(A) Grouped data from stage 48–9 tadpoles in control condition and in the presence of GABAA receptor blocker PTX. In a given cell, data are averaged over trials at each ISI, to determine the MSI ratio. Plotted here are the population means of these trial-averaged MSIn ratios, at each ISI tested. Error bars show +/– S.E.M. (B) Histogram bars show the ISI at which each cell exhibited maximal MSIn values in control and PTX groups. Notice that PTX shifts the preferred ISI values to shorter intervals. (C) Maximum MSIn ratios are compared across control and PTX conditions, and show increase in MSI gain with inhibitory blockade. Error bars indicate +/– S.E.M. (D) The identity of the ISI's responsible for the maximal response (and thus the maximum MSI ratio) in control and PTX groups are compared. Error bars show interquartile range. *p<0.05. (E) MSI-versus-ISI tuning curves from crossmodal pairs were aligned at their peak values, and then averaged across cells. Solid lines connect the population means. Shaded areas demarcate +/– S.E.M. PTX treated cells are more broadly tuned around their preferred ISI. (F) For each cell, the maximum raw spike counts after paired stimulation (combined response) is plotted against the spike count predicted by the linear sum of individual modality responses. Dashed line represents the line of unity. Notice that in PTX treated group most points cluster to the left of the dashed line. (C) Cumulative Frequency Distributions of Z-score values are plotted for the comparison between predicted and actual number of action potentials recorded after paired stimulation, in each cell. Vertical dashed lines indicate Z = +/–1.97, the point at which actual responses are +/–2 S.D.'s away from the respective predicted response. For greater detail on generation of the predicted responses and calculation of Z-scores, see Materials and methods. Z-scores in the PTX treated group are significantly shifted toward the right (p<0.0001, Kolmogorov-Smirnov test).