Frequency-specific brain network architecture in resting-state fMRI

Abstract

The analysis of brain function in resting-state network (RSN) models, ascertained through the functional connectivity pattern of resting-state functional magnetic resonance imaging (rs-fMRI), is sufficiently powerful for studying large-scale functional integration of the brain. However, in RSN-based research, the network architecture has been regarded as the same through different frequency bands. Thus, here, we aimed to examined whether the network architecture changes with frequency. The blood oxygen level-dependent (BOLD) signal was decomposed into four frequency bands—ranging from 0.007 to 0.438 Hz—and the clustering algorithm was applied to each of them. The best clustering number was selected for each frequency band based on the overlap ratio with task activation maps. The results demonstrated that resting-state BOLD signals exhibited frequency-specific network architecture; that is, the networks finely subdivided in the lower frequency bands were integrated into fewer networks in higher frequency bands rather than reconfigured, and the default mode network and networks related to perception had sufficiently strong architecture to survive in an environment with a lower signal-to-noise ratio. These findings provide a novel framework to enable improved understanding of brain function through the multiband frequency analysis of ultra-slow rs-fMRI data.

Network representation of multicellular activity in pancreatic islets: Technical considerations for functional connectivity analysis

“…However, due to relatively demanding computational approaches, encompassing both data extraction and subsequent analyses of coordinated functioning, obtaining patterns of functional connectivity is not straightforward and can easily become ambiguous. In neuroscience, where the greatest progress in this field has been made, it has become evident that objectively assessing connectivity patterns is challenged by various objective reasons tied to experimental variations and computational methodologies, such as thresholding techniques [82][83][84], techniques used for data pooling [85,86], number of sensors used to record brain activity [87,88], and the selection of frequency intervals [89,90]. Most importantly, similar issues are witnessed in the network-based analysis of spatiotemporal cellular dynamics in islets.…”

Frequency modulation increases the specificity of time-resolved connectivity: A resting-state fMRI study

“…Secondly, much of the important information of rsfMRI activity resides in low frequencies (Biswal et al, 1995;Kalcher et al, 2014;Van Dijk et al, 2010). Although recent studies have increasingly emphasized the presence of rich information in higher frequencies within rsfMRI, challenging our prior understanding (Chen & Glover, 2015;DeRamus et al, 2021;Kajimura et al, 2023), the low frequencies contribute the greatest power it seems. Based on these two points, the method proposed here that allows us to select smaller window sizes for SWPC while improving its estimation can be quite useful.…”

Distinct functional subnetworks of cognitive domains in older adults with minor cognitive deficits

“… 45 , 46 Contradictory evidence exists regarding spatial patterns of functional networks, where some results suggest that functional connectivity patterns are similar across the frequency range 43 , 45 while other findings indicate compositional differences between frequency bands. 47 , 48 While our study was underpowered to properly investigate frequency subbands and multiband acquisition protocols are preferred, future studies could investigate whether the associations between functional connectivity and distinct cognitive functions differ across the frequency spectrum and what the behavioural consequences of these differences might be to further improve our understanding of brain function in health and disease.…”