McMaster University professor Gary Hasey is part of an expanding group of psychiatrists who are trying to devise more effective patient treatment strategies by interpreting physiological signals collected from patients' brains through electroencephalogram (EEG) and magnetic resonance imaging (MRI).
Hasey is also testing software that mines the EEGs of schizophrenia patients to determine whether they will respond to clozapine. The developers of the software trained two programs to determine critical EEG features, and one learned to forecast the success of depression treatment options with 80 percent to 85 percent accuracy while the other ascertained whether or not a schizophrenia patient would respond to clozapine with up to 89 percent accuracy.
The developers have completed clinical trials of an EEG-based system that can help predict which type of depression drug a patient will respond to with 65 percent accuracy, but they say using a machine-learning algorithm would boost that accuracy.
Meanwhile, researchers headed by University of New Mexico professor Vince D. Calhoun are developing a machine-learning system that can draw distinctions between people suffering from schizophrenia and those with bipolar disorder on the basis of functional MRI brain scans.
From IEEE Spectrum
View Full Article