Figure 5

Selective change in resolution in human listeners is captured by the hemispheric balance model only. (a) Distribution of reported location values for the 15 test ITDs during the control condition for a representative subject. The resolvability of nearby test ITDs can be approximated by the difference in the corresponding location values (indicated by horizontal bars). (b) Location values for the test condition for the same subject show pronounced shifts for test ITDs in the hemisphere ipsilateral to the adapter ITD, yet only marginal shifts around midline. These non-uniform shifts result in a compression, i.e. decreased differences between location values around midline (compare horizontal bar lengths near 0µs ITD in (a) and (b)). Simultaneously, the perception of auditory space close to the adapter ITD is dilated, indicated by increased differences between location values for nearby test ITDs (compare horizontal bar lengths near 634.9µs ITD in (a) and (b)). (c) Schematic depicting a trial sequence to investigate the listeners’ ITD resolution without and with adapter (ITD of +634.9 µs). One of the two consecutive target intervals was always presented with a fixed ITD (grey); the other interval was presented with a random ITD (white). The interval containing the fixed ITD was chosen randomly. (d) Normalized resolution (re. control) across 8 subjects (colored bars, data are represented as median; grey lines display individual subjects) is plotted as a function of ITD positions. For each subject, ITD resolution at each position was normalized to the resolution at −635 µs for the control and test condition independently. The change relative to the control condition was calculated subsequently. Positive values indicate improved resolution, whereas negative values indicate deteriorated resolution due to the preceding adapter. Overall, the listeners’ resolution increased significantly for locations closer to the adapter (p = 0.00019, Friedman test). Asterisks represents significantly improved or deteriorated normalized resolution for ITD positions close to the adapter position and at 0 µs, respectively (p < 0.05, two-sided Wilcoxon signed rank test). (e) The combined compression (deterioration) and dilation (improvement) in spatial resolution at 0 µs and close to the adapter location, respectively, was only captured correctly by the hemispheric balance model. Both other models did not predict the deterioration around midline, due to the missing hemispheric specificity of their fits to the different scores for the lateral adapter (see Fig. 4).