13 - Automated neuroscience decision support framework

Abstract

Psychophysiological chronic disorders lead to both physical disorders and emotional factors, such as anxiety and stress. This requires patients to live under medical treatments for the rest of their lives. Most of them have a genetic influence, which is observable in children at an early age. Therefore, the early identification of a disorder is important to reducing the negative consequences of adulthood. Machine learning and deep learning techniques have been actively applied in recent years based on neuroimaging inputs, such as electroencephalogram and magnetic resonance imaging, to find feasible computational solutions. However, recent studies lack the support for multiple disorders by rather focusing only on a single disorder even though there are commonalities among many of the psychophysiological chronic disorders. This chapter addresses this research hindrance by proposing a neuroscience decision support system model with the aid of machine learning and deep learning to identify these disorders when neuroimaging inputs are given. The experiment-based evaluation has shown an accuracy of over 0.86 F1 measure for different types of classifiers based on three functional connectivity types.