MIT neuroscientists map four brain-like oscillation regimes in recurrent networks
New theory reveals how neural circuits switch between regular oscillations and chaotic Up-Down states
Deep Dive
Bowen W. Zheng, Earl K. Miller, and Ila R. Fiete develop a dynamical mean-field theory for low-rank recurrent networks with adaptation-driven oscillations. When random connectivity triggers chaos, increasing adaptation sequentially produces four regimes: static coherent state, noise-sustained regular-to-irregular oscillations, stochastic switching between symmetric wells, and a global limit cycle. These dynamics replicate brain activity during wakefulness, sleep, and anesthesia.
Key Points
- Four regimes identified: static coherent, noise-sustained oscillations, stochastic switching, and global limit cycle, driven by adaptation strength
- Two instability mechanisms: chaos onset from random connectivity and a Hopf bifurcation of the coherent mode
- Biological relevance: waxing-waning rhythms, state switching, and Up-Down alternations match observations under wakefulness, sleep, and anesthesia
Why It Matters
Bridges theory and neuroscience—could inspire more brain-like AI and help understand consciousness, anesthesia, and sleep disorders