Online unsupervised performance-based cognitive testing: A feasible and reliable approach to scalable cognitive phenotyping of Parkinson's patients

Abstract

Introduction

A better understanding of the heterogeneity in the cognitive and mood symptoms of Parkinson's disease will require research conducted in large samples of patients. Fully online and remote research assessments present interesting opportunities for scaling up research but the feasibility and reliability of remote and fully unsupervised performance-based cognitive testing in individuals with Parkinson's disease is unknown. This study aims to establish the feasibility and reliability of this testing modality in Parkinson's patients.

Methods

Sixty-seven Parkinson's patients and 36 older adults completed two sessions of an at-home, online battery of five cognitive tasks and three self-report questionnaires. Feasibility was established by examining completion rates and data quality. Test-retest reliability was evaluated using the Intraclass Correlation Coefficient (ICC (2,1)).

Results

Overall completion rates and data quality were high with few participant exclusions across tasks. With regards to test-retest reliability, intraclass correlation coefficients were quite variable across measures extracted from a task as well as across tasks, but at least one standard measure from each task achieved moderate to good reliability levels. Self-report questionnaires achieved a higher test-retest reliability than cognitive tasks. Feasibility and reliability were similar between Parkinson's patients and older adults.

Conclusion

These results demonstrate that remote and unsupervised testing is a feasible and reliable method of measuring cognition and mood in Parkinson's patients that achieves levels of test-retest reliability that are comparable to those reported for standard in-person testing.