UAlberta researchers marry AI-learning and psychiatry for schizophrenia diagnosis
Research out of Edmonton published in Nature could pave the way for an objective diagnosis of schizophrenia using machine learning.
The research team from the Alberta Machine Intelligence Institute (Amii) created a model called Ensemble algorithm with Multiple Parcellations for Schizophrenia prediction, or EMPaSchiz for short (read as ‘emphasis’). The study found an 87 percent accuracy in diagnosis, compared to the baseline of 53 percent when left to chance.
The University of Alberta researchers also believe their model is the first to be tested in patients using only data from patients who haven’t used antipsychotic medication before.
Sunil Kalmady, lead author on the study, worked on the tool under the supervision of both Russ Greiner, professor in the Department of Computing Science, and Andrew Greenshaw, professor and associate chair in the Faculty of Medicine and Dentistry’s Department of Psychiatry.