introduction 14 & memory abilities and fluid intelligence. Journal of NeuroEngineering and Rehabilitation , 12 , 87. doi: 10.1186/s12984-015-0077-6 Araújo, I., Paúl, C., & Martins, M. (2011). Viver com mais idade em contexto familiar: Dependência no auto cuidado. Revista Da Escola de Enfermagem Da USP , 45 (4), 869–875. doi: 10.1590/S0080-62342011000400011 Au, J., Sheehan, E., Tsai, N. et al. Psychon Bull Rev (2015). Improving fluid intelligence with training on working memory: a meta-analysis 22 (2): 366-377. doi: 10.3758/s13423-014-0699-x Baddeley, A. D. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4, 417-423. doi: 10.1016/S1364-6613(00)01538-2 Baddeley, A. D., & Hitch, G. J. (1974). Working memory. In G. H. Bower (Ed.), The psychology of learning and motivation (Vol.8, pp. 47–89). New York: Academic Press Bahar-Fuchs, A., Martyr, A., Goh, A. M., Sabates, J., & Clare, L. (2019). Cognitive training for people with mild to moderate dementia. Cochrane Database of Systematic Reviews , (3). doi: 0.1002/14651858.CD013069.pub2 Bazana, P. G., & Stelmack, R. M. (2002). Intelligence and information processing during an auditory discrimination task with backward masking: An event-related potential analysis. Journal of Personality and Social Psychology, 83 (4), 998-1008. doi: 10.1037/0022-3514.83.4.998 Beauchamp, C. M., & Stelmack, R. M. (2006). The chronometry of mental ability: An event-related potential analysis of an auditory oddball discrimination task. Intelligence , 34 (6), 571–586. doi: 10.1016/j.intell.2006.03.007 Belleville, S., & Bherer, L. (2012). Biomarkers of cognitive training effects in aging. Current Translational Geriatrics and Experimental Gerontology Reports . https://doi.org/10.1007/s13670-012-00145 Berryhill, M. E., & Jones, K. T. (2012). tDCS selectively improves working memory in older adults with more education. Neuroscience Letters , 521 (2), 148–151. https://doi.org/10.1016/j.neulet.2012.05.074 Bisiacchi, P. S., Tarantino, V., & Ciccola, A. (2008). Aging and prospective memory: The role of working memory and monitoring processes. Aging Clinical and Experimental Research . https://doi.org/10.1007/BF03324886 Introduction Chapter II 15 Bopp, K. L., & Verhaeghen, P. (2005). Aging and verbal memory span: A meta-analysis. Journals of Gerontology - Series B Psychological Sciences and Social Sciences . https://doi.org/10.1093/geronb/60.5.P223 Borella, E., Carbone, E., Pastore, M., De Beni, R., & Carretti, B. (2017). Working memory training for healthy older adults: The role of individual characteristics in explaining shortand long-term gains. Frontiers in Human Neuroscience . https://doi.org/10.3389/fnhum.2017.00099 Borella, E., Carretti, B., Cantarella, A., Riboldi, F., Zavagnin, M., & De Beni, R. (2014). Benefits of training visuospatial working memory in young–old and old–old. Developmental Psychology , 50 (3), 714–727. https://doi.org/10.1037/a0034293 Borella, E., Carretti, B., Sciore, R., Capotosto, E., Taconnat, L., Cornoldi, C., & De Beni, R. (2017). Training working memory in older adults: Is there an advantage of using strategies? Psychology and Aging , 32 (2), 178–191. doi: 10.1037/pag0000155 Bourisly, A. K., & Shuaib, A. (2018). Neurophysiological effects of aging: A P200 ERP study. Translational Neuroscience , 9 (1), 61–66. doi: 10.1515/tnsci-2018-0011 Boyke, J., Driemeyer, J., Gaser, C., Buchel, C., & May, A. (2008). Training-induced brain structure changes in the elderly. Journal of Neuroscience, 28 (28), 7031-7035. doi: 10.1523/JNEUROSCI.0742-08.2008 Brunoni, A. R., Amadera, J., Berbel, B., Volz, M. S., Rizzerio, B. G., & Fregni, F. (2011). A systematic review on reporting and assessment of adverse effects associated with transcranial direct current stimulation. The International Journal of Neuropsychopharmacology , 14 (8), 1133– 1145. doi: 10.1017/S1461145710001690 Bürki, C. N., Ludwig, C., Chicherio, C., & de Ribaupierre, A. (2014). Individual differences in cognitive plasticity: An investigation of training curves in younger and older adults. Psychological Research , 78 (6), 821–835. doi: 10.1007/s00426-014-0559-3 Buschkuehl, M., Jaeggi, S. M., & Jonides, J. (2012). Neuronal effects following working memory training. Developmental Cognitive Neuroscience, 2 (S1), S167-S179. doi: 10.1016/j.dcn.2011.10.001 Cantarella, A., Borella, E., Carretti, B., Kliegel, M., Mammarella, N., Fairfield, B., & De Beni, R. (2017). The influence of training task stimuli on transfer effects of working memory training in aging. Psychologie Française . Advance online publication. doi: 10.1016/j.psfr.2017.04.005 Carvalho, S., Boggio, P. S., Gonçalves, Ó. F., Vigário, A. R., Faria, M., Silva, S., …Leite, J. (2015). Chapter II Introduction 16 Transcranial direct current stimulation based metaplasticity protocols in working memory. Brain Stimulation, 8 (2), 289-294. doi: 10.1016/j.brs.2014.11.011 Chaves, A. S., Santos, A. M. dos, Alves, M. T. S. S. de B., & Salgado Filho, N. (2015). Associação entre declínio cognitivo e qualidade de vida de idosos hipertensos. Revista Brasileira de Geriatria e Gerontologia , 18 (3), 545–556. doi: 10.1590/1809-9823.2015.14043 Chen, T., & Li, D. (2007). The roles of working memory updating and processing speed in mediating age-related differences in fluid intelligence. Aging, Neuropsychology, and Cognition, 14 (6), . doi: 10.1080/13825580600987660 Collette, F., & Van Der Linden, M. (2002). Brain imaging of the central executive component of working memory. Neuroscience and Biobehavioral Reviews, 26 (2), 105-125. doi: 10.1016/S01497634(01)00063-X Corrada, M. M., Brookmeyer, R., Paganini-Hill, A., Berlau, D., & Kawas, C. H. (2010). Dementia incidence continues to increase with age in the oldest old the 90+ study. Annals of Neurology, 67 (1), 114-121. doi: 10.1002/ana.21915 Cowan, N. (2017). The many faces of working memory and short-term storage. Psychonomic Bulletin and Review, 24 (4), 1158-1170. doi: 10.3758/s13423-016-1191-6 Cramer, S. C., Sur, M., Dobkin, B. H., O’Brien, C., Sanger, T. D., Trojanowski, J. Q., …Vinogradov, S. (2011). Harnessing neuroplasticity for clinical applications. Brain, 134 (Pt 6), 1591-1609. doi: 10.1093/brain/awr039 Crego, A., Cadaveira, F., Parada, M., Corral, M., Caamaño-Isorna, F., & Rodríguez Holguín, S. (2012). Increased amplitude of P3 event-related potential in young binge drinkers. Alcohol, 46 (5), 415425. doi: 10.1016/j.alcohol.2011.10.002 Dahlin, E., Neely, A. S., Bäckman, L., & Larsson, A. (2008). Transfer of learning after updating training mediated by the striatum. Science , 320 (5882), 1510–1512. doi: 10.1126/science.1155466 D’Esposito, M., Postle, B. R., & Rypma, B. (2000). Prefrontal cortical contributions to working memory: Evidence from event-related fMRI studies. Experimental Brain Research, 133 (1), 3-11. doi: 10.1007/s002210000395 De Pascalis, V., Varriale, V., & Matteoli, A. (2008). Intelligence and P3 components of the event-related potential elicited during an auditory discrimination task with masking. Intelligence, 36 (1), 3547. doi: 10.1016/j.intell.2007.01.002 Introduction Chapter II 17 Deary, I. J., Corley, J., Gow, A. J., Harris, S. E., Houlihan, L. M., Marioni, R. E., …Starr, J. M. (2009). Age-associated cognitive decline. British Medical Bulletin, 92 , 135-152. doi: 10.1093/bmb/ldp033 Diamond, B. J., Deluca, J., Rosenthal, D., Vlad, R., Davis, K., Lucas, G., …Richards, J. A. (1999). Information processing in older versus younger adults: Accuracy versus speed. International Journal of Rehabilitation and Health, 5 (1), 55-64. doi: 10.1023/A:1012911203468 Dienes, Z. (2011). Bayesian versus orthodox statistics: Which side are you on? Perspectives on Psychological Science, 6 (3), 274-290. doi: 10.1177/1745691611406920 Dinteren, R., Arns, M., Jongsma, M. L. A., & Kessels, R. P. C. (2014). P300 development across the lifespan: A systematic review and meta-analysis. PLoS ONE, 9 (2), e87347. doi: 10.1371/journal.pone.0087347 Du, X., Ji, Y., Chen, T., Tang, Y., & Han, B. (2018). Can working memory capacity be expanded by boosting working memory updating efficiency in older adults? Psychology and Aging , 33 (8), 1134–1151. doi: 10.1037/pag0000311 Duan, X., Shi, J., Sun, S., Zhang, X., & Wu, J. (2009). Neural mechanisms of 1-back working memory in intellectually gifted children. In 3rd International Conference on Bioinformatics and Biomedical Engineering, iCBBE 2009 . doi: 10.1109/ICBBE.2009.5163101 Falkenstein, M., Gajewski, P. D., & Getzmann, S. (2014). The electrophysiology of cognitive aging. Journal of Psychophysiology , 28 (3), 101–104. doi: 10.1027/0269-8803/a000118 FFMS. (2018). Retrato de Portugal, Edição 2018 . Lisboa: Pordata. Retrieved from https://www.pordata.pt/ebooks/PT2018v20180713/mobile/index.html Fregni, F., Boggio, P. S., Nitsche, M. A., Bermpohl, F., Antal, A., Feredoes, E., …Pascual-Leone, A. (2005). Anodal transcranial direct current stimulation of prefrontal cortex enhances working memory. Experimental Brain Research, 166 (1), 23–30. doi: 10.1007/s00221-005-2334-6 Friedman, N. P., Miyake, A., Corley, R. P., Young, S. E., DeFries, J. C., & Hewitt, J. K. (2006). Not all executive functions are related to intelligence. Psychological Science , 17 (2), 172–179. doi: 10.1111/j.1467-9280.2006.01681.x Fritsch, B., Reis, J., Martinowich, K., Schambra, H. M., Ji, Y., Cohen, L. G., & Lu, B. (2010). Direct current stimulation promotes BDNF-dependent synaptic plasticity: Potential implications for motor learning. Neuron , 66 (2), 198–204. doi: 10.1016/j.neuron.2010.03.035 Chapter II Introduction 18 Fujiyama, H., Hyde, J., Hinder, M. R., Kim, S. J., McCormack, G. H., Vickers, J. C., & Summers, J. J. (2014). Delayed plastic responses to anodal tDCS in older adults. Frontiers in Aging Neuroscience, 6 , 115. doi: 10.3389/fnagi.2014.00115 Gandiga, P. C., Hummel, F. C., & Cohen, L. G. (2006). Transcranial DC stimulation (tDCS): A tool for double-blind sham-controlled clinical studies in brain stimulation. Clinical Neurophysiology , 117 (4), 845–850. doi: 10.1016/j.clinph.2005.12.003 Gbadeyan, O., Steinhauser, M., Hunold, A., Martin, A. K., Haueisen, J., & Meinzer, M. (2019). Modulation of adaptive cognitive control by prefrontal high-definition transcranial direct current stimulation in older adults. The Journals of Gerontology Series B Psychological Sciences and Social Sciences, gbz048. doi: 10.1093/geronb/gbz048 Getzmann, S., Hanenberg, C., Lewald, J., Falkenstein, M., & Wascher, E. (2015). Effects of age on electrophysiological correlates of speech processing in a dynamic “cocktail-party” situation. Frontiers in Neuroscience, 9 , 341. doi: 10.3389/fnins.2015.00341 Gil, A. P., Kislaya, I., Santos, A. J., Nunes, B., Nicolau, R., & Fernandes, A. A. (2015). Elder abuse in Portugal: Findings from the first national prevalence study. Journal of Elder Abuse and Neglect , 27 (3), 174–195. doi: 10.1080/08946566.2014.953659 Gil, A. P., Santos, A. J., Kislaya, I., Santos, C., Mascoli, L., Ferreira, A. I., & Vieira, D. N. (2015). Estudo sobre pessoas idosas vítimas de violência em Portugal: Sociografia da ocorrência. Cadernos de Saúde Pública , 31 (6), 1234–1246. doi: 10.1590/0102-311x00084614 Gottfredson, L. S., & Deary, I. J. (2004). Intelligence predicts health and longevity, but why? Current Directions in Psychological Science , 13 (1), 1–4. doi: 10.1111/j.0963-7214.2004.01301001.x Gu, L., Chen, J., Gao, L., Shu, H., Wang, Z., Liu, D., …Zhang, Z. (2018). Cognitive reserve modulates attention processes in healthy elderly and amnestic mild cognitive impairment: An event-related potential study. Clinical Neurophysiology, 129 (1), 198-207. doi: 10.1016/j.clinph.2017.10.030 Halford, G. S., Cowan, N., & Andrews, G. (2007). Separating cognitive capacity from knowledge: a new hypothesis. Trends in Cognitive Sciences, 11 (6), 236-242. doi: 10.1016/j.tics.2007.04.001 Hanley, C. J., & Tales, A. (2019). Anodal tDCS improves attentional control in older adults. Experimental Gerontology, 115 , 88-95. doi: 10.1016/j.exger.2018.11.019 Hebb. (1949). The organization of behavior . New York, NY: McGraw-Hill. Heise, K. F., Niehoff, M., Feldheim, J. F., Liuzzi, G., Gerloff, C., & Hummel, F. C. (2014). Differential Introduction Chapter II 19 behavioral and physiological effects of anodal transcranial direct current stimulation in healthy adults of younger and older age. Frontiers in Aging Neuroscience, 6, 146. doi: 10.3389/fnagi.2014.00146 Humayun, H., & Yao, J. (2019). Imaging the aged brain: Pertinence and methods Hannah. Quantitative Imaging in Medicine and Surgery , 9 (19), 842–857. doi: 10.21037/qims.2019.05.06 Instituto Nacional de Estatística (INE). (2017). Projecções de População Residente 2015-2080 . Instituto Nacional de Estatística . Lisboa: Instituto Nacional de Estatística. Retrieved from https://www.ine.pt/xportal/xmain?xpid=INE&xpgid=ine_destaques&DESTAQUESdest_boui=2 77695619&DESTAQUESmodo=2&xlang=pt Jaeggi, S. M., Buschkuehl, M., Jonides, J., & Perrig, W. J. (2008). Improving fluid intelligence with training on working memory. Proceedings of the National Academy of Sciences , 105 (19), 6829– 6833. doi: 10.1073/pnas.0801268105 Jaušovec, N., & Jaušovec, K. (2001). Differences in EEG current density related to intelligence. Cognitive Brain Research, 12 (1), 55-60. doi: 10.1016/S0926-6410(01)00029-5 Jaušovec, N., & Jaušovec, K. (2012). Working memory training: Improving intelligence - Changing brain activity. Brain and Cognition, 79 (2), 96-106. doi: 10.1016/j.bandc.2012.02.007 Jones, K. K. T., Stephens, J. A., Alam, M., Bikson, M., & Berryhill, M. E. M. (2015). Longitudinal neurostimulation in older adults improves working memory. PLoS ONE , 10 (4), e0121904. doi: 10.1371/journal.pone.0121904 Jorm, A. F., & Jolley, D. (1998). The incidence of dementia: A meta-analysis. Neurology, 51 (3), 728733. doi: 10.1212/WNL.51.3.728 Just, M. A., & Carpenter, P. A. (1992). A capacity theory of comprehension: Individual differences in working memory. Psychological Review , 99 (1), 122–149. doi: 10.1037/0033-295X.99.1.122 Kane, M.J. & Engle, R.W. (2002). The role of prefrontal cortex in working-memory capacity, executive attention, and general fluid intelligence: An individual-differences perspective. Psychonomic Bulletin & Review 9 : 637.doi: 10.3758/BF03196323 Kappenman, E. S., & Luck, S. J. (2012). Oxford library of psychology. The Oxford handbook of eventrelated potential components . New York, NY, US: Oxford University Press. doi: 10.1093/oxfordhb/9780195374148.001.0001 Karbach, J., & Verhaeghen, P. (2014). Making working memory work: A meta-analysis of executive- Chapter II Introduction 26 Solé, J. (2006). Repetitive transcranial magnetic stimulation effects on brain function and cognition among elders with memory dysfunction. A randomized sham-controlled study. Cerebral Cortex , 16 (10), 1487–1493. doi: 10.1093/cercor/bhj083 Stepankova, H., Lukavsky, J., Buschkuehl, M., Kopecek, M., Ripova, D., & Jaeggi, S. M. (2014). The malleability of working memory and visuospatial skills: A randomized controlled study in older adults. Developmental Psychology , 50 (4), 1049–1059. doi: 10.1037/a0034913 Sur, S., & Sinha, V. K. (2009). Event-related potential: An overview. Industrial Psychiatry Journal , 18 (1), 70–73. https://doi.org/10.4103/0972-6748.57865 Takeuchi, H., Taki, Y., Nouchi, R., Yokoyama, R., Kotozaki, Y., Nakagawa, S., … Kawashima, R. (2018). General intelligence is associated with working memory-related brain activity: new evidence from a large sample study. Brain Structure and Function . United Nations, Department of Economic and Social Affairs, & Population Divison. (2017). World Population Prospects: The 2017 Revision, Key Findings and Advance Tables. Retrieved from https://esa.un.org/unpd/wpp/Publications/Files/WPP2017_KeyFindings.pdf Valtorta, N. K., Kanaan, M., Gilbody, S., Ronzi, S., & Hanratty, B. (2016). Loneliness and social isolation as risk factors for coronary heart disease and stroke: Systematic review and meta-analysis of longitudinal observational studies. Heart , 102 (13), 1009–1016. doi: 10.1136/heartjnl-2015308790 Verhaeghen, P., Marcoen, A., & Goossens, L. (1992). Improving memory performance in the aged through mnemonic training: A meta-analytic study. Psychology and Aging , 7 (2), 242–251. doi: 10.1037/0882-7974.8.3.338 Verhaeghen, P., & Salthouse, T. A. (1997). Meta-analyses of age-cognition relations in adulthood: Estimates of linear and nonlinear age effects and structural models. Psychological Bulletin, 122 (3), 231-249. doi: 10.1037/0033-2909.122.3.231 Waninger, S., Berka, C., Meghdadi, A., Karic, M. S., Stevens, K., Aguero, C., …Verma, A. (2018). Eventrelated potentials during sustained attention and memory tasks: Utility as biomarkers for mild cognitive impairment. Alzheimer’s and Dementia: Diagnosis, Assessment and Disease Monitoring, 10, 452-460. doi: 10.1016/j.dadm.2018.05.007 Wager, T. D., & Smith, E. E. (2003). Neuroimaging studies of working memory: A meta-analysis. Cogn Affect Behav Neurosci, 3 ( 4), 255-274. doi: 10.3758/CABN.3.4.255 Introduction Chapter II 27 Watson, C. E., & Chatterjee, A. (2012). A bilateral frontoparietal network underlies visuospatial analogical reasoning. NeuroImage , 59 (3), 2831–2838. doi: 10.1016/J.NEUROIMAGE.2011.09.030 West, R. (1999). Visual distraction, working memory, and aging. Memory and Cognition, 27 (6), 10641072. doi: 10.3758/BF03201235 Wolk, D. A., Sen, N. M., Chong, H., Riis, J. L., McGinnis, S. M., Holcomb, P. J., & Daffner, K. R. (2009). ERP correlates of item recognition memory: Effects of age and performance. Brain Research , 1250 , 218–231. doi: 10.1016/j.brainres.2008.11.014 World Health Organization. (2015). World report on Ageing And Health . Geneva. Retrieved from https://www.who.int/ageing/events/world-report-2015-launch/en/ Wronka, E., Kaiser, J., & Coenen, A. M. L. (2013). Psychometric intelligence and P3 of the event-related potentials studied with a 3-stimulus auditory oddball task. Neuroscience Letters , 535 (1), 110– 115. doi: 10.1016/j.neulet.2012.12.012 Zaehle, T., Sandmann, P., Thorne, J. D., Jäncke, L., & Herrmann, C. S. (2011). Transcranial direct current stimulation of the prefrontal cortex modulates working memory performance: Combined behavioural and electrophysiological evidence. BMC Neuroscience , 12 (1), 2. doi: 10.1186/1471-2202-12-2 Zhang, Q., Shi, J., Luo, Y., Liu, S., Yang, J., & Shen, M. (2007). Effect of task complexity on intelligence and neural efficiency in children: An event-related potential study. NeuroReport, 18 (15), 15991602. doi: 10.1097/WNR.0b013e3282f03f22 Zhang, Q., Shi, J., Luo, Y., Zhao, D., & Yang, J. (2006). Intelligence and information processing during a visual search task in children: An event-related potential study. NeuroReport, 17 (7), 747-752. doi: 10.1097/01.wnr.0000215774.46108.60 Zinke, K., Zeintl, M., Eschen, A., Herzog, C., & Kliegel, M. (2011). Potentials and limits of plasticity induced by working memory training in old-old age. Gerontology , 58 (1), 79–87. doi: 10.1159/000324240 Zinke, K., Zeintl, M., Rose, N. S., Putzmann, J., Pydde, A., & Kliegel, M. (2014). Working memory training and transfer in older adults: Effects of age, baseline performance, and training gains. Developmental Psychology, 50 (1), 304-315. doi: 10.1037/a0032982 Zurrón, M., Lindín, M., Cespón, J., Cid-Fernández, S., Galdo-álvarez, S., Ramos-Goicoa, M., & Díaz, F. Chapter II Introduction 28 (2018). Effects of mild cognitive impairment on the event-related brain potential components elicited in executive control tasks. Frontiers in Psychology, 9 , 842. doi: 10.3389/fpsyg.2018.00842 CHAPTER I INTRODUCTION Chapter II Working memory training gains in healthy older adults 30 Reviewing working memory training gains in healthy older adults: A meta-analytic review of transfer for cognitive outcomes3 Abstract The objective of this meta-analytic review was to systematically assess the effects of working memory training on healthy older adults. We identified 552 entries, of which 27 experiments met our inclusion criteria. The final database included 1130 participants. Nearand far-transfer effects were analysed with measures of short-term memory, working memory, and reasoning. Small significant and long-lasting transfer gains were observed in working memory tasks. Effects on reasoning was very small and only marginally significant. The effects of working memory training on both near and far transfer in older adults were moderated by the type of training tasks; the adopted outcome measures; the training duration; and the total number of training hours. In this review we provide an updated review of the literature in the field by carrying out a robust multi-level meta-analysis focused exclusively on WMT in healthy older adults. Recommendations for future research are suggested. Keywords: meta-analysis; working memory training; cognitive plasticity; training transfer; healthy older adults; healthy ageing. 