LibriBrain100 Provides 100 Hours of MEG Data for Speech Decoding
August 24, 2026
LibriBrain100 is a new dataset featuring over 100 hours of high-quality Magnetoencephalography (MEG) data recorded during naturalistic speech. It offers significantly more depth than previous releases, particularly for within-subject neural data scaling.
HOW THIS AFFECTS YOU
●
researcherYou can leverage a much larger dataset for training neural speech decoding models.
●
healthThis advances the development of noninvasive brain-to-text interfaces.