Authors: Bo-Gyeom Kim; Gakyung Kim; Yoshinari Abe; Pino Alonso; Stephanie Ameis; Alan Anticevic; Paul D. Arnold; Srinivas Balachander; Nerisa Banaj; Nuria Bargalló; Marcelo C. Batistuzzo; Francesco Benedetti; Sara Bertolín; Jan Carl Beucke; Irene Bollettini; Silvia Brem; Brian P. Brennan; Jan K. Buitelaar; Rosa Calvo; Miguel Castelo-Branco; Yuqi Cheng; Ritu Bhusal Chhatkuli; Valentina Ciullo; Ana Coelho; Beatriz Couto; Sara Dallaspezia; Benjamin A. Ely; Sónia Ferreira; Martine Fontaine; Jean-Paul Fouche; Rachael Grazioplene; Patricia Gruner; Kristen Hagen; Bjarne Hansen; Gregory L. Hanna; Yoshiyuki Hirano; Marcelo Q. Höxter; Morgan Hough; Hao Hu; Chaim Huyser; Toshikazu Ikuta; Neda Jahanshad; Anthony James; Fern Jaspers-Fayer; Selina Kasprzak; Norbert Kathmann; Christian Kaufmann; Minah Kim; Kathrin Koch; Gerd Kvale; Jun Soo Kwon; Luisa Lazaro; Junhee Lee; Christine Lochner; Jin Lu; Daniela Rodriguez Manrique; Ignacio Martínez-Zalacaín; Yoshitada Masuda; Koji Matsumoto; Maria Paula Maziero; Jose M. Menchón; Luciano Minuzzi; Pedro Silva Moreira; Pedro Morgado; Janardhanan C. Narayanaswamy; Jin Narumoto; Ana E. Ortiz; Junko Ota; Jose C. Pariente; Chris Perriello; Maria Picó-Pérez; Christopher Pittenger; Sara Poletti; Eva Real; Y. C. Janardhan Reddy; Daan van Rooij; Yuki Sakai; João Ricardo Sato; Cinto Segalas; Roseli G. Shavitt; Zonglin Shen; Eiji Shimizu; Venkataram Shivakumar; Noam Soreni; Carles Soriano-Mas; Nuno Sousa; Mafalda Machado Sousa; Gianfranco Spalletta; Emily R. Stern; S. Evelyn Stewart; Philip R. Szeszko; Rajat Thomas; Sophia I. Thomopoulos; Daniela Vecchio; Ganesan Venkatasubramanian; Chris Vriend; Susanne Walitza; Zhen Wang; Anri Watanabe; Lidewij Wolters; Jian Xu; Kei Yamada; Je-Yeon Yun; Mojtaba Zarei; Qing Zhao; Xi Zhu; Paul M. Thompson; Willem B. Bruin; Guido A. van Wingen; Federica Piras; Fabrizio Piras; Dan J. Stein; Odile A. van den Heuvel; Helen Blair Simpson; Rachel Marsh; Jiook Cha · Research
Can Brain Scans Help Diagnose OCD? Findings from a Large International Study
A large international study examines whether brain scans can help diagnose obsessive-compulsive disorder (OCD).
Source: Kim, B.-G., Kim, G., Abe, Y., Alonso, P., Ameis, S., Anticevic, A., Arnold, P. D., Balachander, S., Banaj, N., Bargalló, N., Batistuzzo, M. C., Benedetti, F., Bertolín, S., Beucke, J. C., Bollettini, I., Brem, S., Brennan, B. P., Buitelaar, J. K., Calvo, R., Castelo-Branco, M., ... Cha, J. (2024). White matter diffusion estimates in obsessive-compulsive disorder across 1653 individuals: machine learning findings from the ENIGMA OCD Working Group. Molecular Psychiatry, 29(4), 1063–1074. https://doi.org/10.1038/s41380-023-02392-6
What you need to know
- Researchers used brain scans from over 1,600 people across 18 international sites to see if machine learning could distinguish between those with and without obsessive-compulsive disorder (OCD).
- The machine learning models had low to moderate accuracy in identifying OCD, with better performance when looking at unmedicated OCD patients versus healthy controls.
- Certain brain areas involved in OCD, like the corpus callosum and internal capsule, were important for the models’ predictions.