3 Publications derived from this study: Peer reviewed publications in print or other media Teixeira-Santos, A. C., Moreira, C. S., Magalhães, R., Magalhães, C., Pereira, D. R., Leite, J., Carvalho, S., & Sampaio, A. (2019). Reviewing working memory training gains in healthy old people: A meta-analytic review of transfer for cognitive outcomes. Neuroscience & Biobehavioral Reviews . (103): 163-177. doi: https://doi.org/10.1016/j.neubiorev.2019.05.009 Abstracts, Poster Presentations and Exhibits Presented at Professional Meetings Teixeira-Santos, A.C., Magalhães, R., Magalhães, C., Pereira, D.R., Carvalho, S., Sampaio, A. (2017). Is working memory training in elderly people effective? A meta-analytic review, Poster session presented at the II International Convention of Psychological Science, Vienna, Austria. Working memory training gains in healthy older adults Chapter II 31 Introduction Ageing of the world population is a major public health concern that has captured the attention of the general public. Overall, more than 962 million people were over the age of 60 in 2017. It is estimated that this number will more than double to 2.1 billion people by the year 2050 (United Nations, Department of Economic and Social Affairs, 2017). Specifically, it is estimated that the population of people over the age of 80 will triple by the year 2050, increasing from 137 million to 425 million (United Nations, Department of Economic and Social Affairs, 2017). Therefore, much effort has been made to promote optimal ageing to avoid both declines in cognitive functioning and dependence on others, which are factors associated with ageing. Specifically, much has been done to try to reverse age-related cognitive decline and prevent or delay pathological cognitive disorders. This movement represents a significant attempt to improve the quality of life of older adults and to relieve the burden on medical care systems that has resulted from a substantial increase in the elderly population. Efforts to address the issue include non-pharmacological interventions, such as the Advanced Cognitive Training for Independent and Vital Elderly (ACTIVE) trial (Ball et al., 2002, Ball et al., 2002; Rebok et al., 2014). The promising findings in the field encouraged researchers to further investigate the benefits of cognitive training in older people. Different cognitive training approaches are reported in the literature (Jolles & Crone, 2012). They can be classified into two major categories: “strategy-based training” and “process-based approaches”. Strategy-based training consists of the development of specific adaptations and strategies, such as mnemonics, which can be used to ameliorate daily struggles (Lustig, Shah, Seidler, & Reuter-Lorenz, 2009), whereas process-based approaches focus on the training of specific cognitive abilities (Clare & Woods, 2004). More specifically, core process-based training focuses on training central mechanisms with the purpose of producing more substantial effects in functions that depend upon this central processor and that share a common neural substrate (Morrison & Chein, 2011). Notably, working memory training (WMT) has emerged as a proxy for improving cognitive functions (Neely & Nyberg, 2015). Working memory (WM) refers to the components responsible for maintain temporally a limited amount of information in an available state to allow the processing of ongoing information (Cowan, 2017). WM performance declines markedly with ageing, and this has been associated with abnormalities Chapter II Working memory training gains in healthy older adults 32 on the frontoparietal networks involved in WM, as well as neuromodulatory (dopamine) and neuroanatomical alterations (Bäckman et al., 2017; Bäckman, Lindenberger, Li, & Nyberg, 2010; Lubitz, Niedeggen, & Feser, 2017; Park & Reuter-Lorenz, 2009; Raz, 2005; Rottschy et al., 2012; Salthouse, 1990). This reduction in WM capacity in older adults, along with a decrease in processing speed, seem to underlie age-related cognitive decline (Braver & West, 2008), primarily because WM is associated with higher-order cognitive functions (Unsworth, Heitz, & Engle, 2005), including reasoning (Shakeel & Goghari, 2017), reading (Just & Carpenter, 1992), prospective memory (Bisiacchi, Tarantino, & Ciccola, 2008), processing speed (Diamond et al., 1999), attention (West, 1999), perceptual organization (Ko et al., 2014), and general language (Kemper, Herman, & Liu, 2004). Therefore, given the decrease in WM performance with ageing and its putative role in higher-order cognitive functions, WMT has been studied extensively to enhance cognition in older adults, and positive effects of WMT on both cognition and neural plasticity have been found (Constantinidis & Klingberg, 2016; Karbach & Verhaeghen, 2014). Experimental studies of WMT typically include an experimental group, whose members participate in a WMT, and a control group. The control group can be a no-contact control group (passive control group) or an active control group that completes a non-related activity or a low-level WMT. Participants in active control group are exposed to a training setting (i.e., number of sessions, contact with the experimenter, a style of intervention) that is similar to that of the experimental group, but they are not exposed to the experimental WM condition. This design with active control condition allows the researcher to control for effects that may result from social contact during the experiment or a participant´s expectations. However, participants from both groups (passive and active control groups) undergo the same testing before and after the intervention as the participants of the experimental groups. There is abundant literature on WMT (see App. table 4). They may include computerized tasks and can be visual, auditory or both visual and auditory. Trained tasks usually consist of complex or simple span tasks or updating tasks. In complex span tasks, participants must recall a sequence of stimuli, which is interleaved with a concurrent activity. In simple span tasks, participants must remember the sequence of stimuli in forward (fwd) or backward (bwd) order. Updating includes tasks in which participants hold specific content in memory, continually updating the information to be remembered and dropping information that is no longer needed. Training is usually adaptive, i.e., the task difficulty adjusts based on the individual’s performance (von Bastian & Eschen, 2016). Working memory training gains in healthy older adults Chapter II 33 Several studies have been designed to study the effects of WMT by comparing the preand posttest results of experimental and control groups immediately after training (posttest) and at a delayed post-training assessment (follow-up). Additionally, studies have investigated the transfer effects, i.e., whether training gains can be generalized to other tasks involving different cognitive abilities (e.g., Borella et al., 2010) such as fluid intelligence (Beatty & Vartanian, 2015). Although there are no clear criteria to define transfer distance, most authors locate the generalization of the effects along a continuum of near to far transfer (Noack, Lövdén, Schmiedek, & Lindenberger, 2009). Near transfer consists of an improvement on tasks that are like the trained task and that share the same mechanisms or components, while far transfer represents an improvement on tasks that measure abilities that are not like the abilities trained. Near-transfer effects are commonly observed (Borella et al., 2010; Li et al., 2008), although this is not always the case (Dahlin, Nyberg, Bäckman, & Neely, 2008). Results regarding far transfer are controversial with limited or no evidence (Borella, Carretti, Zanoni, Zavagnin, & De Beni, 2013). Previous narrative and systematic reviews have debated the potentialities and controversies of WMT (Constantinidis & Klingberg, 2016; Karbach & Verhaeghen, 2014; Lampit, Hallock, & Valenzuela, 2014; Melby-Lervåg & Hulme, 2013, 2016; Morrison & Chein, 2011; Schwaighofer, Fischer, & Bühner, 2015; von Bastian & Oberauer, 2013; Weicker, Villringer, & Thöne-Otto, 2016), yet the results are inconclusive (see App. Table 1App. Table ). Therefore, the current meta-analysis aims to contribute to this debate by examining the generalization of training effects to non-trained tasks (near and far transfer) (aim 1) and the maintenance of the effects over time (i.e., at follow-up) (aim 2) by using a meta-analysis approach that is different from the ones used in previous reviews. Additionally, previous meta-analyses (Karbach & Verhaeghen, 2014; Melby-Lervåg, Redick, & Hulme, 2016) and experimental studies (e.g., Bürki, Ludwig, Chicherio, & de Ribaupierre, 2014; Stepankova et al., 2014; Zinke et al., 2014) have suggested that variables such as type of control group (Melby-Lervåg et al., 2016), age (Borella et al., 2014; Borella, Carbone, Pastore, De Beni, & Carretti, 2017; Zinke et al., 2014), education (Borella, Carbone, et al., 2017), general cognitive ability (Borella, Carbone et al., 2017), baseline performance (Zinke et al., 2014; Zinke, Zeintl, Eschen, Herzog, & Kliegel, 2011) and training dosage (Bürki et al., 2014; Lilienthal, Tamez, Shelton, Myerson, & Hale, 2013; Stepankova et al., 2014) might moderate training gains and transfer effects. For instance, in relation to the type of control group, a meta-analisys from Melby-Lervåg et al. (2016) reported that the type of Chapter II Working memory training gains in healthy older adults 34 control group predicted transfer effects. In particular, studies showed more significant effects when using a passive control group than when using an active control group. However, other meta-analitical studies (Karbach & Verhaeghen, 2014; Weicker et al., 2016) did not find influence of type of control group (active or passive) in transfer effects. Regarding the age, an experimental study performed by Borella et al. (2014) found transfer effects of a visuospatial WMT for measures of STM (short-term memory), WM, inhibition, processing speed, and reasoning only in young-old adults but not in old-old adults. In accordance, together with an age-related difference in the transfer effects, Borella et al. (2017) also documented the role of age as an important moderator of the effects in WMT, although the results varied according to the type of transfer task. In addition, Zinke et al. (2014) evidenced that old-old participants had less gains than young-old participants, except for fluid intelligence in which the reverse pattern was verified. Borella, Carbone, et al. (2017) have also shown that vocabulary and baseline performance influenced WMT. In this study participants with higher vocabulary scores and poor pretest performance benefited more from training, although this pattern was not the same in all outcomes (e.g., in fwd digit span, lower vocabulary score was related to more benefit in training). Moreover, participants with low levels of baseline performance in WM tasks were likely to benefit more from WMT (Zinke et al., 2014, 2011). Related to session length/duration, Jaeggi et al. (2008) documented a significant growth in far transfer throughout the sessions (from 8 to 19 sessions). Other researchers showed that a group which trained for 20-day outperformed a 10-day training group in a visuospatial measure (Stepankova et al., 2014), while a small positive significant moderator effect for small training dose in comparison to large training dose was observed in a meta-analysis (Melby-Lervåg et al., 2016). Taken together, in the current study, we verified if the variables as type of control (active/passive), mean age of participants, total number of training hours, number of training sessions, training length in weeks, training type (single training - complex span, simple span, updating, or mixed training: more than one type of WM task), years of formal education, general cognitive ability (operationalized by vocabulary score), and baseline performance would moderate the training effect (aim 3). In addition, we also verified if the type of the outcome adopted (e.g., Cattell; Raven’s Advanced Progressive Matrices - RAPM; complex span) would moderate the transfer effect. Previous meta-analytical work merged the results of different age groups (Mansur-Alves & Silva, 2017; Melby-Lervåg et al., 2016; Melby-Lervåg & Hulme, 2013) or did not include older adults (Au et Working memory training gains in healthy older adults Chapter II 35 al., 2015). This review focuses on only older adults, as WM is markedly affected by ageing (Salthouse, 2000), and WMT is proposed as an innovative approach to counteract age-related cognitive declines (Constantinidis & Klingberg, 2016; Karbach & Verhaeghen, 2014). While merging different ages and conditions may yield sample heterogeneity, this practice can pose some problems for the internal and external validity of the findings (Rothwell, 2006). Additionally, to better isolate the effects of WMT, this meta-analysis addresses the specificity of the training delivered to the experimental groups by including studies whose experimental groups participated in trainings focused exclusively on WM and excluding studies whose experimental groups participated in trainings targeting cognitive functions other than WM. We also excluded papers whose active control groups participated in a non-adaptive WMT that remained always in a lower level of WMT (Brehmer et al., 2011; Chan, Wu, Liang, & Yan, 2015; Loosli et al., 2016; Shing, Schmiedek, Lövdén, & Lindenberger, 2012; Simon et al., 2018; Wayne, Hamilton, Huyck, & Johnsrude, 2016), specific examples include: comparing an adaptive WMT with a WMT whose load (e.g., N = 2 or N = 3) is held constant throughout the training (Brehmer et al., 2011; Chan et al., 2015; Wayne et al., 2016); training both experimental and control groups with a recent-probe and an n -back task, with the experimental group receiving trials with higher proactive interference when compared to the control group (Loosli et al., 2016); the participants performed a numerical memory updating task, however different groups were exposed to distinct rates of stimuli presentation (750 ms, 1500 ms or 3000 ms) (Shing et al., 2012). Considering that our aim was to contrast WMT with a placebo training not related to WM (e.g., questionnaire, quiz, visual search) or a non-training condition, in the present review, the above-mentioned studies were not included in the analysis. The rationale behind this is the fact that even a low-level of WM performance activates similar brain areas as high-level of WM processing (Braver et al., 1997; Kawagoe et al., 2015; Ragland et al., 2002). Since we do not have enough information to determine a suboptimal dosage of WMT that would work solely as placebo (Huitfeldt, Danielson, Ebbutt, & Schmidt, 2001), comparing different loads of WMT could lead to less interpretable data as these WM tasks might produce similar effects. As a consequence, we would not be able to isolate gains that are due to WMT (ICH Harmonised Tripartite Guideline, 2000). In fact, as suggested by Brehmer and colleagues (2011), both adaptive WMT and training at low WM load might lead to neural changes. Additionally, although many researchers classify executive function tasks as WM we did not include training of executive functions, such as Stroop interference, verbal fluency or task switching. As claimed by Oberauer et al. (2018) in the Benchmarks for Models of Short Term and Working Memory, Chapter II Working memory training gains in healthy older adults 42 plots with the effect sizes of the included studies in all comparisons can be found in the supplementary material section (see figure A). Results The results are described in four major sections. First, we describe the different studies that were included in the analysis. Second, we present the small-study effect analyses. The third section targets the main aim of this review which was to verify the WMT effectiveness at posttest and follow-up together with the moderator analysis. Finally, the risk and publication bias results are presented. Characteristics of included studies We identified 300 studies (after removal of duplicates), from which 217 were excluded after reading the abstract and 59 after the full-text analysis. Criteria for paper exclusion: a) review paper; b) sample of non-human animals; c) young participants or elderly but not cognitively healthy participants; d) training does not exclusively target WM; e) the active control group performed a WM task; f) absence of control group; g) studies whose sample has been previously used in a another study already included in the meta-analysis; h) WMT coupled with transcranial direct current stimulation (tDCS); i) incomplete data. Twenty-four articles (27 experiments) met the inclusion criteria (for a list of the included papers, see table B in the supplementary material) and were selected for the quantitative analysis, which included data for up to 1130 participants. All trials were published in the last ten years, with Psychology and Aging as the journal with the highest number of publications. The mean age of the participants ranged from 62.9 to 87.1 years ( M = 69.5, SD = 4.9), and years of formal education ranged from 6 to 17 ( M = 12,7 years, SD = 2.85). Of these studies, 79% were carried out in Europe ( n = 19), with the remainder conducted in North America ( n = 3; 13%) and Asia ( n = 2; 8%). On average, studies implemented 12 training sessions ( SD = 8.59; range = 3 - 40), corresponding to seven total hours ( SD = 4.36; range = 1.5 - 20), with a mean session duration of 42 minutes ( SD = 13.8; range = 20-60), and an average of three days of sessions per week ( SD = 1.36; range = 2 - 7). Follow-up was reported in eight papers, with a mean of eight months after training ( SD = 4.4; range = 3 - 18). The completion rate for the whole sample ranged from 70 to 100%. Most of the training was performed in laboratory settings ( n = 16); however, six trials were conducted at participants’ homes. This information was not detailed in three papers (Richmond, Morrison, Chein, & Olson, 2011; Working memory training gains in healthy older adults Chapter II 43 Xin, Lai, Li, & Maes, 2014). In eight studies, participation was voluntary, one study included both pay and voluntary participation, ten articles reported financial compensation, and five papers did not mention this information. Regarding the type of trained task (see App. table 4), studies were grouped into three major categories (Schmiedek, Hildebrandt, Lövdén, Wilhelm, & Lindenberger, 2009; Shipstead et al., 2012): complex or simple span task; updating; mixed (i.e., participants were trained on more than one type of WM task). Eight studies included a complex span task, participants were trained on a simple span task in one study (Zinke et al., 2011), and updating training was observed in ten studies. Five studies had mixed training. Regarding the modality of training (verbal vs. visuospatial), 10 studies included training with verbal stimuli, five included training with visuospatial stimuli, and the remaining nine were crossmodal. All studies, except Pergher, Wittevrongel, Tournoy, Schoenmakers, & Van Hulle (2018), Xin et al. (2014), Zając-Lamparska & Trempała (2016), had adaptive training. Fourteen articles had an active control group, while ten had a passive control group (PCG). As seen in App. table 3, characteristics regarding type of training and control, outcomes and follow-up varied across studies. Heterogeneity indexes among studies in the different analyses were low to moderate (Higgins, Thompson, Deeks, & Altman, 2003). However, we opted for the random model considering the clinical and methodological heterogeneity found among studies (Higgins & Green, 2008). Before proceeding to the meta-analysis, small-studies effects were explored. The comparison between random-effect modelling, fixed-effect modelling and the Henmi and Copas method were conducted to address this issue. The results of this analysis are summarized in App. Table 9. The conclusions of the three models produced very similar results, and in 71% of the cases the difference was ≤ 0.001, not affecting the significance of the results. The most distinct case happened for verbal complex span at posttest, for which the mean effect from the random-effects model was 0.34, 95% CI = [0.09, 0.58], and the common effect from the fixed-effects model was 0.31, 95% CI = [0.14, 0.49]. In both cases, confidence interval (CI) did not include zero, confirming its statistical significance. Additionally, sensitivity analysis confirmed that the meta-analytic findings were robust regarding the tested correlation coefficients. Indeed, by visual inspection of the table C in the supplementary material, it is possible to observe that when the correlation is assumed to be lower, at r = 0.3, or higher, at r = 0.7, the estimated summary effect varies by no more than 0.04. Chapter II Working memory training gains in healthy older adults 44 WMT efficacy and moderator analysis In this section the results from the effect of WMT on transfer task immediately after training (aim 1) and at follow-up (aim 2), as well as, a moderator analysis (aim 3) will be presented. Results from the classical p -value or those corrected for small samples (Skovgaard’s and RVE) did not differ considerably, so we reported the multi-level p -value in the text and all the values in Table 1. The comparisons only had a small difference between the multi-level p-value ( p = .03) and the RVE ( p = .06) for visuospatial WM in posttest and the multi-level p -value ( p = .04) and RVE p-value ( p = .08) for verbal WM at follow-up. Therefore, the results regarding visuospatial WM in posttest and verbal WM at follow-up should be interpreted with more caution. We did not find any significant difference between the control types (passive versus active control groups) in the moderation analysis (see App. Table 10), except for visuospatial WM at posttest. Additionally, we performed a sensitivity analysis, running the analysis separately for passive and active control groups. The comparison with both passive and active control group merged did not yield an effect size greater than when we performed the comparison of experimental group with studies that included only an active control group, except for visuospatial WM at posttest. Many of the included trials had passive control group ( n = 10). If we had excluded those trials from the analyses, some of the comparisons would have a very few studies, decreasing the power of the analyses. Accordingly, the results from both control groups were merged into a single control condition. The effect sizes were calculated comparing the experimental condition with the merged control condition. Aim 1: examining the generalization of training effects to non-trained tasks (near and far transfer). WMT effects were examined on near transfer constructs (visuospatial and verbal WM, and visuospatial and verbal STM) as well as on a far transfer construct (reasoning) immediately post-training. Verbal WM: A significant transfer effect was identified for verbal WM (0.23; 95% CI [0.07, 0.39]). Visuospatial WM: A significant transfer effect was identified for WM in the visuospatial modality (0.23; 95% CI [0.03, 0.43]). Verbal and visuospatial STM: No significant transfer effects were identified for verbal (0.16; 95% CI [- 0.05, 0.36]) or visuospatial STM (-0.03; 95% CI [- 0.39, 0.32]). Reasoning: For reasoning, the effects were not significant ( p = .08) at posttest (0.10; 95% CI [- 0.03, 0.23]). Working memory training gains in healthy older adults Chapter II 45 Aim 2: verifying the maintenance of the effects at follow-up. Concerning the long-term effects of WMT, we observed that the effects were also observed during follow-up to verbal WM (0.23; 95% CI [0.01, 0.46]). However, in visuospatial WM analysis, the effect was not significant (0.14; 95% CI [- 0.09, 0.37]). Regarding reasoning, results were also not significant (0.13; 95% CI [-0.09, 0.35]), as well