A deformable attractor manifold organizes human resting-state brain dynamics

Area of Science:

• Neuroscience
• Dynamical Systems Theory
• Cognitive Science

Background:

• Intrinsic brain activity is complex and its organizing principles are not fully understood.
• Previous models often describe brain dynamics as movement within a continuous space.
• The constraints and governing rules of spontaneous neural activity remain elusive.

Purpose of the Study:

• To investigate the organizing principles of spontaneous human brain activity during rest.
• To model the geometric and dynamical structure of intrinsic brain dynamics.
• To link these dynamics to cognitive function and neurological conditions.

Main Methods:

• Utilized large-scale functional magnetic resonance imaging (fMRI) datasets.
• Applied a latent dynamical model to analyze brain activity patterns.
• Developed a two-parameter energy landscape to explain observed dynamics.

Main Results:

• Identified spontaneous brain activity as structured by a deformable attractor manifold.
• Discovered two reproducible activity regimes: low-coherence (unimodal) and high-coherence (bimodal/bistable).
• Found that transitions between regimes result from continuous manifold deformation, not discrete state switching.

Conclusions:

• Resting brain activity dynamics are governed by a geometric principle of a deformable attractor manifold.
• This manifold structure explains transitions between brain states, cognitive fluidity, and network coordination.
• Manifold parameters can differentiate between healthy controls and individuals with mild cognitive impairment.