Classification of movements of the rat based on intra-cortical signals using artificial neural network and support vector machine

A BCI aims at creating a communication pathway between the brain and an external device. This is possible by decoding signals from the primary motor cortex and translating them into commands for a prosthetic device. The experimental design was developed starting from intra-cortical signal recorded in the rat brain. The data pre-processing included denoising with wavelet technique, spike detection, and feature extraction. Artificial neural network and support vector machine were applied to classify the rat movements into two possible classes, Hit or No Hit. The misclassification error rates from denoised and not denoised data were statistically different (p<0.05), proving the efficiency of the denoising technique. ANN and SVM gave comparable classification results