as for verbal STM (0.18; 95% CI [- 0.10, 0.45]) and visuospatial STM (-0.04; 95% CI [-0.33, 0.25]). Aim 3: testing moderator variables. Here we examined if the variable age, training dose, number of sessions, training type, training duration, years of formal education, vocabulary score, baseline performance and type of outcome might moderate training effects. The results are presented in Table . The moderator analysis was significant ( p < .05) for number of sessions, training length (in weeks) and training dose (in hours), i.e., the gains in reasoning and verbal STM immediately after training are small when training duration increases. Additionally, while the effect of WMT on verbal STM was linearly moderated by training hours and training length, the effect of WMT on Reasoning-posttest was also moderated by the former factors together with the number of sessions. Table 2 outlines these moderator roles. Indeed, the approximation by higher polynomial degrees were also assessed but, in each case, no significance advantage over the linear approach was observed. Specifically, no asymptotic behaviour was detected, as such characteristic would imply a significant variation in the rate of change of the WMT effect with respect to the corresponding independent variable. Chapter II Working memory training gains in healthy older adults 46 Regarding the training type, we observed that the studies that included mixed training (i.e., having more than one type of WM tasks) had smaller effects on reasoning immediately after training than the training of updating or complex span tasks alone. Additionally, studies having the Cattell Test as an outcome displayed a higher gain than studies that used other measures in posttest (RAPM; RSPM; LPS). For verbal WM, the gains were higher in complex span tasks than in simple span and updating tasks at posttest. Type of control group was a significant moderator for verbal WM at posttest, with the effect size of studies using a passive control group being higher than studies that used an active control group. Finally, baseline performance moderated the effects on visuospatial STM at immediate posttest, with participants with lower performance showing more benefits with the training. Construct No. of effects (k) No. of studies (clusters) RE mean Q-test I2 (%) τ2       Estimate 95% CI pvalue Skovgaard’s p-value RVE pvalue P O S T T R A I N I N G Reasoning 33 24 0.10 [-0.026,0.233] .12 NA .13 28.53 11.51 NA 0.01 <0.01 Verbal WM 40 20 0.23 [0.065,0.392] .006 ** NA .01* 88.79 *** 56.13 NA <0.01 0.09 Visuospatial WM 13 10 0.23 [0.029, 0.426] .025 * NA .06^ 16.03 17.83 NA 0.02 <0.01 Verbal STM 12 11 0.16 [-0.045,0.363] .13 NA .16 12.41 14.07 NA <0.01 0.01 Visuospatial STM 6 5 -0.03 [-0.388, 0.324] .86 NA .74 09.06 45.24 NA <0.01 0.08 F O L L O W - U P Reasoning 12 10 0.13 [-0.085, 0.347] .24 NA .27 9.36 6.37 NA 0.01 <.01 Verbal WM 17 9 0.23 [0.006, 0.457] .04 * NA .08^ 18.59 16.35 NA 0.01 0.01 Visuospatial WM 11 8 0.14 [-0.089, 0.368] .23 NA .14 6.04 <0.01 NA <0.01 <0.01 Verbal STM 6 6 0.18 [-0.097, 0.452] .205 .983 .19 3.85 <0.01 <0.01 NA NA Visuospatial STM 6 5 -0.04 [ -0.334, 0.245] .763 NA .72 3.17 NA NA <0.01 <0.01 Note . ^ p <.1, * p <.05, ** p <.01, *** p <.001. NA – Not applicable (only for groups from the same experiment). I2 – total heterogeneity / total variability;  – estimated amount of total heterogeneity;   – Variance component of the 3-level model for the between-studies heterogeneity;  – Variance component of the 3-level model for the within-studies (effects within studies) heterogeneity. RVE – Robust variance estimation. Number of studies may be smaller than number of effects because each study may have more outcomes for the same construct. RVE and Skovgaard’s (only for 2-level random effects) were applied as a sensitivity analysis to check the robustness of the model. P -values did not differ substantially across these analyses indicating the validity of the model. Table 1 Effects of Working Memory Training Compared with Control Group by Construct Working memory training gains in healthy older adults Chapter II 47 Table 2 Moderator Effects (Significant Results) Construct Moderator effect Estimate SE p -value QE QM - Test of moderators   Reasoning at immediate posttest Measure - Cattell 0.39 0.14 .005** 20.70 7.82 ** <0.01 <0.01 Training dose (hours) -0.04 0.01 .001 ** 17.46 11.040 *** <0.01 <0.01 Number of training sessions -0.02 0.01 .004** 20.17 8.35** <0.01 <0.01 Training length (in weeks) -0.11 0.04 .004 ** 20.38 8.15 ** <0.01 <0.01 Training Type - Mixed -0.41 0.13 .001** 18.36 10.16 ** <0.01 <0.01 Verbal WM at immediate posttest Measure – Complex span 0.27 0.13 .046 *** 80.67 4.00 * <0.01 0.08 Visuospatial WM at immediate posttest Control – PC – AC 0.54 0.24 .023* 10.86 5.17 * <0.01 <0.01 Verbal STM at immediate posttest Training dose (in hours) -0.04 0.02 .043* 8.33 4.08* <0.01 <0.01 Training length (in weeks) -0.11 0.05 .033* 7.89 4.53* <0.01 <0.01 Visuospatial STM at immediate posttest Baseline performance -0.06 0.02 .01* 2.33 6.73** <0.01 <0.01 Note . * p < .05, ** p < .01, *** p < .001;  – Variance component of the 3-level model for the between-studies heterogeneity;  – Variance component of the 3-level model for the within-studies heterogeneity. QE – test for residual heterogeneity when moderators are included. QM – test statistic for the omnibus test of coefficients. Moderator effects with non-significant results were not presented, they were mean age of the participants, years of formal education, vocabulary performance. Analyses of follow-up did not have any significant moderator. In summary, WMT had a small significant and long-lasting effect on verbal WM (specifically on complex span outcomes). For visuospatial WM, gains were only observed at posttest, but not at followup. Far transfer for reasoning was not observed. Training length, number of sessions, training dose (total training duration in hours), type of training and adopted outcomes (Cattell; and complex span), type of control group and baseline performance appeared as significant moderator variables at posttest assessment. Publication and risk of bias Assessment of risk of bias is important when performing a review because it is an index of the quality of included data, and it could also explain heterogeneity when it is highly observed (Viswanathan et al., 2008). Two authors independently assessed the risk of bias. In general, we observed a substantial absence of information for most studies, which limited the ability to classify the risk of bias. Considering the randomization processes (selection bias), 19% of the studies presented risk of bias, whereas in 74% the risk of bias was not clear. Seven percent of the studies adequately reported random sequence generation. Regarding allocation concealment, 22% presented a high risk of bias, 7% adequately reported data, and the remaining 70% did not report on allocation concealment. For blinding (performance bias), 30% of the studies had low risk of bias (compared with 30% with high risk), and 40% of the studies did not mention blindness procedures. Seventy percent of the studies did not exclude data from participants Chapter II Working memory training gains in healthy older adults 48 who dropped out or with missing data. Fifteen percent had high risk of incomplete outcome data, while this was not clear in 15% of the studies. Generally, the studies had high completion rates (ranging from 86% to 100%), although the completion rate was not clear for all studies. Similarly, most articles (93%) reported all outcomes, although they did not state which outcome was the primary. Seven percent presented high risk of selective reporting. Additionally, the lack of adequate correction for multiple comparisons and for baseline group differences were other potential bias observed here. Another possible source of bias was the lack of appropriate screening measures of cognitive decline and of affect disorders such as anxiety and depression. A summary graph of the risk of bias is displayed in App. Table 3 Analysis of publication bias assesses if the set of evidence is biased due to the fact that positive findings are more likely to be published. The analysis of several methods of publication bias (trim-andfill, leave-one-out, asymmetric tests, and Hemni and Copas) suggested a small presence of publication bias, although it did not seem to substantially alter the results. Trim-and-fill is a method that estimates the number of studies missing in the funnel plot (Duval & Tweedie, 2000b). It was only used in analyses with at least 10 studies; otherwise, the test would not have sufficient power to verify asymmetry (Sterne, Egger, & Moher, 2008; Zhou et al., 2017). This analysis suggested the presence of publication bias in only two cases (simple span and complex span at posttest). Additionally, given that the big issue of publication bias is that the positive results are more representative in the published literature (Mlinarić, Horvat, & Smolčić, 2017), it is important to highlight that trim-and-fill method identified only two cases of missing studies (verbal simple span STM and verbal updating WM, both at posttest), however the effect sizes of the corresponding categories were not significant in verbal simple span STM and verbal updating WM at posttest. The leave-one-out method was performed by a sensitivity analysis where one study at a time was removed from the analysis to verify the influence of a single study in the finding. This method showed sensitivity of results to individual studies in three cases (verbal fwd simple span at posttest; Cattell and verbal complex span at follow-up). However, in the first two cases, the elimination of a unique experiment would cause a significant pooled effect size, while only for complex span the elimination of a study (among three) would cause a nonsignificant result. Asymmetric tests indicated publication bias in only one case (verbal simple span at posttest), the same comparison already identified with the trim-and-fill method. Finally, in all cases, the Hemni and Copas robust estimation was not significantly different from Working memory training gains in healthy older adults Chapter II 49 the random-effects results, showing that publication bias did not change the overall meta-analytic effects in a significant manner. Therefore, the positive effect of publication bias was not a big issue here. Overall, the presence of bias did not seem to influence the results as supported by the former publication bias methods (see App. table 9), as well as, by the similarity between effect sizes of studies that presented more criteria classified as high risk of bias (see app. figure 3) (e.g., Goghari & LawlorSavage, 2017; Heinzel et al., 2016; Stepankova et al., 2014; Zinke et al., 2011) and those having a lower risk of bias (e.g., Borella et al., 2013; Borella, Carretti et al., 2017b; Guye & von Bastian, 2017; Lange & Süß, 2015; Weicker et al., 2018). Discussion This meta-analytical review aimed to verify the gains of WMT on transfer measures in healthy older adults. In contrast to previous meta-analyses, we used different analytical methods to address multiple outcomes and the lack of correlation reports. Additionally, a description of the studies included in the review is provided along with a comprehensive overview of different studies in the WMT field. The high variability between the experiments challenged data aggregation and, consequently, data interpretation. The studies presented different experimental and control tasks (see App. table 4 and App. Table 5), different outcomes (see App. Table 7), and training protocols. Follow-up also varied broadly across trials, although it was seldom included in the experimental protocol (see App. Table 3. Regarding the results of the effectiveness of WMT at posttest (aim 1), participants assigned to a WMT group displayed a small significant near transfer effect size of 0.2 for verbal and visuospatial WM, compared to the participants who received a placebo or non-intervention. These results are in line with previous meta-analyses that have shown small to medium near effect sizes immediately after training (Karbach & Verhaeghen, 2014; Melby-Lervåg & Hulme, 2013; Melby-Lervåg et al., 2016). For example, Karbach and Verhaeghen (2014) observed a small near effect size of 0.3 after removal of publication bias (trim-and-fill method). We also observed that WMT had no significant impact on STM, which conflicts with the results of previous research (Schwaighofer et al., 2015). These differences among studies may be due to methodological differences, as Schwaighofer and colleagues (2015) included older adults as well as children and young adults. Moreover, it might be the case that the lack of effect in STM may be due to a preservation of this ability with age (Nittrouer, Lowenstein, Wucinich, & Moberly, 2016; Olson et al., 2004). Therefore, there is less room for transfer in this ability after WMT. Nevertheless, this Chapter II Working memory training gains in healthy older adults 50 hypothesis needs to be further explored as there was one study showing a strong positive effect of WMT on STM (Heinzel et al., 2013). As we observed in the moderator analysis, variables such as the training dose and length, as well as, baseline performance interfered with the effects, which may cause heterogeneity across studies. For the reasoning, there was no significant transfer effect. In fact, a previous meta-analysis (Karbach & Verhaeghen, 2014) only yielded a “marginally significant” far transfer effect that was not fully corroborated by our study with a greater number of WMT trials included in the analysis. With respect to the WMT long-term effects (aim 2), only ten studies reported follow-up assessments; therefore, the results should be considered with caution. Near transfer effects seem to be maintained at follow-up only for verbal WM. These results are in agreement with Schwaighofer et al. (2015) and partially consistent with Melby-Lervåg et al. (2016, 2013), who only observed a significant maintenance effect in WM outcomes. We performed a moderator analysis with the following variables as moderators of transfer effects on STM, WM and reasoning at posttest and follow-up (aim 3): 1) type of control (active/passive); 2) the mean age of participants; 3) training dose (total number of training in hours); 4) training length (in weeks); 5) number of training sessions; 6) training type (single: complex span or updating; mixed training: more than one type of WM task); 7) years of formal education; and 8) category of the outcome (e.g., Cattell; RAPM; complex span); 9) vocabulary score; 10) baseline performance. The variables that explained heterogeneity of the effect sizes in reasoning at posttest were the category of the outcome (i.e., Cattell), training length/dose, number of training sessions, and training type (i.e., mixed training). For verbal WM at posttest, the category of the outcome (i.e., complex span) was the variable that explained heterogeneity of the effect sizes. This means that studies having complex span as outcome found more positive effects than studies using another WM measures. For visuospatial WM at posttest, the type of control group (active versus passive) was a significant moderator, with studies using passive control groups presenting higher effect sizes. For verbal STM at posttest, training length and hours were the significant moderators. For visuospatial STM at posttest, baseline performance moderated the results, with participants with lower performance gaining more with the training. The fact that some measures (i.e., Cattell Test and Complex Span Task) displayed more significant effect sizes than others in the moderator analysis highlights the role of the measures to evaluate the training effects. For reasoning, the effect size on the Cattell Test was significant, showing a positive Working memory training gains in healthy older adults Chapter II 51 moderation effect of this test on far transfer. This result is in line with the results of previous reviews which showed a slightly larger effect of the Cattell Test compared to Raven’s Test (Mansur-Alves & Silva, 2017). This finding could be explained by the fact that the Cattell Test consists of different subtests (series, analogies, matrices and classification), which may position it as a more complete indicator of reasoning compared to tests that only have figural type items (e.g., Raven’s), as postulated by Gignac (2015). Furthermore, this result is consistent with the claim of Shipstead et al. (2012) regarding the importance of having different instruments to assess transfer effects in the experiments, ensuring that all facets of the construct are assessed. Considering the moderation effect of training duration/length, either in reasoning or verbal STM, we found unexpected results. For both variables, the results showed a significant negative effect, i.e., that more training duration (total number of hours and length) produced smaller effect sizes. Other variables probably influenced this analysis, such as the type of training performed: most of the shorter duration studies applied the same training task which may be more effective than the training adopted by the long-duration studies (Borella, Carbone, et al., 2017). It is also noteworthy that only one study had higher dosages of training (more than 15 hours) (Goghari & Lawlor-Savage, 2017), whereas six out of twenty had only three sessions (Borella et al., 2014, 2010, 2013; Borella, Carretti, et al., 2017; Cantarella, Borella, Carretti, Kliegel, & De Beni, 2017; Cantarella, Borella, Carretti, Kliegel, Mammarella, et al., 2017). Previously, Karbach and Verhaeghen (2014) and Melby-Lervåg and Hulme (2013) failed to find a significant influence of total training duration in effect size, except for one measure, the Stroop task in Melby-Lervåg and Hulme (2013). In contrast, Schwaighofer et al. (2015) found a positive influence of total training duration on visuospatial STM and of session duration on verbal STM. Weicker et al. (2016) documented a positive correlation between the number of sessions and the effect sizes. In this case, the authors compared two groups (> 20 sessions vs. < 20 sessions) and observed that more training sessions produced larger effect sizes. Nonetheless, the total number of hours was not related to the effect size (> 10 hours vs. < 10 hours). Finally, similar to our results, a previous meta-analysis on video-game training have shown that short training produced stronger effects than long training (Toril, Reales, & Ballesteros, 2014). These discrepant findings need to be further addressed in new randomized controlled trials. Other factors such as motivation and performance anxiety should also be considered (DelphinCombe et al., 2016; Jaeggi, Buschkuehl, Shah, & Jonides, 2014). As participants are older adults, some 58 App. Table 2. Inclusion and Exclusion Criteria Inclusion criteria • Experiments must be randomized controlled trials, or quasi-experimental trials with treatment and either a placebo or a passive control condition tested at preand post-intervention. • Participants must be healthy older adults, without cognitive decline or any type of dementia. • The intervention of the experimental group must consist of repeated training, computerized or not, focused exclusively on WM skills. • Presenting average and standard deviation/error for preand post-training outcomes. • Paper had to present an outcome measure following one of the constructs presented in App. Table 6. Exclusion criteria • Theoretical and review articles, book chapters, and research protocols. • Studies conducted with non-human animals. • Participants with any neurological condition. • Young participants. • Training not exclusively targeting WM (e.g., multi-component training/game or training coupled with physical-exercise interventions). • Active control group with participants performing a WM task. To be included, the active control group must perform a placebo task to facilitate the isolation of the WMT effect in the training group. • Publications based on the same (or part of a) sample of an experiment already included in the analysis (in this case, the one with more detailed data was kept). • Absence of preand post-training outcomes with average, standard deviation/error and sample size. 59 App. Table 3. Characteristics of the Included Studies in Alphabetical Order Study Intervention condition (n) Comparison (n) Computerized? Measures/ Outcomes Number and duration of training sessions Near gains in posttest Far gains posttest Follow-up: time and effect Completion rate (reason for dropout) Borella et al. (2010) CWMS (n = 20) Questionnaires (n = 20) X Cattel test; Stroop; bwd/fwd DS; CWMS; Dot matrix; Pattern comparisons 3x60 min  Dot Matrix  Fwd/bwd DS  Cattell  Pattern comparison  Stroop test 8 months X all from posttest 100% Borella et al. (2013) CWMS (n = 18) Questionnaires (n = 18) X Cattel; Stroop; bwd/fwd DS; CWMS; Dot matrix task; Pattern comparison 3x60 min  Fwd DS X Dot Matrix X Bwd DS  Stroop incongruent errors X Cattell X Stroop interference index X Pattern comparison 8 months  all from posttest 100% Borella et al. (2014) Matrix task (n = 20) Questionnaires (n = 20)  CWMS; Cattel; Stroop Color; Fwd /bwd Corsi; Pattern comparison. 3x60 min  CWMS  Fwd/bwd Corsi  Pattern comparison X Stroop X Cattell 8 months  CWMS X Fwd/bwd Corsi X Pattern comparison 100% Borella et al. (2017) CWMS (n = 18) Questionnaires (n = 18) X LST; The Jigsaw Puzzle test; Fwd/bwd DS; Pattern Comparison; The letter sets 3 x 30 min X Jigsaw puzzle X LST X DS  Pattern Comparison X The letter sets 6 months  Jigsaw Puzzle  LST  Bwd DS X Pattern comparison n.c. Bürki et al. (2014) Verbal N-back (n = 22) Implicit sequence learning or passive control group (n = 20)  Spatial n-back; Number updating; Reading Span task; RSPM; Stroop; Letter and pattern comparison; Simple reaction time. 10 x 30 min n.a. n.a. n.a n.c. Cantarella, Borella, Carretti, Kliegel, & De Beni (2017) CWMS Task (n = 18) Questionnaires (n = 18) X EPS; TIADL; Cattell; RSPM. 3x 30min n.a.  EPT  Cattell  RSPM X TIADL n.a. 100% Cantarella, Borella, Carretti, Kliegel, Mammarella, et al., 2017 Exp. 1 Matrix task (n = 18) Questionnaires (n = 17)  CWMS; Fwd/ bwd Corsi span; Pattern comparison task; Cattell 3 x 60 min X X X n.c. Cantarella, Borella, Carretti, Kliegel, Mammarella, et al., 2017 Exp. 2 Matrix task (n = 16) Questionnaires (n = 19)  CWMS, Fwd/Bwd Corsi; Pattern comparison; Cattell. 