Results are shown in Figure 6. Figure 6A plots the proportion of saccades for each bird to the lagging source, against delay, in the standard paradigm where the lead and lag sounds were of equal lengths (Fig. 1B).
Indeed, circuitry has been proposed that moderates the inhibition of responses to the lagging source to explain this observation [39], [40].
By lengthening this segment, we were able to evoke neural responses in the space map and saccades to the lagging source, even at short delays that evoked localization dominance in the standard experimental paradigm, i.e., when lead and lag sounds were equally long.
Correspondingly, space map neurons respond weakly to lagging sources at delays less than ∼10 ms, but the responses increase for longer delays [12], [17].
One possible scenario is that the ratio with which saccades are lead- or lag-directed (Nlag/Nlead) is proportional to the ratio of the neural responses to the leading and lagging sources respectively (Rlag/Rlead).
As is commonly done in human lateralization studies [e.g.], [ 34], [35] [37], we determined the proportion of saccades toward the leading and lagging sources while manipulating the lag-alone segment independently of the delay (lead-alone segment).
Unlike in this previous study, saccades to lag sources were comparable in error to those of leading and single sources for lag-alone segments of 12 and 24 ms. Errors in trials with shorter lag-alone segments (<3 ms) were larger, but because there were few lag-directed turns (due to localization dominance), this observation must be viewed with caution.
If the lag-alone segment is crucial for the localization of the lag source, did we render the owl "unaware" of the lagging source when we removed the lag-alone segment?
Responses evoked during the lag-alone and superposed segments are plotted against delay in Figure 3 (red, filled, markers) to determine, quantitatively, which segment best accounts for the recovery to the lag source.
Viewed in this way, the weaker representation of the lag source need not be attributed to a lateral inhibition-like process [11], [38], although such a process is not excluded.
As shown in Figure 3B, the lag source does, in fact, still evoke a response during the superposed segment, and this response may provide the basis for an "awareness" of a reflection, without the need for a specialized circuit.
