Correlated firing of cat retinal ganglion cells. II. Responses of X- and Y-cells to single quantal events.

2. The correlated firing of neighboring cat retinal ganglion cells was examined at low background levels and in darkness. At these backgrounds, cross-correlograms showed a strong, broad peak for a pair of on-center cells, a weak, broad peak for a pair of offcenter cells, and a broad well for two cells of opposite center sign. The much narrower peaks and wells described in the preceding paper (30) were superimposed on these broad peaks and wells. 2. An examination of these correlations and of the firing patterns of individual cells provides evidence for the following conclusions. The broad peaks and wells are caused by a single type of common input event lasting 40-50 ms. This input event is excitatory to on-center cells and often causes 2 or 3 spikes in a single on-center cell, giving these cells a tendency to fire in bursts at low backgrounds. The input event is inhibitory to offcenter cells and causes visible gaps in their maintained discharge at low backgrounds. Two neighboring on-center cells have a strong tendency to fire together because they are both excited by shared input events. Two offcenter cells show a weak tendency to fire together because they are both restricted to firing at the times when they are not inhibited together by shared input events. An onand an off-center cell have a tendency not to fire together because the off-center cell is inhibited when the on-center cell is excited by their shared input events. 3. The rate of the input events received by ganglion cells at the lowest backgrounds was linearly related to the quantity of background light and matched the calculated rate of effective photon absorptions in the rods. The events appear to originate as individual quanta1 events in the rods. 4. The characteristic shapes of the correlograms for the different types of pairs at low light levels imply that the average excitatory response of on Y-cells to these quanta1 events is more transient than that of on X-cells, and that the average inhibition of off Y-cells by quanta1 events is more transient than that of off X-cells. 5. Further analysis of the correlograms indicates that the primary source of the more sustained response of on X-cells to quanta1 events is a relatively sustained excitation that does not pass through the active inputs described in the preceding paper and that is considerably stronger for on X-cells than for on Y-cells. The analysis provides evidence for an analogous sustained inhibition of offcenter cells that is considerably stronger for off X-cells. The analysis suggests that on active inputs have a rather transient response to quanta1 events and provide an initial transient excitation of on-center cells and inhibition of off-center cells. 6. It is proposed that cone bipolar cells are the final pathway for the relatively sustained effects of quanta1 events and that tonic signals from these bipolar cells are a major source of maintained activity for X-cells and a minor source for Y-cells at higher backgrounds. 7. Two hypotheses are discussed as follows: that light-adapted X-cells show a greater