3 x 60 min X X X 88% (medical issues) Dahlin, Nyberg, et al. (2008) Letter/ number/ colors/ spatial location memory task; Keep-track task (n = 13) PCG (assumed) (n = 16)  Digit symbol substitution; 3-back; Computation span; Recall of concrete nouns; Paired associate learning; Verbal fluency (FAS); RAPM 15 x 45min  3-back (for young-old adults)  Recall of concrete nouns (for young-old adults) XOther outcomes X 18 months  3-back (for young oldgroup) X Concrete nouns (for young old-group) 86% (reason n.c.) Du et. al. (2019) Digit running memory task / chess game (n = 14) Mental health-related lecture sessions (n = 9)  Running memory task; matrix updating; simple span; complex span; Corsi; Digit Symbol; Number cancellation; Pattern comparison; Association; Recognition; RAPM; Cattell; Paper folding; Spatial relationship; Everyday problems test. 12 x 45 min  Color updating  Matrix updating X Keep Track X 2-back X Fwd digit span X Corsi block X Corsi spatial span X Operational span X Processing speed X Episodic memory X Figural reasoning 3 months 70% (hospitalization, moving out of the city) Goghari et al. (2017)* Multi-memory game; moving memory game and n-back – BrainGymmer (n = 36) PCG (n = 29)  Aospan; DS; Tower task; RAPM; Symbol search; Letter fluency; DFT3; TMT-4; CWIT 40 x 30 min X X n.a. 90% (disliking the training, health difficulty, low training dosage) Guye & von Bastian (2017) Figural-spatial complex span task / local-recognition task / memory updating task (n=68) Visual search (n = 74)  Brown-Peterson; binding; memory updating; RAPM; Relationships; Locations; Shifting; Flanker; Stroop; Simon. 25 x 30 min X X n.a. 74% (lack of time, technical problems with: diary, installing software; illness; differences in the assessment protocols pre and post-assessment) Heinzel et al. (2013) N-back (n = 15) PCG (n = 15)  Fwd/bwd DS; CERAD (Imm recall; Del recall); digit symbol; verbal fluency; RSPM; LPS 12 x 45 min  Fwd DS X Bwd DS  CERAD del recall  Digit symbol X CERAD imm recall X Verbal fluency X RSPM X LPS n.a. 97% (missing in more than two consecutive training sessions) Heinzel et al. (2016) N-back (n = 15) PCG (n = 14)  Fwd/bwd DS; D2 test, Digit symbol substitution; Verbal fluency; Stroop interference; RSPM; LPS 12 x 45min X  Stroop  D2 test  LPS X RSPM X DS X Verbal fluency XDigit symbol substitution n.a. 91% (technical failure during the fMRI scanning) Lange et al. (2015) Aospan; dot span; numerical memory updating; figural running; multiple switching (n = 31) Non-adaptive computerized quizzes and simple board game (n = 31)  Reading span; Swaps; Switching; Numerical running memory span; Berlin intelligence structure test; DS; Word span; Electronic questionnaire for cognitive failures in everyday life; Cognitive failures questionnaire 12 x 60 min X X All, except for subjective changes n.a. 86% (illness or time conflicts) Payne et al. (2014) Semantic category span; lexical decision span; sentence reading span (n = 22) Component control design (n = 18)  Reading span; Sentence listening span; Aospan; Minus-2 span; Sentence memory; Discourse memory; Reading comprehension; Verbal fluency (FAS) 15 x 30 min   Verbal fluency  Sentence recall X Rivermead X The Nelson-Denny n.a. 98% (absence in posttest) Pergher et al., 2018 N-back (n= 14) PCG (n= 14)  TOVA; Corsi block-tapping test; RAPM 10 x 30 min X  TOVA  RAPM n.a. 100% Richmond et al. (2011) Complex WM span task – spatial and verbal subtests (n = 21) PCG (n = 19)  Reading span; Fwd/bwd DS; RSPM; TEA; CVLT 20 x 20min  Reading Span X Fwd/bwd DS  CVLT (repetition) X TEA X CVLT (total correct; intrusions) n.a. 87% (n.c.) Salminen et al. (2016) Dual n-back (n = 25) PCG (n = 21)  WM updating; Task switching; Attentional blink 14 x 35 min  VS WM updating X AV WM updating X Tasking switching X Attentional blink n.a. 100% Stepankova et al. (2014) Verbal n-back (n = 20) PCG (n = 25)  DS; Letter-number sequencing; Block design (WAIS-III); Matrix reasoning (WAIS-III) 20 x 25 min  WM composite score  Visuospatial skills composite score n.a. 100% von Bastian et al. (2013) Numerical complex span; figural task switching; tower of frame (n = 27) Quiz; visual search; counting (n = 30)  Verbal complex span; Kinship integration; Verbal task switching; Binding; n-back; RAPM 20 x 30 min  Verbal complex span X Kinship integration X Biding X Verbal task switching X RAPM n.a. (n.c. Weicker et al. (2018) WOME intervention (n =20) PCG (n = 20)  Fwd/bwd DS; Fwd/bwd span board; Spatial addition; Symbol span; PASAT; Stroop; VLMT; LPS-3; TMT-B; TAP (alertness, mental flexibility and go-no-go); Operation Span; N-back 12 x 45min  Span board (bwd) X Other WM tasks X 3 months X 90% (illness, one moved away, car accident, traumatic brain injury, death). Xin et al. (2014) Letter, animal, and location running (n = 15) Computer games (n = 14)  Numerical updating; Fwd/bwd DS; RAPM 20 x 20 min  Numerical updating  Bwd DS X Fwd DS X RAPM n.a. 97% (health problem) Zinke et al. (2010) DS fwd/ bcw; Corsi block tapping fwd/bwd and K=ABC icons (n = 20) PCG (n = 16) X RCPM; Stroop color-word interference 10 x 30 min n.a. X Stroop X RCPM n.a. n.c. Note. Aospan = Automated Operation Span task; AV WM = auditory-verbal working memory; CERAD = Consortium to Establish a Registry for Alzheimer's Disease; CWIT = Color-word Interference Task; CWMS = Categorization Working Memory Span; CVLT = California Verbal Learning Test; DFT3 = Design Fluency Test 3; DS = Digit Span; EPS = Everyday Problem Solving; KAB-Icons = Kaufmann Assessment Battery for Children; LPS = Leistungsprüfsystem; LPS-3 - Leistungsprüfsystem Subtest 3; LST = Listening Span Test; n.a. = Not applicable; n.c.= Not Clear; PCG = Passive Control Group; PASAT = Paced Auditory Serial Addition Test; RAPM = Raven’s Advanced Progressive Matrices; RCPM = Raven’s Colored Progressive Matrices; RSPM = Raven’s Standard Progressive Matrices; RSPM = Raven’s Standard Progressive Matrices; TAP = Test of attentional performance; TEA = Test of Everyday attention; TMT = Trail Making Test; TOVA = Test of variables of attention; TIADL = Timed Instrumental Activities of Daily Living; VLMT = Verbal Learning Memory Test; VS WM = visuospatial WM. * These studies also had an active control group. However, they also trained WM or executive function. Therefore, we chose to report data of a control group that was not submitted to a WMT. The symbol  represents a gain in the outcome while the symbol X represents a non-significant gain (except for the column computerized where =yes and X=no. App. Table 3 (cont.) Working memory training gains in healthy older adults Chapter II 61 App. Table 4 Description of the Trained Tasks Task Description CWMS Participants listen to lists of five-words containing common and animal words. They tap their hand on the table each time they hear an animal word, and they also have to memorize the last word of each list. In the end, participants are prompted to recall the memorized words in serial order. Chess game Participants had to listen to a sequence of chess movements and demonstrate, in a 5x5 chess board shown on a computer screen, the final position of the pieces in the correct sequence. Complex span task – spatial subtest or dot span Participants are asked to analyze the symmetry of partially filled matrices (or any other interference task) and, at the same time, to encode a sequence of locations on a grid for later recall. Corsi block-tapping/ span board task In a wooden board with nine blocks, first the experimenter taps a group of blocks in a specific order, then the participant needs to reproduce the same sequence in fwd or bwd order. The length of the sequence increases until the participant had two consecutive errors in the same level. Digit Span Participants listen to a sequence of digits and are instructed to recall them in the fwd or bwd order. The length of the sequence increases until participants have two errors in the same length. Dual n-back Participants perform two n-back tasks (visual and auditory) concomitantly. In other words, participants must decide whether the current visual stimulus matches the stimulus N positions back in a sequence presented one by one and, at the same time, they had to decide if the auditory stimulus also matches the one N positions back in the sequence. Figural task switching Participants categorize different geometrical shapes following two different rules in an alternate manner. The categorization rule cue is presented simultaneously with the stimuli. K-ABC Icons Visuospatial span task, in which participants need to memorize a spatial arrangement of multiple stimuli and then reproduce the icons and their positions on an empty grid after three seconds. Keep-track task 15 words from various semantic categories are presented in a random order, and participants are asked to mentally place the words into categories (animals, clothes, countries, relatives, sports, professions). In the end, participants are asked to remember the last presented word from each category. Running task or updating Lists containing stimuli in a specific order (letter, number, color, animal or spatial location) are presented one by one. The participants are asked to monitor and update the X (e.g., three) last presented stimuli during the list presentation. At the end of the list, participants need to recall them in the correct order. Lexical decision span task Participants must classify a set of letters presented one at a time. After a response of the participant, a single letter appears to be later remembered. Local-recognition task Participants had to recognize a sequence of stimuli (both color and position) and later recognize if the stimuli are in the correct position and color. Matrix task Series of 4 x 4 matrices with white and grey squares are presented. In each series, three black dots appear in different positions in the matrix, one after the other, separated by an empty matrix display. Participants press the space bar whenever a dot occupy a gray cell and, at the end of each series, participants have to recall the position of the last dot seen in an empty matrix. Memory update task Participants are presented with colored circles in a 4x4 grid. Later, the circles are presented one at a time and participant have to update its position mentally according to an arrow displayed on the screen (up, down, left or right). Movie-memory game Pairs of cards with the same image, but different numbers, are scrambled with only the number visible. Participants pick the two cards with the same image until no cards are left. Multi-memory game Participants memorize different tiles placed on a square grid. The tiles disappear and are replaced by a distractor pattern. Participants are then asked to recreate the original pattern. The grid size and tiles’ number are adjusted according to the performance. Multiple switching task Participants are required to shift attention between number-letter pairs that can be vocal/odd/black or consonant/even/green. A switching cue, such as a red bar moving from one instruction to another or a 10-second countdown, is given to the participants. Working memory training gains in healthy older adults Chapter II 62 N-back Participants must decide whether the current stimulus matches the stimulus N positions back in a sequence presented one by one. Numerical complex span A sequence of two-digit numbers followed by one digit is presented. Participants judge if the digit is odd or even (concurrent task) and, at the same time, memorize the two-digit number for later free recall in the correct serial order. Operation span - Aospan Participants must memorize a list of words while analyzing the veracity of solution of simple equations. Reading span Sequences of sentences are presented, and participants are required to make semantic judgments. After a group of sentences, participants recall the last word of each sentence. Semantic Category Span Participants make semantic category judgments for series of words. After a set of words, participants recall each word previously categorized in the same order of presentation. Tower of fame Participants are asked to imagine a tower with six floors, each one with four apartments. Then they listen sentences describing the location of a famous persons’ apartment in this tower. At the end of the task, participants are prompt to recall in which apartment a given famous person lives. WOME intervention This training focus on storage (remembering playing cards), selective attention (remembering only the hearts and the diamonds cards), and manipulation (sorting the cards in the same order as seen previously). Note. CWMS = Categorization Working Memory Span Task; Aospan: = Automated Operation Span Task. App. Table 5 Description of the Control Tasks Tasks Description Component control design Participants completed three tasks: semantic categorization; lexical decision; judgment of sentence acceptability. They do not have to recall anything after completing these tasks. Computer game Computer games not related to the training task. Counting Digits between one and six are presented on a screen. Digits should be repeated in accordance with the presented number. For example, five should appear five times in a row (5 5 5 5 5). If this rule is broken, participants press the wrong number on the keyboard. If all digits are correct, the participant should press zero. Implicit sequence leaning training Four squares are presented horizontally and aligned in the center of the screen. One stimulus is pink and the other three grays. The participants respond by pressing the key matching the position of the pink square as fast and accurately as possible. Mental health-related lectures Participants attended to mental health-related lecture sessions Visual search Participants identify, as fast as possible, a target stimulus and its position within a display of different stimuli. Not all trials have target items and, in this case, the participants should press another key. Questionnaires Participants complete questionnaires about different matters (e.g., autobiographic memory; memory sensitivity; psychological wellbeing, life satisfaction; emotional competencies; coping strategies). Quiz General knowledge quiz with multiple choice. Simple board game Participants played simple board games such as Ludo. App. Table 4 (cont.) Working memory training gains in healthy older adults Chapter II 63 App. Table 6 Outcome Constructs and Categories Used in the Analysis Construct Categories Definition (measures) Subcategories Reasoning Involves problem-solving (Tests/tasks: Cattel culture fair; figural reasoning; RAPM; RCPM; RSPM; the letter sets; LPS; matrix reasoning WAIS-III). Cattell; RSPM; RAPM; LPS Verbal WM WM measures using verbal stimuli (Tests/tasks: digit span bwd; categorization working memory span task recall; reading span; verbal complex span; WM updating; letter-number sequencing). Verbal bwd simple span; verbal complex span; updating. Verbal STM STM using verbal stimuli (Tests/Tasks: fwd digit span). Verbal fwd simple span Visuospatial WM WM measures with visuospatial stimuli (Tests/tasks: dot matrix task; spatial 2-back; the jigsaw puzzle test; bwd Corsi span; working memory updating performance; bwd span board from WMS-IV) Visuospatial bwd simple span; visuospatial updating Visuospatial STM STM measure using visuospatial stimuli (Tests/tasks: fwd Corsi task; figural short-term memory; fwd span board from WMS-R) Visuospatial fwd simple span Note. Bwd = backward; Fwd = forward; LPS = Leistungsprüfsystem; RAPM = Raven’s Advanced Progressive Matrices; RCPM = Raven’s Colored Progressive Matrices; RSPM = Raven’s Standard Progressive Matrices; STM = Short-term Memory; WAIS-III = Wechsler Adult Intelligence Scale - III; WM = Working memory; WMS = Wechsler Memory Scale; WMS-IV = Wechsler Memory Scale-IV. Working memory training gains in healthy older adults Chapter II 64 App. Table 7 Description of the Tasks Used to Assess Near and Far Transfer Tasks (cognitive domain) Description of the task NEAR TRANSFER Binding Sequences of stimuli (e.g., words or shapes) are displayed in different positions on the screen. The stimuli are again displayed, and the participants judge if they are in the same position or not. Brown-Peterson Participants have to memorize a series of Gabor patches, followed by a distractor task. CWMS See App. table 4 for a detailed description. Category fluency During a given time, participants are asked to say as many words as possible from a specific category. Computation span Participants solve arithmetic problems while holding the final digit from each problem in memory for later recall. Corsi block tapping/ span board task See App. table 4 for a detailed description. Corsi spatial span Du et al. (2018) modified the Corsi task by presenting a flag inside each block and requiring the participants to identify orientation of the flag. Subsequently, participants have to recall the location of the blocks. Digit Span See App. table 4 for a detailed description. Discourse memory Participants are required to read a whole paragraph and then to reproduce it as accurately as possible. D2 test Participants are asked to cross out any letter ‘d’ with two marks above or below it. The other presented stimuli are distractors and should not be marked. Digit symbol substitution Participants complete as fast and accurately as possible digit-symbol correspondences during a specific time, following a key provided in the top of the page. Dot matrix Participants see a sequence of dots displayed in different locations in a 5 × 5 grid interspersed with equations. At the end of the display, they should point, in a 5 × 5 blank grid, the exact spatial locations of the dots as presented previously. Keep track measure Participants see a list of words from four different categories, and they need to recall the most recent words from each category Kinship integration Descriptions of the kinship between two people are presented sequentially. Afterward, participants indicate the relationship between two people, which is not explicitly mentioned in the descriptions, but could be inferred from the presented information. Letter-number sequencing Participants listen to a mixed sequence of number and letters, and they are asked to recall the sequence items by ordering first the numbers and then the letters in the ascending order. Memory updating Participants have to memorize the orientation of arrays and then update their orientation by rotating them according to an arrow indicating the rotation direction. Minus-2 span After listening to a sequence of digits, participants are instructed to subtract two from each number and then produce the obtained sequence. N-back task See App. table 4 for a detailed description. Operation span - Aospan See App. table 4 for a detailed description. Reading span See App. table 4 for a detailed description. Recall of concrete nouns Participants must freely recall a list of 18 nouns read by the experimenter. There are two subsequent trials in which items not recalled are re-read. Participants are required to remember all the items on the list. Sentence span After listening to a series of sentences with a variable number of words, participants do a free recall test and retrieve the maximum number of sentences they can. Symbol span – WMS Participants are presented with a row of symbols. Then, they recognize these symbols in the correct order in a set of symbols. Swaps Three letters are presented simultaneously on the computer screen. Participants mentally swap the position of two letters as many times as requested. In the end, after conducting all swaps, participants type the three letters in the expected order. The Jigsaw puzzle Participants solve puzzles of inanimate objects. However, they must be done without moving the pieces. Instead, participants indicate where each piece should be moved. The level of difficulty is adapted by manipulating the number of puzzle pieces. Updating / running tasks A series of stimuli are presented, and participants are asked to recall the three last numbers presented in the correct serial order. Verbal complex span A sequence of words followed by one letter is presented to the participants. They evaluate if the letter is a consonant or a vowel and memorize the words for later free recall in the expected serial order. WMS Spatial addition Participants are presented with a grid with dots located in different positions in two separate pages. Participants add or subtract the locations of the dots, holding and manipulating visuospatial information. Working memory training gains in healthy older adults Chapter II 65 Word span Participants listen to a sequence of words, and they are required to recall them in the same serial order. FAR TRANSFER Attentional Blink A stream of letters and digits are displayed one by one on a screen. Participants identify a first visual digit followed by a second auditory digit. Berlin intelligence structure test Intelligence test covering the following domains: verbal; numeric; spatial; processing capacity; creativity; memory; speed. Block design Participants reproduce a model using red-and-white blocks in a given time. Cattell test It contains four subtests: 1) participants see an incomplete series of abstract shapes and figures, and they must choose the one that best fits the series from six options. 2) participants see 14 problems comprising abstract shapes and figures, and they must choose which two out of five differ from the other three. 