Classification of Functional Movement Disorders with Resting-State Functional Magnetic Resonance Imaging.

Area of Science:

• Neuroscience
• Neurology
• Machine Learning in Medicine

Background:

• Functional movement disorder (FMD) is a subtype of functional neurological disorder.
• FMD is characterized by abnormal movements not perceived as self-generated.
• Previous imaging studies suggest altered brain activity in emotional, motor control, and agency regions.

Purpose of the Study:

• To characterize brain connectivity alterations in FMD using machine learning.
• To build a classifier to distinguish FMD patients from healthy controls.
• To identify predictive neuroimaging features of FMD.

Main Methods:

• Used resting-state functional magnetic resonance imaging (fMRI) from 61 FMD patients and 59 controls.
• Selected 66 seed regions and computed the full correlation matrix.
• Employed recursive feature elimination to identify predictive features for a support vector machine classifier.

Main Results:

• Identified 29 highly predictive features for FMD.
• Achieved 80% accuracy in classifying FMD patients versus healthy controls.
• Key predictive regions included the right sensorimotor cortex, left dorsolateral prefrontal cortex, left cerebellum, and left posterior insula.

Conclusions:

• Machine learning successfully identified distinct functional connectivity patterns in FMD.
• Altered functional linkages between emotion, reward, and sensorimotor integration regions are crucial in FMD.
• These findings enhance understanding of FMD pathophysiology and may aid diagnosis and treatment.