Schizophrenia Network Dysfunction During Associative Learning Revealed Using Dynamic Graphical Models

Tristan Attisha, Shahira Baajour, Asadur Chowdury, Vaibhav Diwadkar, Department of Psychiatry and Behavioral Neurosciences, Wayne State University School of Medicine, 540 E Canfield St, Detroit, MI 48201

Schizophrenia is a mental disorder characterized by dysfunctional brain networking, expressed during learning and memory. Interest on its biological correspondence has focused on network profiles of the frontal, unimodal and hippocampal regions due to their hypothesized roles during associative learning. Previous studies have described dysfunctional modulation of brain networks during associative learning, and have presented the importance of frontal-hippocampal and frontal-unimodal network associations during memory encoding and retrieval. We expand on current knowledge by testing network functionality of patients during an object-location associative learning paradigm in order to detect associative memory proficiency. fMRI data were collected for 79 subjects (46 SCZ). The paradigm is comprised of eight memory blocks divided into encoding and retrieval epochs. During encoding, nine objects were presented to the subjects in a 3 columns by 3 rows spatial grid. Following a 9 second rest period, the subsequent retrieval epochs consist of a location being cued and subjects stating the associated object. Following data collection, the fMRI data were preprocessed using standard methods in SPM 12. Coactivated peaks were then used to extract a time series (p<0.05). Dynamic Graphical Models, portraying dynamic, directed functional connectivity between parent and children nodes, were created in MATLAB, and the results were categorized among the right and left hemispheres of 6 nodes for encoding and retrieval: HPC, DLPFC, dACC, SPC, ITG, and FG. Following, in-degree and out-degree centrality models were created to portray the number of parent nodes (in-degree) and children nodes (out-degree) connected to Node X. The SCZ group were observed to have a lower in-degree and out-degree value in the dACC-R, DLPFC-L, ITG-L, SPC-R and HPC-R. Our findings support the primacy of frontal-hippocampal regions, and the lack of interaction between parent-child relationships may evince an inefficient attempt to build associations and maintain information during encoding and recollection in SCZ. 

Additional Abstract Information

Presenter: Tristan Attisha

Institution: Wayne State University

Type: Poster

Subject: Psychology

Status: Approved

Time and Location

Session: Poster 11
Date/Time: Wed 3:00pm-4:00pm
Session Number: 7002