- There was significant variability between research sites, which impacted the models’ performance.
- While not yet accurate enough for clinical use, this research provides insights into the brain patterns associated with OCD and highlights areas for improvement in future studies.
Using brain scans to identify OCD
Obsessive-compulsive disorder (OCD) is a mental health condition that affects 1-1.5% of people worldwide. It involves recurring, unwanted thoughts (obsessions) and repetitive behaviors or mental acts (compulsions). While we know OCD involves certain brain circuits, diagnosing it relies on assessing symptoms rather than biological tests.
Researchers have been exploring whether brain scans could help diagnose OCD more objectively. This study, part of an international collaboration called ENIGMA-OCD, used the largest dataset yet to test if machine learning algorithms could identify OCD based on brain scans.
The study included 1,653 participants (865 with OCD and 788 without) from 18 research sites across the globe. They used a type of brain scan called diffusion tensor imaging (DTI), which looks at the structure of white matter - the “wiring” that connects different brain regions.
How well did the machine learning work?
The researchers used sophisticated machine learning techniques to try to distinguish between people with and without OCD based on their brain scans. Here’s how well the models performed:
- Identifying OCD vs. healthy controls:
- Adults: 57% accuracy
- Children: 60% accuracy
- Identifying unmedicated OCD vs. healthy controls:
- Adults: 63% accuracy
- Children: 49% accuracy
- Distinguishing medicated vs. unmedicated OCD:
- Adults: 77% accuracy
- Children: 72% accuracy
While these results are better than random guessing (50%), they aren’t accurate enough for clinical use in diagnosing OCD. However, the higher accuracy in distinguishing medicated from unmedicated OCD patients is interesting and may reflect how OCD medications affect brain structure.
Which brain areas were important?
The machine learning models relied on certain brain regions to make their predictions. Some key areas included:
The corpus callosum: This is the large bundle of nerve fibers connecting the brain’s left and right hemispheres. It’s involved in communication between different parts of the brain.
The internal capsule: This is a region containing nerve fibers that connect the cerebral cortex (the brain’s outer layer) with other brain areas. It’s part of circuits involved in movement control and cognitive functions.
The posterior thalamic radiation: This connects the thalamus (a relay center for sensory and motor signals) with the cerebral cortex.
The cingulum: A collection of nerve fibers involved in attention, emotion processing, and executive function.
Many of these regions have been implicated in OCD in previous research. They’re part of brain circuits involved in habits, cognitive control, and processing emotions - all of which can be affected in OCD.
Challenges and variability
One of the study’s main findings was the significant variability between different research sites. The accuracy of OCD classification ranged from about 52% to 79% across different adult sites, and 36% to 63% across pediatric sites.
This variability could be due to several factors:
- Differences in brain scanning equipment and procedures
- Variations in the characteristics of OCD patients at different sites (e.g., symptom severity, medication use)
- Cultural or environmental factors that might influence OCD presentation or brain structure
The researchers tried using a technique called “harmonization” to reduce site-related differences, but this didn’t significantly improve the overall results.
What does this mean for OCD research and treatment?
While this study doesn’t provide a brain scan-based test for OCD, it offers valuable insights:
It confirms that OCD is associated with subtle differences in brain structure, particularly in white matter connections.
The ability to distinguish medicated from unmedicated OCD patients suggests that OCD treatments may have detectable effects on brain structure. This could potentially help in monitoring treatment effectiveness in the future.
The study highlights the challenges of using brain scans for psychiatric diagnosis, especially when combining data from multiple sites. This is important for improving future research methods.
The findings support the idea that OCD involves disruptions in brain circuits related to habits, cognitive control, and emotion processing.
Conclusions
- Machine learning analysis of brain scans shows promise for understanding OCD but isn’t yet accurate enough for clinical diagnosis.
- The study highlights specific brain regions and connections involved in OCD, supporting existing theories about the disorder’s neurobiology.
- Significant variability between research sites emphasizes the need for standardized methods in brain imaging studies of psychiatric disorders.
While we’re not yet at the point of diagnosing OCD with a brain scan, this research brings us closer to understanding the biological basis of the disorder. Future studies building on these findings may lead to more objective diagnostic tools and personalized treatment approaches for OCD.