Abstract
Purpose :
Previously we demonstrated that the mouse retinal ganglion cell (GC) space-time receptive field can be separated into five subfilters. To elucidate the anatomic underpinnings of each subfilter, we studied their functional alteration with removal of photoreceptor to ON-bipolar cell (BC) signaling using the drug LAP4. This approach allows us to answer fundamental questions such as the nature of the biphasic temporal filtering and antagonistic surround.
Methods :
Flat-mount retinal preparations from three-month-old dark-adapted C57/B6 mice were placed onto a multi-electrode array (MEA) for multicellular recording. Retinas were first perfused with oxygenated solution, followed by 20 μM LAP4 solution. Receptive fields (RFs) were mapped using a binary white noise checkerboard stimulus with 50 μm squares presented at 15 Hz. Spike-triggered averages (STA) were used to identify the average stimulus preceding a spike in each condition. Subsequent model fitting identified tuning properties. Surround strength was obtained by regional averaging of STA data. Paired comparisons between conditions were compared with a student’s t-test unless specified otherwise. ON (N=63) and OFF (N=54) GCs were identified from their linear temporal filter.
Results :
The median of surround strength of OFF GCs shifted from -0.158 to -0.05 with LAP4 (p<.05, KS Test). Similarly, before LAP4 55.6% OFF GCs had an antagonistic surround by model fit, whereas after LAP4 only 16.7% had an antagonistic surround (p<.01, χ2). These results indicate that in OFF GCs the antagonistic surround is driven by ON BC-dependent mechanisms.
The biphasic shape of the temporal STA for OFF GCs is maintained in the presence of LAP4, indicating that the mechanism driving these are largely ON BC-independent.
Lastly, we used spike-triggered covariance (STC) to identify GCs that had linear filters containing both ON and OFF components. LAP4 selective removed the ON component.
Conclusions :
Combining our novel subfilter approach with circuit dissection allowed us to determine that the antagonistic surround of OFF GCs is primarily driven by crossover inhibition from ON BCs. Also, the biphasic nature of OFF RGC temporal filtering is driven by biphasic filtering of the upstream OFF BC, rather than contribution of ON and OFF BCs. Lastly, GCs with both ON and OFF linear responsivity are likely bistratified GCs that receive independent ON and OFF BC input.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.