Learning and dysfunctional higher order functional connectivity in schizophrenia
Poster No:
1583
Submission Type:
Abstract Submission
Authors:
Hady Saad1, John Kopchick2, Patricia Thomas2, Usha Rajan2, Dalal Khatib2, Caroline Zajac-Benitez1, Luay Haddad2, Alireza Amirsadri2, Jeffrey Stanley2, Vaibhav Diwadkar1
Institutions:
1Wayne State University, Detroit, MI, 2Wayne State University, Department of Psychiatry, Detroit, MI
First Author:
Co-Author(s):
Introduction:
Traditional methods for assessing second order functional connectivity (FC) rely on the quantification of bivariate relationships between pairs of nodes (Silverstein, Bressler, & Diwadkar, 2016). These pairwise characterizations typically rely on zero log correlations that can then be used in the service of understanding connectivity and dysconnectivity in conditions like schizophrenia (SCZ)(Meram et al., 2023). In a large connectomic space, these pairwise characterizations bely relationships between individual nodes and all other nodes in the network. Thus, in a connectivity matrix (of size k), the connectivity vector of any node i consists of k elements. Between any two nodes i and j, it is possible to estimate their Higher Order Functional Connectivity (HOFC) in the network from their vector cross product (Zhang et al., 2016). This measure captures similarities (or dissimilarities) in the cross-matrix similarity of the connectivity profiles of nodes. Here, we provide the first application of HOFC to recover dysfunctional learning induced profiles in SCZ (compared to Healthy Controls, HC).
Methods:
fMRI data (Siemens Verio 3T) were acquired in 88 participants (49 SCZ) while they engaged in object-location learning (Stanley et al., 2017) separated into epochs for Encoding (objects shown in their associated locations) and Retrieval (locations cued for retrieving object identity). fMRI data were processed in SPM 12 (standard methods). In preparation for HOFC analyses, in each participant, time series were extracted from 246 functionally defined cerebral parcels (Fan et al., 2016) for the computation of pairwise functional connectivity in each of Encoding and Retrieval (30,135 pairs). Next, HOFC was computed between each of the 246 FC vectors for data in each participant and condition. Finally, inter-group differences in HOFC in each condition were identified (p<.05) and separated by direction (HC > SCZ, SCZ > HC).
Results:
The results are depicted for each of Encoding (Figure 1) and Retrieval (Figure 2). In each figure, the central brain maps are frequency maps where the shading indicates the frequency with each that region is part of a pathway with a significant HOFC differences (Red: HC greater; Blue: SCZ greater). These frequencies are redundantly depicted in the surrounding circular bar plots. The chord diagrams in the satellites break down the observed effects by lobe (frontal, temporal, parietal, insular, limbic, visual and sub-lobar nuclei). Here, each chord connects a pair with significant differences in HOFC. As seen, during Encoding SCZ are characterized by decreased HOFC involving frontal, temporal, parietal and visual regions, but a converse (and presumably compensatory) increase in HOFC involving sub-lobar nuclei. This pattern largely repeats itself for Retrieval.
Conclusions:
HOFC focuses on the correlation of spatial or topographical FC properties rather than actual temporal correlations, and while the neuronal correlates are obscure HOFC applied to resting state has captured salient differences between patients with autism (Zhao, Zhang, Rekik, An, & Shen, 2018) and mild cognitive impairment (Zhang et al., 2019). The current application in task-based fMRI data in schizophrenia is unique in terms of both the class of fMRI data (task-based) and the target population (schizophrenia). In our ongoing work, we are attempting to clarify the level of information embedded in HOFC that exceeds that embedded in conventional second order FC data.
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 2
Learning and Memory:
Learning and Memory Other
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 1
Keywords:
FUNCTIONAL MRI
Learning
Memory
Schizophrenia
1|2Indicates the priority used for review
Provide references using author date format
Meram, E. D., Baajour, S., Chowdury, A., Kopchick, J., Thomas, P., Rajan, U., . . . Diwadkar, V. A. (2023). The topology, stability, and instability of learning-induced brain network repertoires in schizophrenia. Network Neuroscience, 7(1), 184-212. doi: 10.1162/netn_a_00278
Silverstein, B., Bressler, S., & Diwadkar, V. A. (2016). Inferring the dysconnection syndrome in schizophrenia: Interpretational considerations on methods for the network analyses of fMRI data. Front Psychiatry, 7, 132. doi: 10.3389/fpsyt.2016.00132
Stanley, J. A., Burgess, A., Khatib, D., Ramaseshan, K., Arshad, M., Wu, H., & Diwadkar, V. A. (2017). Functional dynamics of hippocampal glutamate during associative learning assessed with in vivo 1H functional magnetic resonance spectroscopy. Neuroimage, 153, 189-197. doi: 10.1016/j.neuroimage.2017.03.051
Zhang, H., Chen, X., Shi, F., Li, G., Kim, M., Giannakopoulos, P., . . . Shen, D. (2016). Topographical Information-Based High-Order Functional Connectivity and Its Application in Abnormality Detection for Mild Cognitive Impairment. J Alzheimers Dis, 54(3), 1095-1112. doi: 10.3233/JAD-160092
Zhang, H., Giannakopoulos, P., Haller, S., Lee, S. W., Qiu, S., & Shen, D. (2019). Inter-Network High-Order Functional Connectivity (IN-HOFC) and its Alteration in Patients with Mild Cognitive Impairment. Neuroinformatics, 17(4), 547-561. doi: 10.1007/s12021-018-9413-x
Zhao, F., Zhang, H., Rekik, I., An, Z., & Shen, D. (2018). Diagnosis of Autism Spectrum Disorders Using Multi-Level High-Order Functional Networks Derived From Resting-State Functional MRI. Front Hum Neurosci, 12, 184. doi: 10.3389/fnhum.2018.00184