Technical Program

Paper Detail

Paper: PS-1B.38
Session: Poster Session 1B
Location: Symphony/Overture
Session Time: Thursday, September 6, 18:45 - 20:45
Presentation Time:Thursday, September 6, 18:45 - 20:45
Presentation: Poster
Publication: 2018 Conference on Cognitive Computational Neuroscience, 5-8 September 2018, Philadelphia, Pennsylvania
Paper Title: Brain age prediction for post-traumatic stress disorder patients with convolutional neural networks: a multi-modal neuroimaging study
Manuscript:  Click here to view manuscript
Authors: Xin Niu, Hualou Liang, Fengqing Zhang, Drexel university, United States
Abstract: Brain imaging shed lights on brain development which is closely related to various cognitive abilities. Immaturity and accelerated aging of the brain are typical consequence of developmental brain disorders. We developed a deep learning method to use convolutional neural networks (CNNs) to predict age of patients with post-traumatic stress disorders (PTSD) and healthy controls based on multi-modal brain imaging. N-fold cross validation was conducted to evaluate the prediction accuracy of age on healthy controls. Then the CNNs were trained with data of healthy controls and tested with PTSD group and another healthy control group with traumatic experiences, but no long-lasting PTSD symptoms. Our result showed that CNNs can be used to predict age with accuracy comparable to state- of-the-art machine learning methods such as ridge regression. Importantly, we found that the predicted age for PTSD patients are older than that of the control group, indicating an accelerated aging process of the brain in PTSD patients relative to healthy population.