Graph Convolutional Networks Reveal Neural Connections Encoding Prosthetic Sensation

TL;DR

This study applies graph convolutional networks to neuronal data from a rat with a neuroprosthesis, successfully predicting stimulation parameters and revealing neural circuitry involved in encoding artificial sensory information.

Contribution

It introduces the use of GCNs to interpret neuronal ensembles in neuroprosthetic applications, demonstrating improved decoding of sensory stimuli and insights into neural connectivity.

Findings

GCNs achieved 73.5% accuracy in classifying stimulation frequency.

Inferred neural adjacency matrix reflects underlying circuitry.

Method enhances decoding and understanding of artificial sensory encoding.

Abstract

Extracting stimulus features from neuronal ensembles is of great interest to the development of neuroprosthetics that project sensory information directly to the brain via electrical stimulation. Machine learning strategies that optimize stimulation parameters as the subject learns to interpret the artificial input could improve device efficacy, increase prosthetic performance, ensure stability of evoked sensations, and improve power consumption by eliminating extraneous input. Recent advances extending deep learning techniques to non-Euclidean graph data provide a novel approach to interpreting neuronal spiking activity. For this study, we apply graph convolutional networks (GCNs) to infer the underlying functional relationship between neurons that are involved in the processing of artificial sensory information. Data was collected from a freely behaving rat using a four infrared (IR) sensor, ICMS-based neuroprosthesis to localize IR light sources. We use GCNs to predict the stimulation frequency across four stimulating channels in the prosthesis, which encode relative distance and directional information to an IR-emitting reward port. Our GCN model is able to achieve a peak performance of 73.5% on a modified ordinal regression performance metric in a multiclass classification problem consisting of 7 classes, where chance is 14.3%. Additionally, the inferred adjacency matrix provides a adequate representation of the underlying neural circuitry encoding the artificial sensation.