Brain scans reveal two distinct biological profiles of migraine

A recent analysis of brain scans has revealed that people who experience migraines can be grouped into two distinct biological categories. These categories are based on how the brain is wired and physically structured, offering a new way to understand the disorder beyond traditional symptom checklists. The findings were published in the journal Cephalalgia.

Migraine is a neurological condition that causes severe head pain, sensitivity to light, and other debilitating symptoms. Doctors currently classify the disorder based on how often attacks occur and whether a patient experiences an aura, which refers to visual or sensory disturbances preceding the headache. This symptom-based approach, outlined in the International Classification of Headache Disorders, often fails to predict which treatments will work best for individual patients.

The biological differences between people with migraines remain largely unmapped. Researchers suspect that categorizing patients based on brain biology, rather than just symptom frequency, might eventually improve treatment strategies. Stanford University researchers Jaiashre Sridhar and Danielle D. DeSouza led a team to investigate whether patterns in brain imaging could identify hidden biological subgroups.

To do this, the research team used two types of magnetic resonance imaging, or MRI. Structural MRI measures the physical dimensions of the brain, such as the volume and thickness of the outer layer known as the cerebral cortex, as well as deeper subcortical structures. Functional MRI tracks blood flow to observe how different brain regions communicate when a person is at rest, a metric called functional connectivity.

The researchers first analyzed combined structural and functional brain scan data from 111 individuals with migraines and 51 healthy controls. They used a mathematical algorithm to simplify the massive amount of data and group the patients based on shared biological patterns. This exploratory approach was designed to let the data dictate the groups rather than relying on prior clinical labels.

This combined analysis identified two biological subgroups with distinct brain profiles and clinical experiences. One group tended to be older, had lived with migraines longer, and reported higher levels of daily disability. This higher-burden group also experienced longer individual headache durations and lower confidence in their ability to manage pain.

In this higher-burden group, functional MRI scans showed elevated connectivity between deeper brain structures and cortical networks responsible for attention, movement, and visual processing. Structurally, these individuals also exhibited reduced brain volume across several cortical regions, including the frontal, parietal, and temporal lobes, compared with the other subgroup. Many of these heightened functional connections were also elevated relative to the healthy control group.

The second subgroup presented a milder biological profile. Their brain structure was largely preserved in comparison to the first group. Their functional connectivity patterns and brain volumes were not statistically significant when compared to the healthy control group.

After identifying the combined groups, the researchers conducted a secondary analysis using only the functional connectivity data. They applied the same mathematical grouping process to see how the patients would cluster based solely on how different brain regions communicate.

This functional-only model produced two subgroups that closely matched the groups found in the initial combined analysis. Patients with higher clinical burden again clustered together, exhibiting similar patterns of elevated brain connectivity. When grouped this way, the resulting clusters did not display any differences in brain structure, indicating that functional connectivity drove most of the initial subgroupings.

Next, the team ran a third clustering model using exclusively structural MRI data. They grouped the same patients based entirely on the thickness and volume of their brain tissue.

This structural-only analysis generated two entirely different patient clusters that had almost no overlap with the groups formed by the combined or functional data. While these two new groups showed widespread differences in brain volume, they exhibited no differences in functional connectivity. This divergence indicates that structural variations represent a completely separate dimension of migraine biology than functional variations.

To verify the stability of their findings, the researchers performed a final sensitivity analysis. Instead of looking at broad functional networks, they repeated the combined analysis using a much more detailed map that divided the brain into over a hundred smaller, specific regions.

The results of this fine-grained analysis strongly mirrored the original combined model. Between 90 and 95 percent of the participants were assigned to the exact same subgroups as before. This consistency suggests that the biological groups are robust, regardless of the scale used to map the brain.

While these biological groupings provide a new perspective on migraines, the research relies on data collected at a single point in time. It is not possible to know whether prolonged migraines alter the brain over the years, or if these brain differences exist first and influence how the condition develops. The clinical differences between the two subgroups were also relatively subtle, and the groups did not align with traditional categories like chronic or episodic migraine.

The researchers noted that this was a modestly sized study, meaning the results will need to be verified in larger populations. The study also did not track the exact phase of the patients’ migraine cycle during the brain scans, such as whether they were actively having a migraine or in a resting phase. Additionally, the researchers did not account for all preventive medications the participants might have been taking at the time.

Future research will need to track larger groups of patients over extended periods to see how these biological profiles evolve and whether they can eventually guide medical care.

The study, “Neuroimaging-based subtyping of migraine identifies clinically distinct phenotypes,” was authored by Jaiashre Sridhar, Mahsa Babaei, Bharati M. Sanjanwala, Robert P. Cowan, and Danielle D. DeSouza.