How to Probe Dynamics of Brain Function: A Narrative Review
Helen Pushkarskaya1, Smita Krishnaswamy2, Christopher Pittenger3
1Department of Psychiatry, Yale School of Medicine, New Haven, Connecticut; Yale Biomedical Imaging Institute, Yale School of Medicine, New Haven, Connecticut.
Biological Psychiatry
|December 10, 2025
Summary
Dynamic modeling of brain activity offers deeper insights than static analyses, crucial for understanding complex brain function and psychiatric illness. These advanced methods quantify temporal complexity and causal influences for theory-informed research.
Area of Science:
- Neuroscience
- Computational Neuroscience
- Systems Neuroscience
Background:
- The brain is a complex dynamic system, but static analytical approaches limit understanding of its temporal processes.
- Static methods like time-averaged connectivity fail to capture dynamic processes essential for brain function and dysfunction.
- Psychiatric illnesses involve state fluctuations, necessitating dynamic analysis for deeper clinical insight.
Purpose of the Study:
- To review and organize current dynamic analytic approaches for studying brain function over time.
- To highlight methods for describing temporal brain activity, inferring causality, decoding latent dynamics, and simulating neural processes.
- To provide a framework for testing theories of brain function in clinical populations using dynamic modeling.
Main Methods:
- Review of dynamic modeling approaches including sliding-window correlations, temporal ICA, dynamic causal modeling, recurrent neural networks, and neural differential equations.
- Categorization of methods based on four research goals: describing temporal patterns, inferring causal mechanisms, decoding latent dynamics, and simulating neural processes.
- Discussion of method assumptions, clinical applications, limitations, and provision of links to open-access tools.
Main Results:
- Dynamic modeling quantifies temporal complexity, identifies causal influences, compresses activity into latent trajectories, and simulates data.
- Methods are organized by research goals, detailing representative techniques, their assumptions, clinical uses, and limitations.
- Open-access tools are linked for each method, facilitating application in clinical research.
Conclusions:
- Dynamic modeling represents a conceptual shift from static, data-driven descriptions to theory-informed tests of brain processes.
- These approaches enable direct testing of brain function theories in clinical populations.
- Aligning analytical tools with systems-level theories advances the study of brain function and dysfunction over time.
