A profile of auditory‐responsive neurons in the larval zebrafish brain

Abstract

Many features of auditory processing are conserved among vertebrates, but the degree to which these pathways are established at early stages is not well explored. In this study, we have observed single cell activity throughout the brains of larval zebrafish with the goal of identifying the cellular responses, brain regions, and brain-wide pathways through which these larvae perceive and process auditory stimuli. Using GCaMP and selective plane illumination microscopy, we find strong responses to auditory tones ranging from 100 Hz to 400 Hz. We also identify different categories of auditory neuron with distinct frequency response profiles. Auditory responses occur in the medial octavolateral nucleus, the torus semicircularis, the medial hindbrain, and the thalamus, and the flow of information among these regions resembles the pathways described in adult fish and mammals. The details of these patterns, however, indicate that auditory processing is still rudimentary in larvae. The range of frequencies detected is small, and while different neurons have distinct response profiles, most are sensitive to multiple frequencies, and distinct categories show substantial overlap in their responses. Likewise, while there are signs of nascent spatial representations of frequency in the larval brain, this only faintly resembles the clear tonotopy seen in adult fish and mammals. Overall, our results show that many fundamental properties of the auditory system are established early in development, and suggest that zebrafish will provide a good model in which to study the development and refinement of these pathways.

Diffuse light‐sheet microscopy for stripe‐free calcium imaging of neural populations

“…Automatically identifying regions of interest roughly corresponding to individual neurons, a process referred to as segmentation, is an important step in describing activity across populations of neurons . To gauge the impacts of stripes on segmentation, we used the MorphoLibJ segmentation tool in FIJI , a widely used approach for segmenting based on the neurons' morphology . Applying this segmentation to images from SPIM and dSPIM configurations, we found that the neurons within dark stripes of SPIM images could not be segmented while the dSPIM images could be segmented across the entire brain with matching segmentation parameters (Figure C, D).…”

“…To test this, we imaged neuronal activity in response to auditory stimulation using both SPIM and dSPIM. Auditory stimulation was chosen because it produces highly repeatable responses from trial to trial, thus providing a stable response with which to compare the performance of SPIM and dSPIM . Fluorescent images were recorded at 10 Hz over 10 seconds experiments in which a pure 300 Hz audio tone (within the range of peak sensitivity ) played for 1 second (Figure ; Video S3).…”

“…Auditory stimulation was chosen because it produces highly repeatable responses from trial to trial, thus providing a stable response with which to compare the performance of SPIM and dSPIM . Fluorescent images were recorded at 10 Hz over 10 seconds experiments in which a pure 300 Hz audio tone (within the range of peak sensitivity ) played for 1 second (Figure ; Video S3). As described previously , clear audio responses were seen in the medial octavolateral nucleus and the medial hindbrain (Figure A‐D).…”

“…Fluorescent images were recorded at 10 Hz over 10 seconds experiments in which a pure 300 Hz audio tone (within the range of peak sensitivity ) played for 1 second (Figure ; Video S3). As described previously , clear audio responses were seen in the medial octavolateral nucleus and the medial hindbrain (Figure A‐D). Moving stripes appear as bands across the local correlation image from SPIM (Figure C), which are absent in the dSPIM local correlation image (Figure D).…”

“…Automatically identifying regions of interest roughly corresponding to individual neurons, a process referred to as segmentation, is an important step in describing activity across populations of neurons [5][6][7][8][9]. To gauge the impacts of stripes on segmentation, we used the MorphoLibJ segmentation tool in FIJI [24], a widely used approach for segmenting based on the neurons' morphology [9,25,26].…”