Mathematical methods of signal analysis applied in medical diagnostic (Q2023595)

The paper is focused on the precise localization of the area of the subthalamic nucleus (STN) through the use of signal processing techniques and classification algorithms. The paper is comprehensive and self-contained by delivering a detailed discussion of the main phases of signal processing, namely filtering: specialized spike activity filters, background activity-based filters, moving average filters, wavelet decomposition, and artifact filtering. A classifier with an objective of discriminating between recordings registered within STN and outside STN is Adaboost while a comparative analysis is realized with the use of support vector machines endowed with a collection of kernels. The study covers detailed algorithms and provides a suite of experimental results.