Tinnitus classification based on resting-state functional connectivity using a convolutional neural network architecture

• Published in: NeuroImage (Orlando, Fla.) Vol. 290; p. 120566
• Main Authors: Xu, Qianhui, Zhou, Lei-Lei, Xing, Chunhua, Xu, Xiaomin, Feng, Yuan, Lv, Han, Zhao, Fei, Chen, Yu-Chen, Cai, Yuexin
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 15.04.2024; Elsevier Limited; Elsevier
• Subjects: Artificial intelligence; Brain mapping; Brain research; Classification; Convolutional neural network; Deep learning; Diagnosis; Functional connectivity; Functional magnetic resonance imaging; Machine learning; Magnetic resonance imaging; Medical imaging; Neural networks; Neuroimaging; Otolaryngology; Resting state fMRI; Sensorimotor system; Support vector machines; Tinnitus; Transfer learning; Functional connectivity; Resting state fMRI; Tinnitus; Convolutional neural network
• DOI: 10.1016/j.neuroimage.2024.120566
• ISSN: 1053-8119; 1095-9572

•A decomposed convolutional neural network model was established based on rs-fMRI connectivity.•The model paired with the Dos_160 atlas can be effectively applied to the diagnosis of tinnitus.•This study pinpointed key brain regions for subjective tinnitus using a data-driven approach. Many studies have investigated aberrant functional connectivity (FC) using resting-state functional MRI (rs-fMRI) in subjective tinnitus patients. However, no studies have verified the efficacy of resting-state FC as a diagnostic imaging marker. We established a convolutional neural network (CNN) model based on rs-fMRI FC to distinguish tinnitus patients from healthy controls, providing guidance and fast diagnostic tools for the clinical diagnosis of subjective tinnitus. A CNN architecture was trained on rs-fMRI data from 100 tinnitus patients and 100 healthy controls using an asymmetric convolutional layer. Additionally, a traditional machine learning model and a transfer learning model were included for comparison with the CNN, and each of the three models was tested on three different brain atlases. Of the three models, the CNN model outperformed the other two models with the highest area under the curve, especially on the Dos_160 atlas (AUC = 0.944). Meanwhile, the model with the best classification performance highlights the crucial role of the default mode network, salience network, and sensorimotor network in distinguishing between normal controls and patients with subjective tinnitus. Our CNN model could appropriately tackle the diagnosis of tinnitus patients using rs-fMRI and confirmed the diagnostic value of FC as measured by rs-fMRI.