New transport equation method models neural firing rate fluctuations more accurately
Firing rate fluctuations emerge from initial voltage distributions, not just steady states
A new paper by Wilten Nicola and Sue Ann Campbell presents a novel analytical approach for understanding firing rate fluctuations in spiking neural networks. Traditional mean field models often rely on asynchronous or constant-flux steady-state solutions to the Fokker-Planck system. This work instead uses the transport solution to the advection equation, assuming slow time-varying inputs and an excitation-driven regime. The result is an approximation for the instantaneous population rate (flux) as a function of the initial voltage distribution.
The transport mean field system reveals how firing rate fluctuations arise from the dynamic interplay of time-varying inputs, initial densities, and coupling between neurons. This offers a more accurate way to model neural activity across multiple time scales—from single neurons to populations. The approach could enhance neuromorphic computing, brain-computer interfaces, and our fundamental understanding of neural variability in response to stimuli.
- Derives an analytical approximation for instantaneous population firing rate using the transport solution to the Fokker-Planck equation
- Accounts for initial voltage distribution and slow time-varying inputs, unlike steady-state mean field approaches
- Predicts emergence of firing rate fluctuations from dynamic interaction of inputs, initial densities, and coupling
Why It Matters
Could improve brain-inspired computing and neural variability models by capturing dynamic, non-steady-state firing patterns.