Publication Details
Overview
 
 
N.D. Kien, Fahimeh Akbarian, D.Q. Manh, , S. Engelborghs, Jeroen Van Schependom, N. Linh Trung,
 

Contribution to journal

Abstract 

Introduction: Cognitive deterioration is an important symptom of multiple sclerosis (MS). The golden standard for evaluating cognition is the neuropsychological examination. With regard to biomarkers, periodic components in electroencephalography (EEG) have previously been shown to carry information about cognitive state. Recently the aperiodic component of the EEG has gained interest. In this study, we examined the value of both the periodic and aperiodic components for detecting cognitive impairment in MS patients. Methods: We included 307 people with MS (PwMS) and assessed cognition using the brief repeatable neuropsychological test battery. Subjects were assigned to the “decreased cognition” (MSd) group if their score was 0, 1 or 2, which was the case for 129 PwMS. The “intact cognition” (MSi) group counted 178 PwMS. EEGs were recorded during a cognitive oddball task and converted from the time domain to power spectra. Subsequently, we used the Fitting Oscillations \& One Over F (FOOOF-tool) to extract both periodic and aperiodic components. These were analysed with logistic regression and random forest classifiers. Results: Higher aperiodic exponents were associated with impaired cognition. Optimal classification performance was 66.7\% in the standard frequency range of [2-40] Hz. Additionally, the aperiodic information showed a better performance in our classification tasks when we calculated it across a wider frequency band. Topographic EEG maps showed that cognitive deterioration was associated with a steeper 1/f slope in multiple scalp areas. Conclusions: Our findings underscore the complex relationship between non-periodic and periodic EEG parameters, highlighting their potential to distinguish cognitive states in MS patients.

Reference