2026-10-14T15:58:18Z https://www.scidb.cn/oai

10.57760/sciencedb.38440 2026-06-03T18:10:20Z

2026-06-03 NeuBCI-AAD doi:10.57760/sciencedb.38440 en 1. Dataset OverviewExperimental ParadigmThis dataset contains EEG data collected from 20 normal-hearing participants during an auditory selective attention task. The experimental setup involves spatially separated competing speakers: one interfering speaker and one attended speaker. The audio stimuli consist of one male and one female speaker. The speech signals are synthesized using head-related transfer functions (HRTFs) and presented through headphones.During stimulus presentation, the two speakers are initially positioned on the left and right sides of the participant. Over time, their positions rotate around the participant, with one moving clockwise and the other counterclockwise. Each condition includes 60 trials, and each trial lasts approximately 1 minute.EEG Acquisition and PreprocessingEEG signals were recorded using a NeuroScan system with a 64-channel cap arranged according to the international 10/20 system. The sampling rate is 1000 Hz, and electrode impedance is maintained below 20 kΩ.2. Dataset File OrganizationThe stimuli folder contains the audio stimuli used in the experiment. The file trial1m.wav represents the speech signal of the male speaker in the first trial, while trial1f.wav represents the speech signal of the female speaker (similarly for other trials).There are 20 folders named sub1 to sub20, corresponding to the 20 participants.For each participant, the processed_data folder contains preprocessed EEG data files named in the format trial_{trial_id}.npz.Each EEG file for a single trial includes a field named eeg_trial, which is a NumPy array of shape [n_channels, n_samples], representing the EEG recordings.The raw_data folder for each participant contains the original EEG data in .cnt format. If a log file is present, it records events such as trial termination or repetition, which should be considered during preprocessing.The answer folder contains files named sub{id}.xlsx, which store participants’ responses. These responses are used to assess attention levels. Additionally, the trial column in the Excel file indicates which audio stream the participant was instructed to attend to during each trial. For example, 5f indicates that in the 5th trial, the participant was instructed to attend to the female speaker. auditory attention decoding; EEG; speech chen xiao yu He Huiguang PUBLIC https://creativecommons.org/licenses/by-sa/4.0/ dataset Science Data Bank