A parameter-free statistical test for neuronal responsiveness

Abstract

Neurophysiological studies depend on a reliable quantification of whether and when a neuron responds to stimulation. Simple methods to determine responsiveness require arbitrary parameter choices, such as binning size, while more advanced model-based methods require fitting and hyperparameter tuning. These parameter choices can change the results, which invites bad statistical practice and reduces the replicability. New recording techniques that yield increasingly large numbers of cells would benefit from a test for cell-inclusion that requires no manual curation. Here, we present the parameter-free ZETA-test, which outperforms t-tests, ANOVAs, and renewal-process-based methods by including more cells at a similar false-positive rate. We show that our procedure works across brain regions and recording techniques, including calcium imaging and Neuropixels data. Furthermore, in illustration of the method, we show in mouse visual cortex that 1) visuomotor-mismatch and spatial location are encoded by different neuronal subpopulations; and 2) optogenetic stimulation of VIP cells leads to early inhibition and subsequent disinhibition.

Triple dissociation of visual, auditory and motor processing in primary visual cortex

“…The onset latency of spiking activity for individual neurons was estimated using ZETA, a recently developed bin-less statistical test for determining whether a neuron shows a time-locked modulation of spiking activity (Montijn et al, 2021). We opted for this as visual and auditory stimuli can elicit very different neural dynamics in visual and auditory cortex (specifically, spiking responses in A1 can be very brief (DeWeese et al, 2003)) and ZETA prevents confounds related to different temporal dynamics by avoiding the need to bin spikes.…”

“…Top histogram shows the distribution of onset latencies of all significantly auditory responsive neurons (red) and visually responsive neurons (blue). Significance and onset latency were assessed using a binning-free algorithm, ZETA (Montijn et al, 2021). Spiking onset was significantly earlier for auditory versus visual stimuli (55.3 ms (31.4 - 108.5 ms) versus 80.3 ms (61.5 - 98.5 ms); median and interquartile range; F(1,411)=5.37, p=0.0209), similar to our earlier population-averaged approach (Fig.…”

Primary Somatosensory Cortex Bidirectionally Modulates Sensory Gain and Nociceptive Behavior in a Layer-Specific Manner

“…Units were considered modulated in a given stimulus condition if a significant difference was measured between baseline and stimulus-evoked spiking either in terms of absolute spike counts (p<0.05, signed-rank test for paired baseline and stimulus-evoked trials) or spike timing detected by using the ZETA test 72 .…”

A universal pipeline for the alignment of electrode tracks with slice histology and electrophysiological data

“…The Probe Finder calculates three complementary metrics that provide different types of insight into where area boundaries might occur along the recording probe (fig. 6): 1) A normalized spiking rate for each recording location gives a strong indication of the presence of white matter bundles; 2) Multi-unit activity (MUA; here aggregate spike counts over nearby clusters/channels) correlation at a 10 ms time window is strong for noisy contact points not inside the brain; 3) Responsiveness ZETA z-scores shows a clear delineation between areas that respond to the experiment versus those that do not (Montijn et al 2021). An example experiment with visual stimuli creates a sharp boundary in responsiveness values between visually-responsive and non-visually responsive regions (fig.…”

“…Here we present the Universal Probe Finder, a histology processing and probe alignment pipeline that uses the recently developed zeta-test to calculate neuronal responsiveness based only on single-unit or multi-unit spike-times and event onsets (Montijn et al 2021). The Universal Probe Finder supports atlases from multiple species, such as mouse, rat, and macaque, and can use various electrophysiological data formats, including kilosort output and raw SpikeGLX files (Pachitariu et al 2016; billkarsh [2016] 2022; Allen Institute for Brain Science 2017).…”