Sign in Subscribe
Machine learning

Deep Learning and Artificial Intelligence in Personalized Medicine

Strategic Insights

— 1 min read

Researchers from Sichuan University in Chengdu, People's Republic of China, have made significant advancements in using deep learning and artificial intelligence to diagnose migraine, a neuropsychiatric disorder. By combining multiple functional magnetic resonance imaging (fMRI) indicators with explainable artificial intelligence (XAI) techniques, the researchers aimed to improve the accuracy of migraine classification and identify discriminative brain regions. The study analyzed resting-state fMRI data from 64 participants, including patients with migraine without aura, with aura, and healthy controls.

Key Takeaways:

  • The GoogleNet model combined with regional functional connectivity strength (RFCS) indicators achieved the best classification performance, with an accuracy of 98.44% and an area under the curve of 0.99 for the test set.
  • The RFCS indicator improved accuracy by approximately 8% compared with the amplitude of low-frequency fluctuation.
  • Brain activation heat maps revealed that the precuneus and cuneus were the most discriminative brain regions, with slight activation also observed in the frontal gyrus.
  • The use of XAI technology combined with brain region features provides visual explanations for the progression of migraine in patients.
  • This study demonstrates the potential of XAI in clinical applications, hindering biomarker discovery and personalized treatment.
  • The findings of this study can aid in the development of new diagnostic techniques and enhance diagnostic accuracy.

Statistics:

  • 98.44% accuracy rate achieved by the GoogleNet model combined with RFCS indicators.
  • 0.99 area under the receiver operating characteristic curve achieved by the GoogleNet model combined with RFCS indicators.
  • 8% improvement in accuracy by the RFCS indicator compared with the amplitude of low-frequency fluctuation.
  • 64 participants included in the study, comprising 21 patients with migraine without aura, 15 patients with migraine with aura, and 28 healthy controls.
  • 3 fMRI metrics-amplitude of low-frequency fluctuation, regional homogeneity, and RFCS-were extracted and classified using GoogleNet, ResNet18, and Vision Transformer.

Sources:

  • NewsRx. New Personalized Medicine Data Have Been Reported by Researchers at Sichuan University (Interpretable Artificial Intelligence Analysis of Functional Magnetic Resonance Imaging for Migraine Classification: Quantitative Study). Drug Week. September 19, 2025; p 3851.
  • Huang Y, et al. (2025) Interpretable Artificial Intelligence Analysis of Functional Magnetic Resonance Imaging for Migraine Classification: Quantitative Study. JMIR Medical Informatics, 13.

Read more