Multivariate pattern classification of pediatric Tourette syndrome using functional connectivity MRI.

Area of Science:

• Neuroscience
• Psychiatry
• Medical Imaging

Background:

• Tourette syndrome (TS) is a neurodevelopmental disorder presenting with motor and vocal tics.
• Predicting symptom trajectory and treatment response in TS is crucial for patient care.
• Current diagnostic methods may not fully capture the complexity of brain network alterations in TS.

Purpose of the Study:

• To investigate if resting-state functional connectivity (RSFC) patterns can predict diagnostic group membership in children with TS using machine learning.
• To evaluate the efficacy of multivariate pattern analysis (support vector machine - SVM) compared to univariate methods for TS classification.
• To introduce a novel SVM adaptation for individual classification confidence assessment.

Main Methods:

• Recruitment of 42 children with TS and 42 age-, IQ-, and movement-matched healthy controls.
• Acquisition and analysis of resting-state functional connectivity (RSFC) MRI data.
• Application of support vector machine (SVM) classification to differentiate between TS and control groups.

Main Results:

• Univariate statistical tests did not reveal significant group differences in RSFC.
• SVM classification achieved approximately 70% accuracy in distinguishing between children with TS and controls (p < .001).
• A novel SVM adaptation provided a confidence measure for individual classification accuracy.

Conclusions:

• Multivariate methods, such as SVM, can effectively identify complex brain network patterns associated with Tourette syndrome.
• These findings suggest that RSFC patterns hold potential for predicting TS prognosis and treatment outcomes.
• This study highlights the utility of advanced machine learning techniques in understanding and potentially diagnosing neurodevelopmental disorders.