A common neural architecture for encoding finger movements

Read the full article See related articles

Listed in

This article is not in any list yet, why not save it to one of your lists.
Log in to save this article

Abstract

Finger tapping is widely used to investigate sensorimotor brain responses, yet it remains unclear whether shared neural codes for complex finger movements exist across individuals, given the highly individual-specific topography of the sensorimotor cortex. We combined spherical searchlight Procrustes hyperalignment on 7-Tesla fMRI data with between-subject classification to investigate the existence of a shared latent neural architecture encoding sequential finger movements in the human sensorimotor cortex. Our model afforded above-chance level accuracy in both the contralateral and ipsilateral areas, suggesting that such a high-dimensional neural information space exists, is traceable, and spreads over the sensorimotor cortex. Critically, these high-dimensional neural codes were identified without task-specific training and cannot be attributed to error-related activity. These findings establish universal organizational principles of motor sequence encoding and pave the way for personalized neurotechnologies, scalable brain-computer interfaces, and cross-subject models for motor learning and rehabilitation.

Article activity feed