A statistical significance test for spatio-temporal receptive field estimates obtained using spike-triggered averaging of binary pseudo-random sequences

TL;DR

This paper introduces a statistical significance test for spatio-temporal receptive field estimates obtained via spike-triggered averaging, enabling identification of significant features and inference of receptive field size in neural data.

Contribution

The authors derive an analytical significance testing method for STRF pixels under the null hypothesis of independence, enhancing the interpretability of spike-triggered averaging results.

Findings

The method accurately identifies significant STRF pixels in mouse retinal ganglion cells.

Significant pixels can be used to infer the size of the receptive field.

The approach provides a statistical framework for validating STRF estimates.

Abstract

Spatio-temporal receptive fields (STRF) of visual neurons are often estimated using spike-triggered averaging of binary pseudo-random stimulus sequences. The spike train of a visual neuron is recorded simultaneously with the stimulus presentation. The neuron's STRF is estimated by averaging the stimulus frames that coincide with spikes at fixed latencies. Although this is a widely used technique, an analytical method for determining the statistical significance of the estimated value of the STRF pixels seems to be lacking. Such a significance test would be useful for identifying the significant features of the STRF and investigating their relationship with experimental variables. Here, the distribution of the estimated STRF pixel values is derived for given spike trains, under the null hypothesis that spike occurrences and stimulus values are statistically independent. This distribution is then used for computing amplitude thresholds to determine the STRF pixels where the null hypothesis can be rejected at a desired two-tailed significance level. It is also proposed that the size of the receptive field may be inferred from the significant pixels. The application of the proposed method is illustrated on spike trains collected from individual mouse retinal ganglion cells.