Cognitive Symptoms in Cross-Sectional Parkinson Disease Cohort Evaluated by Human-in-the-Loop Machine Learning and...

Area of Science:

• Neurology
• Cognitive Science
• Data Science

Background:

• Cognitive impairment affects up to 80% of individuals with Parkinson disease (PD).
• The subjective experience and frequency of cognitive problems in PD patients remain under-researched.
• Understanding patient-reported cognitive symptoms is crucial for comprehensive PD care.

Purpose of the Study:

• To describe the types and frequency of bothersome cognitive symptoms reported by PD patients in their own words.
• To analyze the verbatim patient-reported data using advanced NLP and machine learning techniques.
• To identify associations between cognitive symptoms and demographic/disease-related factors.

Main Methods:

• Utilized the online Fox Insight study and Parkinson Disease Patient Report of Problems.
• Collected verbatim self-reported bothersome problems from PD patients.
• Employed human-in-the-loop curation, NLP, and machine learning to categorize 8 cognitive symptom domains.
• Conducted multivariate logistic regression to examine symptom associations.

Main Results:

• 32% of 25,192 PD participants reported cognitive symptoms.
• Most frequent symptoms: memory (13%), language/word finding (12%), concentration/attention (9%).
• Depression was linked to most cognitive domains; higher education and disease duration also predicted symptom reporting.

Conclusions:

• Nearly one-third of PD patients report cognitive symptoms as highly bothersome, even early in the disease.
• Large-scale online verbatim reporting is feasible for detailed cognitive symptom analysis in PD.
• Findings highlight the significant impact of cognitive issues on PD patients' lives.