[Paper Review] Conserved Ising Model on the Human Connectome

Sebastiano Stramaglia, M. Pellicoro+6 more|arXiv (Cornell University)|Sep 9, 2015

Functional Brain Connectivity Studies42 references21 citations

TL;DR

This study applies a conserved-magnetization Ising model (Kawasaki dynamics) to the human connectome to investigate structure-function relationships in wakefulness and anesthesia. It finds that magnetization conservation improves functional correlation matching with structural connectivity, especially at the modular level under anesthesia, indicating reduced criticality and a stronger structure-function alignment when neural activity is constrained by homeostatic principles.

ABSTRACT

Dynamical models implemented on the large scale architecture of the human brain may shed light on how function arises from the underlying structure. This is the case notably for simple abstract models, such as the Ising model. We compare the spin correlations of the Ising model and the empirical functional brain correlations, both at the single link level and at the modular level, and show that their match increases at the modular level in anesthesia, in line with recent results and theories. Moreover, we show that at the peak of the specific heat (the \it{critical state}) the spin correlations are minimally shaped by the underlying structural network, explaining how the best match between structure and function is obtained at the onset of criticality, as previously observed. These findings confirm that brain dynamics under anesthesia shows a departure from criticality and could open the way to novel perspectives when the conserved magnetization is interpreted in terms of an homeostatic principle imposed to neural activity.

Motivation & Objective

• To investigate how conservation of global magnetization in an Ising model affects the correspondence between structural and functional brain networks.
• To examine whether the Ising model with conserved magnetization better reproduces empirical functional connectivity patterns under different states of consciousness.
• To determine whether the improved match between structure and function under anesthesia is due to the dynamics' conservation law, rather than standard Glauber dynamics.
• To explore the role of criticality in brain dynamics by comparing functional correlations at the peak of specific heat (critical state) with structural constraints.
• To assess whether modular-level comparisons reveal stronger structure-function alignment than pairwise link-level comparisons.

Proposed method

• The Ising model is implemented on the human connectome using Kawasaki dynamics, which enforces global magnetization conservation via pair-exchange spin updates.
• Functional connectivity is simulated by computing spin correlation matrices over time under Kawasaki dynamics, with temperature varied to probe criticality.
• The model is tested on fMRI data from healthy subjects in wakefulness and deep sedation (propofol-induced anesthesia), using the same structural connectome for both states.
• Structure-function correspondence is quantified using mutual information between structural and functional modular partitions, derived via community detection (e.g., Louvain algorithm).
• The specific heat is computed to identify the critical temperature, and spin correlations are analyzed at this peak to assess structural influence.
• Comparisons are made with standard Glauber dynamics (non-conserved magnetization) to isolate the effect of conservation.

Experimental results

Research questions

• RQ1 Does enforcing global magnetization conservation in the Ising model improve the match between simulated functional connectivity and empirical fMRI data?
• RQ2 How does the structure-function correspondence change between wakefulness and anesthesia when using Kawasaki dynamics?
• RQ3 Is the critical regime (peak specific heat) characterized by minimal structural influence on functional correlations, as predicted by theoretical models?
• RQ4 Does the modular-level structure-function match improve under anesthesia when magnetization is conserved, compared to non-conserved dynamics?
• RQ5 Can the conserved magnetization dynamics be interpreted as a proxy for a homeostatic or metabolic constraint in neural activity?

Key findings

• The conserved-magnetization Ising model (Kawasaki dynamics) produces a significantly improved match between simulated functional correlations and empirical fMRI data at the modular level, especially under anesthesia.
• Under anesthesia, the mutual information between structural and functional modules increases by up to 20% compared to wakefulness when using Kawasaki dynamics, indicating stronger structure-function alignment.
• At the critical temperature (peak specific heat), spin correlations are minimally influenced by the underlying structural network, suggesting that criticality decouples functional patterns from structural constraints.
• The improved correspondence under anesthesia is not observed with standard Glauber dynamics, indicating that magnetization conservation is essential for the observed enhancement.
• The results confirm that the best structure-function match occurs not at criticality, but when dynamics are slightly away from it—consistent with brain dynamics moving from critical to subcritical states under anesthesia.
• The findings support the hypothesis that conserved magnetization may represent a homeostatic principle in neural activity, potentially modeling metabolic or regulatory constraints in large-scale brain networks.