This repository contains the source code of the following paper:
Ahn, G. (*), Kim, B. (*), Kim, K.K., Kim, H., Lee, E., Ahn, W.Y., Kim, J.W. and Cha, J., 2021, February. Identifying prepubertal children with risk for suicide using deep neural network trained on multimodal brain imaging. In International Workshop on Health Intelligence (pp. 75-86). Springer, Cham. https://link.springer.com/chapter/10.1007/978-3-030-93080-6_7
Fig1. ROC curve for each dataset. (a: resting state fMRI, b: stop signal task fMRI, c: Emotional N-back task fMRI
Fig2. Top 10 features of combined (fMRI + Psychosocial features) models. Red indicates features from fMRI, blue indicates features from psychosocial.