Classification of tic disorders based on functional MRI by machine learning: a study protocol.

Area of Science:

• Neuroscience
• Pediatrics
• Computational Psychiatry

Background:

• Tic disorders (TD) are common neurodevelopmental conditions in children, with subtypes including provisional tic disorder (PTD), chronic motor or vocal TD (CMT/CVT), and Tourette syndrome (TS).
• Early diagnostic classification of TD subtypes is challenging based solely on new-onset symptoms.
• Machine learning (ML) shows promise for early diagnostic classification using functional MRI (fMRI), but models for TD subtypes are scarce.

Purpose of the Study:

• To develop and validate a machine learning model for the early diagnostic classification of TD subtypes.
• To differentiate between PTD, CMT/CVT, and TS in children using neuroimaging data.
• To establish a protocol for early TD subtype identification.

Main Methods:

• Recruitment of 200 children (aged 6-9) with new-onset tic symptoms and 100 healthy controls.
• Resting-state fMRI data acquisition for all participants.
• Development of a support vector machine (SVM) model based on functional connectivity for classification.

Main Results:

• The study protocol outlines the methodology for building an SVM model using fMRI data.
• The model aims to classify TD subtypes, aiding in early diagnosis.
• Specific results regarding classification accuracy will be available upon study completion.

Conclusions:

• This study protocol provides a framework for utilizing ML and fMRI for early TD subtype classification.
• The developed SVM model has the potential to improve diagnostic accuracy and timely intervention for children with TD.
• Publication of trial results in peer-reviewed journals is planned.