3) 13 incomplete matrices containing four to nine boxes of abstract figures and shapes plus an empty box are presented. Participants must select the best fitting option from six possible alternatives; 4) 10 sets of abstract figures, lines, and a single dot are shown to the participants, along with five alternatives. Participants assess the relationship between the figures and choose the best fitting option. CERAD immediate and Delayed Recall Participants are asked to free recall ten words presented on printed cards. There are three learning trials in which the words are presented in different orders. Participants recall the learned items after each trial immediately and with a delayed interval of 15 minutes. CFQ A self-report questionnaire with 32 items that evaluates perception, memory, and motor function in everyday life. CVLT Participants are asked to free recall a 16-word list. This list is repeated five times or until participants recall all the words. After that, an intrusion list is presented to recall. Delayed recall is also assessed after 20 minutes. CWIT These are two subtests of the DKEFS, one that measures inhibition control (Stroop effect), and other that measures cognitive flexibility (switching between inhibitory and non-inhibitory responses). DF It is a subtest from the DKEFS in which participants are instructed to draw as many designs as possible by connecting five dots. Some dots are filled, while others are not. Participants must alternate between filled and unfilled dots. Digit symbol Participants are presented to nine numbers associated with symbols. Then they have to fulfill, as fast as possible, a sequence of number with their respective symbol. eKFA It is a 13-items questionnaire of daily cognitive failures. EPS Hypothetical real-life situations questionnaire. Flanker Participants indicate the orientation of a central arrow presented between two other arrows or two other stimuli. LST Participants listen to sets of simple sentences and judge its plausibility. Later, they recall the last word of each sentence. Letter and pattern comparison Two pages containing one column of 30 items are presented, and participants decide whether the arrangements are identical or not. LPS Participants are asked to analyse patterns of symbols and to mark the one that does not match the pattern. Locations Participants have to identify the rule of a spatial distribution of a x in four lines. Then they place a ‘x’ in a new line according to the rule previously identified. Matrix reasoning Participants select the best fitting response to complete a missed element in a matrix. Memory association Participants learn and remember the relationship between two unrelated items. Memory recognition Participants had to recognize previously learned items (meaningless-images and words) Number Cancellation Participants are presented with lines having different digits. They have to cross off, as fast as possible, two specific digits presented on the top of the form. Paired-associate learning Participants are asked to learn word pairs. After studying a list of 18 pairs, when given the first word of the pair as a cue, participants are asked to recall the corresponding word from the pair. Paper folding Participants have to do, mentally, a sequence of steps related to the action of folding a paper, and then recognize the pattern that corresponds to the outcome of the previous folding steps. Pattern comparison Participants check if two figures presented side-by-side are the same or not, as fast as possible. RAPM; RCPM; RSPM By choosing from six to eight options, participants are asked to complete a missing part from a figure composed of lines and shapes. Reading comprehension Participants read prose passages and then answer questions about it. Relationships Participants identify the correct Venn diagram representing the relationship among three stimuli. Shifting It is a shifting task, in which participants categorize a stimulus according to two different classification rules. Simon Participants indicate the color of a stimulus presented on the left, right or in the center of the screen by pressing an arrow key congruent, incongruent or neutral to the stimulus position. Simple reaction time Participants press a button box as quickly as possible when a cross is displayed in one of five positions on a black screen. Spatial relationship Reasoning test, in which participants have to rotate objects mentally. App. Table 7 (cont.) Working memory training gains in healthy older adults Chapter II 66 Stroop task It is a task that contains a congruent condition wherein colour names are presented in congruent ink colour and an incongruent condition in which colour names are printed in incongruent ink colour. A third control condition is presented, containing colour patches. Participants are instructed to say the ink colour of each stimulus. Switching Participants switch between different instructions, and the switch is prompted by specific cues. Symbol search In each trial, individuals are required to check if at least one of two symbols initially presented can be found amid a sequence of other symbols. If they can find at least one corresponding symbol, they draw a cross in yes, otherwise they signal the answer no. They have 120 seconds to complete the task. TAP alertness Participants see a series of stimulus and give a response when: a specific stimulus is presented or a specific cue precedes a stimulus. TAP mental flexibility Participants are simultaneously presented with a letter and a number on random sides of a screen. They have to press a key on the side where the letter or the number is displayed. TAP go-no-go Participants are randomly presented with two different stimuli, they have to press a key, as fast as possible, when the target is shown, while suppressing the response to the non-target stimuli. TEA Subtests that cover selective attention, sustained attention, and attentional switching. The letter sets Participants identify a set of letters that deviates from a pattern. TIAD Participants simulate the actions involved in different activities: communication; use of money; cooking; shopping; use of medicine. TMT Participants connect numbers or letters in ascending order or alternate between numbers and letters. In the DKEFS version, they had also to connect dots to assess motor speed. TOVA In this test participants have to respond to a target that is shown first in an infrequent condition and later in a more frequent condition. Tower Task (DKEFS) Participants build series of towers with five coloured disks using an apparatus with three-pegs varying in size. The aim is to complete a tower like the one presented in a picture model, using the least number of movements. Verbal fluency Letter fluency: participants are asked to produce as many words as possible beginning with the letters F, A, and S during 90 s, excluding proper names and places. Category fluency: participants are asked to say the maximum of words as possible about specific categories and by complying to specific constraints, such as provisions, animal names beginning with the letter S, and professions beginning with the letter B. VLMT Participants are asked to free recall a word list read by the experimenter. The list is repeated five times. After that, an intrusion list is presented to recall. Delayed recall, after 20 minutes, is also performed. In the end, a list containing all the words and distractors is shown to the participants, and they are required to recognize the words from the first list. Note. CFQ = Cognitive Failures Questionnaire; CERAD = Consortium to Establish a Registry for Alzheimer's Disease; CWIT = Colour-word interference task; CWMS = Categorization Working Memory Span Task; CVLT = California Verbal Learning Test; DFT = Design Fluency YRDY; DKEFS = DelisKaplan Executive Function System; eKFA = electronic Questionnaire for Cognitive Failures in Everyday Life; EPS = Everyday problem solving; RAPM = Raven’s Advanced Progressive Matrices; RSPM = Raven’s Standard Progressive Matrices; LSP = Leistungsprüfsystem; LST = Listening Span Test; TEA = Test of Everyday Attention; TIAD = Timed Instrumental Activities of Daily Living; TMT = Trail Making Test; TOVA = Test of Variables of Attention; VLMT = Verbal Learning Memory Test. App. Table 7 (cont.) Working memory training gains in healthy older adults Chapter II 67 App. Table 8 Influential Studies in Each Group (Divided by Measure) Construct Measure Influential Study Outcome Id Final No. of experiments Final No. Outcomes P O S T T E S T Reasoning Cattell No 8 8 RSPM 7 – Cantarella et al., 2016 19 – Richmond et al., 2011 8 28 4 4 RAPM 13 – Guye & von Bastian, 2017 17 5 5 LPS NA 3 3 Others 6 14 Total 13 – Guye & von Bastian, 2017 18 25 33 Verbal WM Bwd Simple span 1 – Borella et al., 2010 38 9 9 Complex span 3 – Borella et al., 2014 – Exp1 40 13 16 Updating 25 – Xin et al., 2014 80, 81 7 11 Others 3 4 Total 20 41 Visuospatial WM Bwd Simple span 3 – Borella et al., 2014 – Exp1 84 4 4 Updating 11 – Du et al., 2018 13 – Guye & von Bastian, 2017 90 94 4 4 Others 4 5 Total 1 – Borella et al., 2010 82 10 13 Verbal STM Simple Span 1 – Borella et al., 2010 99 11 12 Visuospatial STM Simple span 3 – Borella et al., 2014 – Exp1 112 5 6 F O L L O W - U P Reasoning Cattell 11 – Du et al., 2018 128 6 6 Others 4 6 Total 10 12 Verbal WM Bwd Simple span 24 – Weicker et al., 2018 146 4 4 Complex span 3 – Borella et al., 2014 – Exp1 134 7 7 Updating NA 3 6 Total 9 17 Visuospatial WM Bwd Simple span 3 – Borella et al., 2014 – Exp1 153 5 5 Updating NA 2 3 Others 3 3 Total 8 11 Verbal STM Simple span NA 6 6 Visuospatial STM Simple span NA 5 6 TOTAL 27 156 Note. NA – Not applicable due to the existence of a small number ( <4 ) of experiments in the group (measure). Influential studies indicate the studies that are outliers and have strong influence in the results. These studies were excluded from the analysis. Chapter II Working memory training gains in healthy older adults 74 App. Figure 3. Risk of bias summary graph. Rows correspond to the articles included in the metaanalysis, and each column corresponds to the risk of bias assessment items. Each circle represents the risk of bias assessment a particular parameter in each study: +low risk; - high risk; ?unclear. Working memory training gains in healthy older adults C hapter II 75 References Au, J., Sheehan, E., Tsai, N., Duncan, G. J., Buschkuehl, M., & Jaeggi, S. M. (2015). Improving fluid intelligence with training on working memory: A meta-analysis. Psychonomic Bulletin & Review, 22 (2), 366–377. doi: 10.3758/s13423-014-0699-x Bäckman, L., Lindenberger, U., Li, S. C., & Nyberg, L. (2010). Linking cognitive aging to alterations in dopamine neurotransmitter functioning: Recent data and future avenues. Neuroscience and Biobehavioral Reviews, 34 (5), 670-677. doi: 10.1016/j.neubiorev.2009.12.008 Bäckman, L., Waris, O., Johansson, J., Andersson, M., Rinne, J. O., Alakurtti, K., …Nyberg, L. (2017). Increased dopamine release after working-memory updating training: Neurochemical correlates of transfer. Scientific Reports, 7 , 7160. doi: 10.1038/s41598-017-07577-y Ball, K., Berch, D. B., Helmers, K. F., Jobe, J. B., Leveck, M. D., Marsiske, M., …Willis, S. L. (2002). Effects of cognitive training interventions with older adults: a randomized controlled trial. The Journal of the American Medical Association, 288 (18), 2271–2281. doi: 10.1001/jama.288.18.2271. Basak, C., & O’Connell, M. A. (2016). To switch or not to switch: Role of cognitive control in working memory training in older adults. Frontiers in Psychology, 7 , 230. doi: 10.3389/fpsyg.2016.00230 Beatty, E. L., & Vartanian, O. (2015). The prospects of working memory training for improving deductive reasoning. Frontiers in Human Neuroscience, 9 , 56. doi: 10.3389/fnhum.2015.00056 Becker, B. J. (1988). Synthesizing standardized mean-change measures. British Journal of Mathematical and Statistical Psychology, 41 (2), 257-278. doi: 10.1111/j.20448317.1988.tb00901.x Bisiacchi, P. S., Tarantino, V., & Ciccola, A. (2008). Aging and prospective memory: The role of working memory and monitoring processes. Aging Clinical and Experimental Research, 20 (6), 569-577. doi: 10.1007/BF03324886 Borella, E., Cantarella, A., Carretti, B., De Lucia, A. & , De Beni, R. (2019). Improving everyday functioning in the old-old with a working memory training. The American Journal of Geriatric Psychiatry . Advance online publication. doi: 10.1016/j.jagp.2019.01.210 Borella, E., Carbone, E., Pastore, M., De Beni, R., & Carretti, B. (2017). Working memory training for healthy older adults: The role of individual characteristics in explaining shortand long-term Chapter II Working memory training gains in healthy older adults 76 gains. Frontiers in Human Neuroscience, 11 , 99. doi: 10.3389/fnhum.2017.00099 Borella, E., Carretti, B., Cantarella, A., Riboldi, F., Zavagnin, M., & De Beni, R. (2014). Benefits of training visuospatial working memory in young–old and old–old. Developmental Psychology , 50 (3), 714–727. doi: 10.1037/a0034293 Borella, E., Carretti, B., Riboldi, F., & De Beni, R. (2010). Working memory training in older adults: Evidence of transfer and maintenance effects. Psychology and Aging , 25 (4), 767–778. doi: 10.1037/a0020683 Borella, E., Carretti, B., Sciore, R., Capotosto, E., Taconnat, L., Cornoldi, C., & De Beni, R. (2017). Training working memory in older adults: Is there an advantage of using strategies? Psychology and Aging , 32 (2), 178–191. doi: 10.1037/pag0000155 Borella, E., Carretti, B., Zanoni, G., Zavagnin, M., & De Beni, R. (2013). Working memory training in old age: An examination of transfer and maintenance effects. Archives of Clinical Neuropsychology , 28 (4), 331–347. doi: 10.1093/arclin/act020 Borenstein, M. (2009). Effect sizes for continuous data. In H. Cooper, L. V. Hodges, & J. C. Valentine (Eds.), The handbook of research synthesis and meta-analysis (pp. 221–235). New York, NY, US: Russell Sage Foundation. Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Effect sizes based on means. In Introduction to meta-analysis (pp. 21–32). Chichester: Wiley. doi:10.1002/9780470743386.ch4 Braver, T. S., Cohen, J. D., Nystrom, L. E., Jonides, J., Smith, E. E., & Noll, D. C. (1997). A parametric study of prefrontal cortex involvement in human working memory. NeuroImage, 5 (1), 49-62. doi: 10.1006/nimg.1996.0247 Braver, T. S., & West, R. (2008). Working memory, executive control, and aging. In F. I. M. Craik & T. A. Salthouse (Eds.), The handbook of aging and cognition (pp. 311-372). New York, NY, US: Psychology Press. Brehmer, Y., Rieckmann, A., Bellander, M., Westerberg, H., Fischer, H., & Bäckman, L. (2011). Neural correlates of training-related working-memory gains in old age. NeuroImage , 58 (4), 1110–1120. doi: 10.1016/j.neuroimage.2011.06.079 Brucki, S. M. D. (2010). Illiteracy and dementia. Dementia & Neuropsychologia , 4 (3), 153–157. doi: 10.1590/S1980-57642010DN40300002 Working memory training gains in healthy older adults C hapter II 77 Bürki, C. N., Ludwig, C., Chicherio, C., & de Ribaupierre, A. (2014). Individual differences in cognitive plasticity: An investigation of training curves in younger and older adults. Psychological Research , 78 (6), 821–835. doi: 10.1007/s00426-014-0559-3 Buschkuehl, M., Jaeggi, S. M., & Jonides, J. (2012). Neuronal effects following working memory training. Developmental Cognitive Neuroscience, 2 (S1), S167-S179. doi: 10.1016/j.dcn.2011.10.001 Cantarella, A., Borella, E., Carretti, B., Kliegel, M., & De Beni, R. (2017). Benefits in tasks related to everyday life competences after a working memory training in older adults. International Journal of Geriatric Psychiatry, 32 (1), 86-93. doi: 10.1002/gps.4448 Cantarella, A., Borella, E., Carretti, B., Kliegel, M., Mammarella, N., Fairfield, B., & De Beni, R. (2017). The influence of training task stimuli on transfer effects of working memory training in aging. Psychologie Française . Advance online publication. doi: 10.1016/j.psfr.2017.04.005 Chan, J. S. Y., Wu, Q., Liang, D., & Yan, J. H. (2015). Visuospatial working memory training facilitates visually-aided explicit sequence learning. Acta Psychologica, 161 , 145-153. doi: 10.1016/j.actpsy.2015.09.008 Clare, L., & Woods, R. T. (2004). Cognitive training and cognitive rehabilitation for people with earlystage Alzheimer’s disease: A review. Neuropsychological Rehabilitation, 14 (4), 385-401. doi: 10.1080/09602010443000074 Constantinidis, C., & Klingberg, T. (2016). The neuroscience of working memory capacity and training. Nature Reviews Neuroscience, 17 (7), 438-449. doi: 10.1038/nrn.2016.43 Conway, A. R. A., Kane, M. J., & Engle, R. W. (2003). Working memory capacity and its relation to general intelligence. Trends in Cognitive Sciences , 7 (12), 547–552. doi: 10.1016/j.tics.2003.10.005 Cowan, N. (2017). The many faces of working memory and short-term storage. Psychonomic Bulletin and Review, 24 (4), 1158-1170. doi: 10.3758/s13423-016-1191-6 da Silva, H. S., & Yassuda, M. S. (2009). Memory training for older adults with low education: Mental images versus categorization. Educational Gerontology , 35 (10), 890–905. doi: 10.1080/03601270902782487 Dahlin, E., Nyberg, L., Bäckman, L., & Neely, A. S. (2008). Plasticity of executive functioning in young and older adults: Immediate training gains, transfer, and long-term maintenance. Psychology and Aging , 23 (4), 720–730. doi: 10.1037/a0014296 Chapter II Working memory training gains in healthy older adults 78 Dahlin, E., Neely, A. S., Bäckman, L., & Larsson, A. (2008). Transfer of learning after updating training mediated by the striatum. Science , 320 (5882), 1510–1512. doi: 10.1126/science.1155466 Delphin-Combe, F., Bathsavanis, A., Rouch, I., Liles, T., Vannier-Nitenberg, C., Fantino, B., …KrolakSalmon, P. (2016). Relationship between anxiety and cognitive performance in an elderly population with a cognitive complaint. European Journal of Neurology , 23 (7), 1210–1217. doi: 10.1111/ene.13004 Diamond, B. J., Deluca, J., Rosenthal, D., Vlad, R., Davis, K., Lucas, G., …Richards, J. A. (1999). Information processing in older versus younger adults: Accuracy versus speed. International Journal of Rehabilitation and Health, 5 (1), 55-64. doi: 10.1023/A:1012911203468 Duval, S., & Tweedie, R. (2000a). A nonparametric “trim and fill” method of accounting for publication bias in meta-analysis. Journal of the American Statistical Association , 95 (449), 89–98. doi: 10.1080/01621459.2000.10473905 Duval, S., & Tweedie, R. (2000b). Trim and fill: A simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics , 56 (2), 455–463. doi: 10.1111/j.0006-341X.2000.00455.x Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in metaanalysis detected by a simple, graphical test. BMJ, 315 (7109), 629-634. doi: 10.1136/bmj.315.7109.629 Gignac, G. E. (2015). Raven’s is not a pure measure of general intelligence: Implications for g factor theory and the brief measurement of g . Intelligence , 52 , 71–79. doi: 10.1016/j.intell.2015.07.006 Goghari, V. M., & Lawlor-Savage, L. (2017). Comparison of cognitive change after working memory training and logic and planning training in healthy older adults. Frontiers in Aging Neuroscience , 9 (2), 39. doi: 10.3389/fnagi.2017.00039 Golino, M. T. S., & Flores-Mendoza, C. E. (2016). Development of a cognitive training program for the elderly. Revista Brasileira de Geriatria e Gerontologia , 19 (5), 769–785. doi: 10.1590/180998232016019.150144 Gordon, M., & Lumley, T. (2016). forestplot: Advanced forest plot using “grid” Graphics. Retrieved from https://cran.r-project.org/web/packages/forestplot/index.html Guolo, A., & Varin, C. (2012). The R package \pkgmetaLik for likelihood inference in meta-analysis. Journal of Statistical Software , 50 (7), 1–19. doi: 10.18637/jss.v050.i07 Working memory training gains in healthy older adults C hapter II 79 Guye, S., & von Bastian, C. C. (2017). Working memory training in older adults: Bayesian evidence supporting the absence of transfer. Psychology and Aging, 32 (8), 732-746. doi: 10.1037/pag0000206 Harbord, R. M. (2011). Commentary on 'Multivariate meta-analysis: Potential and promise'. Statistics in Medicine, 30 (20), 2507-2508. doi: 10.1002/sim.4278 Hedges, L. V. (1989). An unbiased correction for sampling error in validity generalization studies. Journal of Applied Psychology , 74 (3), 469–477. doi: 10.1037/0021-9010.74.3.469 Hedges, L. V., Tipton, E., & Johnson, M. C. (2010). Robust variance estimation in meta-regression with dependent effect size estimates. Research Synthesis Methods , 1 (1), 39–65. doi: 10.1002/jrsm.5 Heinzel, S., Lorenz, R. C., Pelz, P., Heinz, A., Walter, H., Kathmann, N., …Stelzel, C. (2016). Neural correlates of training and transfer effects in working memory in older adults. NeuroImage , 134 , 236–249. doi: 10.1016/j.neuroimage.2016.03.068 Heinzel, S., Schulte, S., Onken, J., Duong, Q. L., Riemer, T. G., Heinz, A., …Rapp, M. A. (2013). Working memory training improvements and gains in non-trained cognitive tasks in young and older adults. Neuropsychology, Development, and Cognition. Section B, Aging, Neuropsychology and Cognition , 21 (2), 146–173. doi: 10.1080/13825585.2013.790338 Henmi, M., & Copas, J. B. (2010). Confidence intervals for random effects meta-analysis and robustness to publication bias. Statistics in Medicine , 29 (29), 2969–2983. doi: 10.1002/sim.4029 Higgins, J. P. T., & Altman, D. G. (2008). Assessing risk of bias in included studies. In J. P. T. Higgins, J.P.T. & Green, S. (Eds.), Cochrane handbook for systematic reviews of interventions: Cochrane book series (pp. 187–241). Chichester, UK: John Wiley & Sons. doi: 10.1002/9780470712184.ch8 Higgins, J. P. T., & Green, S. (2008). Cochrane handbook for systematic reviews of interventions: Cochrane book series . Chichester, UK: John Wiley & Sons. Higgins, J. P. T., Thompson, S. G., Deeks, J. J., & Altman, D. G. (2003). Measuring inconsistency in meta-analyses. BMJ , 327 (7414), 557–560. doi: 10.1136/bmj.327.7414.557 Huitfeldt, B., Danielson, L., Ebbutt, A., & Schmidt, K. (2001). Choice of control in clinical trials - Issues and implications of ICH-E10. Therapeutic Innovation and Regulatory Science , 35 (4), 1147– 1156. doi: 10.1177/009286150103500411 Chapter II Working memory training gains in healthy older adults 80 ICH Harmonised Tripartite Guideline. (2000). Choice of control group and related issues in clinical trials E10. Retrieved from https://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Efficacy/E10/Ste p4/E10_Guideline.pdf Jaeggi, S. M., Buschkuehl, M., Jonides, J., & Perrig, W. J. (2008). Improving fluid intelligence with training on working memory. Proceedings of the National Academy of Sciences , 105 (19), 6829– 6833. doi: 10.1073/pnas.0801268105 Jaeggi, S. M., Buschkuehl, M., Shah, P., & Jonides, J. (2014). The role of individual differences in cognitive training and transfer. Memory & Cognition , 42 (3), 464–480. doi: 10.3758/s13421013-0364-z Jolles, D. D., & Crone, E. A. (2012). Training the developing brain: A neurocognitive perspective. Frontiers in Human Neuroscience , 6 , 76. doi: 10.3389/fnhum.2012.00076 Just, M. A., & Carpenter, P. A. (1992). A capacity theory of comprehension: Individual differences in working memory. Psychological Review , 99 (1), 122–149. doi: 10.1037/0033-295X.99.1.122 