Spatio-spectral patterns based on Stein kernel for EEG signal classification

Attention-Deficit/Hyperactivity Disorder (ADHD) is a childhood-onset neurological disorder that can persist in adolescence and adult life, reducing concentration, memory, and productivity. The main drawback with mental health abnormalities of this type is the traditional diagnostic technique. Since...

Descripción completa

Detalles Bibliográficos
Autor: Galindo Noreña, Steven
Tipo de recurso: tesis de maestría
Estado:Versión publicada
Fecha de publicación:2021
País:Colombia
Institución:Universidad Tecnológica de Pereira
Repositorio:Repositorio Institucional UTP
Idioma:inglés
OAI Identifier:oai:repositorio.utp.edu.co:11059/13163
Acceso en línea:https://hdl.handle.net/11059/13163
Access Level:acceso abierto
Palabra clave:Neurofisiología
Electroencefalografía
Procesamiento de imágenes
Descripción
Sumario:Attention-Deficit/Hyperactivity Disorder (ADHD) is a childhood-onset neurological disorder that can persist in adolescence and adult life, reducing concentration, memory, and productivity. The main drawback with mental health abnormalities of this type is the traditional diagnostic technique. Since this is based exclusively on a symptomatological description without considering any biological data, leading to high overdiagnosis rates. To address the above problem, clinical researchers are attempting to extract ADHD biomarkers from recorded electroencephalographic (EEG) signals. Among the most common biomarkers are Theta/Beta Ratio and P300, of which recent studies have shown a lack of significance on the differences between ADHD and control subjects. Besides, another great challenge in EEG processing is given by the sensitivity of the signals, since they can be easily affected by background noise, muscle artifacts, head movements and flickering that greatly impair their quality, which limits its introduction into real world applications. This work proposes an EEG signal representation methodology for identifying subject-wise discrepancies of inhibitory responses, decoding the data structure, and supporting diagnosis of mental disorders. For this, first we develop a feature extraction approach based on the common spatial patterns (CSP) from EEG signals to support the ADHD diagnosis as show in Chapter...