Brain scans could predict future weight gain in people with mental disorders

A simple standard MRI brain scan could predict which people with mental illnesses will gain weight after their initial diagnosis, thereby increasing their risk of physical diseases, and which ones will not.

Visualisation and performance of the BMI predictor. Credit: Nature Mental Health (2025). DOI: 10.1038/s44220-025-00522-3

"This would allow us to start targeted prevention against the weight gain frequently observed in this patient group," explained Professor Nikolaos Koutsouleris from the Department of Psychiatry and Psychotherapy of the LMU University Hospital Munich following the results of a new study involving numerous cooperation partners from Germany and abroad, including the University of Cologne.

In Germany, almost 18 million people are affected by a mental illness, including depression in particular, but also anxiety disorders or schizophrenia. It is little known that people with severe mental illness die, on average, ten to 15 years earlier than the general population. The problem is largely due to physical illnesses, especially cardiovascular diseases, which are more common than average among people with severe mental illness.

"This is why it's important for patients to be mindful of risk factors such as lack of exercise, smoking or being overweight or obese," explains Koutsouleris. "Beyond the well-known side effects of certain medications, we assume, based on some findings, that weight gain may be related to brain changes that are in turn associated with the mental disorder.”

Could these brain changes be used at the time of initial diagnosis to predict, which patients will later develop a higher body mass index (BMI)? To examine this hypothesis, the research team first developed a machine learning model. They trained this artificial intelligence using MRI scans of the brains of healthy people. The goal was for the model to learn, on its own, how to predict a person's body weight based solely on their brain scans.

The working group led by Professor Joseph Kambeitz from the Clinic and Polyclinic for Psychiatry and Psychotherapy at University Hospital Cologne contributed a patient database as training material, the so-called PRONIA cohort. The team was also significantly involved in the study design, data analysis and interpretation. As an expert in AI-supported analyses in psychiatry, Kambeitz has helped to shape the research area of neuroscientific findings informed by artificial intelligence.

In a second step, the researchers applied their system to the MRI brain scans of patients with mental disorders.

"In these cases, our prognosis model made systematic errors and incorrectly determined the weight of the corresponding patients," added Koutsouleris. "This system significantly controls our eating behaviour. Our prediction model had previously learned from healthy people: Less volume in these brain regions means higher weight.”

For example, in people with schizophrenia, the model overestimated the weight, because certain brain areas such as the anterior cerebral cortex, which contains parts of the reward system, are smaller than usual. Although schizophrenia patients have smaller brain volumes when they are first diagnosed, they do not necessarily have a higher body mass index (BMI).

In the final step, the researchers tracked the patients' BMI for a year after the initial diagnosis and initial weight assessment.

"We observed that there is actually a sharp increase in those patients for whom our AI model had misjudged their BMI to be too high. "The difference between the estimated and the actually observed BMI, the so-called BMI gap, has a predictive power for the future weight development of the patients," said Koutsouleris.

This is especially true for people with schizophrenia, but it also applies to those with depression.

"We can try to encourage those affected to adopt a healthier lifestyle by participating in weight-loss programmes, exercising more regularly, and making healthier food choices," Kambeitz added. "Alternatively, we can prescribe medication such as metformin to reduce or prevent the risk of metabolic diseases. This would be a major benefit, especially since there is evidence that less weight gain is associated with reduced inflammatory activity in the brain and, consequently, fewer psychiatric symptoms as the disease progresses."

As soon as the new tool has been refined with additional parameters such as the patient's individual genetics or blood values such as cholesterol, making it even more accurate, it will be made available to all doctors for determining the BMI gap.

The findings were reported in the paper, ‘The BMIgap tool to quantify transdiagnostic brain signatures of current and future weight’, published in Nature Mental Health. To access this paper, please click here