Longitudinal BOLD variability change coincides with executive function change
Poster No:
2350
Submission Type:
Abstract Submission
Authors:
Tzu-Chen Lung1, Ekarin Pongpipat1, Karen Rodrigue1, Kristen Kennedy1
Institutions:
1The University of Texas at Dallas, Dallas, TX
First Author:
Co-Author(s):
Introduction:
Variability in the fluctuations of the brain's functional activity has become a sensitive metric for explaining age-related cognitive performance [2,3]. Blood-oxygen-level-dependent (BOLD) variability decreases with increasing task demands and advanced age [7]. Most studies interpret BOLD variability as the dynamic range of the brain resources or coherence between regions [2,3]. However, our previous study showed the opposite result of greater BOLD variability associated with older age and worse performance [1]. One possible factor that could explain these divergent findings is the regions/networks utilized to retrieve BOLD variability signals [5]. Additionally, a recent longitudinal study on resting state BOLD variability indicated both cortical and subcortical contribution, with BOLD variability loss associated with loss of cognition. However, how BOLD variability and cognition may longitudinally change together during a task paradigm are still unknown.
The current study explored BOLD variability change during an n-back paradigm over a period of 4-years, and examined how changes in BOLD variability were linked to age and cognition changes utilizing the behavioral partial least squares (PLS) method [6].
The current study explored BOLD variability change during an n-back paradigm over a period of 4-years, and examined how changes in BOLD variability were linked to age and cognition changes utilizing the behavioral partial least squares (PLS) method [6].
Methods:
Participants included 87 healthy adults aged 20-86 yrs at baseline who returned for follow-up assessment. The mean square of successive differences (MSSD) was calculated as the proxy of BOLD variability. Participants' cognitive performance was assessed via 3 out-of-scanner tasks: fluid intelligence (Cattell culture fair intelligence test, CFIT), switching and inhibition (Delis-Kaplan executive function system, DKEFS) scores, and processing speed (digit symbol substitution task); and one in-scanner task: updating (mean accuracy on the n-back task). All change metrics from BOLD variability and cognition were calculated as Time 2 minus Time 1, and partialed out baseline age before the PLS analysis. Significance of the latent variable from the PLS analysis was assessed via 1000 permutations followed by 1000 bootstraps. The whole-brain bootstrap ratios (BSR; each voxel's weight/bootstrapped standard error) were thresholded at values of ±3. Each of the cognitive change scores were then correlated with the latent brain score, and the significance of the correlations were determined by 95% confidence intervals.
Results:
PLS results indicated one significant latent component. The whole-brain, BSR-thresholded map of the first component resulted in widespread cortical and subcortical areas involving lingual gyrus, cingulate, cerebellum, superior frontal, pre/post central gyrus, and thalamus (Figure 1). The latent brain score was positively correlated with the latent behavioral score (r(85) = 0.37, p < .001). However, over time, changes in cognition per se indicated both positive and negative significant associations with the latent brain score of changes in BOLD MSSD. Specifically, increased BOLD variability was associated with better switching and inhibition performance, and worse updating and processing speed performance (Figure 2).
Conclusions:
The current study explored longitudinal changes in task-based BOLD variability and cognition using PLS. By correlating the change in BOLD variability with change in age and cognition, widespread cortical and subcortical regions demonstrated significant change-change associations. Partially consistent with the previous longitudinal study [4], loss of switching and inhibition was associated with BOLD variability loss. Additionally, over time, loss of updating and processing speed with increasing BOLD variability was found, which supports our previous argument that increased BOLD variability might be a detrimental effect of aging [1]. These findings reveal the possibility that BOLD variability could be both beneficial and detrimental to aging, depending on the type of cognitive resources needed.
Higher Cognitive Functions:
Executive Function, Cognitive Control and Decision Making 2
Lifespan Development:
Aging
Modeling and Analysis Methods:
Multivariate Approaches
Novel Imaging Acquisition Methods:
BOLD fMRI 1
Keywords:
Aging
Other - longitudinal; BOLD variability, executive function
1|2Indicates the priority used for review
Provide references using author date format
Garrett, D.D. (2011), 'Moment-to-moment signal variability in the human brain can inform models of stochastic facilitation now', Nature Reviews Neuroscience,12(10).
Garrett, D.D. (2013), 'Moment-to-moment brain signal variability: A next frontier in human brain mapping?', Neuroscience and Biobehavioral Reviews, 37(4): p. 610-624.
Garrett, D.D. (2021), 'Lost Dynamics and the Dynamics of Loss: Longitudinal Compression of Brain Signal Variability is Coupled with Declines in Functional Integration and Cognitive Performance', Cerebral Cortex, 31(11): p. 5239-5252.
Goodman, Z.T. (2023), 'Brain Signal Variability and Executive Functions Across the Lifespan', Network Neuroscience, 2023: p. 1-36.
McIntosh, A.R. (2004), 'Partial least squares analysis of neuroimaging data: applications and advances', Neuroimage, 23: p. S250-S263.
Rieck, J.R. (2022), 'Reduced modulation of BOLD variability as a function of cognitive load in healthy aging', Neurobiology of Aging, 112: p. 215-230.