Figure 5.

In a population of neurons, reverberant stimuli had a coding advantage over anechoic stimuli after controlling for stimulus modulation depth. A, MIOF from an IC neuron (BF = 4 kHz) for f_m = 16 Hz, showing RMD (mean ± 2 SD) as a function of stimulus modulation depth. Thick black line indicates least-squares fit with an incomplete beta function (see Materials and Methods). Green cross represents prediction of reverberant RMD from the MIOF. Red dot indicates measured strong reverberant RMD. In this example, measured reverberant RMD is significantly greater than the prediction from the MIOF. B, RMD to reverberant stimuli against RMD for diotic, anechoic stimuli with matched modulation depths in the contralateral ear for 99 neurons. In 59 neurons, RMDs were measured in response to depth-matched anechoic stimuli interleaved with the reverberant stimuli. In the 40 other neurons, anechoic data were not collected at the depth of the reverberant stimulus, so the depth-matched anechoic RMD was calculated by taking the value from a curve fit to the MIOF. In both cases, reverberant and depth-matched anechoic RMDs were computed in the steady-state portion of the stimulus (>250 ms). In 39% of the neurons, reverberant RMDs were significantly larger than RMDs for depth-matched anechoic stimuli (blue). In 10% of the neurons, reverberant RMDs were significantly smaller than RMDs for depth-matched anechoic stimuli (red).