Investigating the neural basis of representation-mediated learning in humans
Poster Session B - Sunday, March 30, 2025, 8:00 – 10:00 am EDT, Back Bay Ballroom/Republic Ballroom
Xiaolin Qiao1 (xiaolinqiao@brandeis.edu), Lauren A. Wolters1, Liam P. McMahon1, Jared G. Newell1, James D. Howard1; 1Brandeis University
The representation-mediated learning (RML) task has been used in rodents to demonstrate that mental representation of chemosensory stimuli can form associations with aversive stimuli, even if the chemosensory stimulus is not physically present. This behavioral effect is amplified in animal models of the positive symptoms of schizophrenia, suggesting that RML holds promise for identifying neural biomarkers of psychosis. We recently developed a human behavioral version of the RML task by showing that expected, but not delivered, odors can form associations with an aversive sound. In the current study, we aim to address the underlying brain mechanism by conducting an adapted RML task while participants undergo fMRI. Participants first learn associations between visual symbols and two distinct appetitive food odors. We then acquire pleasantness ratings for symbols and odors before and after one of the symbols is paired with an aversive sound. Preliminary results (n = 9) showed a selective decrease in pleasantness for the odor previously paired with the aversively conditioned symbol, similar to our previous study. Multivoxel pattern analysis of fMRI data indicate that mental representations of expected odors are embedded in distributed patterns of orbitofrontal cortex activity. Further planned analyses will test whether these mental representations are reactivated by the visual symbols during the aversive conditioning, and whether this predicts the propensity for odors to enter into association with the sound. Such findings may reveal novel targets for noninvasive stimulation-based treatments of psychotic symptoms.
Topic Area: THINKING: Decision making