EEG Motor Intention Dataset for Rehabilitation Oriented Brain Computer Interfaces

Centro en Ingeniería en Rehabilitación e Investigaciones Neuromusculares CIRINS - FIUNER EEG Motor Intention Dataset for Rehabilitation-Oriented BCIs Dataset overview This dataset contains electroencephalographic (EEG) recordings acquired from 30 healthy volunteers performing motor intention tasks involving both upperand lower-limb movements. The dataset was designed to support the development and evaluation of brain–computer interface (BCI) algorithms for motor rehabilitation under realistic clinical constraints, including a reduced number of EEG channels and limited calibration data. The dataset was acquired under a protocol approved by the Central Bioethics Committee of the Province of Entre Ríos, Argentina, and conducted in accordance with the Declaration of Helsinki. The study was registered in the National Registry of Health Research under number IS004678 and in ClinicalTrials.gov (Clinical Trials ID: NCT06861517). Two complementary EEG datasets are included. The first dataset consists of full multichannel recordings from 30 healthy adult participants, acquired using a 32-channel EEG montage following the international 10–20 system. This dataset preserves the complete spatial information and is intended for methodological development and exploratory analyses. The second dataset is a curated subset derived from the original recordings and includes data from 24 participants selected based on signal quality. This reduced dataset is restricted to five motor-related electrodes (Cz, C3, C4, Fz, and Pz) and was specifically designed to emulate lowdensity EEG setups commonly encountered in clinical rehabilitation environments. This subset is the one used in the accompanying research article. Both datasets include structured motor tasks for upper and lower limbs. Lower-limb tasks involved ankle dorsiflexion movements, while upper-limb tasks consisted of repeated opening and closing of the hand. Tasks were presented using visual and auditory cues and repeated across multiple trials, with randomized rest intervals to reduce anticipatory effects. Participants The dataset includes EEG recordings from thirty healthy adult volunteers, with a mean age of 28 ± 6.6 years, including sixteen female and fourteen male participants. All participants provided written informed consent prior to data acquisition. Inclusion criteria comprised the absence of known neurological or motor disorders, while exclusion criteria included a history of neurological disease or conditions affecting motor control. (See Dataset_Info.xlsx) EGG acquisition EEG signals were recorded using an eego™ amplifier equipped with a 32-channel EEG cap arranged according to the international 10–20 system. Data were sampled at 500 Hz, and electrode impedances were kept below 5 KΩ throughout the recording sessions. The amplifier Centro en Ingeniería en Rehabilitación e Investigaciones Neuromusculares CIRINS - FIUNER also featured eight independent TTL trigger input lines to enable precise synchronization of experimental events. (Left, Right and Rest) Experimental protocol Lower-Limb Motor Tasks Protocol Informative video: https://youtu.be/ERjb_XSgAcY General posture Participants were instructed to sit comfortably in front of the screen with both feet placed flat on the floor. This position was defined as the resting posture and was maintained throughout the experiment unless movement was explicitly instructed. Task 1 – Visually cued ankle dorsiflexion In Task 1, participants performed ankle dorsiflexion movements guided by visual cues. The movement consisted of lifting the tip of the foot as much as possible while keeping the heel in contact with the floor and then returning to the resting position. Three visual cues were presented on the screen to indicate whether the participant should move the left foot, the right foot, or remain at rest. Each trial began with a rest period of random duration uniformly distributed between 5 and 6 seconds, followed by a movement period lasting 2.5 seconds. A total of ten trials per foot were presented in a randomized order, always interleaved with rest periods between consecutive trials. This task sequence was repeated four times, resulting in multiple repetitions of each condition. Task 2 – Auditory-cued ankle dorsiflexion Task 2 focused on ankle dorsiflexion movements executed in response to an auditory cue. Participants were instructed to perform the movement only with the foot indicated by the system. The onset of each movement was triggered by a beep sound. As in Task 1, each trial started with a rest period of random duration between 5 and 6 seconds. The dorsiflexion movement was then performed during a 2.5-second movement window. For each repetition, the movement was executed ten times, and the entire task was repeated three times. Rest intervals were included between repetitions to minimize fatigue and maintain signal quality. Centro en Ingeniería en Rehabilitación e Investigaciones Neuromusculares CIRINS - FIUNER Task 3 – Auditory-cued ankle dorsiflexion with feedback Task 3 followed the same movement and auditory cue structure as Task 2, with the addition of visual feedback. Participants were instructed to perform the dorsiflexion movement exclusively with the indicated foot upon hearing the beep. In this task, the rest period preceding each movement had a longer duration of approximately 8 seconds, accounting for the presence of the visual animation and feedback processing. Following the rest period, participants performed continuous ankle dorsiflexion movements for a duration of 2.5 seconds. As in Task 2, the movement sequence consisted of ten repetitions and the task was repeated three times. During the movement period, an animation was displayed on the screen whenever the system detected the participant’s movement, providing real-time feedback. Summary Across all tasks, ankle dorsiflexion movements of the lower limb were performed under controlled visual and auditory cueing conditions, with structured repetitions and rest periods. This design allowed the acquisition of EEG signals associated with rest, motor preparation, and movement execution or attempt, supporting subsequent single-trial analysis and the evaluation of BCI algorithms under realistic experimental conditions. Upper-Limb Motor Tasks Protocol Informative video: https://youtu.be/TTc038NYwMQ The experimental protocol for upper-limb motor tasks followed the same structure as the lower-limb protocol described above, including task organization, visual and auditory cueing, timing of rest and movement periods, and number of repetitions. This design ensured consistency across limb conditions within the dataset. The only difference concerned the type of movement performed. For upper-limb tasks, participants were instructed to repeatedly open and close the dominant hand, instead of performing ankle dorsiflexion. All other aspects of the protocol, including randomized rest durations, movement duration, auditory beeps, and visual feedback conditions, were kept identical. Although both upperand lower-limb recordings are included in the dataset, the accompanying article focuses exclusively on the lower-limb condition. Data format and organization EEG recordings are provided in FIF format compatible with MNE-Python. Metadata are provided in an accompanying Excel file as a lightweight alternative to the BIDS standard. Centro en Ingeniería en Rehabilitación e Investigaciones Neuromusculares CIRINS - FIUNER Intended use The dataset is intended for research and educational purposes, particularly for EEG-based BCI development. Citation Quiroga Forero, A., Acevedo, R. C., Rufiner, H. L. EEG Motor Intention Dataset for RehabilitationOriented BCIs. Zenodo, 2025.