Modeling adaptation to emphasize interaural time differences during rising amplitude

Correctly localizing a target sound in a background of competing sources is vital for survival in many species and important for human communication. Direct sound (emanating directly from a sound source to a listener's ear) is useful for source segregation and localization. Reverberant sound reflecting from multiple surfaces in the environment can obscure direct sound, but during a sound amplitude minimum of sufficient duration both the direct sound and reverberation become small. Then during the subsequent increase in sound amplitude, direct sound is relatively stronger than reverberation because direct sound leads reverberation in time. Noting this acoustic emphasis of source location information during increasing sound amplitude, Dietz and colleagues (2013, 2014) applied the amplitude-modulated binaural beat (AMBB, in which sound amplitude and interaural-phase difference (IPD) are modulated with a fixed mutual relationship), and their investigations produced two key results: (1) the human auditory system uses interaural timing differences in the temporal fine structure of modulated sounds only during the rising portion of each modulation cycle, and (2) the IPD during the rising segment dominates the total response in 78% of MSO neurons (anesthetized gerbils) and 69% of IC neurons (anesthetized guinea pigs), with the remaining neurons predominantly encoding IPD near the modulation maximum. Comparing two phenomenological models, they concluded that emphasis on IPDs during the rising slope of the AM cycle depends on adaptation processes occurring before binaural interaction. This modeling effort continues in the present study with the exploration of adaptation in biophysical models that aim to emphasize the encoding of IPD during rising amplitudes. Here AMBB stimuli will drive inputs to existing model auditory nerve fibers (Zilany et al, 2014), which in turn drive our models for cochlear nucleus (CN) bushy cells and MSO neurons. While the realistic adaptation of model auditory nerve fibers does not appear sufficient to emphasize IPD during the rising amplitude envelope, synaptic depression of inputs to spherical bushy cells (Xu-Friedman and Regehr, 2008) will be modeled as an additional means of adaptation. A recent globular bushy cell model (Rudnicki and Hennert, 2017) that constrains this synaptic depression together with measured entrainment of CN-driven inputs to the MSO (Joris et al., 1994) will be leveraged where practical. Model MSO neurons with slow membranes and synapses will be explored in relation to the minority of neurons that encode IPD near the modulation maximum.