SMORASO-DT : A hybrid machine learning classification model to classify individuals based on working memory load in mental arithmetic task

Abstract

AbstractNonlinear dynamics and chaos theory are being widely used nowadays in neuroscience to characterize complex systems within which the change of the output is not proportional to the change applied at the input. Such nonlinear systems compared to linear systems, often appear chaotic, unpredictable, or counterintuitive, however, yet their behaviour is not mapped out as random. Thus, hidden potential of the dynamical properties of the physiological phenomenon can be detected by these approaches especially to elucidate the complex human brain activity gathered from the electroencephalographic (EEG) signals. As it is known, brain is a chaotic dynamical system and its generated EEG signals are generally chaotic because, with respect to time, the amplitude changes continuously. A reliable and non-invasive measurement of memory load, to measure continuously while performing a cognitive task, is highly desirable to assess cognitive functions, crucial for prevention of decision-making errors. Such measurements help to keep up the efficiency and productivity in task completion, work performance, and to avoid cognitive overload, especially at high mental or physical workload places like traffic control, military operations, and rescue commands. In this work, we have measured the linear and nonlinear dynamics of the EEG signals in subjects undergoing mental arithmetic task. Further, we have also differentiated the subjects who can perform a mental task good or bad, and developed a hybrid machine learning model, the SMORASO-DT (SMOte + Random forest + lASso-Decision Tree), to differentiate good and bad performers during n-back task state with an accuracy rate of 78%.