Deep learning and brain network

Deep learning has great superiority in image analysis and disease prediction, thus may play an important role in neuroimaging. We are seeking applications that deploys deep learning techniques into brain network analysis. For instance, Alzheimer’s Disease (AD) is a typical example to show advantages of deep learning in diagnosing brain diseases and providing clinical decision support. In order to achieve this goal, raw functional magnetic resonance imaging (fMRI) was converted to a matrix to represent activity of 90 brain regions. Secondly, to represent the functional connectivity between different brain regions, a correlation matrix is obtained by calculating the correlation between each pair of brain regions. Furthermore, a targeted autoencoder network is built to perform classification the correlation matrix, which turns out to be extremely sensitive to AD. The experiment results showed that our proposed method for AD prediction achieves better performance than most traditional approaches. Compared to Support Vector Machine (SVM), a 25% improvement is gained in terms of prediction accuracy.

Hu, Chenhui, et al. "Clinical decision support for Alzheimer's disease based on deep learning and brain network." Communications (ICC), 2016 IEEE International Conference on. IEEE, 2016.