The shape of attention reflects flexible filtering of natural speech modulations.

This dataset provides EEG recordings, stimulus features, and behavioral data associated with the study: Huet, M.-P., & Elhilali, M. (2025). The shape of attention reflects flexible filtering of natural speech modulations. The experiment investigates how different listening goals (speech comprehension vs. sound identification) shape cortical representations of speech stimuli presented in noise. The dataset includes:- Preprocessed EEG data (2–30 Hz, downsampled to 128 Hz), segmented and time-aligned to stimulus features- Acoustic stimulus representations (spectrograms and temporal envelopes)- Behavioral responses for two tasks: keyword identification (cpt) and category detection The code to compute the modulation profiles is available on: Huet, M.-P., & Elhilali, M. (2026). MatGaborSTM (v1.0.0). Zenodo. https://doi.org/10.5281/zenodo.19830754