machine learning21 min read
Train Once, Reuse Later: Three Tests of Skipping BCI Calibration
Same person, a labeled earlier session, no new labels: a frozen foundation model scored above a spectral baseline, a target detector ranked less well by the last visit, and a fixed ridge for cursor tracking did worse than a constant.
By Tianyao WuTowards AI
Read the full story
Originally published by Towards AI on Oct 17, 2026.
Comments
Sign in to join the discussion.