Poster No:
316
Submission Type:
Abstract Submission
Authors:
Jeremy Fleming1, Jenna Blujus1, Hwamee Oh1
Institutions:
1Brown University, Providence, RI
First Author:
Co-Author(s):
Introduction:
Aging is associated with cognitive decline across multiple domains, underpinned by disruptions to functional connectivity within and between large-scale networks. Connectivity between brain regions is traditionally determined using a bivariate correlation approach. However, connectivity matrices resulting from bivariate correlations are afflicted with a considerable number of spurious associations, reflecting indirect connections or associations due to confounders. A recent method called combinedFC was developed to eliminate such spurious connections by sequentially applying partial and bivariate correlation methods. Through simulations and application to fMRI data in young adults, combinedFC was shown to remove spurious connections and improve causal inference. In the current study, we implemented bivariate alone and combined FC approaches in a sample of young and older adults to examine the impact of functional connectivity estimation method on resulting connections, or edges retained, in healthy young and older adults.
Methods:
The sample consisted of 30 young adults (age range 18-30; M = 20.03, SD = 2.95; 20 females) and 18 older adults (age range 60-75; M = 65.33, SD = 4.37; 11 females). T1-weighted MRI and resting state fMRI (rs-fMRI) data were collected on a 3T Siemens scanner. The data were preprocessed and denoised using ENIGMA HALFpipe. Average signals from ICA noise components, white matter, and CSF were removed. The Power atlas was utilized to extract the average time series from 264 regions of interest. Functional connectivity matrices were calculated using two methods: (1) bivariate correlation alone, and (2) combinedFC. At the subject level, edges retained were determined using an alpha cutoff at 0.01. A two-way ANCOVA model was conducted using R to examine the interaction of age group (young, older) and functional connectivity method (bivariate, combinedFC) on the proportion of edges retained in functional connectivity matrices, controlling for sex and years of education. Posthoc simple contrasts were conducted within each level of functional connectivity estimation method.
Results:
As expected, there was a significant main effect of method on the proportion of edges retained (F(1, 90)=352.83, p<.001), with the bivariate approach resulting in a greater proportion of edges retained (p<.001). There was also a significant interaction between age group and functional connectivity estimation method on the proportion of edges retained (F(1, 90)=6.07, p=.016). Simple contrasts showed a trend that old retained a greater proportion of edges than young when the bivariate method was applied (p=.076), but this age difference was no longer evident when the combinedFC method was utilized (p=.244).
Conclusions:
In line with past work, our results showed that compared to a bivariate approach, combinedFC eliminated a significant portion of potentially spurious edges, regardless of age group. We extend past findings by demonstrating that age group differences in edges retained via bivariate approaches were eradicated using the combined FC approach. The combinedFC approach may provide a complementary perspective to traditional bivariate estimates of functional connectivity and uncover alterations in direct causal connections, estimated from resting state data, which underlie age-related cognitive decline.
Disorders of the Nervous System:
Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s) 1
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 2
Keywords:
Aging
FUNCTIONAL MRI
1|2Indicates the priority used for review
Provide references using author date format
Sanchez-Romero, R., and Cole, M.W. (2021) "Combining multiple functional connectivity methods to improve causal inferences" Journal of Cognitive Neuroscience. https://doi.org/10.1162/jocn_a_01580
Jenna Blujus, Michael W. Cole, Elena Festa, Stephen Correia, Stephen Salloway, William Heindel, Hwamee Oh. Functional Redundancy of the Posterior Hippocampi, but not Anterior Hippocampi or Left Frontal Cortex, is Disrupted in Pathological Brain Aging. [preprint] bioRxiv 2022.06.18.496543; doi: https://doi.org/10.1101/2022.06.18.496543
Jenna Blujus & Hwamee Oh. Denoising approach affects diagnostic differences in brain connectivity across Alzheimer's continuum. [preprint] bioRxiv 2022.06.16.496466; doi: https://doi.org/10.1101/2022.06.16.496466