The effect of experience and instructions on learned attentional biases

Cobos-Cano, Pedro Luis; Vadillo, Miguel Ángel; Luque-Ruiz, David; Le Pelley, Mike E.

The effect of experience and instructions on learned attentional biases

Cobos-Cano, Pedro Luis,Vadillo, Miguel Ángel,Luque-Ruiz, David,Le Pelley, Mike E.

Published: 2015-09-01
Repository: riuma.uma.es
Language: en

Abstract

It has been shown that selective attention is allocated to the best available predictor of an outcome, which is known as learned predictiveness. Mitchell et al. (2012) have shown that instructions about the ‘relevance’ of each stimulus can influence (and even reverse) the learned predictiveness attentional bias, suggesting that propositional reasoning plays a crucial role in this phenomenon. Our experiment further explores the effects of instructions on this learned attentional bias. As a difference with previous work, we measured attentional capture through spatial cueing effects, which have been found to rely on rapid attentional processes (Le Pelley et al., 2013). Participants responded faster to events presented in the spatial location cued by stimuli that had previously been trained as predictive through trial-by-trial learning. However, verbal instructions regarding relevance failed to speed up participants’ responses or to modulate the effect of learned predictiveness on spatial cueing. These results suggest that predictive stimuli produce an attentional bias which is not (always) under voluntary control.

Full text

The effect of experience and instructions on learned attentional biases XXVII Congreso de la Sociedad Española de Psicología Comparada. Sevilla, 9-11 septiembre 2015 Pedro L. Cobos, Miguel A. Vadillo David Luque & Mike E. Le Pelley The relationship between predictive learning and attentional capture We learn from experience which stimuli are predictive of relevant outcomes and which are nonpredictive Predictive stimulus Nonpredictive stimulus The relationship between predictive learning and attentional capture Selective attention prioritizes predictive over nonpredictive stimuli The relationship between predictive learning and attentional capture We learn more about attended stimuli than about nonattended stimuli Our main concern What mechanism underlies the effect of learned predictiveness on attentional capture? Top-down mechanism Mitchell et al. (2012) Voluntary control of attention Based on reasoning processes Can be flexibly altered Bottom-up mechanism Le Pelley et al. (2013) Automatic control of attention Triggered by stimulus properties Rather inflexible Evidence based on the effect of verbal instructions Mitchell et al. (2012): The effect of learned predictiveness can be reversed through verbal instructions Instructions provided between learning phases 1 and 2 Continuity group Change group Those stimuli that were predictive during Phase 1 will continue to be predictive during Phase 2 Those stimuli that were predictive during Phase 1 will be nonpredictive during Phase 2 Evidence based on the effect of verbal instructions Mitchell et al. (2012) Evidence based on the effect of verbal instructions Mitchell et al. (2012) concluded that: The effect of learned predictiveness on attentional capture is better explained by a top-down mechanism of selective attention. Bottom-up processes play no role in the effect of learned predictiveness on attentional capture. A possible limitation of Mitchell et al.’s (2012) study The amount of time spent looking at each stimulus may be insensitive to bottom-up processes of attentional capture. Limited to overt attention. Insensitive to fast, covert attentional shifts. Method Stimuli Target stimuli Filler stimuli Method Procedure 1000 ms 500 ms 250 ms or 1000 ms Until participant’s response Until participant’s response 6 6,02 6,04 6,06 6,08 6,1 6,12 6,14 InstrucRel NoInstruc Predictive NonPred 6 6,02 6,04 6,06 6,08 6,1 6,12 6,14 InstrucRel NoInstruc Predictive NonPred N = 122 Predictiveness x SOA: F(1, 120) = 3.37, p = .069 SOA 250 SOA 1000 Log RTs AC DBAC DB SOA 250 Predictiveness: F(1, 62) = 6.6, p = .013 Results Learning phase 2: RTs in the dot probe task (old stimuli) Compounds in learning phase 2: AC BD * 6 6,02 6,04 6,06 6,08 6,1 6,12 6,14 SOA 250 SOA 1000 InstrucRel NoInstruc N = 122 Instructed relevance x SOA: F(1, 120) = 2.59, p = .110 SOA: F(1, 120) = 2.76, p = .022 SOA 250 Instructed relevance: F(1, 62) = 5.43, p = .023 F/G E/H F/G E/H Log RTs Results Learning phase 2: RTs in the dot probe task (new stimuli) * * Compounds in learning phase 2: EF GH N = 122 Confidence ratings Instructed relevance: F(1, 121) = 65.59, p < .001 Compound: F(2, 242) = 11, p < .001 Instructed relevance x Compound: F(2, 242) = 8.41, p < .001 * * * ACFG EH DB Results Memory of instructions 0 1 2 3 4 5 6 7 AC EFGH BD InstrucRel NoInstruc Compound conditions Learned predictiveness produced a covert attentional bias towards predictive stimuli a very few milliseconds after the onset of stimuli. This effect vanished quickly. •The results from the dot probe revealed an attentional bias only when an SOA of 250 ms was used. Instructions could not revert or even modulate the effect of learned predictiveness. •But an attentional bias due to instructions was found for new stimuli that did not form part of any previous learning experience, which is consistent with previous demonstrations of top-down influences on rapid attentional capture (see Nordfang, Dyrholm, & Bundesen, 2012, JEP:G). Our results suggest that the learned predictiveness effect on attentional capture is (to great extent) produced by bottom-up processes out of participants' volitional control. Discussion Thank you