A Novel Machine Learning Approach for Classifying Resting EEG Data to Detect Possible Propensities for Executive Dysfunction
Keywords:Resting electroencephalogram (EEG), Random Forest, Executive Dysfunction, Schizophrenia, Trail Making Test
Executive dysfunction defines the lack of ability to plan, sequence, or temporally structure behavior, which is essential in daily life and thus drastically reduces the quality of life of affected individuals. Early detection assumes an immense role in treatment, and this work demonstrates that resting-state electroencephalogram (EEG) data can be used to reveal potential tendencies of executive dysfunction at an early stage. An improved machine learning (ML) algorithm was successfully used in combination with deconvolution of EEG bandwidths and the results of a neuropsychological test - the Trail Making Test (TMT) - based on a finely graded equidistant electroencephalographic subband spectrum to develop and evaluate a classification model for identifying signs of executive dysfunction. The machine learning algorithm used achieved an accuracy of 75.00%, and to the best of my knowledge, this result set a new standard. Based on this preliminary study, a means of early diagnosis of potential schizophrenia patterns from resting-state EEG data may develop to aid intervention for early executive dysfunction, suggesting potential application in the medical field.
How to Cite
The papers published in the journal are licensed under a Creative Commons Attribution (CC BY 4.0) International License.