AI is playing a critical role in depression therapy

Posted Feb 24, 2020 by Tim Sandle
By applying artificial intelligence to assess patients with clinical depression, psychiatrists can gain a new understanding into which therapeutic treatment stands the best chance of working for an individual patient.
Doctors working in displacement camps say hormonal changes caused by anxiety and depression  the lac...
Doctors working in displacement camps say hormonal changes caused by anxiety and depression, the lack of privacy and physical comfort as well as of balanced nutrition all affect a mother's ability to breastfeed
The research comes after a series of trials conducted at the UT Southwestern Medical Center, using patients living in the U.S. The focus of the trial was to examine if an algorithm can assess mood disorders to the extent that it can accurately predict whether a specific antidepressant will be effective for a given patient. This is based on the individual patient's brain activity.
The longer-term aim of the project is to develop a platform that can aid psychiatrists with the diagnosis of depression and recommend prescriptions for depression treatments. This is to be achieved through a combination of artificial intelligence, brain imaging, and blood tests, together with the professional expertise of the medical doctor in charge.
Commenting on how the trial has progressed so far, lead researcher Dr. Madhukar Trivedi says: “These studies have been a bigger success than anyone on our team could have imagined.”
The medical researcher adds that: “We provided abundant data to show we can move past the guessing game of choosing depression treatments and alter the mindset of how the disease should be diagnosed and treated."
The research group have examined the output of studies involving some 300 participants, each of whom was diagnosed with a form of depression. For the purposes of study, the subjects were divided up into two groups, through being randomly assigned. In one group, the people received a placebo; in the other group, the subjects were given a form of medication called a selective serotonin reuptake inhibitor, which is a common type of antidepressant.
To assess differences between the two groups, an electroencephalogram (EEG) was deployed to measure electrical activity in the participants' cortex. These measurements were taken before any treatment was given.
At this stage, a machine-learning algorithm was put to task to analyze the EEG data. The algorithm proceeded to predict which patients would benefit from the medication. The actual data from the patients, in terms of their response to the antidepressant or the placebo, was then assessed against the predictive model. Here it was found that the algorithm accurately predicted most patient outcomes.
Such research is perhaps the start of machines making recommendations as to who might benefit from a specific antidepressant, thereby aiding the psychiatric profession with clinical-based decision making.
The research has been published in the journal Nature Biotechnology. The associated research paper is titled “An electroencephalographic signature predicts antidepressant response in major depression.”