Objective: To evaluate the validity and reliability of cognitive markers to identify patients with PD

Background: Parkinson’s disease (PD) is classically diagnosed by the clinical detection of motor symptoms, including bradykinesia, rigidity, tremor, and postural instability.  Clinicians rarely rely on non-motor symptoms, such as cognitive impairment, sleep disturbances, and mood changes, in the diagnostic criteria for PD. Ample evidence suggest that PD is associated with a unique cognitive profile in terms of learning accuracy and response time.

Method: Using a 10-minute computer-based game that utilizes learning from positive and negative feedback, we tested 63 patients with PD and 59 healthy individuals.  We processed the cognitive results using neurocomputational models: (1) a Q-learning algorithm for learning accuracy, and (2) a drift-diffusion algorithm for response times.

Results: Our classifier (80% training set, 20% testing set) can identify patients with PD in 98% of cases.  Our results provide a potential novel approach for the diagnosis of PD using cognitive computer games.

Conclusion: The computer-based game could be of immediate clinical relevance as an initial screening tool for PD, even before a professional neurological consult. Further, this ought to introduce non-motor symptoms of PD as an irreplaceable component of the diagnostic criteria for PD.

« Back to MDS Virtual Congress 2021

MDS Abstracts - https://www.mdsabstracts.org/abstract/cognitive-diagnostics-for-parkinsons-disease-a-novel-digitized-approach-using-non-motor-symptoms/