A simple white noise analysis of neuronal light responses

Abstract

A white noise technique is presented for estimating the response properties of spiking visual system neurons. The technique is simple, robust, efficient and well suited to simultaneous recordings from multiple neurons. It provides a complete and easily interpretable model of light responses even for neurons that display a common form of response nonlinearity that precludes classical linear systems analysis. A theoretical justification of the technique is presented that relies only on elementary linear algebra and statistics. Implementation is described with examples. The technique and the underlying model of neural responses are validated using recordings from retinal ganglion cells, and in principle are applicable to other neurons. Advantages and disadvantages of the technique relative to classical approaches are discussed.

Theoretical analysis of reverse-time correlation for idealized orientation tuning dynamics

“…RTC is only used to determine the characteristics of the linear filter (Chichilnisky 2001;Nykamp and Ringach 2002;Nykamp 2003). The characteristics of the (assumed form of the) nonlinearity are computed from these and the experimental data.…”

Spike history neural response model

“…An example of a model that can be used for the controller design is a linear-nonlinear model (Chichilnisky 2001). This model has been successfully applied in many applications, including to describe the response of retinal ganglion cells (RGCs) to light stimulation.…”

“…A key advantage of this form of model is that there are well-established experimental protocols for accurately estimating the linear filters and the nonlinear function using reverse correlation and maximum likelihood methods (Chichilnisky 2001;Lesica et al 2007;Nirenberg et al 2010). Briefly, the simplest of these methods involves recording the RGC's response to a long sequence of pixelised images, in which the brightness of each pixel is chosen randomly at every frame from a Gaussian distribution.…”

“…Alternative methods that maximise the likelihood of predicting the recorded spike train are based on optimisation techniques and allow the linear-nonlinear model to be fitted to data based on stimuli other than Gaussian white noise, such as natural scenes (Nirenberg et al 2010). Together, these models and methods have demonstrated a remarkable ability to predict the response of RGCs to a large variety of stimuli (Chichilnisky 2001;Lesica et al 2007;Nirenberg et al 2010). These linear-nonlinear models are often referred to as "generalized linear model" (Pillow et al 2005).…”

“…It has been shown that RGCs typically respond with brief clusters of spikes separated by a period of silence in response to naturalistic movies, random visual flicker, and white noise visual stimulation (Chichilnisky 2001;Keat et al 2001; van Hateren et al 2002). The phenomena of spike clusters separated by periods of suppression is characteristic of a memory in the system; i.e., the response of the neuron depends on its previous response.…”