Precision Targeting Before Surgery

The challenge facing neurosurgeons who implant brain-computer interfaces has been geographic: which square millimeters of cortex contain the richest signal for decoding movement? A new study using 7-Tesla fMRI scans provides the most detailed answer yet, mapping how well different regions of sensorimotor cortex can distinguish between 20 different hand gestures performed by ten able-bodied participants.

The results point clearly to the hand region of the sensorimotor cortex, confirming what many clinicians suspected but quantifying it with unusual precision. More valuable than this confirmation, though, are two secondary findings that could reshape how research teams prepare for implantation.

First, the researchers found that six carefully chosen gestures could predict the optimal decoding location for the full set of twenty. This matters because pre-surgical mapping sessions are exhausting for patients with severe motor impairments. The ability to identify ideal electrode targets using a abbreviated gesture set could reduce mapping time by more than half while maintaining accuracy.

Surface Recordings May Suffice

The second finding challenges assumptions about electrode depth. Using support vector machines to measure decodability across cortical regions, the team discovered that sulcal areas, the valleys between brain folds, contributed no unique information beyond what adjacent gyral regions (the peaks) already provided. The same pattern held in postcentral cortex, where decoding was primarily driven by gyral activity.

This has immediate practical implications. Electrodes that rest on the cortical surface are surgically simpler to place and carry lower risk of tissue damage than those penetrating into sulci. If surface recordings capture the same information, the risk-benefit calculation shifts.

The study focused on able-bodied participants, which means the findings require validation in clinical populations where motor cortex may have reorganized after injury or disease. But the methodological approach itself, high-resolution fMRI combined with machine learning decodability analysis, offers a template for individualized pre-surgical planning.

For individuals with locked-in syndrome or high-level spinal cord injuries, the difference between adequate and optimal electrode placement can determine whether a BCI provides functional communication or falls short. These findings narrow the target.