Probing depressing using deep learning on functional MRI data
Title: Probing depressing using deep learning on functional MRI data
SNIC Project: Berzelius-2022-13
Project Type: LiU Berzelius
Principal Investigator: Paul Hamilton <paul.hamilton@liu.se>
Affiliation: Linköpings universitet
Duration: 2022-02-03 – 2022-09-01
Classification: 30599
Keywords:

Abstract

We are planning on creating dual regression maps (measures of the brains connectivity with itself during rest) for several different brain locations and then feed them into several different 3D CNN classification networks using 20k-40k neuroimaging MRI datasets. With this approach we an both probe if there are connectivity differencies between healthy individuals and MDD patients by seeing if the network can tell them apart and with the same approach also see which brain nodes give the best preformance. This will cast light on which brain regions / networks that are more or less affected by MDD.