Discriminating cognitive status in Parkinson's disease through functional connectomics and machine learning

Area of Science:

• Neuroscience
• Medical Imaging
• Machine Learning

Background:

• Neurodegenerative diseases like Parkinson's disease (PD) pose challenges for early diagnosis.
• Non-invasive biomarkers are crucial for monitoring disease progression and cognitive status.
• Functional connectivity patterns from neuroimaging show potential for identifying neurological disorders.

Purpose of the Study:

• To investigate the utility of connection-wise functional connectivity patterns in distinguishing Parkinson's disease patients based on cognitive status.
• To develop and validate a machine learning model for classifying PD patients with and without mild cognitive impairment (MCI).

Main Methods:

• Resting-state functional magnetic resonance imaging (fMRI) was employed on two independent cohorts (training and validation).
• The Brainnetome atlas was used for functional connectome reconstruction.
• A support vector machine (SVM) classifier, trained with features selected via randomized logistic regression and leave-one-out cross-validation, was utilized.

Main Results:

• The machine learning model achieved 82.6% accuracy in discriminating PD patients with MCI from those without in the training sample.
• The model demonstrated 80.0% accuracy when applied to the independent validation sample.
• Key connectivity patterns identified by the model correlated with memory and executive function performance in PD patients.

Conclusions:

• Connection-wise functional connectivity patterns are a promising non-invasive biomarker for differentiating Parkinson's disease patients with cognitive deficits.
• This neuroimaging approach holds potential for improving early detection and management of cognitive impairment in Parkinson's disease.
• Machine learning applied to functional connectivity data can aid in understanding and classifying neurological conditions.