線形相関モデルを用いた運動野神経信号によるラット歩行状態推定

概要

論文の詳細を見る
We are developing a brain-machine interface (BMI) called “RatCar," a small vehicle controlled by the neural signals of a rats brain. An unconfined adult rat with a set of bundled neural electrodes in the brain rides on the vehicle. Each bundle consists of four tungsten wires isolated with parylene polymer. These bundles were implanted in the primary motor and premotor cortices in both hemispheres of the brain. In this paper, methods and results for estimating locomotion speed and directional changes are described. Neural signals were recorded as the rat moved in a straight line and as it changed direction in a curve. Spike-like waveforms were then detected and classified into several clusters to calculate a firing rate for each neuron. The actual locomotion velocity and directional changes of the rat were recorded concurrently. Finally, the locomotion states were correlated with the neural firing rates using a simple linear model. As a result, the abstract estimation of the locomotion velocity and directional changes were achieved.
社団法人電気学会の論文
2007-10-01