Direct Cortical Control of Primate Whole-Body Navigation in a Mobile Robotic Wheelchair

TL;DR

This study demonstrates that rhesus monkeys can learn to control a robotic wheelchair using cortical activity, showing potential for restoring mobility in paralyzed humans through brain-machine interfaces.

Contribution

First demonstration that cortical ensembles can control a whole-body robotic device for navigation in primates.

Findings

Monkeys learned to navigate toward a food reward using cortical signals.

Neuronal activity modulated with whole-body displacements and distance to target.

Cortical control enabled effective and improved wheelchair navigation over training.

Abstract

We and others have previously developed brain-machine-interfaces (BMIs), which allowed ensembles of cortical neurons to control artificial limbs (1-4). However, it is unclear whether cortical ensembles could operate a BMI for whole-body navigation. Here we show that rhesus monkeys can learn to navigate a robotic wheelchair while seated on top of it, and using their cortical activity as the robot control signal. Two monkeys were chronically implanted with multichannel electrode arrays which simultaneously sampled activity of roughly 150 premotor and sensorimotor cortex neurons per monkey. This neuronal ensemble activity was transformed by a linear decoder into the robotic wheelchair's translational and rotational velocities. During several weeks of training, monkeys significantly improved their ability to navigate the wheelchair toward the location of a food reward. The navigation was enacted by ensemble modulations attuned to the whole-body displacements, and also to the distance to the food location. These results demonstrate that intracranial BMIs could restore whole-body mobility to severely paralyzed patients in the future.