资 源 简 介
A common question is neuroscience addresses the receptive field of neurons: what makes the neuron spike? Generalized linear models (GLM) are good at describing many kinds of data, including Poisson distributed point processes. The code here uses a time varying Poisson process to relate arbitrary facts ("features") to the action potentials of a neuron ("spikes"). The model is geared towards time varying features like a flickering luminance. The fit co-efficients of the model relate to the temporal profile of a receptive field as measured by the spike triggered average (STA) of white noise.
This code is a tutorial for educational purposes and should not be used "as is" for any application, especially if the authors/coders don"t understand the assumptions. It is written in Matlab and requires a license from Mathworks to run. A motivated individual could port it to Octave.
