Biophysics of Computation II

on YouTube ↗01h 08m

" summary

Glutamate uncaging plus detailed compartmental simulations show dendrites act as location-dependent nonlinear subunits rather than simple adders: two spatially separated NMDA-bearing inputs produce an asymmetric 2D sigmoid; a pared-down five-bin predictor reconstructs dendritic output with R^2 = 0.94 (and branch-sum predictions approach R^2 ≈ 0.9999 at the soma). Inhibition produces opponent, location-specific effects, circuits can turn complex dendritic interactions into multiplicative gain, and the computed 2‑D nonlinearities match a contour‑detection statistic in natural images.

AI-written summary · about

" chapters (10)

1. 0:00Intro
2. 0:11A progression of models
3. 6:42Measuring proximal-distal interactions
4. 9:08Comparing model and data
5. 11:37Same basic story for firing rates
6. 16:54The anti-paralel nature of E-llocation effects
7. 35:02What's the dimensionality of dendritic integration
8. 47:231 input to 2 subunits
9. 49:072 inputs to 2 subunits
10. 52:27Natural contour statistics vs. dendritic integration