Abstract
Purpose :
Large-scale micro-electrode array (MEA) recordings can provide a comprehensive characterization and classification of retinal ganglion cells (RGCs) but do not supply anatomical or molecular identities of the recorded cells. We developed a robust technique for the anatomical and molecular identification of RGCs detected on an extracellular MEA.
Methods :
Genetically modified mice that expressed channelrhodopsin (ChR2) with a fused fluorescent protein tag in a subset of RGCs (Grik4- and CRH- crossed with ChR2-tdTom reporter) were used.
The retinas were removed and placed ganglion-side down on a 512 MEA.
First, we measured RGCs responses to visual stimuli (spatio-temporal white noise, moving bars) using low photopic light levels that did not activate ChR2. After pharmacologically blocking normal transmission of signals within the retina, we stimulated RGCs by projecting a 16x16 high-intensity blue (450 nm) microLED array directly into the RGCs. LEDs were optically reduced to 20x20 μm2 in size with 5-10 mW/mm2 intensity. Frames of the sparse-white-noise LED stimulation consisted of a single LED turned on at a random position. We calculated the Spike-Triggered Average (STA) response to the stimulus for each detected RGC. Fluorescent cell bodies were localized during the recordings by imaging the retina on the MEA mounted on an inverted microscope for the entire experiment, after which we prepared the retina for follow-up high-resolution imaging.
Results :
STA responses to microLED stimulation showed time dynamics consistent with direct ChR2 activation. Individual fluorescent RGC cell bodies overlapped the highly-localized optogenetically-induced spatial receptive fields. Around 70% of the ChR2-expressing cells were matched to the cells identified on the MEA. The unique action-potential waveform of optogenetically-identified cells allowed matching with their visual response properties.
Conclusions :
We successfully matched mouse RGCs detected on a high-density microelectrode array to genetically labeled RGCs. The described technique will allow for comprehensive functional characterization of already described and newly discovered anatomical and molecular RGC types.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.