Karbach, J., & Verhaeghen, P. (2014). Making working memory work: A meta-analysis of executivecontrol and working memory training in older adults. Psychological Science , 25 (11), 2027– 2037. doi: 10.1177/0956797614548725 Kawagoe, T., Suzuki, M., Nishiguchi, S., Abe, N., Otsuka, Y., Nakai, R., … Sekiyama, K. (2015). Brain activation during visual working memory correlates with behavioral mobility performance in older adults. Frontiers in Aging Neuroscience , 7 , 186. https://doi.org/10.3389/fnagi.2015.00186 Kelly, M. E., Loughrey, D., Lawlor, B. A., Robertson, I. H., Walsh, C., & Brennan, S. (2014). The impact of cognitive training and mental stimulation on cognitive and everyday functioning of healthy older adults: A systematic review and meta-analysis. Ageing Research Reviews , 15 (1), 28-43. doi: 10.1016/j.arr.2014.02.004 Kemper, S., Herman, R. E., & Liu, C. J. (2004). Sentence production by young and older adults in controlled contexts. The Journals of Gerontology Series B Psychological Sciences and Social Sciences, 59 (5), 220-224. doi: 10.1093/geronb/59.5.P220 Ko, P. C., Duda, B., Hussey, E., Mason, E., Molitor, R. J., Woodman, G. F., & Ally, B. A. (2014). Understanding age-related reductions in visual working memory capacity: Examining the stages Working memory training gains in healthy older adults C hapter II 81 of change detection. Attention, Perception, and Psychophysics, 76 (7), 2015-2030. doi: 10.3758/s13414-013-0585-z Laguna, K., & Babcock, R. L. (1997). Computer anxiety in young and older adults: Implications for human-computer interactions in older populations. Computers in Human Behavior, 13 (3), 317326. doi: 10.1016/S0747-5632(97)00012-5 Lampit, A., Hallock, H., & Valenzuela, M. (2014). Computerized Cognitive training in cognitively healthy older adults: A systematic review and meta-analysis of effect modifiers. PLoS Medicine , 11 (11), e1001756. doi: 10.1371/journal.pmed.1001756 Landi, D., & Rossini, P. M. (2010). Cerebral restorative plasticity from normal ageing to brain diseases: A “never ending story.” Restorative Neurology and Neuroscience, 28 (3), 349-366. doi: 10.3233/RNN-2010-0538 Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data data for categorical of observer agreement the measurement. Biometrics , 33 (1), 159–174. doi: 10.2307/2529310 Lange, S., & Süß, H. M. (2015). Experimental evaluation of nearand far-transfer effects of an adaptive multicomponent Working memory training. Applied Cognitive Psychology , 29 (4), 502–514. doi: 10.1002/acp.3126 Li, S. C., Schmiedek, F., Huxhold, O., Röcke, C., Smith, J., & Lindenberger, U. (2008). Working memory plasticity in old age: Practice gain, transfer, and maintenance. Psychology and Aging , 23 (4), 731–742. doi: 10.1037/a0014343 Lilienthal, L., Tamez, E., Shelton, J. T., Myerson, J., & Hale, S. (2013). Dual n-back training increases the capacity of the focus of attention. Psychonomic Bulletin & Review , 20 (1), 135–141. doi: 10.3758/s13423-012-0335-6 Loosli, S. V., Falquez, R., Unterrainer, J. M., Weiller, C., Rahm, B., & Kaller, C. P. (2016). Training of resistance to proactive interference and working memory in older adults: A randomized doubleblind study. International Psychogeriatrics, 28 (3), 453-467. doi: 10.1017/S1041610215001519 Lubitz, A. F., Niedeggen, M., & Feser, M. (2017). Aging and working memory performance: Electrophysiological correlates of high and low performing elderly. Neuropsychologia , 106 , 42– 51. doi: 10.1016/j.neuropsychologia.2017.09.002 Chapter II Working memory training gains in healthy older adults 82 Lustig, C., Shah, P., Seidler, R., & Reuter-Lorenz, P. A. (2009). Aging, training, and the brain: A review and future directions. Neuropsychology Review, 19 (4), 504-522. doi: 10.1007/s11065-0099119-9 Mansur-Alves, M., & Silva, R. S. (2017). Treinar memória de trabalho promove mudanças em inteligência fluida? Temas Em Psicologia , 25 (2), 787–807. doi: 10.9788/TP2017.2-19Pt Melby-Lervåg, M., & Hulme, C. (2013). Is working memory training effective? A meta-analytic review. Developmental Psychology , 49 (2), 270–91. doi: 10.1037/a0028228 Melby-Lervåg, M., & Hulme, C. (2016). There is no convincing evidence that working memory training is effective: A reply to Au et al. (2014) and Karbach and Verhaeghen (2014). Psychonomic Bulletin & Review , 23 (1), 324–330. doi: 10.3758/s13423-015-0862-z Melby-Lervåg, M., Redick, T. S., & Hulme, C. (2016). Working memory training does not improve performance on measures of intelligence or other measures of “far transfer”: Evidence from a meta-analytic review. Perspectives on Psychological Science , 11 (4), 512–534. doi: 10.3837/tiis.0000.00.000 Mlinarić, A., Horvat, M., & Smolčić, V. Š. (2017). Dealing with the positive publication bias: Why you should really publish your negative results. Biochemia Medica, 27 (3), 030201. doi: 10.11613/BM.2017.030201 Moeyaert, M., Ugille, M., Natasha Beretvas, S., Ferron, J., Bunuan, R., & Van den Noortgate, W. (2017). Methods for dealing with multiple outcomes in meta-analysis: A comparison between averaging effect sizes, robust variance estimation and multilevel meta-analysis. International Journal of Social Research Methodology , 20 (6), 559–572. doi: 10.1080/13645579.2016.1252189 Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & PRISMA Group. (2009). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. BMJ , 339 , b2535. doi: 10.1136/bmj.b2535 Morris, S. B. (2008). Estimating effect sizes from pretest-posttest-control group designs. Organizational Research Methods , 11 (2), 364–386. doi: 10.1177/1094428106291059 Morrison, A. B., & Chein, J. M. (2011). Does working memory training work? The promise and challenges of enhancing cognition by training working memory. Psychonomic Bulletin & Review , 18 (1), 46– 60. doi: 10.3758/s13423-010-0034-0 Nittrouer, S., Lowenstein, J. H., Wucinich, T., & Moberly, A. C. (2016). Verbal working memory in older Working memory training gains in healthy older adults C hapter II 83 adults: The roles of phonological capacities and processing speed. Journal of Speech, Language, and Hearing Research, 59 (6), 1520-1532. doi: 10.1044/2016_jslhr-h-15-0404 Noack, H., Lövdén, M., Schmiedek, F., & Lindenberger, U. (2009). Cognitive plasticity in adulthood and old age: Gauging the generality of cognitive intervention effects. Restorative Neurology and Neuroscience , 27 (5), 435–453. doi: 10.3233/RNN-2009-0496 Oberauer, K., Lewandowsky, S., Awh, E., Brown, G. D. A., Conway, A., Cowan, N., … Ward, G. (2018). Benchmarks for models of short-term and working memory. Psychological Bulletin, 144 (9), 885958. doi: 10.1037/bul0000153 Oberauer, K., Süß, H. M., Wilhelm, O., & Wittmann, W. W. (2008). Which working memory functions predict intelligence? Intelligence , 36 (6), 641–652. doi: 10.1016/j.intell.2008.01.007 Olson, I. R., Zhang, J. X., Mitchell, K. J., Johnson, M. K., Bloise, S. M., & Higgins, J. A. (2004). Preserved spatial memory over brief intervals in older adults. Psychology and Aging, 19 (2), 310-317. doi: 10.1037/0882-7974.19.2.310 Oswald, W. D., Gunzelmann, T., Rupprecht, R., & Hagen, B. (2006). Differential effects of single versus combined cognitive and physical training with older adults: The SimA study in a 5-year perspective. European Journal of Ageing , 3 (4), 179–192. doi: 10.1007/s10433-006-0035-z Park, D. C., & Reuter-Lorenz, P. (2009). The adaptive brain: Aging and neurocognitive scaffolding. Annual Review of Psychology , 60 (1), 173–196. doi: 10.1146/annurev.psych.59.103006.093656 Payne, B. R. (2014). The effects of verbal working memory training on language comprehension in older adulthood (Doctoral dissertation) . Retrieved from https://www.ideals.illinois.edu/handle/2142/50700. Penner, I.-K., Vogt, A., Stöcklin, M., Gschwind, L., Opwis, K., & Calabrese, P. (2012). Computerised working memory training in healthy adults: A comparison of two different training schedules. Neuropsychological Rehabilitation , 22 (5), 716–733. doi: 10.1080/09602011.2012.686883 Pergher, V., Wittevrongel, B., Tournoy, J., Schoenmakers, B., & Van Hulle, M. M. (2018). N-back training and transfer effects revealed by behavioral responses and EEG. Brain and Behavior, 8 (11), e01136. doi: 10.1002/brb3.1136 Pustejovsky, J. (2017). Package “clubSandwich”: Cluster-robust (sandwich) variance estimators with small-sample corrections. Retrieved from https://rdrr.io/cran/clubSandwich/ APPENDIX A Supplementary material (study I) 90 Table C Sensitive Analysis - Posttest Construct r ρ RVE mean Q-test   Estimate 95% CI p -value RVE p-value Reasoning .3 .3 .097 [-.041, .234] .1676 .165 21.19 <.01 <01 .5 .098 [-.043, .239] .1734 .160 21.29 <.01 <.01 .7 .097 [-.047, .242] .1856 .165 22.51 <.01 <.01 .5 .3 .105 [-.023, .234] .1087 .125 28.40 .01 <.01 .5 .103 [-.026, .233] .1175 .134 28.53 .01 <.01 .7 .100 [-.030, .231] .1316 .148 30.18 .01 <.01 .7 .3 .119 [-.006, .245] .0627 ^ .0766 ^ 43.27 ^ .04 <.01 .5 .118 [-.009, .245] .0694 ^ .0834 ^ 43.46 ^ .04 <.01 .7 .114 [-.014, .242] .0816 ^ .0958 ^ 46.02 ^ .04 <.01 Verbal WM .3 .3 .223 [.065, .380] .0055 ** .0115 * 55.06 * <.01 .05 .5 .233 [.058, .408] .0090 ** .0078 ** 66.26 ** <.01 .07 .7 .240 [.049, .432] .0137 * .0062 ** 96.23 *** <.01 .10 .5 .3 .223 [.069, .377] .0045 ** .0122 * 73.77 *** <.01 .08 .5 .228 [.065, .392] .0062 ** .0095 ** 88.79 *** <.01 .09 .7 .235 [.059, .411] .0090 ** .0075 ** 128.94 *** <.01 .11 .7 .3 .228 [.072, .384] .0042 ** .0114 * 112.21 *** .02 .10 .5 .225 [.070, .379] .0045 ** .0119 * 135.06 *** <.01 .12 .7 .227 [.067, .388] .0055 ** .0103 * 196.02 *** <.01 .13 Visuospatial WM .3 .3 .223 [.016, .431] .0345 * .0879 ^ 11.29 <.01 <.01 .5 .214 [.008, .419] .0414 * .0929 ^ 12.02 <.01 <.01 .7 .220 [-.006, .445] .0564 * .0725 ^ 14.55 <.01 .02 .5 .3 .228 [.035, .441] .0215 * .0564 ^ 15.10 .02 <.01 .5 .228 [.029, .426] .0245 * .0619 ^ 16.03 .02 <.01 .7 .232 [.013, .450] .0379 * .0592 ^ 19.33 <.01 .04 .7 .3 .247 [.047, .446] .0154 * .0407 * 22.85 * .04 <.01 .5 .247 [.044, .450] .0173 * .0463 * 24.16 * .01 .05 .7 .245 [.035, .454] .0223 * .0488 * 28.88 ** <.01 .07 Verbal STM .3 .3 .151 [-.070, .371] .1798 .181 8.84 <.01 <.01 .5 .153 [-.070, .376] .1797 .184 9.03 <.01 <.01 .7 .155 [-.072, .382] .1801 .185 9.50 <.01 <.01 .5 .3 .159 [-.045, .362] .1257 .161 12.15 .02 <.01 .5 .159 [-.045, .363] .1258 .161 12.41 <.01 .01 .7 .162 [-.046, .369] .1269 .158 13.06 <.01 .02 .7 .3 .179 [-.024, .382] .0846 ^ .117 19.41 ^ .04 .01 .5 .179 [-.024, .382] .0846 ^ .117 19.85 ^ .03 .02 .7 .179 [-.024, .382] .0846 ^ .117 20.92 * .02 .03 Visuospatial STM .3 .3 -.030 [-.366, .306] .8596 .746 5.22 <.01 .02 .5 -.030 [-.410, .351] .8782 .750 6.76 <.01 .06 .7 -.029 [-.448, .389] .8912 .755 10.46 ^ <.01 .10 .5 .3 -.033 [-.354, .288] .8408 .729 6.99 <.01 .05 .5 -.032 [-.388, .324] .8619 .739 9.06 <.01 .08 .7 -.031 [-.417, .356] .8764 .746 14.00 * <.01 .11 .7 .3 -.036 [-.341, .269] .8181 .713 10.61 ^ <.01 .07 .5 -.034 [-.363, .295] .8391 .725 13.73 * <.01 .09 .7 -.033 [-.383, .318] .8543 .733 21.28 *** <.01 .11 Note . ^p<.1, *p<.05, **p<.01, ***p<.001. NA = Not Applicable (only for groups from the same measure); I2 – total heterogeneity / total variability;  – estimated amount of total heterogeneity;   – Variance component of the 3-level model for the between-studies heterogeneity;   – Variance component of the 3-level model for the within-studies heterogeneity; RVE = Robust Variance Estimation; WM = Working Memory. This table represents the sensitivity analysis performed with three different correlational values ( r = 0.3, 0.5, 0.7 and ρ=0.3, 0.5, 0.7) due to the fact that correlations between pre-and post-test scores and between-studies were not reported in the original studies. r is the pre-posttest correlation and ρ is the intrastudy measures correlation. Results are consistent between the different correlations. Supplementary material (study I) APPENDIX A 91 Table C (cont.) Sensitive Analysis - Follow-up Construct r rho RVE mean Q-test τ2   Estimate 95% CI p-value Skovgaard’s p-value RVE p-value Reasoning -3 .3 .113 [-.129, .355] .3612 .383 7.54 <.01 <.01 .5 .130 [-.115, .376] .2979 .276 6.90 <.01 <.01 .7 .143 [-.105, .391] .2583 .210 6.54 <.01 <.01 .5 .3 .121 [-.108, .349] .2997 .331 10.22 .02 <.01 .5 .131 [-.085, .347] .2355 .268 9.36 .01 <.01 .7 .142 [-.070, .353] .1891 .209 8.86 <.01 <.01 .7 .3 .131 [-.086, .349] .2373 .270 15.92 .05 <.01 .5 .138 [-.072, .347] .1972 .230 11.58 .04 <.01 .7 .145 [-.057, .346] .1596 .194 13.76 .03 <.01 Verbal WM .3 .3 .209 [-.022, .439] .0760 ^ .105 12.80 <.01 <.01 .5 .232 [-.012, .477] .0626 ^ .083 ^ 13.97 <.01 <.01 .7 .248 [-.015, .510] .0640 ^ .073 ^ 17.58 * <.01 .01 .5 .3 .218 [-.005, .440] .0553 ^ .092 ^ 16.99 .02 <.01 .5 .231 [.006, .457] .0446 * .079 ^ 18.59 .01 .01 .7 .237 [.001, .476] .0491 * .075 ^ 23.46 <.01 .03 .7 .3 .225 [.002, .448] .0481 * .083 ^ 25.39 ^ .05 .01 .5 .226 [.004, .448] .0461 * .081 ^ 27.89 * .03 .03 .7 .227 [.006, .449] .0443 * .079 ^ 35.43 ** .01 .04 Visuospatial WM .3 .3 .140 [-.012, .400] .2915 .124 4.01 <.01 <.01 .5 .142 [-.126, .409] .2990 .134 4.48 <.01 <.01 .7 .140 [-.133, .413] .3147 .152 5.61 <.01 <.01 .5 .3 .139 [-.084, .361] .2228 .129 5.43 <.01 <.01 .5 .140 [-.089, .368] .2313 .140 6.04 <.01 <.01 .7 .137 [-.096, .370] .2500 .137 7.53 <.01 <.01 .7 .3 .136 [-.041, .313] .1332 .137 8.43 <.01 <.01 .5 .136 [-.046, .317] .1423 .151 9.33 <.01 <.01 .7 .136 [-.064, .335] .1832 .160 11.51 <.01 .01 Verbal STM .3 - .178 [-.145, .501] .2804 .2102 .187 2.79 <.01 .5 - .178 [-.097, .452] .2048 .9828 .186 3.85 <.01 .7 - .176 [-.053, .405] .1317 .2476 .189 6.20 .01 Visuospatial STM .3 .3 -.048 [-.380, .284] .7787 .689 2.20 <.01 <.01 .5 -.045 [-.383, .293] .7926 .711 2.32 <.01 <.01 .7 -.045 [-.387, .298] .7977 .721 2.62 <.01 <.01 .5 .3 -.046 [-.331, .238] .7488 .698 3.01 <.01 <.01 .5 -.044 [-.334, .245] .7632 .718 3.17 <.01 <.01 .7 -.045 [-.338, .249] .7661 .726 3.57 <.01 <.01 .7 .3 -.044 [-.270, .183] .7053 .719 4.75 <.01 <.01 .5 -.043 [-.276, .191] .7211 .733 4.99 <.01 <.01 .7 -.044 [-.292, .205] .7319 .732 5.58 <.01 .01 Note . ^p<.1, *p<.05, **p<.01, ***p<.001. NA = Not Applicable (only for groups from the same measure); I2 – total heterogeneity / total variability;  – estimated amount of total heterogeneity;   – Variance component of the 3-level model for the between-studies heterogeneity;  – Variance component of the 3-level model for the within-studies heterogeneity; RVE = Robust Variance Estimation; WM = Working Memory. This table represents the sensitivity analysis performed with three different correlational values ( r = 0.3, 0.5, 0.7) since correlations between pre and posttest scores and between-studies were not reported in the original studies. r is the pre-posttest correlation and ρ is the between-study correlation. Results are consistent between the different correlations. APPENDIX A Supplementary material (study I) 92 Table D Sensitive Analysis for Control Groups Estimate 95% CI pvalue P O S T T E S T Reasoning Merged Active control Passive control 0.10 0.14 0.08 [-0.03,0.23] [-0.04, 0.33] [-0.11,0.27] .118 .147 .418 Verbal WM Merged Active control Passive control 0.23 0.25 0.22 [0.07,0.39] [0.03, 0.48] [0.005,0.43] .006 ** .030 * .045 * Visuospatial WM Merged Active control Passive control 0.23 0.12 NA [0.03, 0.43] [-0.08, 0.31] NA .025 * .234 NA Verbal STM Merged Active control Passive control 0.16 0.16 0.18 [-0.05,0.36] [-0.09, 0.41] [-0.21,0.58] .126 .212 .358 Visuospatial STM Merged Active control Passive control -0.03 0.03 NA [-0.39, 0.32] [-0.27, 0.35] NA .862 .803 NA F O L L O W - U P Reasoning Merged Active control Passive control 0.13 0.13 NA [-0.09, 0.35] [-0.14, 0.39] NA .236 .341 NA Verbal WM Merged Active control Passive control 0.23 0.30 0.06 [0.01, 0.46] [-0.01, 0.60] [-0.34, 0.46] .045 * .055 ^ .756 Visuospatial WM Merged Active control Passive control 0.14 0.14 NA [-0.09, 0.37] [-0.11, 0.38] NA .231 .287 NA Visuospatial STM Merged Active control Passive control 0.18 0.28 NA [-0.10, 0.45] [-0.05, 0.62] NA .205 .099 ^ NA Visuospatial STM Merged Active control Passive control -0.04 -0.01 NA [-0.33, 0.25] [-0.34, 0.31] NA .763.934 NA Note . ^ p <.1, * p <.05, ** p <.01. NA – Not applicable (analyses were performed only for constructs with more than 4 studies). Estimates, CI and P-values would not differ substantially if only studies with active control were included, except for Visuospatial WM in immediate posttest. Supplementary material (study I) Appendix B 93 Figure A. Trim-and-fill Plots by Group (measure). Verbal WM – Updating. Note. For groups having at least two outcomes from the same trial, all possible combinations of subgroups, including exactly one outcome per trial, were considered to assess publication bias and the “leave-one-out” method. Updating at posttest (a) Standard Error 0.396 0.297 0.198 0.099 0 −1 −0.5 0 0.5 1 Updating at posttest (b) Standard Error 0.396 0.297 0.198 0.099 0 −0.5 0 0.5 1 Updating at posttest (c) Standard Error 0.396 0.297 0.198 0.099 0 −0.5 0 0.5 1 Updating at posttest (d) Standard Error 0.396 0.297 0.198 0.099 0 −1 −0.5 0 0.5 Updating at posttest (e) Standard Error 0.396 0.297 0.198 0.099 0 −1 −0.5 0 0.5 Updating at posttest (f) Standard Error 0.396 0.297 0.198 0.099 0 −1 −0.5 0 0.5 APPENDIX A Supplementary material (study I) 94 Figure A. (cont.). Verbal WM – Updating Updating at posttest (g) Standard Error 0.388 0.291 0.194 0.097 0 −1 −0.5 0 0.5 Updating at posttest (h) Standard Error 0.388 0.291 0.194 0.097 0 −1 −0.5 0 0.5 Updating at posttest (i) Standard Error 0.388 0.291 0.194 0.097 0 −0.5 0 0.5 1 Updating at posttest (j) Standard Error 0.388 0.291 0.194 0.097 0 −1 −0.5 0 0.5 Updating at posttest (k) Standard Error 0.388 0.291 0.194 0.097 0 −1 −0.5 0 0.5 Updating at posttest (l) Standard Error 0.388 0.291 0.194 0.097 0 −1 −0.5 0 0.5 Supplementary material (study I) APPENDIX A 95 Figure A. (cont.). Verbal WM – complex span Complex span at posttest (a) Standard Error 0.383 0.287 0.191 0.096 0 −0.5 0 0.5 1 Complex span at posttest (b) Standard Error 0.383 0.287 0.191 0.096 0 −0.5 0 0.5 1 1.5 Complex span at posttest (c) Standard Error 0.383 0.287 0.191 0.096 0 −0.5 0 0.5 1 Complex span at posttest (d) Standard Error 0.383 0.287 0.191 0.096 0 −0.5 0 0.5 1 APPENDIX A Supplementary material (study I) 96 Figure A. (cont.). Verbal WM – STM Simple span at posttest (a) Standard Error 0.406 0.305 0.203 0.102 0 −0.5 0 0.5 1 Simple span at posttest (b) Standard Error 0.406 0.305 0.203 0.102 0 −0.5 0 0.5 1 CHAPTER III EFFECTS OF tDCS ON WORKING MEMORY IN HEALTHY OLDER ADULTS Effects of tDCS on working memory in healthy older adults Chapter III 98 Effects of transcranial direct current stimulation on working memory in healthy older adults: a systematic review4 Abstract This mini systematic review aimed to investigate the effects of transcranial direct current stimulation (tDCS) on working memory in older adults without cognitive impairment. The search was carried out in three different databases for all human trials published from 2005 to 2015, assessing the effects of tDCS on working memory in healthy older adults. The screening was conducted by two independent reviewers. Four studies were included. All studies combined anodal tDCS (applied to pre-frontal or parietal cortex) with working memory training. Anodal tDCS seems to be able to modulate working memory performance. Nonetheless, there is evidence that suggests that variables, such as level of education, working memory task and time of assessment can moderate the effect. Recommendations for futures studies are also provided. Keywords: working memory; older adults; transcranial direct current stimulation; tDCS. 4 Teixeira-Santos, A. C., Nafee, T., Sampaio, A., Leite, J., & Carvalho, S. (2015). Effects of transcranial direct current stimulation on working memory in healthy older adults: a systematic review. Principles and Practice of Clinical Research, 1 (3), 73-81. Effects of tDCS on working memory in healthy older adults Chapter III 99 Introduction Aging is associated with structural and functional loss, affecting a wide range of cognitive skills, such as memory, language and executive function (Salthouse, 1996). These changes can have a negative impact on activities of daily living and quality of life and may result in disorders such as depression, mild cognitive impairment and dementia (e.g.: Alzheimer’s disease and fronto-temporal dementia), ultimately becoming a significant burden on health-care systems (Christensen, Doblhammer, Rau, & Vaupel, 2009). Hence, growing interest emerges in an attempt to promote healthy aging, optimizing cognitive skills and remediating cognitive impairment. Among the cognitive skills affected by the aging process, working memory (WM) stands out due its notable decline throughout the individual’s lifespan. The decline begins in the mid-20s and concerns both visuospatial and verbal aspects of WM (Park et al., 2002). WM is a mental workspace in which information is maintained and processed over a short period of time while a task is being performed (Baddeley, 2003). WM is related to several higher order cognitive functions such as reading (Daneman & Carpenter, 1980), mathematics (Gathercole, Pickering, Knight, & Stegman, 2003), intelligence (Conway, Kane, & Engle, 2003; Kyllonen & Raymond, 1990; Miyake, Friedman, Rettinger, Shah, & Hegarty, 2001), prospective memory (Braver, Paxton, Locke, & Barch, 2009), processing speed (Fry & Hale, 2000), attention (Engle & Kane, 2004), perceptual organization (Woodman, Vecera, & Luck, 2003) and general language (Kemper, Herman, & Liu, 2004). The mechanisms underlying WM decline are unclear. Normal Functional brain alterations have been reported in healthy older adults; greater bilateral activation has been found in healthy older adults during a WM task compared to younger adults. This phenomenon is thought to represent a functional reorganization and compensation mechanism by the recruitment of additional resources in order to maintain cognitive performance (Rajah & D'Espósito, 2005; Schulze et al., 2011). Also normal aging is followed by structural loss in brain tissue (Ge et al., 2002; Good et al., 2001; Raz, Rodrigue, & Haacke, 2007), mainly in prefrontal brain regions (Raz & Rodrigue, 2006). Given the centrality of WM in these higher order cognitive functions and its substantial decline over aging, new strategies to reduce the impact of WM loss in this population are sorely necessary. In the past few years, there has been a growing interest in non-invasive brain stimulation techniques and the development