Harmonizing Healthy Cohorts to Support Multicenter Studies on Migraine Classification using Brain MRI Data.

Area of Science:

• Neuroimaging
• Machine Learning
• Data Science

Background:

• Multicenter and multi-scanner brain MRI studies are crucial for developing accurate predictive models but face challenges with data generalizability due to heterogeneity.
• Subtle differences in participant characteristics, MRI scanners, and acquisition protocols can limit the applicability of machine learning models across different datasets.
• Ensuring the generalizability of classification models is vital for reproducible results in large-scale neuroimaging research.

Conclusions:

• A "healthy core" derived from harmonized data can mitigate unwanted heterogeneity in multicenter studies.
• This approach facilitates the development of more accurate and generalizable machine learning models for neuroimaging analysis.
• The proposed strategy offers a viable solution for enhancing the reliability and reproducibility of machine learning applications in migraine research.