6°, 95% confidence interval

[−13.3°, 20.5°], circular one-sample t test). Most cells also received DS inhibitory inputs, whose PD was significantly different from the PD of spike output: (mean[PDSpike – PDInh] = 175.9°, 95% confidence interval [98.1°, 253.6°], circular t test; Figure 4E), suggesting that inhibitory inputs were tuned to nonpreferred directions. In identified type 1 and type 2 cells, the absolute angular Apoptosis inhibitor separation between PDSpike and PDExc (|PDSpike – PDExc|) was 18.1° ± 4.7° (n = 12). If inhibitory input tuning was the dominant factor in controlling spike output, we would expect PDInh to be antiparallel to PDSpike. However, PDInh was often not strictly opposite to PDSpike (Figure 4E). The absolute angular separation of PDInh from the null direction of spike output (|PDInh − [PDSpike − 180°]|) was 61° ± 15°, which was larger than the angular separation between PDSpike and PDExc (p = 0.002, n = 12, see more Wilcoxon signed-rank test). Together, this suggests

that the tuning of excitatory inputs largely determines PDSpike in these neurons. Furthermore, comparing PDSpike, PDExc, and PDInh between type 1 and type 2 cells corroborated our earlier observation that their directional tuning is different (p < 0.001 for PDSpike and PDExc, p = 0.028 for PDInh; Watson-Williams test for equal means). The excitatory charge transfer during bar stimulation was 7.3 ± 1.2 pC and 3.6 ± 0.9 pC in type 1 and type 2 cells, respectively, when averaged across all directions. The inhibitory charge transfer was 2.8 ± 0.6 pC and 4.1 ± 1.0 pC in type 1 and type 2 cells, respectively. In addition, we observed that the mean DSI for spiking was similar to that of excitatory inputs in type 1 cells (p = 0.063) and type 2 cells (p = 0.93, Wilcoxon signed-rank tests), while it was somewhat larger in deep cells (p = 0.04) (Figure 4F). The mean DSI of inhibitory currents was not different from that of spike output tuning curves for the three cell types

(p > 0.29 for all cell types, Wilcoxon signed-rank test). In summary, this suggests that directionally tuned excitatory synaptic currents determine the PD of these morphologically identified DS cells, and differently Astemizole tuned synaptic inhibition contributes to sharpening the directional response. Whole-cell recordings showed that DS type 1 and type 2 cells in our transgenic lines received strongly tuned excitatory inputs in response to moving bars. We next searched for the source of this DS excitatory drive by imaging Ca2+ transients in postsynaptic and presynaptic compartments of the tectal neuropil. Specifically, we asked whether RGC axonal compartments exhibit DS signals that functionally colocalize with postsynaptic dendrites of type 1 and type 2 cells, which would provide strong evidence for retinal DS axons being the source of DS excitatory drive in these cells.