Spike Train decoding Summary Decoding of stimulus from response –Two choice case Discrimination...

Spike Train decoding

Spike Train decoding

Spike Train decoding

Summary

• Decoding of stimulus from response– Two choice case

• Discrimination• ROC curves

– Population decoding• MAP and ML estimators• Bias and variance• Fisher information, Cramer-Rao bound

– Spike train decoding

Chapter 4

Entropy

Entropy

Mutual information

H_noise< H

Mutual information

KL divergence

Continuous variables

Entropy maximization

Entropy maximization

Population of neurons

Retinal Ganglion Cell Receptive Fields

Retinal Ganglion Cell Receptive Fields

Retinal Ganglion Cell Receptive Fields

Retinal Ganglion Cell Receptive Fields

Retinal Ganglion Cell Receptive Fields

Temporal processing in LGN

Temporal processing in LGN

Temporal processing in LGN

Temporal vs spatial coding

Entropy of spike trains

Entropy of spike trains

Entropy of spike trains

Entropy of spike trains

• Spike train mutual information measurements quantify stimulus specific aspects of neural encoding.

• Mutual information of bullfrog peripheral auditory neurons was estimated– 1.4 bits/sec for broadband noise stimulus– 7.8 bits/sec for bullfrog call-like stimulus

Summary

• Information theory quantifies how much a response says about a stimulus– Stimulus, response entropy– Noise entropy– Mutual information, KL divergence

• Maximizing information transfer yields biological receptive fields– Factorial codes– Equalization– Whitening

• Spike train mutual information