Origin of the Time Lag Phenomenon and the Global Signal in Resting-State fMRI

Abstract

The global mean signal of resting-state fMRI (rs-fMRI) shows a characteristic spatiotemporal pattern that is closely related to the pattern of vascular perfusion. Although being increasingly adopted in the mapping of the flow of neural activity, the mechanism that gives rise to the BOLD signal time lag remains controversial. In the present study, we compared the time lag of the global mean signal with those of the local network components obtained by applying temporal independent component analysis to the resting-state fMRI data, as well as by using simultaneous wide-field visual stimulation, and demonstrated that the time lag patterns are highly similar across all types of data. These results suggest that the time lag of the rs-fMRI signal reflects the local variance of the hemodynamic responses rather than the arrival or transit time of the stimulus, whether the trigger is neuronal or non-neuronal in origin as long as it is mediated by local hemodynamic responses. Examinations of the internal carotid artery signal further confirmed that the arterial signal is tightly inversely coupled with the global mean signal in accordance with previous studies, presumably reflecting the blood flow or blood pressure changes that are occurring almost simultaneously in the internal carotid artery and the cerebral pial/capillary arteries, within the low-frequency component in human rs-fMRI.

A partially nested cortical hierarchy of neural states underlies event segmentation in the human brain

“…While estimating the association between neural state boundaries and event boundaries, we also identified the delay between the neural state boundaries and event boundaries within each searchlight. These delay differences across searchlights were similar to BOLD signal delays described previously (Spearman r=0.43; see figure S2B; Amemiya et al, 2020), further supporting our interpretation that the neural state boundaries across all these brain regions are associated with the experience of an event boundary.…”

“…The delay for which the optimal correspondence between neural state boundaries and event boundaries was observed. The scatterplot shows that the observed delay in each searchlight was similar to the BOLD signal time lag that was previously identified based on the time lag of the global signal (Amemiya et al, 2020). Power et al (2011) that showed the highest overlap and the percentage of searchlights in the network that overlapped with that particular Power (2011) Power (2011) manuscript.…”

“…The delay for which the optimal correspondence between neural state boundaries and event boundaries was observed. The scatterplot shows that the observed delay in each searchlight was similar to the BOLD signal time lag that was previously identified based on the time lag of the global signal (Amemiya et al, 2020). Figure S3.…”

Human CSF Movement Influenced by Vascular Low Frequency Oscillations and Respiration

“…The left and right ICAs and IJVs were identified using T1weighted and T2-weighted structural images and these vessel masks (figure 3B), registered on to the functional space, were used to extract the corresponding fMRI time series (figure 3E). These methods have been used and validated in previous studies [22,23,26]. As we demonstrated in our previous work [20], the changes of CBV rather than CBV itself, are the driving force of CSF movement.…”

Imaging of the pial arterial vasculature of the human brain in vivo using high-resolution 7T time-of-flight angiography

“…Although signal changes in fMRI data using a blood-oxygenation-level-dependent (BOLD) contrast arise predominantly on the venous side of the vascular hierarchy (Ogawa et al, 1990), the strongest vascular response to neural activity is located in intracortical arteries and arterioles (Hillman et al, 2007;Vanzetta, 2005), and significant diameter changes in upstream arteries have been observed (Bizeau et al, 2018;Cho et al, 2012Cho et al, , 2008. With the recent interest in cerebral blood volume-based fMRI (Huber et al, 2014) and the numerous accounts of vascular contributions found in BOLD fMRI signals (Amemiya et al, 2020;Bright et al, 2020;Chen et al, 2020;Drew et al, 2020), including a detailed, subject-specific depiction of the underlying angio-architecture would immensely augment forthcoming modelling initiatives (Havlicek and Uludağ, 2020;Tak et al, 2014).…”