Real-world MRI data confirm shared brain signatures of mental health disorders

Over 1 billion people worldwide are living with one or more mental health disorders that affect their mood, thinking processes and behavior, impacting their daily functioning to varying degrees. Identifying variations in the brain's structure and organization that are commonly linked with mental health disorders could help to devise more effective tools to diagnose these conditions or create personalized treatment plans.

Researchers at Copenhagen University Hospital and University of Copenhagen recently analyzed thousands of brain scans and medical records collected in Denmark to identify structural brain variations associated with mental health disorders. Their findings, published in Molecular Psychiatry, were aligned with some earlier observations, showing that mental health disorders were associated with a smaller thalamus and amygdala, larger ventricles and a thinner outer brain layer (i.e., cortex).

"For years, brain imaging research in psychiatry has relied on carefully curated datasets from smaller research studies, with varying inclusion criteria," Stefano Cerri and Prof. Michael Eriksen Benros, first and senior author of the paper, respectively, told Medical Xpress.

"A consortium called ENIGMA has done amazing work pooling these datasets worldwide, but it was hampered by its reliance on summary statistics from prior studies, with different inclusion criteria and varying availability of clinical data. We kept asking ourselves a simple question: what about all the MRI scans that are already being acquired every day in hospitals as part of routine clinical care?"

Analyzing thousands of clinical MRI scans

In Denmark, brain scans collected with a non-invasive imaging technique called magnetic resonance imaging (MRI) are linked to the electronic health records of the corresponding patients. This makes it easier for researchers to explore the connections between the structure of patients' brains and specific aspects of their clinical history.

"We wanted to see whether these 'real-world' scans, which can be messier and of lower quality than those collected specifically for research purposes, could still reveal the brain differences shown by the ENIGMA Consortium," said Cerri and Benros. "If they could, it would open the door to studying mental disorders at a scale that was previously unimaginable, paving the way for novel biological insights into mental disorders."

Cerri, Benros and their colleagues initially analyzed all clinical MRI scans collected in the Eastern part of Denmark throughout 2019, using electronic health records to identify patients who were diagnosed with a mental health disorder shortly before or after the scan was collected. They also selected patients who had no known psychiatric or neurological conditions, so that they could compare their brain scans to those of patients with mental health disorders.

The team analyzed data collected from over 23,000 patients living in Denmark, selecting 4,800 patients fulfilling the inclusion criteria for their study. The data they analyzed was stored on a secure research platform using pseudonyms that protect the privacy of patients.

"To measure the brain, we used a new tool called 'recon-all-clinical," which was specifically designed to handle the kind of lower-resolution, varied scans you get in real clinical settings," said Cerri and Benros. "It gave us measurements of cortical thickness and the volumes of different brain regions, which we then compared between patients and controls while accounting for important factors such as age, sex, and the scanner used."

The study's key observations and their implications

Cerri, Benros and their colleagues observed specific variations in brain structure that were more prevalent in people diagnosed with mental health disorders. Most notably, they found that people with severe mental disorders presented a smaller thalamus, a smaller amygdala, larger ventricles (i.e., fluid-filled cavities at the center of the brain) and a thinner cerebral cortex.

"What struck us most is that the brain differences reported in decades of research studies also showed up in our clinical data," said Cerri and Benros.

"That's important validation: it means that routine hospital scans are also good at detecting these subtle changes, when used at scale. As each scan can be linked to a full medical history, we can also start asking questions that traditional research cohorts simply can't answer, like how medications, comorbidities, or the stage of illness shape what we see in the brain."

The results of this recent study could soon be combined with earlier observations to create a clear and consistent picture of how mental health disorders manifest in the brain. This could potentially guide the development of AI-based computational tools for predicting these disorders from brain scans, along with other diagnostic and treatment solutions.

"The bigger picture is that clinical data, when combined with health records, may actually be more valuable than purely research-grade data for understanding mental disorders, because it reflects the real patients that clinicians see every day," said Cerri and Benros. "That's a step toward better stratifying patients and, eventually, supporting diagnosis and treatment decisions."

Future research directions

The team's study only focused on MRI scans collected in 2019, yet the Danish medical database contains data spanning across several more years. In the future, the researchers would like to use all available relevant data to carry out longitudinal analyses, to better understand how people's brains change before and after they are diagnosed with mental health disorders.

"We've really only scratched the surface," said Cerri and Benros. "Following the same patients over time to track how their brains change, and ideally catching early signals before a disorder fully develops or worsens."

As part of their future research, Cerri and Benros would also like to extend their analyses to include a broader range of mental health disorders. In fact, some conditions, such as bipolar disorder and attention deficit disorder (ADHD) were excluded from their recent study, as the sample of data available for these disorders was too small.

"There is also a lot more to explore around how factors like duration of illness, age at onset, and cumulative medication exposure shape the brain," added the researchers. "The long-term goal is to bridge real-world imaging with clinical trajectories, so we can move toward more biologically grounded subgroups of patients, facilitating more precision psychiatry approaches leading to treatment improvements."

This article was originally published on MedicalXpress Breaking News-and-Events.