A new Markovian approach towards neural spike sorting

Brain is the most complicated organ of body. It controls the activity of all other organs. Understanding its function and its language could give us a direct communication pathway for connecting with injured motor organ and it could be the core of functional repairing. Neurons are the vertices of a vast network that generates the brain signals. Neuronal recordings capture brain activity signatures. The processing of these signals can help to translate brain's language. Usually it follows three main stages: spike detection and extraction, spike sorting, and intention extraction from the encoded signal. In this work, we introduce an original idea based on Hidden Markov Models (HMM) which helps to improve the spike sorting stage. Our idea is a fast and simple method which uses Inter Spike Interval information besides spike waveforms to define a Hidden Markov Model that consecutive spikes should track.