Application of quantitative EEG analysis in machine learning research on cognitive impairment in Parkinson's disease: A systematic review.

OBJECTIVE

Cognitive impairments (CI) is a common non-motor symptom in Parkinson's disease (PD). With the increasing prevalence of PD, early diagnosis and therapeutic interventions for cognitive decline are crucial. This article reviews major advances in the application of machine learning (ML) to quantitative electroencephalography (qEEG) for identifying alterations associated with CI in PD.

METHODS

The publications for this review were selected according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyzes (PRISMA) guidelines with the applied search key for publications from January 2016 to March 2024.

RESULTS

Nine studies were included in the review. We summarized the characteristics of the articles, including objective, study group description, analyzed parameters of qEEG, ML models, and analysis results.

CONCLUSIONS

These findings indicate that ML using qEEG patterns has proven its effectiveness in detecting and monitoring changes connected with CI in patients with PD. Finding highly specific qEEG patterns would enable the implementation of qEEG analysis into clinical algorithms to enhance the diagnostic process and apply beneficial treatments.

SIGNIFICANCE: This review highlights the future potential and methodological challenges of integrating ML with qEEG analysis and neuropsychological diagnostics in advancing the early detection and monitoring of cognitive impairment in PD.