With machine learning, you no longer have to experiment for months with different drugs to help you fight depression.
Deep learning can predict how patients will respond to different antidepressant medications. Based on these predictions, the most appropriate treatment is selected.
The process includes scanning patients’ brains; those with depression show various patterns of depressed brain activity.However, responses to different drugs also vary among patients. Neural networks can determine, based on studied brain patterns, the medications that are expected to be effective.
The researchers trained separate Vanilla neural networks to predict changes in depression levels in patients, using several medications and placebos. The research results demonstrate the network’s ability to accurately determine, based on brain patterns, which drug is beneficial for a given patient.
Millions of adults suffer from clinical depression, and one-third try at least three medications before settling on one. In addition, many physicians, by outcomes they observe in a handful of patients, can’t analyze the overwhelming data from a large group. Reliable predictions about which drugs work best—even if far from accurate—can make a difference.