Robust Scaling in Human Brain Dynamics Despite Correlated Inputs and Limited Sampling Distortions.

Area of Science:

• Neuroscience
• Complex Systems Analysis
• Statistical Physics

Background:

• The question of whether brain dynamics operate near criticality is central to understanding information processing and computational flexibility.
• Conventional methods for assessing criticality are prone to artifacts from temporal/spatial correlations and subsampling, potentially misidentifying noncritical systems as critical.
• Existing challenges necessitate more robust analytical and numerical frameworks for empirical data analysis.

Purpose of the Study:

• To introduce a novel analytical and numerical framework for testing criticality in brain dynamics.
• To address limitations of conventional approaches by incorporating temporal/spatial correlations and colored inputs.
• To provide reliable tools for empirical data inference and control in neuroscience.

Main Methods:

• Development of a framework centered on the covariance matrix and its spectrum.
• Integration of a phenomenological renormalization group (PRG) approach.
• Extension to handle colored inputs, temporal and spatial correlations, and robust inference for resting-state fMRI data.

Main Results:

• Collective brain activity analyzed via resting-state fMRI was found to be slightly subcritical, yet close to criticality.
• Extracted critical exponents demonstrated robustness.
• Results align with predictions from recurrent firing-rate models in the long-time correlation limit.

Conclusions:

• The study provides a more reliable methodological toolkit for testing criticality in neuroscience and complex systems.
• Brain dynamics exhibit proximity to criticality, supporting theories of optimal information processing.
• The developed framework offers improved strategies for analyzing empirical neuroimaging data.