Figure 4.

Decoding of different visual scenes. (A) Response of one example unit to two different congruent bimodal audio-visual pairs (conspecific vocalization and cage door being opened) and one incongruent combination. (B) Performance in discriminating two different visual scenes presented either as unimodal visual stimuli, congruent bimodal pairs or two visual scenes presented in the same acoustic context. For decoding the full time course of the response sampled at 20 ms was considered. The graphs display decoding performance as a function of the response interval considered, which started at stimulus onset and terminated at the corresponding value on the x-axis. Lines denote the mean across units (n = 33), blue area indicates the standard error for the congruent condition. Chance level is 50%. The lower panel displays the p-values (on a negative logarithmic axis) of point-wise t-tests between congruent and incongruent bimodal conditions (black) and visual and congruent bimodal conditions (gray). The dashed line indicates a critical p-value of 0.05 corrected for false discovery rate (for the black trace).