LibriBrain: Over 50 Hours of Within-Subject MEG to Improve Speech Decoding Methods at Scale

Miran Özdogan (University of Oxford) · Gilad Landau (University of Oxford) · Gereon Elvers (University of Oxford) · Dulhan Jayalath (University of Oxford) · Pratik Somaiya (University of Lincoln) · Francesco Mantegna (University of Oxford) · Mark Woolrich (University of Oxford) · Oiwi Parker Jones (University of Oxford)

baseline resultsbrain-computer interfacesdata splitsdecoding performancedeep learning frameworkshigh-quality recordingslibribrainmeg datasetnaturalistic spoken englishneural representationsnon-invasive methodsphoneme classificationreproducibilityspeech decodingspeech detectionword classification

LibriBrain represents the largest single-subject MEG dataset to date for speech decoding, with over 50 hours of recordings