Does a lack of precisely timed input hamper envelope ITD sensitivity when presenting amplitude-modulated low-frequency (< 500hz) carriers?

Sensitivity to interaural time differences conveyed in the temporal fine structure (TFS) of low-frequency sounds is most often assessed using unmodulated sound envelopes. Nevertheless, many natural sounds are modulated in amplitude, and a recent psychoacoustic study highlighted how sensitivity to ITDs conveyed in the TFS is weighted as a function of the location within the envelope modulation cycle (Hu et al., 2017). Whilst perception of ITDs is dominated by ITDs conveyed in the initial, rising phase of an amplitude-modulated 600Hz stimulus (consistent with psychoacoustic/physiological studies employing 500-Hz carriers (Dietz et al., 2013, 2014)), it resides in the middle portion of the modulation cycle for a 200-Hz stimulus. This disparity in envelope ITD processing strategies is surprising: the 200-Hz carrier strategy more closely resembles an "integration of energy across the modulation cycle" approach observed for high-frequency (>1500Hz) carriers (Dietz et al., 2013; Remme et al., 2014). The binaural computations underlying envelope ITD processing for such low-frequency carriers are typically ascribed to principal neurons in the MSO (Dietz et al., 2013, 2014). Their exquisite ability to detect fine structure ITDs arises from biophysical specializations seemingly ubiquitous across the MSO's tonotopic axis (Svirskis et al., 2004; Scott, Mathews and Golding, 2005; Chirila et al., 2007; Mathews et al., 2010; Myoga et al., 2014; Remme et al., 2014). It is therefore more likely that the observed shift in envelope ITD sensitivity from rising phase to peak of modulation cycle for a 200Hz carrier is a result of the synaptic input generated in the rising amplitude envelope being more unreliable for a 200-Hz carrier compared to a 600-Hz carrier (i.e. fewer synaptic input cycles with increased jitter during the rising phase). To test this hypothesis, we presented the amplitude-modulated stimuli employed by Hu et al. (2017) to an auditory nerve computational model (Zilany, Bruce and Carney, 2014) and harnessed its output to generate synaptic input for a linear MSO principal neuron model that included a threshold mechanism for spiking (Remme et al., 2014). Simulated neurometric functions (d primes) were thereby generated from the linear model's suprathreshold responses to the various envelope ITD conditions (i.e. envelope ITDs in rising, middle and end phases). Initial results suggest that it is indeed possible to reproduce the psychoacoustic results within our model for both 200Hz and 600Hz carrier stimuli.