Selective diagnostic discrimination of schizophrenia across cognitive control brain regions using multi-voxel pattern classification

Poster C86, Saturday, October 22, 11:30 am - 1:00 pm, Le Baron

Alan Ceaser1, Benjamin Acland2, Deanna Barch2; 1Stanford University, 2Washington University in St. Louis

Recently we found that altered associative striatum, but not prefrontal, activity during cognitive control positively correlated with aberrant salience symptoms in schizophrenia. This finding suggests that striatal activity may be tied to schizophrenia etiology and could demonstrate selective diagnostic discrimination when compared with other brain regions. To quantify the predictive ability of brain activity for group status we conducted a follow up multi-voxel pattern classification analysis on an existing data set of 22 patients and 20 controls during a task that measured updating and interference control in working memory. We trained the classifier on brain activity within anatomical masks of the basal ganglia, prefrontal, and parietal cortices during correct and incorrect distracter and update trials, with the training focused on distinguishing patients from controls. Classification accuracy was determined for each 2 second frame of the trial to determine if significant classification accuracy occurs throughout the trial, or if accuracy is higher for particular events within the trial (i.e. during distracter presentation versus the presentation of the memory set or probe). Results are forthcoming, but we predict that only activity within the basal ganglia mask during incorrect distracter trials will significantly predict group status, and only in response to distracter presentation.

Topic Area: Translational Research

Back to Poster Schedule