BCIwiki (bciwiki.com) — A 32-electrode electrocorticography (ECoG) grid, placed in the right spot, classified spoken words as accurately as high-density grids of 64 to 128 electrodes. Researchers at University Medical Center Utrecht Brain Center and partner institutions reported the finding in Clinical Neurophysiology on June 10, 2026 (PMID 42263493).
The team analyzed data from 8 able-bodied participants who had high-density ECoG grids over the sensorimotor cortex and repeated 12 Dutch words. Saying the words produced high-frequency activity in 2 foci on the sensorimotor cortex, one ventral and one dorsal.
Rectangular 32-electrode configurations covering 325 to 561 square millimeters reached a mean word classification accuracy of 76 ± 16% across participants, compared with 75 ± 17% for the larger grids. The practical chance level was 16.7%. The best configurations were oriented vertically and centered on the central sulcus.
The authors said an optimally placed 32-electrode grid may be enough for high accuracy on a closed set of words. They said this supports targeted placement of small grids, which would reduce the surgical burden on users, and simpler, more energy-efficient fully implantable brain-computer interface (BCI) devices for people with severe paralysis. The participants were not paralyzed, so the findings have yet to be tested in that population.