Bypassing Direct Reconstruction: Speech Detection from MEG via Large-Scale Audio Retrieval
For researchers decoding speech from non-invasive brain signals, this work shows that leveraging external audio databases can bypass the difficult direct reconstruction problem, but the approach is task-specific and incremental.
The authors propose a two-step framework for speech detection from MEG that retrieves matching audio from a large library and then detects speech from that audio, achieving first place in the LibriBrain 2025 Speech Detection task with an F1-score of 0.962.
Decoding speech from non-invasive brain signals is challenging. For the LibriBrain 2025 Speech Detection task, we propose a novel two-step framework that bypasses direct reconstruction. First, a contrastive learning model retrieves the matching speech segment for the given test MEG from a large-scale audio library (LibriVox). Second, a speech detection model generates the binary silence/speech sequence directly from this retrieved audio. With this approach, our team Sherlock Holmes achieved first place in the extended track (F1-score: 0.962), demonstrating that leveraging external audio databases is a highly effective strategy.