Theoretical Neuroscience Podcast On balanced neural networks - with Nicolas Brunel - #34
Nov 8, 2025
Nicolas Brunel, a computational neuroscientist renowned for his foundational work on balanced cortical networks, dives into the mechanics of how neurons receive balanced inputs. He discusses the importance of understanding spontaneous cortical activity and explores the balance of excitatory and inhibitory signals that leads to irregular firing. Brunel also connects this theory to practical implications in memory storage and coding efficiency, highlighting interesting parallels between neural dynamics and methods from physics while addressing potential breakdowns in balance linked to conditions like epilepsy.
AI Snips
Chapters
Books
Transcript
Episode notes
Fluctuation-Driven Cortical Activity
- Balanced networks receive large excitatory and inhibitory inputs that approximately cancel, so spiking is driven by fluctuations.
- This fluctuation-driven regime explains irregular, low-rate cortical firing observed in vivo.
Rome Postdoc Shaped Research Direction
- Nicolas Brunel recalled moving to Rome and starting work with Daniel Amit, which shaped his shift toward biologically grounded models.
- That period led him to read experimental literature and focus on spontaneous cortical activity.
Diffusion Approximation Enables Analytic Rates
- With many weak inputs, synaptic input can be approximated by Gaussian diffusion (diffusion approximation).
- That yields an Ornstein–Uhlenbeck description and a solvable first-passage firing-rate formula.