of new combined interventions that can be used as rehabilitation strategies. Transcranial direct current stimulation (tDCS) is one such NIBS. Given the safety profile, high Chapter III Effects of tDCS on working memory in healthy older adults 106 Outcome measures Several outcome measures were used in the selected studies: verbal 2 back ( Berryhill & Jones, 2012; Park et al., 2014; Seo et al., 2011), visual 2-back (Berryhill & Jones, 2012; Jones et al., 2015), visuospatial WM task (Jones et al., 2015; Seo et al., 2011), Ospan (Jones et al., 2015), Stroop (Jones et al., 2015), digit span forward (Jones et al., 2015; Park et al., 2014), digit span backward (Park et al., 2014), verbal learning test (Park et al., 2014), visual span test (Park et al., 2014), Continuous Performance Task (CPT) (Park et al., 2014), word-color test (Park et al., 2014) and trail making test (Park et al., 2014). In general, the literature indicated that tDCS had a positive effect on working memory by improving verbal and visual working memory performance. Interestingly, Berryhill and Jones (2012) did not find significant effect of anodal tDCS on WM, when comparing to sham, immediately after stimulation. However, subgroup analysis demonstrated that older adults with higher levels of education had significant improvement in working memory performance after stimulation, having no difference between the stimulation in F4 and F3 or between verbal and visual tasks. In the group of lower educational levels, the stimulation had negative effects on visual WM performance and had no effects on verbal WM. Jones et al. (2015) modulated the current flow to identify the spatial extent of brain stimulation after anodal tDCS to the PFC and PC and they found that tDCS to the PFC supplied current to PFC regions and also to orbitofrontal and ventro-temporal regions. The PC stimulation targeted PC and posterior occipital and ventral temporal regions. There was considerable overlap of current flow in both areas. They have also identified that both active and control tDCS groups (PFC, posterior parietal cortex - PPC, PFC/PPC and sham) showed equivalent improvement immediately after 10 sessions of training. However only the active tDCS group maintained significant improvements at the 1-month follow-up for both trained and non-trained tasks. All active tDCS groups (PFC, PC, PFC altering with PC) resulted in equivalent improvements. Jones and colleagues also reported that the more challenging and adaptive tasks (recall and Ospan tasks) showed greater improvements when compared to recognition tasks. The largest transfer effect was observed in the most difficult near transfer task, the spatial 2-back. The two other near transfer measures, the Stroop task and the digit span showed no transfer effects. Park et al. (2014) showed that improved verbal WM accuracy was sustained for up to 28 days, after 10 sessions of computer-based cognitive training combined with bilateral anodal tDCS of the PFC (F3 and F4). The Effects of tDCS on working memory in healthy older adults Chapter III 107 reaction time of the verbal WM task was significantly shortened in the real stimulation group only in the last day of stimulation and not in the follow up. They also reported a near transfer effect, namely improvement in digit span forward in the active group that was observed only 7 days after stimulation. Finally, Seo et al. (2011) failed to find differences in visual working memory performance following tDCS, but they reported verbal WM improvements in the active group. There was no significant effect of tDCS in reaction time in both visuospatial and verbal working memory performance. Adverse effects Two studies reported the following adverse effects: minimal skin discoloration on the arms for a few days (Park et al., 2014) and transient aching and redness on the arm (Seo et al., 2011). Variables mediating the tDCS effect Berryhill and Jones (2012) demonstrated that the educational level has been a potential effect modifier, with participants with higher levels of education benefiting more from the intervention. Modality of working memory (verbal or visual) can also be influenced differently by tDCS. Seo et al. (2011) found positive effect of intervention only on verbal WM performance of the active group, having no difference in visuospatial WM performance. Berryhill and colleagues (2012) reported impairment in visual working memory of lower educational group. Time of assessment can also mediate the effects. Jones et al. (2015) reported that both active (PFC,PPC, PFC/PPC) and sham tDCS groups showed equivalent improvement immediately after 10 session of training. However only the active tDCS group maintained significant improvements on trained and non-trained tasks at follow-up a month later. Discussion The aim of this paper was to review the literature on tDCS effects on the WM performance of healthy elderly people. We found four papers that met our criteria. Most of the studies were randomized and included sham controlled blinded trials. The included studies showed that WM training administered with anodal tDCS over the PFC and PC can enhance WM, and these positive effects can be transferred to tasks similar to those used in the WM training. In the elderly, the effects of tDCS on WM seem to have a similar pattern to the one showed with young adults, in which anodal tDCS over the left DLPFC improves WM (Richmond, Wolk, Chein, & Olson, Chapter III Effects of tDCS on working memory in healthy older adults 108 2014). However, this enhancement was found only in the verbal component of working memory, as was the case in Seo et al. (2011). Indeed, Berryhill and colleagues (2012) reported an impairment in visual WM performance after stimulation in older people with lower educational level, which was more evident during stimulation of the right PFC (F4). Richmond et al. (2014) argue that this absence of results and negative effects of tDCS on visuospatial WM could be due to stimulation of the left hemisphere, since the left side is associated with verbal contents and the right side is responsible for visuospatial processing (Reuter-Lorenz et al., 2000). The WM task used in training can be adaptive; the difficulty level adapts to match the participant ability, and it can be increased throughout training according to the improvement of the participant’s proficiency. All the studies in this review adopted a non-adaptive WM task; the task had the same level of difficulty for all participants and was not adjusted according to the performance of the subject. One meta-analysis of WM training in an elderly population (Karbach & Verhaeghen, 2014) failed to recognize a difference between adaptive and non-adaptive training paradigms, which suggests that utilizing an adaptive structure of WM training not improve the quality of WMT. Finally, a neuroimaging study showed that a single session of anodal tDCS administered to the left inferior frontal gyrus can temporarily reverse changes in brain activity and connectivity in older adults (Meinzer, Lindenberg, Antonenko, Flaisch, & Flöel, 2013). In that study, a decrease in bilateral hyperactivity related to the intervention was observed, suggesting a “youth-like” connectivity pattern during resting state fMRI (Meinzer et al., 2013). In one of the studies (Berryhill & Jones, 2012), tDCS was beneficial in older adults if they had a higher level of education. However, in the study by Berryhill and Jones (2012), the group with relatively lower education was in school for an average 13.5 years (comparing to 16.9 on the higher level of education). These effects may be due to differences on cognitive reserve. Thus, older adults with higher educational level may present differences in the flexibility and adaptation of cognitive networks (Stern, 2013). It would be interesting to further examine the effect of educational level within groups with lower educational level or even with illiterate participants. Additionally, it is important to verify if other variables such as genetic factors, gender, age and personality can mediate the intervention effect. Half of the studies included in this review had an online cognitive training (Park et al., 2014; Seo et al., 2011), which means that the training was performed simultaneously with application of tDCS. The other half of the studies used an offline design, meaning that the task was carried out after Effects of tDCS on working memory in healthy older adults Chapter III 109 stimulation (Berryhill & Jones, 2012; Jones et al., 2015). However, in the two offline studies, participants performed practice trials while receiving tDCS. Both kinds of stimulation (online and offline) showed similar results, which would be expected since the physiological effects of tDCS have been reported to last for more than one hour after several minutes of stimulation, and the participants of the offline studies performed the practice trial during stimulation (Nitsche & Paulus, 2000). Nevertheless, it is worth pointing out that there is evidence that online and offline tDCS can have differential effects. For instance, anodal tDCS over the motor cortex increases motor learning when applied during the task, while offline tDCS has the opposite effect (Stagg et al., 2011). Moreover, online anodal tDCS over the left DLPFC is more effective on skill acquisition following two days of WM training than offline tDCS (Martin, Liu, Alonzo, Green, & Loo, 2014). In line with this finding, a neuroimaging study reported greater brain activation during stimulation compared to the period following stimulation (Stagg et al., 2013). Further studies should explore the effects of tDCS timing during WM training in elderly people. Repetitive sessions of tDCS are thought to boost the effects of stimulation, since single stimulation has relatively short after-effects (Nitsche et al., 2008; Nitsche & Paulus, 2000). The main assumption underlying the effects of repetitive sessions is that it will change the mechanisms of synaptic plasticity, such as long-term potentiation and long term depression (Fritsch et al., 2010; Nitsche, Fricke, et al., 2003; Nitsche et al., 2004; Nitsche, Schauenburg, et al., 2003; Rroji, van Kuyck, Nuttin, & Wenderoth, 2015). Long-term potentiation is activity-dependent plasticity that induces an increase of synaptic transmission, while long-term depression reduces the efficacy of synaptic transmission (Bliss & Cooke, 2011). Therefore, the use of repetitive sessions of tDCS may induce learning in the neural networks which will ultimately benefit cognitive training (Brunoni et al., 2013). Among the papers analyzed, we identified only two studies showing the effect of repeated tDCS sessions on WM in older people (Jones et al., 2015; Park et al., 2014) which is in line with findings reported with younger people (Baker, Rorden, & Fridriksson, 2010; Glisky et al., 1986; Hamilton, Chrysikou, & Coslett, 2011; Iuculano & Kadosh, 2013; Kim, Han, Ahn, Kim, & Kim, 2012; Loo et al., 2012; Martin et al., 2013; Reis et al., 2009). In both studies (Jones et al., 2015; Park et al., 2014) participants received intervention five days a week for 2 weeks. There is no concrete evidence of the optimal sessions frequency and duration of tDCS, however the cumulative effects of motor cortical excitability in daily sessions of anodal tDCS seem to be greater compared to sessions separated by a two day interval (Alonzo, Brassil, Taylor, Martin, & Loo, 2012). Similar results were found in stroke patients in which the cumulative effects of motor function Chapter III Effects of tDCS on working memory in healthy older adults 110 was associated with five daily sessions, but not associated with weekly sessions of tDCS (Boggio et al., 2007). Additionally, tDCS has been reported to have different effects depending on the duration of stimulation. Three minutes have been reported as the minimum required time to induce an after-effect; and longer periods (i.e. more than 30-min) have produced mixed results (Batsikadze et al., 2013; MonteSilva et al., 2013). There is also the possibility that the baseline level of cortical activity in a given neural network can modify subsequent modification to that network (Carvalho et al., 2015). Although it is costlier and logistically difficult to carry out studies with multiple sessions compared to single sessions, investment in this area is warranted and may significantly contribute to development of this field. This would go a long way to validate the effectiveness of this type of intervention, as well as standardize the tDCS intervention protocol in terms of the number of sessions, interval between sessions and duration of stimulation. The optimized site of tDCS is another issue that needs consideration. Based on computer modeling, the largest effects induced by tDCS polarity are elicited beneath the stimulation electrode (Wagner et al., 2007). However, tDCS over different stimulation locations (such as bi-hemispheric, unihemispheric, prefrontal, parietal, prefrontal alternating with parietal and right and left) can lead to similar effects on WM. It is important to have active stimulation targeting an area that is not related to WM in order to determine whether similar effects can be observed by stimulating any given area of brain (Jones et al., 2015). Moreover, as most of tDCS effects so far have been on the verbal subcomponent of WM, it will be important to test different targets in order to increase other subcomponents, such as the visual. Finally, our results provide evidence of the safety of tDCS in elderly people, as only minor adverse effects were reported among studies. Conclusion In sum, anodal tDCS over the PC and PFC seems to improve WM in healthy elderly subjects, and those improvements can be sustained up to one-month post-intervention. However, better parameters of stimulation are still required before mainstream use. Effects of tDCS on working memory in healthy older adults Chapter III 111 References Alonzo, A., Brassil, J., Taylor, J. L., Martin, D., & Loo, C. K. (2012). Daily transcranial direct current stimulation (tDCS) leads to greater increases in cortical excitability than second daily transcranial direct current stimulation. Brain Stimulation, 5 (3), 208-213. doi: 10.1016/j.brs.2011.04.006 Andrews, S. C., Hoy, K. E., Enticott, P. G., Daskalakis, Z. J., & Fitzgerald, P. B. (2006). Improving working memory: The effect of combining cognitive activity and anodal transcranial direct current stimulation to the left dorsolateral prefrontal cortex. Brain Stimulation, 4 (2), 84-89. doi:10.1016/j.brs.2010.06.004 Baddeley, A. (2003). Working memory: Looking back and looking forward. Nature Reviews Neuroscience, 4 (10), 829-839. doi: 10.1038/nrn1201 Baker, J. M., Rorden, C., & Fridriksson, J. (2010). Using transcranial direct-current stimulation to treat stroke patients with aphasia. Stroke, 41 (6), 1229-1236. doi: 10.1161/STROKEAHA.109.576785 Batsikadze, G., Moliadze, V., Paulus, W., Kuo, M. F., & Nitsche, M. A. (2013). Partially non-linear stimulation intensity-dependent effects of direct current stimulation on motor cortex excitability in humans. The Journal for Physiology, 591 (7), 1987-2000. doi: 10.1113/jphysiol.2012.249730 Berryhill, M. E., & Jones, K. T. (2012). tDCS selectively improves working memory in older adults with more education. Neuroscience Letters, 521 (2), 148-151. doi: 10.1016/j.neulet.2012.05.074 Bliss, T. V. P., & Cooke, S. F. (2011). Long-term potentiation and long-term depression: A clinical perspective. Clinics (Sao Paulo), 66 (1), 3-17. doi: 10.1590/S1807-59322011001300002 Boggio, P. S., Ferrucci, R., Rigonatti, S. P., Covre, P., Nitsche, M., Pascual-Leone, A., & Fregni, F. (2006). Effects of transcranial direct current stimulation on working memory in patients with Parkinson's disease. Journal of the Neurological Sciences, 249 (1), 31-38. doi: 10.1016/j.jns.2006.05.062 Boggio, P. S., Nunes, A., Rigonatti, S. P., Nitsche, M. A., Pascual-Leone, A., & Fregni, F. (2007). Repeated sessions of noninvasive brain DC stimulation is associated with motor function improvement in stroke patients. Restorative Neurology and Neuroscience, 25 (2), 123-129. Braver, T. S., Paxton, J. L., Locke, H. S., & Barch, D. M. (2009). Flexive neural mechanisms of cognitive control within human prefrontal cortex. Proceeding of the National Academy of Sciences, 106 (18), 7351-7356. doi: 10.1073/pnas.0808187106 Chapter III Effects of tDCS on working memory in healthy older adults 112 Brunoni, A. R., Amadera, J., Berbel, B., Volz, M. S., Rizzerio, B. G., & Fregni, F. A. (2011). A systematic review on reporting and assessment of adverse effects associated with transcranial direct current stimulation. The International Journal of Neuropsychopharmacology, 14 (8), 1133-1145. doi:10.1017/S1461145710001690 Brunoni, A. R., Nitsche, M. A., Bolognini, N., Bikson, M., Wagner, T., Merabet, L., …Fregni, F. (2012). Clinical research with transcranial direct current stimulation (tDCS): Challenges and future directions. Brain Stimulation, 5 (3), 175-195. doi: 10.1016/j.brs.2011.03.002 Carvalho, S., Boggio, P. S., Gonçalves, Ó., Vigário, A. R., Faria, M., Silva, S., …Leite, J. (2015). Transcranial direct current stimulation based metaplasticity protocols in working memory. Brain Stimulation, 8 (2), 289-294. doi:10.1016/j.brs.2014.11.011 Christensen, K., Doblhammer, G., Rau, R., & Vaupel, J. W. (2009). Ageing populations: The challenges ahead. The Lancet, 374 (9696), 1196-1208. doi: 10.1016/S0140-6736(09)61460-4 Conway, A. R., Kane, M. J., & Engle, R. W. (2003). Working memory capacity and its relation to general intelligence. Trends in Cognitive Sciences, 7 (12), 547-552. doi: 10.1016/j.tics.2003.10.005 Daneman, M., & Carpenter, P. A. (1980). Individual differences in working memory and reading. Journal of Verbal Learning and Verbal Behavior, 19 (4), 450-466. doi: 10.1016/S0022-5371(80)90312-6 Engle, R. W., & Kane, M. J. (2004). Executive attention, working memory capacity, and a two-factor theory of cognitive control. Psychology of Learning and Motivation, 44 , 145-199. doi: 10.1016/S0079-7421(03)44005-X Ferrucci, R., Marceglia, S., Vergari, M., Cogiamanian, F., Mrakic-Sposta, S., Mameli F., …Priori, A. (2008). Cerebellar transcranial direct current stimulation impairs the practice-dependent proficiency increase in working memory. Journal of Cognitive Neuroscience, 20 (9), 1687-1697. doi: 10.1162/jocn.2008.20112 Fregni, F., Boggio, P. S., Nitsche, M., Bermpohl, F., Antal, A., Feredoes, E., …Pascual-Leone, A. (2005). Anodal transcranial direct current stimulation of prefrontal cortex enhances working memory. Experimental Brain Research, 166 (1), 23-30. doi: 10.1007/s00221-005-2334-6 Fregni, F., & Pascual-Leone, A. (2007). Technology insight: Noninvasive brain stimulation in neurologyperspectives on the therapeutic potential of rTMS and tDCS. Nature Clinical Practice Neurology, 3 (7), 383-393. doi:10.1038/ncpneuro0530 Effects of tDCS on working memory in healthy older adults Chapter III 113 Fritsch, B., Reis, J., Martinowich, K., Schambra, H. M., Ji, Y., Cohen, L. G., & Lu, B. (2010). Direct current stimulation promotes BDNF-dependent synaptic plasticity: Potential implications for motor learning. Neuron, 66 (2), 198-204. doi: 10.1016/j.neuron.2010.03.035 Fry, A. F., & Hale, S. (2000). Relationships among processing speed, working memory, and fluid intelligence in children. Biological Psychology, 54 (1-3), 1-34. doi: 10.1016/S03010511(00)00051-X Gandiga, P. C., Hummel, F. C., & Cohen, L. G. (2006). Transcranial DC stimulation (tDCS): A tool for double-blind sham-controlled clinical studies in brain stimulation. Clinical Neurophysiology, 117 (4), 845-850. doi:10.1016/j.clinph.2005.12.003 Gathercole, S. E., Pickering, S. J., Knight, C., & Stegman, Z. (2003). Working memory skills and educational achievement: Evidence from national curriculum assessments at 7 and 14 years of age. Applied Cognitive Psychology, 18 (1), 1-16. doi: 10.1002/acp.934 Ge, Y., Grossman, R. I., Babb, J. S., Rabin, M. L., Mannon, L. J., & Kolson, D. L. (2002). Age-related total gray matter and white matter changes in normal adult brain. Part I: volumetric MR imaging analysis. American Journal of Neuroradiology, 23 (8), 1327-1333. Glisky, E. L., Schacter, D. L., & Tulving, E. (1986). Computer learning by memory-impaired patients: Acquisition and retention of complex knowledge. Neuropsychologia, 24 (3), 313-328. doi: 10.1016/0028-3932(86)90017-5 Good, C. D., Johnsrude, I. S., Ashburner, J., Henson, R. N., Friston, K. J., & Frackowiak, R. S. (2001). A voxel-based morphometric study of ageing in 465 normal adult human brains. Neuroimage, 14 (1), 21-36. doi: 10.1006/nimg.2001.0786 Hamilton, R. H., Chrysikou, E. G., & Coslett, B. (2011). Mechanisms of aphasia recovery after stroke and the role of noninvasive brain stimulation. Brain and Language, 118 (1-2), 40-50. doi: 10.1016/j.bandl.2011.02.005 Hsu, W. Y., Ku, Y., Zanto, T. P., & Gazzaley, A. (2008). Effects of noninvasive brain stimulation on cognitive function in healthy aging and Alzheimer’s disease: A systematic review and metaanalysis. Neurobiology of Aging, 36 (8), 2348–2359. doi:10.1016/j.neurobiolaging.2015.04.016 Iuculano, T., & Kadosh, R. C. (2013). The mental cost of cognitive enhancement. The Journal of Neuroscience, 33 (10), 4482-4486. doi: 10.1523/JNEUROSCI.4927-12.2013 Chapter III Effects of tDCS on working memory in healthy older adults 114 Jasper, H. H. (1958). The ten-twenty electrode system of the International Federation. Electroencephalography and Clinical Neurophysiology, 10 , 371-375. Jo, J. M., Kim, Y. H., Ko, M. H., Ohn, S. H., Joen, B., & Lee, K. H. (2009). Enhancing the working memory of stroke patients using tDCS. American Journal of Physical Medicine & Rehabilitation, 88 (5), 404-409. doi: 10.1097/PHM.0b013e3181a0e4cb Jones, K. T., Stephens, J. A., Alam, M., Bikson, M., & Berryhill, M. E. (2015). Longitudinal neurostimulation in older adults improves working memory. PLoS One, 10 (4), e0121904. doi: 10.1371/journal.pone.0129751 Karbach, J., & Verhaeghen, P. (2014). Making working memory work: A meta-analysis of executive control and working memory training in older adults. Psychological Science, 25 (11), 2027–2037. doi: 10.1177/0956797614548725 Kemper, S., Herman, R. E., & Liu, C. J. (2004). Sentence production by young and older adults in controlled contexts. The Journals of Gerontology Series B Psychological Sciences and Social Sciences, 59 (5), 220-224. doi: 