China: A new study published in eBioMedicine-The Lancet has revealed promising results for identifying individuals at risk for severe mental illness (SMI) before the onset of the illness using a digital model called Multiple Instance learning (MIL) with clinical MRI scans.
This study aims to develop an efficient and practical model for mental health screening among at-risk populations for severe mental illness.
The researchers used a deep learning model known as Multiple Instance Learning (MIL) to train and test an SMI detection model with clinical MRI scans of 14,915 patients with SMI and 4538 healthy controls in the primary dataset. The validation analysis was conducted in an independent dataset with 290 patients and 310 healthy participants. Three other machine-learning models were also used for comparison (ResNet, DenseNet, and EfficientNet).
The results of the study revealed the following findings:
1.The study found that the MIL model and other machine learning models were similarly effective in identifying individuals with SMI and healthy controls, with an AUC (Area under the ROC curve) of 0.82 for the MIL model.
2.The MIL model had better generalization in the validation test and performed better on lower-powered MRI scanners.
3.The MIL model also performed better in predicting clinician ratings of distress than self-ratings with questionnaires.
4.The right precuneus, bilateral temporal areas, left precentral/postcentral gyrus, bilateral medial prefrontal cortex, and right cerebellum were discovered to contribute to SMI recognition.
The findings suggest that the MIL model offers a potentially useful aid for early identification and intervention to prevent illness onset in vulnerable at-risk populations. With incremental improvements, the approach may become a practical model for mental health risk monitoring.
The researchers of the study added that “This study provides promising results for identifying individuals at risk for severe mental illness and highlights the potential of deep learning models for mental health screening.”
References:
Zhang, W., Yang, C., Cao, Z., Li, Z., Zhuo, L., Tan, Y., He, Y., Yao, L., Zhou, Q., Gong, Q., Sweeney, J. A., Shi, F., & Lui, S. (2023, April 1). Detecting individuals with severe mental illness using artificial intelligence applied to magnetic resonance imaging. eBioMedicine. https://doi.org/10.1016/j.ebiom.2023.104541