Imagined speech

This dataset contains electroencephalography (EEG) recordings collected for imagined speech classification research. Participants were instructed to silently imagine predefined words without any vocalization or physical movement while EEG signals were recorded under controlled experimental conditions. The dataset includes eight imagined speech classes: Water, Toilet, Pain, Doctor, Move, Silent, Yes, and No. EEG signals were acquired using an RMS SuperSpec 21-channel EEG acquisition system from 10 healthy participants aged 18–21 years. The recordings preprocessed and are provided in CSV format, with each file corresponding to an individual imagined speech trial. The dataset is intended for research in imagined speech decoding, brain-computer interfaces (BCIs), EEG signal processing, and machine learning/deep learning applications.