Poster No:
552
Submission Type:
Abstract Submission
Authors:
Brian Kraus1, Richard Zinbarg1, Robin Nusslock1, Michelle Craske2, Caterina Gratton3
Institutions:
1Northwestern University, Evanston, IL, 2University of California, Los Angeles, Los Angeles, CA, 3Florida State University, Tallahassee, FL
First Author:
Co-Author(s):
Introduction:
Internalizing psychopathology is characterized by elevated negative affect, such as major depressive disorder (MDD) which shows significant symptom fluctuations over time. "Precision" fMRI has demonstrated that functional connectivity (FC) is reliable within individuals across sessions with >40 minutes of artifact free data at each session. This method can be used to identify idiosyncratic areas of FC which greatly deviate from canonical network FC, which we call network variants. However, while network variants are highly stable in neurotypical individuals (r > .9; Kraus et al., 2021), they have not yet been explored in psychiatric contexts where temporal changes in symptoms occur. Here, we sought to evaluate the stability of network variants in individuals diagnosed with internalizing disorders versus healthy controls.
Methods:
The current study followed participants longitudinally from ages 18-19 to 21-22 (see Young et al., 2021). At baseline and the 3-year follow-up visits, 150 participants completed approximately 65 minutes of resting-state and task fMRI scans. The data from these scans was processed using fMRIPrep (Esteban et al., 2019). FC processing was performed using custom scripts which included nuisance regression and censoring high motion frames (fFD > .1; (Gratton et al., 2020)), and mapping functional data to surface space. Building off our past work (Kraus et al., 2021), task scans were modeled with a GLM and the residuals were combined with rest to estimate network variants for each participant. Only participants with >40 minutes of fMRI data after censoring at each timepoint were analyzed.
To estimate network variants, the pairwise correlation between the timeseries of each vertex on the surface and every other vertex was calculated. Next, a row-wise correlation was performed between each participant's matrix and an independent group-average matrix. In the resulting spatial map, a low value denoted that a given vertex had a very dissimilar FC profile from what would typically be expected in the group-average at that location, and vice-versa. For analysis, variant maps were estimated at baseline and 3-year follow-up for individuals who had met lifetime criteria for anxiety or depressive disorders versus healthy controls.
To quantify similarity over time, each individual's variant map was correlated longitudinally with their own map as well as every other individual's map. Then, the mean value for the correlation between each individual and every other individual was compared to each individual's actual longitudinal correlation using a paired t-test.
Results:
The longitudinal stability of network variants was not significantly different between healthy controls (r = .85) and the lifetime depressive disorder group (r = .85, t(87) = .12, p = .91, d = .03), or the lifetime disorder anxiety group (r = .86, t(89) = .92, p = .36, d = .19). In each group, network variant maps for an individual looked significantly more similar to the same person at a 3-year longitudinal follow up than to other individuals in the study: healthy controls (t(71) = 50.46, p < .001, d = 5.94), the lifetime depressive disorder group (t(16) = 28.1, p < .001, d = 6.82), and the lifetime disorder anxiety group (t(18) = 26.78, p < .001, d = 6.14). Thus, variant maps showed similar stability over time for those in the anxiety and depressive groups versus the healthy controls, and all three groups' variant maps were much more similar to themselves over time versus other individuals in their group.
Conclusions:
These findings suggest that network variants are trait-like in their stability within individuals, regardless of psychiatric history. This suggests that network variants, and likely FC in general, are more sensitive to longitudinally stable factors and less to shorter-term changes associated with fluctuations in psychopathology symptoms. Future work will have to determine the temporal relationship between FC and psychiatric symptoms.
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 1
Modeling and Analysis Methods:
fMRI Connectivity and Network Modeling 2
Keywords:
Affective Disorders
Anxiety
FUNCTIONAL MRI
Psychiatric Disorders
Systems
Other - Functional Connectivity
1|2Indicates the priority used for review
Provide references using author date format
Esteban, O. (2019). fMRIPrep: A robust preprocessing pipeline for functional MRI. Nature Methods, 16(1), 111–116.
Gratton, C. (2020). Removal of high frequency contamination from motion estimates in single-band fMRI saves data without biasing functional connectivity. NeuroImage, 116866.
Kraus, B. T. (2021). Network variants are similar between task and rest states. NeuroImage, 229, 117743.
Young, K. S. (2021). Dysregulation of threat neurociruitry during fear extinction: The role of anhedonia. Neuropsychopharmacology, 1–8.