Baseline Brain Gray Matter Volume as a Predictor of Acupuncture Outcome in Treating Migraine

This pioneering study explores one of the major challenges of modern acupuncture: predicting which patients will respond better to treatment. Migraine without aura, known as MwoA, is a complex neurological condition that affects more than one billion people globally and represents the second leading cause of disability worldwide. Characterized by recurrent headaches, often unilateral, and accompanied by symptoms such as nausea, vomiting, light sensitivity, and sound sensitivity, this condition significantly affects patients' quality of life. Although acupuncture is widely used to treat migraine in many countries, approximately half of patients do not experience substantial improvement, creating the need to identify in advance who will benefit from treatment.

Chinese investigators conducted an innovative study with 41 patients diagnosed with migraine without aura, who underwent four weeks of acupuncture treatment at specific points traditionally used for this condition. The central goal was to develop an artificial intelligence model capable of predicting therapeutic success based on brain structural features observed before the start of treatment. To do so, they used an advanced machine learning technique called support vector machine, combined with magnetic resonance imaging that allowed analysis of cerebral gray matter volume. The methodology included detailed patient follow-up through pain diaries, recording migraine days, intensity, and medications used.

Those who showed a reduction of at least fifty percent in migraine days were classified as responders.

Results revealed promising findings for personalized medicine. After four weeks of acupuncture, nineteen patients were classified as responders, representing forty-six percent of the sample. The artificial intelligence model developed demonstrated remarkable discrimination capacity, with eighty-three percent accuracy in identifying responders versus non-responders. More specifically, the model showed seventy-three percent sensitivity for detecting true responders and eighty-five percent specificity for correctly identifying non-responders.

The analyses identified ten specific brain regions whose baseline gray matter volume contributed to this prediction, primarily located in the frontal, temporal, and parietal gyri, precuneus, and cuneus. Interestingly, the investigators found that the reduction in migraine days correlated with baseline gray matter volume in four specific regions: left cuneus, right middle frontal gyrus, left inferior parietal gyrus, and right superior parietal gyrus.

The clinical implications of these findings are significant for patients and clinicians. For patients, this advance offers the prospect of more targeted treatments, avoiding prolonged periods of ineffective therapy and reducing unnecessary costs. Physicians can use this information to make more informed decisions about therapeutic strategies, directing patients with lower probability of response toward more suitable treatment alternatives. The study also revealed interesting structural brain changes in responders, particularly an increase in gray matter volume in the region of the left cuneus after treatment, suggesting neural plasticity as a possible mechanism of action of acupuncture.

This finding contributes to scientific understanding of how acupuncture may modify brain structures related to pain processing.

Despite encouraging results, the study presents important limitations that must be considered. The relatively small sample of forty-one patients and the high dropout rate limited long-term follow-up analyses, preventing conclusions about the durability of predictive effects. Additionally, the study used only one type of brain imaging information, when multimodal approaches including cortical thickness, white matter structure, and brain function could improve the accuracy of the predictive model. Validation of these results in larger and more diverse populations will be essential before clinical implementation.

The investigators also acknowledge that future studies should incorporate multiple research centers to confirm reproducibility of findings and establish truly reliable predictive models for routine clinical use. Nevertheless, this work represents an important step toward personalized medicine in acupuncture, offering hope for more efficient and individualized treatments for patients with migraine.