Early Career Award for Neuroscience Researcher Jason Smucny
Grant is targeted at identifying biomarkers for psychotic illness using deep learning
- Post-doctoral fellow Jason Smucny has received a prestigious Mentored Research Scientist Career Development Award from the National Institutes of Health. The award is designed to provide support and protected time for supervised career development. Smucny
Jason Smucny. an assistant adjunct professor in the Department of Psychiatry and Behavioral Sciences, has received a prestigious Mentored Research Scientist Career Development Award from the National Institutes of Health. Smucny is also a fellow in the Center for Neuroscience, working with Cameron Carter.
The award provides support and protected time for intensive supervised career development in the biomedical, behavioral or clinical sciences leading to research independence.
Smucny’s research uses computational psychiatry, which combines different data types to identify ways to improve the understanding and treatment of mental illness.
The award will allow him to undergo a multi-year training and mentoring program to learn machine learning and deep learning skills. Smucny will also study with Ian Davidson, a professor in the Department of Computer Science and an expert in machine learning and data mining algorithm development.
“The goal is to use deep learning with data from the ABCD Study dataset to look for biomarkers that might be able to predict the longitudinal trajectory of psychotic-like experiences in children, as these may be risk factors for future mental illness in adulthood,” Smucny said, referring to the largest long-term study of brain development and child health in the United States.
Biomarkers could help those identified to be at risk receive early treatment.
Another area where deep learning may be able to help people with psychosis is with medication. Currently there are no biomarkers that can predict how a patient with schizophrenia will respond to a particular medication.
“We recently conducted an analysis of first-episode patients enrolled for treatment in the UC Davis EDAPT clinic. Less than 60 percent of the patients experienced a positive clinical response at the end of one year,” said Smucny. “Trying to find the right medication can lead to long delays in treatment, which on average leads to worse outcomes. The whole process is frustrating.”
With deep learning, researchers may be able to match patients with better treatments earlier in the course of illness.
Although Smucny is hoping to find specific biomarkers for psychosis, the nature of deep learning means the results may be unpredictable.
“With deep learning, you don’t have to tell the machine what features of a dataset are important before you start it — the machine figures it out on its own. That’s one of the advantages. You aren’t going to know what you are going to find,” Smucny said.
Smucny begins his training and mentoring program this fall.
This story originally appeared in UC Davis Health Inside Out.