Spontaneous activity changes in large-scale cortical networks in older adults couple to distinct hemodynamic morphology

Abstract

Neurovascular coupling is a dynamic core mechanism supporting brain energy demand. Therefore, even spontaneous changes in neural activity not linked directly to goal-directed behavior are expected to evoke a vascular hemodynamic response (HDR). Here, we developed a novel procedure for estimating transient neural activity states based on source-localized electroencephalogram (EEG) in combination with HDR estimation based on simultaneously acquired functional magnetic resonance imaging (fMRI). We demonstrate a readily apparent spatial correspondence between electrophysiological neural states and time-locked local HDR during rest, describing for the first time how features of neurovascular coupling may differ among unique large-scale brain networks. In the default mode network, the HDR pattern in our sample of older adults was associated with a structural surrogate marker of general cerebrovascular deterioration and predicted temporal disruption in electrophysiological activity linked to memory decline. These results demonstrate the potential of our integrated EEG/fMRI analysis for making inferences about neural and vascular processes in higher-level cognitive networks in healthy and at-risk populations.Intrinsic neural activity not explicitly associated with performing a task is an indicator of brain health. However, our understanding of how natural activity patterns arise is far from complete. The consequences of dysregulated intrinsic neuronal firing on life-sustaining cellular processes, such as gene expression and protein synthesis, are highlighted by animal models 1 . In the human brain, large-scale recordings at the cell population level link abnormal intrinsic neural activity to the progression of brain disorders, such as Alzheimer's dementia [2][3][4] . Much of what we know about intrinsic neural activity in humans comes from functional magnetic resonance imaging (fMRI), which relies on neurovascular coupling to make inferences about neural activity 5 . Inherently, variations both in neural activity and responsiveness of the associated vascular support, which is the signal detected by this imaging technique, can influence patterns in fMRI data, especially in the aging brain prone to cerebrovascular changes and patients with brain disorders accompanied by vascular comorbidities 6,7 . Therefore, new procedures for disambiguating between the neural and hemodynamic features in the human brain are needed.Studies using fMRI have demonstrated that large-scale brain networks recruited by tasks are also spontaneously reactivated during resting state [8][9][10][11] . Two recent analytic approaches, focusing on electrophysiological correlates of neural activity that are temporally resolved on rapid scales of cognition 12 , suggested that such spontaneous reactivations may constitute transient states of network activity coupled to local vascular hemodynamic response (HDR). First, a novel mathematical tool based on the Hidden Markov Model (HMM) discerned recurring transient states of spatially coordinated amplitude chan...

Replay bursts in humans coincide with activation of the default mode and parietal alpha networks

“…Each RSN state can be described by its distinct spectral power and phase-locking profile; these profiles are summarized by averaging over frequency bands in Figure 2B for RSN states 1-4. This highlights that RSN state 1 was associated with activity over the parietal cortex; the equivalent network in simultaneous electroencephalogram (EEG)-fMRI studies has been shown to be anticorrelated with the dorsal attention network (DAN) (Mantini et al, 2007;Sitnikova et al, 2020); thus, activation of RSN state 1 corresponds to the DAN switching off. RSN state 2 combines high-power signals in frontal and temporal regions with coherent oscillations in the lateral parietal cortex, regions that comprise the DMN.…”

“…Activity in the gamma band (>30 Hz) has consistently shown a strong correlation with a subsequent BOLD signal (Niessing et al, 2005;Nir et al, 2007); however, a more complex relationship emerges between the BOLD signal and activity in lower-frequency bands, in which electrophysiological RSNs are largely defined (Brookes et al, 2011;Mantini et al, 2007). It has now been shown that different RSN states show markedly different hemodynamic profiles; in particular, the DMN and DAN evoke BOLD signals that are opposed in polarity and distinct in their temporal decay profile (Sitnikova et al, 2020). If the DMN state visits cluster together in time while being linked to high-frequency power increases, as our results indicate, then this may explain these distinct profiles and provide a key bridge between the understanding of RSNs recorded across distinct modalities.…”

Microstates and power envelope hidden Markov modeling probe bursting brain activity at different timescales

“…To the best of our knowledge, a microstate analysis of MEG data has not yet been developed. The HMM approach has been applied to EEG power envelopes (Hunyadi et al, 2019; Sitnikova et al, 2020), but the focus was on the relationship with fMRI networks rather than MEG or EEG microstates. Here, we assessed the impact of both the state clustering model and the recording modality on temporal and spatial signatures of transient brain states.…”

Replay bursts coincide with activation of the default mode and parietal alpha network

“…2B for RSN-states 1-4. This highlights that RSN-state 1 was associated with activity over parietal cortex; the equivalent network in simultaneous EEG-fMRI studies has been shown to be anticorrelated with the Dorsal Attention Network (DAN) (20,21); thus the activation of RSN-state 1 corresponds to the DAN switching off. RSN-state 2 combines high power signals in frontal and temporal regions with coherent oscillations in lateral parietal cortex, regions that comprise the DMN.…”

“…Activity in the gamma band (>30Hz) has consistently shown a strong correlation with a subsequent BOLD signal (39,40), however a more complex relationship emerges between the BOLD signal and activity in lower frequency bands, in which electrophysiological RSNs are largely defined (10,20). It has now been shown that different RSN-states show markedly different hemodynamic profiles; in particular, the DMN and DAN evoke BOLD signals that are both opposed in polarity and distinct in their temporal decay profile (21). But if the DMN state visits cluster together in time whilst being linked to high frequency power increases, as our results indicate, then this may help to explain these distinct profiles and provide a key bridge between the understanding of RSNs recorded across distinct modalities.…”