Predicting response to tVNS in patients with migraine using functional MRI: A voxels-based machine learning analysis

Area of Science:

• Neuroimaging
• Neurology
• Machine Learning

Background:

• Migraine diagnosis lacks objective biomarkers, leading to misdiagnosis.
• Transcutaneous vagus nerve stimulation (tVNS) shows variable efficacy in migraineurs.
• Identifying biomarkers is crucial for accurate diagnosis and personalized tVNS treatment.

Purpose of the Study:

• To identify objective biomarkers for diagnosing migraine without aura (MWoA).
• To develop a predictive model for tVNS treatment efficacy in MWoA patients.
• To correlate brain activity patterns with migraine attack frequency reduction.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was used to analyze fractional amplitude of low-frequency fluctuation (fALFF) in 70 MWoA patients and 70 controls.
• Support vector machine (SVM) and support vector regression (SVR) were employed for biomarker identification and prediction.
• Brain regions including the trigeminal cervical complex/rostral ventromedial medulla (TCC/RVM), thalamus, medial prefrontal cortex (mPFC), and temporal gyrus were analyzed.

Main Results:

• A biomarker with 3,650 features accurately discriminated MWoA patients (79.3% accuracy).
• Key discriminative regions included TCC/RVM, thalamus, mPFC, and temporal gyrus.
• A predictive model using 70 features showed a correlation coefficient of 0.36 for predicting tVNS efficacy.

Conclusions:

• Machine learning identified potential biomarkers for MWoA diagnosis and tVNS response prediction.
• Pivotal features were located in TCC/RVM, thalamus, mPFC, and temporal gyrus.
• These findings support the use of neuroimaging biomarkers for personalized migraine management.