Dynamics of brain-wide patterns of resting-state fMRI traveling waves is related to arousal levels

Abstract

Brain waves propagate as traveling waves, reflecting the brain’s functional organization and information transmission. Macroscale traveling waves observed in blood-oxygen-level-dependent (BOLD) signals offer insights into large-scale neural coordination. Since multiple brain-wide traveling wave patterns can exist, understanding their temporal dynamics may illuminate how neural information processing unfolds. As infra-slow BOLD activity (<0.1 Hz) is known to correlate with arousal, we hypothesized that temporal dynamics of brain-wide traveling wave patterns in this band are modulated by arousal levels. We analyzed a publicly available resting-state functional MRI (rsfMRI) dataset (N = 27) to examine the relationship between arousal and brain-wide traveling wave patterns. Infra-slow BOLD traveling waves were quantified using the 3-D local phase gradient (LPG) method. Wave patterns were identified by clustering LPG directions via k-means clustering, with the optimal number of clusters (k = 2) determined using the silhouette method. We then examined temporal interactions between these clusters and arousal levels, defined using pupillometry data as the lowest and highest 20 percentiles. The two clusters showed distinct characteristics. Cluster 1, compared to Cluster 2, included more sustained and prominent traveling waves, particularly in the salience ventral attention and frontoparietal control networks. Cluster 1 occurred more frequently during low-arousal periods, whereas Cluster 2 appeared more during high-arousal periods. Moreover, Cluster 2 persisted longer during high- than low-arousal states. These findings suggest that Cluster 1 corresponds to low-arousal dynamics and Cluster 2 to high-arousal dynamics. Low-arousal states were associated with more stable and sustained traveling waves, while high-arousal states involved more transient and variable patterns. Traveling directions differed by network, indicating spatial heterogeneity within each cluster. While prior studies proposed simple top-down vs. bottom-up traveling directions for different arousal levels, our results reveal more complex, brain-wide dynamics tied to arousal, suggesting a richer framework for understanding large-scale neural coordination.