MRI-based deep learning for differentiating between bipolar and major depressive disorders

• Published in: Psychiatry research. Neuroimaging Vol. 345; p. 111907
• Main Authors: Li, Ruipeng, Huang, Yueqi, Wang, Yanbin, Song, Chen, Lai, Xiaobo
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier B.V 01.12.2024
• Subjects: Adult; Bipolar Disorder - diagnostic imaging; Brain - diagnostic imaging; Brain - pathology; Deep Learning; Deep neural networks; Depressive Disorder, Major - diagnostic imaging; Diagnosis, Differential; Female; Humans; Magnetic Resonance Imaging - methods; Male; Middle Aged; Mood disorders differentiation; MRI pattern analysis; Neuroimaging diagnostics; Pattern recognition; Mood disorders differentiation; Deep neural networks; MRI pattern analysis; Pattern recognition; Neuroimaging diagnostics
• ISSN: 0925-4927; 1872-7506; 1872-7506
• DOI: 10.1016/j.pscychresns.2024.111907

•Enhanced attention Mechanism: we augment the conventional SE layer, which traditionally generates channel attention maps, by incorporating a dedicated branch for spatial attention maps. This branch includes a soft pool, a 7 × 7 convolution, and a sigmoid function, enabling the detection of complex spatial patterns effectively.•Integrated attention Maps: by executing an element-wise addition, our framework merges channel and spatial attention maps, aiming to boost the model's capability to discern features.•Soft-pooling technique: diverging from the standard max-pooling, our model employs soft-pooling for downsampling, which preserves a richer representation of input features and reduces data loss. Mood disorders, particularly bipolar disorder (BD) and major depressive disorder (MDD), manifest changes in brain structure that can be detected using structural magnetic resonance imaging (MRI). Although structural MRI is a promising diagnostic tool, prevailing diagnostic criteria for BD and MDD are predominantly subjective, sometimes leading to misdiagnosis. This challenge is compounded by a limited understanding of the underlying causes of these disorders. In response, we present SE-ResNet, a Residual Network (ResNet)-based framework designed to discriminate between BD, MDD, and healthy controls (HC) using structural MRI data. Our approach extends the traditional Squeeze-and-Excitation (SE) layer by incorporating a dedicated branch for spatial attention map generation, equipped with soft-pooling, a 7 × 7 convolution, and a sigmoid function, intended to detect complex spatial patterns. The fusion of channel and spatial attention maps through element-wise addition aims to enhance the model's ability to discriminate features. Unlike conventional methods that use max-pooling for downsampling, our methodology employs soft-pooling, which aims to preserve a richer representation of input features and reduce data loss. When evaluated on a proprietary dataset comprising 303 subjects, the SE-ResNet achieved an accuracy of 85.8 %, a recall of 85.7 %, a precision of 85.9 %, and an F1 score of 85.8 %. These performance metrics suggest that the SE-ResNet framework has potential as a tool for detecting psychiatric disorders using structural MRI data.