Investigation of multi-site micro recordings of subthalamic nucleus neurons using machine learning MER with DBS in Parkinson`s – A simulation study

Abstract

Multineural spikes were acquired with a multisite electrode placed in the hippocampus pyramidal cell layer of non-primate anesthetized snitch animals. If the impedance of each electrode-site is relatively low and the distance amongst electrode sites is appropriately miniatured, a spike generated by a neuron is parallelly recorded at multielectrode sites with different amplitudes. The covariance between the spike of the at each electrode-point and a template was computed as a damping-factor due to the volume conduction of the spike from the neuron to electrode-site. Computed damping factors were vectorized and analyzed by simple but elegant hierarchical-clustering using a multidimensional statistical-test. Since a cluster of damping vectors was shown to correspond to an antidromically identified neuron, spikes of distinct neurons are classified by suggesting to the scatterings of damping vectors. Errors in damping vector computing due to partially overlapping spikes were minimized by successively subtracting preceding spikes from raw data. Clustering errors due to complex-spike-bursts (i.e., spikes with variable-amplitudes) were prevented by detecting such bursts and using only the first spike of a burst for clustering.