10.1093/geronb/59.5.P220 Kim, S. H., Han, H. J., Ahn, H. M., Kim, S. A., & Kim, S. E. (2012). Effects of five daily high-frequency rTMS on Stroop task performance in aging individuals. Neuroscience Research, 74 (3-4), 256-260. doi: 10.1016/j.neures.2012.08.008 Kyllonen, P. C. C., & Raymond, E. (1990). Reasoning ability is (little more than) working-memory capacity?! Intelligence, 14 (4), 389–433. doi: 10.1016/S0160-2896(05)80012-1 Loo, C. K., Alonzo A., Martin, D., Mitchell, P. B., Galvez, V., & Sachdev, P. (2012). Transcranial direct current stimulation for depression: 3-week, randomised, shamcontrolled trial. The British Journal of Psychiatry, 200 (1), 52–59. doi: 10.1192/bjp.bp.111.097634 Martin, D. M., Liu, R., Alonzo, A., Green, M., & Loo, C. K. (2014). Use of transcranial direct current stimulation (tDCS) to enhance cognitive training: Effect of timing of stimulation. Experimental Brain Research, 232 (10), 3345-3351. doi: 10.1007/s00221-014-4022-x Martin, D. M., Liu, R., Alonzo, A., Green, M., Player, M. J., Sachdev, P., & Loo, C. K. (2013). Can transcranial direct current stimulation enhance outcomes from cognitive training? A randomized controlled trial in healthy participants. International Journal of Neuropsychopharmacology, 16 (9), 1927-1936. doi: 10.1017/S1461145713000539 Effects of tDCS on working memory in healthy older adults Chapter III 115 Meinzer, M., Lindenberg, R., Antonenko, D., Flaisch, T., & Flöel, A. (2013). Anodal transcranial direct current stimulation temporarily reverses age-associated cognitive decline and functional brain activity changes. The Journal of Neuroscience, 33 (30), 12470-12478. doi: 10.1523/JNEUROSCI.5743-12.2013 Miyake, A., Friedman, N. P., Rettinger, D. A., Shah, P., & Hegarty, M. (2001). How are visuospatial working memory, executive functioning, and spatial abilities related? A latent-variable analysis. Journal of Experimental Psychology: General , 130(4), 621-640. doi: 10.1037/00963445.130.4.621 Monte-Silva, K., Kuo, M. F., Hessenthaler, S., Fresnoza, S., Liebetanz, D., Paulus, W., & Nitsche, M. A. (2013). Induction of late LTP-like plasticity in the human motor cortex by repeated non-invasive brain stimulation. Brain Stimulation, 6 (3), 424-432. doi: 10.1016/j.brs.2012.04.011 Nitsche, M. A., Cohen, L. G., Wassermann, E. M., Priori, A., Lang, N., Antal, A. …Pascual-Leone, A. (2008). Transcranial direct current stimulation: State of the 2008. Brain Stimulation, 21 (3), 206223. doi:10.1016/j.brs.2008.06.004 Nitsche, M. A., Fricke, K., Henschke, U., Schlitterlau, A., Liebetanz, D., Lang, N., …Paulus, W. (2003) Pharmacological modulation of cortical excitability shifts induced by transcranial direct current stimulation in humans. The Journal of Physiology, 553 , 293-301. doi: 10.1113/jphysiol.2003.049916 Nitsche, M. A., Liebetanz, D., Schlitterlau, A., Henschke U., Fricke, K., & Frommann, K., …Tergau, F. (2004). GABAergic modulation of DC stimulation-induced motor cortex excitability shifts in humans. European Journal of Neuroscience, 19 (10), 2720-2726. doi: 10.1111/j.0953816X.2004.03398.x Nitsche, M. A., & Paulus, W. (2000). Excitability changes induced in the human motor cortex by weak transcranial direct current stimulation. The Journal of Physiology, 527 (3), 633-639. doi: 10.1111/j.1469-7793.2000.t01-1-00633.x Nitsche, M. A., Schauenburg, A., Lang, N., Liebetanz, D., Exner, C., Paulus, W., & Tergau, F. (2003). Facilitation of implicit motor learning by weak transcranial direct current stimulation of the primary motor cortex in the human. Journal of Cognitive Neuroscience, 15 (4), 619-626. doi: 10.1162/089892903321662994 Chapter IV Transfer effects of working memory training coupled with tDCS in older adults 122 Cognitive transfer effects of working memory training coupled with transcranial direct current stimulation in healthy older adults: a double-blinded, randomized, sham controlled experiment6 Abstract Background: Working memory training (WMT) has been used for cognitive enhancement in older adults. Furthermore, transcranial direct current stimulation (tDCS) has been used to boost the effects of WMT. Nevertheless, there is limited evidence on the combination of tDCS and WMT efficacy in older people. Objective: The present study aimed to assess the immediate effects of tDCS coupled with WMT and, whether those effects are maintained at a 15-day follow-up in 54 healthy older. We also explored if baseline performance, age and educational level modified the effects of treatment. Method: In this double-blind randomized placebo-controlled experiment, participants were randomized into three groups: anodal-tDCS+WMT; sham-tDCS+WMT or double-sham. Five-sessions of tDCS (2mA) were applied over the left dorsolateral prefrontal cortex (DLPFC). Near transfer effect was assessed through Digit Span and Corsi Block-tapping Test, while far transfer was measured by Raven Advanced Progressive Matrices (RAPM) and Digit-symbol Coding. Results: Multilevel modeling analysis showed that only the group with anodal-tDCS+WMT displayed a significant improvement from pretest to follow-up in measures of reasoning (RAPM) and short-term memory (forward digit span). Near transfer gains predicted gains in far transfer. Moreover, there was no 6 Publications derived from this study: Peer reviewed publications in print or other media Teixeira-Santos, et al. (2019). Can tDCS enhance transfer effects of working memory training in older adults? Manuscript in preparation. Abstracts, Poster Presentations and Exhibits Presented at Professional Meetings Teixeira-Santos, A.C, Moreira C.S., Pereira, D.R., Carvalho S., Sampaio A. (2019). Transfer effects of anodal transcranial direct current stimulation coupled with working memory training in healthy older adults: a randomized controlled trial. Poster session accepted at the International Conference AGENortC, Viana do Castel, Portugal. Teixeira-Santos, A.C., Pereira, D.R., Alves, S., Leite, J., Carvalho, S., & Sampaio, A. (2018) Can tDCS enhance transfer effects of working memory training in older adults? Poster session presented at the 4º Congresso da Ordem dos Psicólogos Portugueses, Braga, Portugal. Teixeira-Santos, A.C., Pereira, D.R., Alves, S., Leite, J., Carvalho, S., & Sampaio, A. (2018). Can tDCS enhance transfer effects of working memory training in older adults? Abstract available at Abstracts of ongoing projects supported by the Bial Foundation (www.fundacaobial.com). Transfer effects of working memory training coupled with tDCS in older adults Chapter IV 123 strong evidence that supported age, educational level, baseline performance, and general cognitive ability as predictors of reasoning gains. Conclusion: WMT coupled with anodal-tDCS may improve short-term memory and reasoning in older adults, with the effects most strongly observed at follow-up. Keywords: working memory training; aging; tDCS; neuroplasticity; individual differences. Chapter IV Transfer effects of working memory training coupled with tDCS in older adults 124 Introduction Working memory (WM), a temporary set of mental components in which information is held and available for ongoing information processing (Cowan, 2017), is a core process for many higher-order cognitive functions (Glisky, 2007). It is among the most impaired cognitive functions in elderly people (Kirova, Bays, & Lagalwar, 2015; Murman, 2015). Considering that 13% of world population is aged 60 or more and that the prevalence of mild cognitive impairment within this stage ranges from 3% to 42%, it is of paramount importance to develop strategies to preserve cognitive functioning in this population (United Nations, Department of Economic and Social Affairs, & Population Divison, 2017; Ward, Arrighi, Michels, & Cedarbaum, 2012). Therefore, working memory training (WMT) has been proposed as a prominent intervention in the elderly, which may benefit not only WM, but also other cognitive processes related to it. However, the results of different studies are controversial (Karbach & Verhaeghen, 2014; Melby-Lervåg & Hulme, 2013, 2016; Melby-Lervåg, Redick, & Hulme, 2016). In this sense, other techniques have been developed to promote cognitive enhancement in the elderly (Davis, 2017; Strenziok et al., 2014). Among them, transcranial direct current stimulation (tDCS) has been tested as an add-on tool to boost WMT (Teixeira-Santos, Nafee, Sampaio, Leite, & Carvalho, 2015). Most of the studies in which tDCS was used as add-on, have been conducted with younger adults, and report that anodal tDCS (atDCS) over the DLPFC may improve WM performance (Fregni et al., 2005; Ke et al., 2019; Ohn et al., 2008; Zaehle, Sandmann, Thorne, Jäncke, & Herrmann, 2011). However, these studies were performed in a single tDCS session, and recent literature has suggested a beneficial effect of repeated sessions (Alonzo, Brassil, Taylor, Martin, & Loo, 2012; Gálvez, Alonzo, Martin, & Loo, 2013; Hsu, Ku, Zanto, & Gazzaley, 2015; Martin et al., 2013). Few studies have combined WMT with tDCS in elderly people, showing improvements on the trained task, as well as, transfer effects, which were found even a month after training (Jones, Stephens, Alam, Bikson, & Berryhill, 2015; Park, Seo, Kim, & Ko, 2013). On the other hand, Nilsson et al. (2017) failed to find transfer effects in a 20-session tDCS coupled with an executive functioning protocol in older adults. The same team also performed a single session cross-over trial comparing 1mA with 2mA atDCS and a sham condition (stDCS) in older people and failed to find superiority of the stimulation conditions in a n-back task performance (Nilsson, Lebedev, & Lövdén, 2015). Additionally, individual differences (i.e., age, baseline cognitive performance, general cognitive ability and educational level) seem to Transfer effects of working memory training coupled with tDCS in older adults Chapter IV 125 interact with WMT and tDCS (Berryhill & Jones, 2012; Borella, Carbone, Pastore, De Beni, & Carretti, 2017; Gözenman & Berryhill, 2016; Ke et al., 2019; Ruf, Fallgatter, & Plewnia, 2017). Considering this mixed evidence regarding the effects of tDCS coupled with WMT in older adults, in this study, we assessed the effects of 5-day tDCS coupled with dual n-back training, after training and in a 15-day follow-up in healthy older adults. In this experiment, the main outcomes were WM tasks (near transfer), as well as, a reasoning task (far transfer). We expected that both WMT groups will present near transfer when compared to the double-sham group. However, we believed that far transfer for reasoning would be verified only in atDCS+WMT group. Since far transfer has not been demonstrated in WMT alone (Salminen, Frensch, Strobach, & Schubert, 2016), we hypothesized that tDCS could boost training, improving its generalization. Additionally, we expected that near transfer gains would predict far transfer gains, since it is postulated that far transfer is due to plasticity in WM (Melby-Lervåg & Hulme, 2013). An additional analysis explored whether vocabulary, general cognitive ability, age, and educational level modulated WMT effects. Methods Study design A CONSORT (Consolidated Standards of Reporting Trials) diagram is presented in Figure. Participants and assessors were blinded to both stimulation and task conditions. Stimulation sessions and assessments were performed by different researchers. The randomization list was generated in a website (http://www.randomization.com) in blocks of 6 with a ratio of 2:2:2. The allocation list was masked from all investigators. The condition of each participant was described in different excel sheets in a way that the researcher responsible for the randomization had access only to the allocation of the next participant (allocation concealment). Chapter IV Transfer effects of working memory training coupled with tDCS in older adults 126 Figure 1. CONSORT (Consolidated Standards of Reporting Trials) diagram. Participants Fifty-four participants (68.20 ± 5.92 years old) were randomized to one of three groups: 1) atDCS+WMT; 2) stDCS+WMT; 3) double-sham. All participants were right-handed, with normal or corrected-to-normal visual (≥20/40 in both eyes) and auditory acuity, with no history of neurological/psychiatric disorders, substance abuse or recent use of psychotropic medication, nor contraindication for tDCS. Participants were recruited in senior daycare centers and in sport and recreation clubs. See Supplementary Table S1 for group differences at baseline. All included participants scored above MoCA cut off (of 2 standard deviation) for cognitive impairment following the normative score of the Portuguese population, considering age and education (Freitas et al., 2011). Participants scoring above 9 in GDS (Pocinho, Farate, Dias, Lee, & Yesavage, 2009) were excluded. The study was performed in accordance with the Declaration of Helsinki and approval was obtained from the ethics Transfer effects of working memory training coupled with tDCS in older adults Chapter IV 127 subcommittee for life and health sciences of University of Minho (SECVS 012/2016). Participants gave informed consent before their inclusion in the study. Procedure Participants underwent 11 sessions as depicted in Figure. Figure 2. Schematic representation of the sessions. The experimental task was illustrated for a 2-back condition. Note. Participants underwent an EEG session, the results of which will be discussed in another manuscript. Screening session Participants completed a socioeconomic and medical questionnaire, the Jaeger Card (Kniestedt & Stamper, 2003), auditory discrimination of letters, GAI (Pachana et al., 2007; Ribeiro, Paúlac, Simoes, & Firmino, 2011), GDS (Pocinho et al., 2009; Yesavage et al., 1982), and MoCA (Freitas et al., 2011; Nasreddine et al., 2005). Chapter IV Transfer effects of working memory training coupled with tDCS in older adults 128 Pretest In this session, participants performed the vocabulary, the digit span and digit-symbol Coding of Wechsler Adult Intelligence Scale (WAIS) (Wechsler, 2008), Raven’s Advanced Progressive Matrices (RAPM) (Raven, Raven, & Court, 1998), Corsi Block-Tapping Test (Corsi, 1973), and the dual n-back task. Training Pergherm Wittevrongel, Tournoy, Schoenmakers, & Van Hulle (2018) demonstrated that 5 days of WMT were sufficient to improve n-back performance in elderly participants and no difference was observed between five or ten days of training. Accordingly, in the current study a total of 5 sessions were administered, although our experimental design has some differences from that study. Participants answered a Visual Analogue Scale (VAS) to assesses possible tDCS side effects (i.e., levels of discomfort, fatigue, anxiety, pain, itching, humor, tingling, headache and sleepiness), before and after each day of intervention. Blinding was assessed in the last session using a questionnaire asking which condition participants thought they were allocated regarding task and tDCS. Posttest and 15-days follow-up In these sessions, participants performed the digit-symbol code, digit span, RAPM, Corsi BlockTapping Test, and the dual n-back. tDCS parameters tDCS was applied during 20 minutes with an intensity of 2 mA, using two 5x7 cm2 saline-soaked electrode sponges (current density approximately 0.0057 mA/cm2) with anode positioned over the left DLPFC (F3) and the cathode over the contralateral supraorbital area (Fp2) (Jasper, 1958). The current fade in and fade out was 15/15 seconds. The electrode setup was identical for sham condition. However, the stimulation was discontinued after 45 sec of administration (15 sec of fade in/stimulation/fade out). Participants started the task after 3 minutes of tDCS. Transfer effects of working memory training coupled with tDCS in older adults Chapter IV 129 Trained tasks Experimental task The task was displayed using the Presentation software package (Neurobehavioral Systems, Albany, CA). Participants were simultaneously shown visuospatial and auditory-verbal stimuli. The visuospatial stimulus was a square presented in one of eight possible locations in a 3x3 grid with a fixation cross on the central square. The auditory-verbal stimulus was one of nine possible consonants (T, G, X, H, R, S, L, K, J) displayed in a random order, delivered binaurally through Sony MDR-NC6 noise cancelling headphones. Stimuli were presented for 500 ms, with 2500 ms of interstimulus interval. For each trial, participants decided whether the stimulus presented was the same presented n trials before. Participants were instructed to press the ‘spacebar’ every time either a visuospatial or an auditory-verbal target was presented. At the end of each block, a feedback with the participant’s hits was shown. The task consisted of 12 blocks with 25 trials each. In each block, there were 2 auditory-verbal and 2 visuospatial targets, and 1 stimulus that was a target in both modalities. The n level started with n = 1 and increased by 1 if the participants achieved 100% of hits in three consecutives blocks. During training, the n level started in the maximum level achieved by the participant in the previous day. However, if the number of hits in the last 3 blocks of the previous session was inferior to 60%, the n level was decreased by 1. Training lasted approximately 20 minutes per session. Placebo task A visuoperceptual task was used as placebo in order to control confounding variables resulting from the intervention setting and to allow the blinding. The task was presented in Superlab software (The Experimental Laboratory Software, version 5.0.3; Cedrus Corporation, San Pedro, CA). In this task, a 3x3 grid was presented with a fixation cross in the center. Squares were presented in one of four possible locations in the grid. Participants had to press the key (‘t’; ‘f’; ‘v’; ‘g’ – marked with arrows), correspondent to the position where the square showed up. In total, there were 203 trials. Stimuli duration was 700 ms, with 800 ms of interstimulus interval. Transfer measures are important to verify the generalization of the stimulation for other tasks not trained. Therefore, we used transfer tasks to be described in the next sections. Chapter IV Transfer effects of working memory training coupled with tDCS in older adults 130 Near transfer tasks Digit span Participants listened to a sequence of digits and were instructed to recall them in the forward and backward order. Corsi block-tapping test In a board with nine blocks, participants reproduced a sequence of movements in the forward or backward order. In both near transfer tasks, the first two trials consisted of 2 digits, and the length of the sequence increased by 1 every 2 trials until the participant had two consecutive errors at the same level. The outcome was the total number of sequences correctly recalled in the forward and backward modalities separately (maximum score of 16 for each task and order). Far transfer tasks RAPM RAPM (set 1 and 2) (Raven et al., 1998) is composed by 48 figures with a 3x3 matrix of line and geometrical shapes, in which one of the shapes is missing. By choosing from eight options, participants were asked to complete the missing part from the figure. Two parallel forms were set by separating odd and even trials. The versions were randomized and counterbalanced between sessions in a way that in pretest and follow-up participants used one form, while in the posttest they performed the other version. Outcome was the sum of correct responses (one point for each/ maximum score of 6 points for set 1 and 18 for set 2). Digit-symbol code Participants completed as fast and accurately as possible digit-symbol correspondences during 120 sec, following a key provided in the top of the page. Total score is the number of correctly identified pairs. Maximum score is 133. Digit-symbol Code was performed in order to assess gains in processing speed. . Gains in processing speed were expected in all three conditions, since the control group also executed a task Transfer effects of working memory training coupled with tDCS in older adults Chapter IV 131 that targeted this domain. This task was used to control if gains observed in experimental conditions would be restricted to gains in WM and reasoning Data analysis Statistical analyses were conducted using RStudio, version 3.5.2 (R Core Team, 2018), with the following package: lme4 (Bates, Maechler, Bolker, & Walker, 2015); lmerTest (Kuznetsova, Brockhoff, & Christensen, 2017); glmmTMB (Brooks et al., 2017); brms (Bürkner, 2017); ordinal (Christensen, 2019); effects (Fox & Weisberg, 2018) and lsmeans (Lenth, 2016). Mixed-effects models were used due to their flexibility and efficiency in analysing repeated measures, accounting for pretest-differences in the outcomes (Winter, 2013). Level of significance was set at p < 0.05. We supported our results with Bayesian analysis. The effects were interpreted as significant only when the Bayesian analysis confirmed the frequentist results (Dienes, 2011). The models were analysed to verify the relationship between each transfer outcome, in the three testing assessment sessions, considering the stimulation condition. As fixed effects, we entered the interaction group*moment into the model and participants as random effects. The performance of the trained task (probability of maximum level achieved across the 5 days of training) of the WMT+atDCS and WMT+stDCS groups was assessed with a similar modelling approach; however, in this case the “moment” variable was considered a continuous variable (from 1 to 5). To verify whether near transfer gains predicted far transfer gains (Melby-Lervåg et al., 2016), we ran linear mixed models (LMMs) analyses with the RAPM set_1 scores of the atDCS+WMT group, having the gains for near transfer measures (i.e., forward and backward digit span and Corsi Block-Tapping Test) as fixed effect and participants as random effect. Gains were calculated as the difference between posttest/follow-up and pretest scores. In the individual difference analysis, we added the predictors (age, educational level, general cognitive ability operationalised by RAPM-set 2, and vocabulary scores at baseline) as fixed effect, in the 3-way interaction: RAPM-set1~group*testing session*predictor. As random effects, we had intercepts for participants. The effect sizes of post-intervention/follow-up were calculated using the package “metafor” (Viechtbauer, 2010), taken into consideration the pretest performance. They were calculated using