Infra-Slow Neural Oscillations: A Gateway to Understanding Cognition and Brain Aging

Abstract

Infra-slow oscillations (ISO) constitute a dominant component of spontaneous and task-related brain activity, yet their functional and organizational significance is still not fully understood. Traditionally, infra-slow blood-oxygen-level-dependent (BOLD) fluctuations have often been treated as nonspecific drift, physiological noise, or a background feature of resting-state activity. This thesis examines an alternative view: that ISO reflects the structured spatiotemporal brain dynamics, constrained by intrinsic temporal properties of the brain and relevant to cognition and aging. Across three studies, this thesis investigates ISO from complementary perspectives. Study 1 examines the time-domain organization of ISO by testing whether infra-slow phase dynamics are constrained by intrinsic neural timescales (INT). Using resting-state and naturalistic movie-watching functional magnetic resonance imaging (fMRI) data, this study shows that infra-slow phase progression is not uniform across the oscillatory cycle. Instead, the BOLD instantaneous frequency varies systematically across phase components and is negatively related to regional INT, such that regions with longer timescales show slower phase progression. This relationship persists across rest and task states, suggesting that ISO dynamics are related to stable temporal properties of the brain. Study 2 examines the frequency-domain organization of ISO by focusing on scale-free dynamics during sustained attention. It shows that the power-law exponent (PLE), indexing scale-free infra-slow organization, mediates the relationship between BOLD signal variability and behavioral variability. These findings suggest that infra-slow scale-free dynamics provide a background temporal feature through which neural variability becomes behaviorally relevant. Study 3 extends this framework to brain aging by examining global infra-slow connectivity across the adult lifespan. The results demonstrate that global infra-slow functional connectivity becomes less differentiated/more homogenous with age across both spatial and frequency dimensions. Importantly, this spatiotemporal dedifferentiation remains evident after controlling for physiological and vascular confounds, supporting the interpretation that aging reorganizes, rather than simply degrades, the ISO. Together, these findings support the view that ISO is not only nuisance activity but a dynamical structure of the human brain. By linking infra-slow dynamics to intrinsic neural timescales, scale-free spectral organization, behavioral variability, and lifespan reorganization, this thesis provides a spatiotemporal framework for understanding how slow brain activity contributes to cognition and brain aging.