Poster No:
1719
Submission Type:
Abstract Submission
Authors:
John Sampaio1, Norman Farb1, Stephen Strother1
Institutions:
1University of Toronto, Toronto, Ontario
First Author:
Co-Author(s):
Introduction:
Resting-state brain activity exhibits structured spatiotemporal patterns known as resting state networks in human fMRI research. The strength of resting state functional connectivity (rs-FC) may reveal individual differences in brain activity, and seem to be predictive of spatial brain activation patterns during task performance (Tavor, 2016). The study of these networks has focused on static structures, i.e., a consistent set of networks derived from resting state data. Yet resting state networks are not static. They dynamically reconfigure over time; shifting between many different transient network configurations, termed "brain states."
Recent studies have begun to explore the dynamic nature of these resting state networks and their potential role in shaping aspects of human cognition. One proposed function is that the exploration of different configurations during rest allows the brain to rehearse brain states and optimize neural networks in anticipation of future stimulation (Deco, 2011). The dynamic repertoire hypothesis suggests that spontaneous engagement in distinct brain states during rest could predict the frequency of engaging these brain states during cognitive tasks to predict behavioural performance.
Using a sliding time window approach, we tested the dynamic repertoire hypothesis. We predicted that performance on a given task would be correlated with the frequency of entering certain brain states. Additionally, the brain state most correlated with task performance was also expected to be most similar to the functional connectivity participants exhibited while performing said task.
Methods:
We analyzed the fMRI data of 250 participants collected by Ohio State University. In addition to 5 minutes of resting state, participants completed a series of neurocognitive tasks while in the MRI scanner. We examined the go-nogo and 2-back tasks, which were designed to measure inhibitory control and working memory, respectively. Neuroimaging data were preprocessed using the in-house developed PreProcessing Resting Imaging Data with Scrubbing pipeline (Eusebio, 2023).
We extracted region of interest (ROI) time courses using a sliding time window. ROIs were defined using the 246 ROI Brainnetome atlas (Fan, 2016). We used K-means clustering to cluster rs-FC matrices into k brain states. Static task FC was computed across the entire task fMRI time course. Participants' performance metrics were correlated with time spent in each brain state to determine which brain state was associated with better performance. We used Manhattan distance to determine the similarity between each resting brain state and task FC. Shorter Manhattan distances mean the matrices were more alike.
Results:
The results were partially consistent with the hypothesis. As expected, the resting brain state most associated with go-nogo task performance also had the shortest Manhattan distance from the task-based go-nogo FC matrix. However, this was not the case for the 2-back task, as the brain state most correlated with task performance did not have the shortest Manhattan distance from the task-based go-nogo FC matrix.
Conclusions:
The results of this study provide tenuous support for the hypothesis that brain state exploration during rest allows the brain to prepare for future stimulation. While our go-nogo results were consistent with this, the 2-back results were not. Still, the 2-back results were somewhat consistent since the brain state most correlated with 2-back performance did have the second-lowest Manhattan distance. Future studies may expand these findings using a wider variety of neurocognitive tasks. It could be that some neural circuits benefit more from this spontaneous rehearsal than others.
Higher Cognitive Functions:
Executive Function, Cognitive Control and Decision Making 2
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI)
Connectivity (eg. functional, effective, structural)
fMRI Connectivity and Network Modeling 1
Keywords:
Cognition
FUNCTIONAL MRI
Other - dynamic repertoire; clustering; sliding time-windows
1|2Indicates the priority used for review
Provide references using author date format
Deco, G. (2011), 'Emerging concepts for the dynamical organization of resting-state activity in the brain', Nature Reviews. Neuroscience, vol. 12, no. 6, pp. 43–56
Eusebio, J. (2023), 'PreProcessing Resting Imaging Data with Scrubbing (PPRIDS) [Linux]', https://github.com/johneusebio/PPRIDS-pipeline
Fan, L. (2016), 'The Human Brainnetome Atlas: A New Brain Atlas Based on Connectional Architecture', Cerebral Cortex, vol. 26, no. 8, pp. 3508–3526
Tavor, I. (2016), 'Task-free MRI predicts individual differences in brain activity during task performance', Science, vol. 352, no. 6282, 216–220