Performance analysis in a Wavelet-Based algorithm for automatic detection of High-Voltage spindles in parkinson’s disease rat models

Abstract

The waxing-and-waning high-voltage spindles (HVSs) are the electroencephalographic hallmarks of the abnormalities in the synchronization of oscillatory activities in cortical-basal ganglia networks. The HVSs were observed during waking immobility on lesioned Sprague-Dawley rats by unilateral injection of 6-hydroxydopamine (6-OHDA). The local field potentials (LFPs) collected from the lesioned and control rats were analyzed with continuous wavelet transform using a tunable complex Morlet wavelet function identified with a careful choice of design parameters for the efficient detection of HVSs. In this study, an online detection algorithm optimized with suitable wavelet parameters using a window size of 500 and a constant decision threshold, is verified to detect the HVSs lasting 1.15-3.49 seconds from seven lesioned rats with maximum precision, sensitivity and specificity. These results provide further motivation for the real-time implementation of the automatic HVS detection systems with improved performance for pathophysiological and therapeutic applications to the thalamocortical network dysfunctions like Parkinson’s disease.