Bengaluru-based National Institute of Mental Health and Neuro Sciences (Nimhans) has created artificial-intelligence (AI) capabilities to accurately diagnose schizophrenia in collaboration with Canada’s University of Alberta.
The study, published in a journal called npj Schizophrenia, shows how the researchers leverage data from magnetic resonance imaging (MRI) of affected brains to create AI capabilities for diagnosis.
The model, named as EMPaSchiz (read as ‘emphasis’), used MRI data of over 81 patients who were diagnosed with schizophrenia – AI models usually munch on much bigger data sets. The researchers had to ensure that the patients had to stay away from any psychoactive drugs, as that could hamper with the test results.
The study said that EMPaSchiz had an accuracy of 90% and outperformed other models that were built to diagnose schizophrenia.
The researchers report that the future utilisation of AI techniques can effectively help develop markers to better understand schizophrenia, and its origins. It is to be noted that current medicine only targets symptoms of schizophrenia, but facts regarding the disorder as a whole still remain elusive.
The AI study on schizophrenia comes after Microsoft recently partnered Apollo Hospitals to start a heart disease risk score API (application programme interface) powered by AI in Indian villages. The Satya Nadella-led tech company has also partnered with Devi Prasad Shetty-led Narayana Health to provide real-time data analysis of individuals to detect cancer.
AI is being touted as the future of healthcare -- from tasks such as managing medical records, compiling and analysing information, data management has been the most widely used application of AI and digital automation.
However, better technologies and more powerful algorithms have allowed scientists to take on complex tasks such as digital consultations, wherein apps take note of medical history and current symptoms to provide a fairly accurate analysis of the disorder.
Another area in medicine where AI could come on its own is drug testing. Development of pharmaceuticals via clinical trials usually takes years and costs billions of dollars in capital. AI could prove useful in such instances. For example, AI was used to screen medicines that had the potential to be redesigned to fight Ebola.
The Ebola programme found two medicines that could counter the deadly virus.