Spike Detection in Deep Brain Stimulation Surgery with Convolutional Neural Networks

The paper addresses issues concerning the application of deep learning in deep brain stimulation (DBS) neurosurgery. DBS is a functional neurosurgical procedure used to treat conditions in the central nervous system that arise from improper physiology, such as Parkinson’s disease. Electrostimulation, carried out by implanting electrodes into identified regions in the brain, makes it possible to reduce the symptoms of this disease significantly. This paper uses a convolutional neural network to analyze recordings of neuronal activity acquired during DBS neurosurgery and spikes (neuronal activity) detection. The experimental results on real data demonstrate that our method allows for spike detection with high accuracy. It can successfully support neurosurgeons during surgery in the optimal placement of electrodes in the brain and monitor the effects of DBS treatment over time.