Meeting Banner
Abstract #0713

A novel deep learning framework on brain functional networks for diagnosis of psychiatric diseases

Milad Mashhady Ali Poury 1, Ali Ameri 1, Saeed Masoudnia 2, Hosna Tavakoli2,3, Faezeh Ghasemi4,5, Reza Rostami6, and Mohammad-Reza Nazem-Zadeh2,7
1Department of Biomedical Engineering, Shahid Beheshti University of Medical Sciences, Tehran, Iran (Islamic Republic of), 2Research Center for Molecular and Cellular Imaging, Tehran University of Medical Sciences, Tehran, Iran (Islamic Republic of), 3cognitive neuroscience, Institute for Cognitive Science Studies, Tehran, Iran (Islamic Republic of), 4Medical Physics and Engineering, Shahid Beheshti University of Medical Sciences, Tehran, Iran (Islamic Republic of), 5Research Center for Biomedical Technologies and Robotics, Tehran University of Medical Sciences, Tehran, Iran (Islamic Republic of), 6Department of Psychology, University of Tehran, Tehran, Iran (Islamic Republic of), 7Medical Physics and Biomedical Engineering, Tehran University of Medical Sciences, Tehran, Iran (Islamic Republic of)

Synopsis

We developed a predicting algorithm based on brain connectivity to quantify the altered brain regions in schizophrenia, bipolar, and attention deficit hyperactivity disorders, to help diagnose them using neuroimaging biomarkers. Functional connectivity was utilized to construct brain graphs, on which the node2vec framework was applied to produce the node embeddings. The concatenation of embeddings was used to derive the region feature vectors to feed support vector machine (SVM) classifiers. Also, we build a model to assist the diagnosis of disorders using a weighted voting ensemble. The achieved accuracy proved to outperform to the state-of-the-art models.

This abstract and the presentation materials are available to members only; a login is required.

Join Here