Information continuously flows between regions of the human brain, forming patterns that shift across states of consciousness, cognitive modes, and neuropsychiatric conditions. While functional magnetic resonance imaging (fMRI) reveals large-scale activity changes over seconds, the electrophysiological dynamics governing sub-second reconfiguration remain poorly understood. Here, relative phase analysis (RPA), a method leveraging phase lead/lag relationships, is introduced to capture whole-brain dynamics with millisecond precision in real time from electroencephalography (EEG). RPA reveals sub-second alternations, occurring approximately every 200 ms, between two dominant modes of information flow: a top-down mode, where anterior regions drive posterior activity, and a bottom-up mode, characterized by reverse directionality. These dynamics are most prominent during wakefulness, gradually diminish under anesthesia, and exhibit pathological imbalance in attention-deficit/hyperactivity disorder (ADHD). Simultaneous EEG-fMRI recordings demonstrate that top-down dynamics coincide with increased activity of higher-order cognitive networks, whereas bottom-up dynamics correspond to heightened activity in sensory networks. A connectome-based coupled-oscillator model reproduces these transitions, indicating that sub-second fluctuations emerge naturally from inter-regional interactions shaped by underlying structural connectivity. This study establishes RPA as a framework for tracking whole-brain dynamics precisely in real time and identifies sub-second top-down/bottom-up alternations as a fundamental organizing principle of human brain function and consciousness.

Keywords: ADHD; Kuramoto model; cortical traveling waves; coupled-oscillator model; general anesthesia; human brain dynamics; relative phase analysis; simultaneous EEG-fMRI; sub-second transitions; top-down versus bottom-up modes.

Copyright © 2026 Elsevier Inc. All rights reserved.