Automatic Spike-Wave discharge detection within EEG of mice models of human absence epilepsy
Year: 2020
Type: dissertation
Abstract: Recent studies link Alzheimer disease (AD) with epilepsy.Spike-wave discharges (SWD) found in electroencephalographic recordings of mice models of human absence epilepsy are pathological phenomena.Its automatic detection is a key tool for evaluation of present and future drug studies addressing both epilepsy and AD.In this study is proposed a new promising method of automatic SWD detection.Investigated parameters describing SWD features are time and frequency scale based applying morlet wavelet transform.Classification was done implementing random forest method and its proximity matrix.Comparing proximities to positively and negatively annotated observations provided interesting results.Sensitivity and precision achieved on a validation dataset was 85.7 % with 97.4 %, respectively.However, this was estimated using judgement of only one expert.This study revealed a lack in a SWD definition and further investigation is needed.Nevertheless, this study offers a tool to achieve correct classification results if correct annotation is used.Moreover, the proposed algorithm can work as semi-automatic tool requiring human input saving countless of hours by preselecting a small group of candidates sharing SWD-like features.
Author: Michal Demjan
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W3202838336