Aneurovascular Coupling Framework for Resting-State Functional Anatomy in Humans

Thesis Abstract

This study confronts a critical gap in understanding how intrinsic neurovascular dynamics shape resting-state functional architecture in the healthy human brain, addressing inconsistencies in how neurovascular coupling (NVC) is modeled across imaging modalities and analytic pipelines. The aim is to develop an integrated Aneurovascular Coupling Framework (ACF) that links vascular, neuronal, and metabolic signals to resting-state functional anatomy, enabling more accurate inference of functional connectivity and its regional variability. Specific objectives are (1) to quantify the relationships among resting-state fMRI BOLD signals, concomitant cerebral blood flow as measured by arterial spin labeling (ASL), and spontaneous neural activity indices derived from magnetoencephalography (MEG) in a normative sample; (2) to test the applicability of two established theories—the Neurovascular Uncoupling Theory and the BOLD Slow-Wave Synchronization Model—in predicting regional deviations in NVC across cortical networks; (3) to develop a multilevel statistical framework that integrates physiological, hemodynamic, and neural lag parameters within a unified mechanistic model of resting-state anatomy; (4) to assess test–retest reliability of the ACF across sessions and scanners; and (5) to provide actionable implications for interpreting resting-state biomarkers in health and early pathology. The methodology adopts a convergent mixed-methods design combining quantitative multimodal imaging with hierarchical modeling. The study population comprises 120 neurologically healthy adults aged 22–40 years, balanced for sex, with no history of cardiovascular, metabolic, or neurological disorders. A stratified sampling approach ensures representation across hemispheric dominance profiles. Data collection involves high-field 7 Tesla MRI for BOLD resting-state fMRI (10-minute eyes-open protocol), pseudo-continuous ASL for CBF quantification, diffusion tensor imaging (DTI) for structural connectivity, and MEG resting-state recordings (15 minutes) to capture oscillatory activity across alpha, beta, and gamma bands. Physiological monitoring includes continuous heart rate, respiration, and end-tidal CO2 to control for systemic influences on NVC. Instruments consist of a 3D EPI sequence for fMRI, pseudocontinuous ASL labeling, a high-resolution T1-weighted anatomical scan, a 64-channel MEG system, and standardized physiological sensors. Data preprocessing pipelines apply motion and physiological noise correction (AAS and RETROICOR), cerebral spin labeling corrections for ASL, and co-registration to individualized cortical surfaces. Analytical approaches comprise (i) cross-modal regression analyses to quantify coupling between BOLD fluctuations, CBF, and MEG-derived neural indices; (ii) dynamic functional connectivity and Granger causality analyses to characterize directional interactions among core networks; (iii) Bayesian hierarchical modeling to integrate vascular and neural lag parameters into a mechanistic NVC framework; (iv) test–retest reliability assessments via intraclass correlation coefficients (ICCs) across sessions; and (v) sensitivity analyses to evaluate robustness to scanner differences and physiological noise. Expected findings include (a) robust, region-specific associations between BOLD signal amplitude and CBF with corresponding MEG power fluctuations, supporting a spatially heterogeneous NVC profile across default mode, salience, and frontoparietal networks; (b) identification of cortical regions where neurovascular delays or decoupling predict anomalous resting-state connectivity patterns, aligning with the Neurovascular Uncoupling Theory; (c) validation of the ACF as a parsimonious yet comprehensive model that captures the coupling dynamics among neuronal activity, vascular response, and hemodynamic signals, with predictive capacity for resting-state connectivity strength and network topology; and (d) high test–retest reliability of key NVC metrics, supporting their utility as stable biomarkers. The study contributes to knowledge by offering a formalized, cross-modality framework that reconciles hemodynamic signals with neural activity in resting-state analyses, advancing interpretation of functional connectivity in normative aging and early pathology contexts. The ACF has potential applications in refining biomarker development for neurodegenerative and cerebrovascular conditions and informing the design of multimodal imaging protocols. Limitations include the cross-sectional design and resource-intensive data collection, potentially constraining generalizability to broader age ranges. Future work will extend the framework to clinical cohorts with vascular risk factors and explore longitudinal trajectories of neurovascular coupling across aging.