Machine Learning Algorithm Could Help Predict Who Is at Risk of Depression, Anxiety
A team of researchers from University of Texas-Austin, University of Washington School of Medicine and University of California-San Diego has successfully trained a new machine learning algorithm to help in detecting who might be vulnerable to depression and anxiety.
In a proof-of-concept study published in the journal Psychiatry Research: Neuroimaging, the new machine learning algorithm was able to determine individuals who have major depressive disorder with about 75 percent accuracy.
"One of the benefits of machine learning, compared to more traditional approaches, is that machine learning should increase the likelihood that what we observe in our study will apply to new and independent datasets. That is, it should generalize to new data," said Christopher Beevers, a professor of psychology and director of the Institute for Mental Health Research at UT Austin and co-author of the study, in a press release. This is a critical question that we are really excited to test in future studies."
For the study, the researchers used the Stampede supercomputer at the Texas Advanced Computing Center (TACC) to train a Support Vector Machine Learning algorithm that can detect subtle connections among hundreds of patients using Magnetic Resonance Imaging (MRI) brain scans, genomics data and other relevant factors, providing accurate predictions of risk for those with depression and anxiety.
The researchers then asked 50 participants to undergo diffusion tensor imaging (DTI) MRI scans. DTI measures the diffusion of water molecules in multiple spatial directions, generating vectors for each voxel to quantify the dominant fiber orientation. Voxel is the (three-dimensional cubes that represent either structure or neural activity throughout the brain.
They found that one common parameter used to characterize DTI is the so-called fractional anisotropy. However, the brain is represented by roughly 175,000 voxels, making it impossible to compare scans manually. Due to this, the researchers decided to train a machine learning algorithm to automate the discovery process.
This DTI-derived factional anisotropy maps can be used to accurately to classify those who have depression, those who were at risk of depression and healthy individuals in the study sample.