Auditory brain–computer interfaces (BCIs) based on non-invasive electrodes offer significant potential for advancing the understanding of selective auditory attention mechanisms and enabling natural human–computer interaction. Despite growing interest in auditory attention research, publicly available datasets focused on spontaneous auditory attention switching remain limited, particularly those with high-quality electroencephalography (EEG) recordings in realistic listening environments. To address this gap, we present the Auditory Attention Switching Dataset (AASD), a non-invasive EEG dataset designed to investigate spontaneous selective auditory attention switching during naturalistic auditory processing. The dataset captures both sustained attention and spontaneous attention switching events through EEG signals. A baseline decoding model is introduced to verify data integrity and demonstrate its potential for practical applications. This open-access resource lays the foundation for developing algorithms for spontaneous auditory attention switching and advancing research in natural-scenario auditory BCIs.
Wang et al. (Thu,) studied this question.