Statistical models for neural data: from GLMs to latent variables (Part 1)
Part 1 of 2 of a tutorial on statistical models for neural data,. Speaker: Jonathan Pillow.
- Introduction to the neural coding problem: how are stimuli and actions encoded in neural activity?
- What aspects of neural activity carry information?
- Regression models, latent variable models: aim to capture hidden structure underlying neural activity
- Model fitability/tractability vs richness and flexibility, finding a "sweet spot" in between
Spike count models and maximum likelihood
simple example #1 (21:46): linear Poisson neuron, maximum likelihood estimation, log-likelihood
simple example #2 (42:22): linear Gaussian neuron
example #3: (50:17): unknown neuron, basic Poisson GLM
Spike train models (GLMs with spike history)
What is a GLM? (57:20)
Design matrix (1.00.07)
GLM summary (1.20.18)
Multiple spike train models (GLMs with coupling)
GLMs with spike history and coupling (1.23.19)
GLM dynamic behaviour examples (1.34.43)
Latent variable models