A groundbreaking study from Stanford Medicine combines functional MRI with machine learning to identify distinct biological subtypes of depression, paving the way for more personalised treatment approaches.
In a significant advancement towards precision psychiatry, researchers at Stanford Medicine have utilised brain imaging and machine learning techniques to identify six distinct biological subtypes of depression. This novel approach, detailed in a study published in Nature Medicine [1], not only categorises depression into specific ‘biotypes’ but also predicts treatment responses for three of these subtypes. The findings could revolutionise how depression is diagnosed and treated, potentially ending the current trial-and-error approach to medication prescription.
Addressing the treatment-resistant challenge
Depression remains a formidable challenge in mental health, with approximately 30% of patients experiencing treatment-resistant depression. For up to two-thirds of individuals with depression, current treatments fail to fully alleviate symptoms. This high rate of treatment failure is partly due to the lack of a reliable method for deter- mining which antidepressant or therapy will be most effective for a given patient.
The study’s senior author, Leanne Williams, PhD, the Vincent V.C. Woo, Professor and director of Stanford Medicine’s Center for Precision Mental Health and Wellness, emphasised the urgent need for better treatment-matching methods. “The goal of our work is figuring out how we can get it right the first time,” Williams said. “It’s very frustrating to be in the field of depression and not have a better alternative to this one-size-fits-all approach.”
Methodology and findings
The research team assessed 801 participants previously diagnosed with depression or anxiety using functional MRI (fMRI) to measure brain activity. Scans were conducted both at rest and during tasks designed to test cognitive and emotional functioning. The scientists focused on brain regions and connections known to play a role in depression.
Using a machine learning approach called cluster analysis, the researchers identified six distinct patterns of brain activity among the participants. These patterns formed the basis for the six biotypes of depression identified in the study.
Treatment responses linked to biotypes
In a randomised trial involving 250 of the study participants, the researchers found that certain biotypes responded better to specific treatments:
- Patients with overactivity in cognitive brain regions showed the best response to venlafaxine (Effexor).
- Those with higher resting activity in regions associated with depression and problem-solving benefited more from behavioural talk therapy.
- Individuals with lower resting activity in the brain circuit controlling attention were less likely to improve with talk therapy.
Jun Ma, MD, PhD, from the University of Illinois Chicago and a study co-author, noted that these findings align with current understanding of brain function. The biotype responding well to talk therapy, for instance, may be better equipped to adopt new skills due to higher activity in relevant brain regions.
Implications for precision psychiatry
The study represents a significant step towards personalised medicine in mental health. “To our knowledge, this is the first time we’ve been able to demonstrate that depression can be explained by different disruptions to the functioning of the brain,” Williams explained. “In essence, it’s a demonstration of a personalised medicine approach for mental health based on objective measures of brain function.”
In a related study [2], Williams and her team showed that using fMRI improves the ability to identify individuals likely to respond to antidepressant treatment. For the ‘cognitive biotype’ of depression, which affects over a quarter of patients and is less responsive to standard antidepressants, fMRI-based identification increased the accuracy of predicting remission from 36% to 63%.
Further insights and future directions
The study also revealed correlations between biotypes and specific symptoms and task performance. For example, patients with overactive cognitive regions exhibited higher levels of anhedonia (lack of interest, inability to feel pleasure) and performed worse on executive function tasks.
Interestingly, one of the six biotypes showed no significant differences in brain activity compared to non-depressed individuals in the regions studied. This suggests that there may be other types of brain dysfunction in depression that were not captured by the current imaging approach.
Moving forward
The research team is expanding their imaging study to include more participants and explore a wider range of treatments across all six biotypes. They are also working to establish standardised protocols for implementing this approach in clinical practice.
“To really move the field toward precision psychiatry, we need to identify treatments most likely to be effective for patients and get them on that treatment as soon as possible,” Ma emphasised. “Having information on their brain function, in particular the validated signatures we evaluated in this study, would help inform more precise treatment and prescriptions for individuals.”
References:
- Tozzi, L., Zhang, X., Pines, A. et al. Personalized brain circuit scores identify clinically distinct biotypes in depression and anxiety. Nat Med (2024).
https://doi.org/10.1038/s41591-024-03057-9 - Williams, L. M., & Yesavage, J. Cognitive control circuit function predicts antidepressant outcomes: A signal detection approach to actionable clinical decisions. Personalized Medicine in Psychiatry (2024). https://doi.org/10.1016/j.pmip.2024.100126
