ADVANCEMENTS IN ELECTROENCEPHALOGRAPHY AND COGNITIVE PROFILING FOR PARKINSON’S DISEASE

Abstract

Background: Parkinson's disease (PD) is a neurological defect that causes both non-motor (brain-related) and motor (physical) symptoms. As PD progresses, brain function (cognition) may diminish in some cases. Over recent times, quantitative electroencephalographic (QEEG) has emerged as a valuable tool for investigating the growth of cognition, psychological variables and psychopathological illnesses (mental disorders).

Objective: The principal objective of this study is to convey knowledge about analysis methodologies, novel approaches and practical information to those with little expertise, with future directions.

Methods: This paper covers the EEG preprocessing steps with methods, analysis techniques, transformation methods and classification details from the research conducted between 2006 to 2024. There exists no study categorizing the signal processing techniques based on the different domains. Therefore, to fill this gap, this paper focuses particularly on the analysis categories that are time-frequency domain (TFD), frequency domain (FD), time domain (TD), graph theory-based, network dynamics, functional connectivity and connectivity analysis techniques with their strengths and limitations. Additionally, an experiment is carried out to investigate the ability of EEG to categorize cognitive processes and the relevancy of machine learning approaches in determining the best classifier to handle EEG data.

Conclusion: Analysis of 59 studies revealed that non-neural machine learning techniques are the most frequently employed. Additionally, the highest accuracy rates from recent studies are identified and the usage percentages for each model are also calculated. Therefore, the purpose of this study is to provide researchers with an initial base for future research on cognition in PD.