A neuronal model of sound location map generated based on multiplicative binding of ITD and IID information

The sound localization of barn owl with respect to horizontal and vertical direction is made based on analysis of the interaural differences in arrival time and intensity of sound, respectively. The neural information processing of detection of interaural intensity difference (IID) is made on the neural pathway parallel with the pathway of interaural time difference (ITD) detection before both the signals arrive at the shell of central nucleus of the inferior colliculus (ICc shell). We have presented a neural model of ITD detection pathway (Inoue et al, Neurocomputing 2001, vol. 38-40, pp. 675-682). We present a neuronal model of the nucleus ventralis lemnisci lateralis pars posterior (VLVp) in which the signals of sound intensities coming from both the ears are first combined to be compared with each other. We clarify the functional role of mutual inhibitory connections between both VLVps observed experimentally in detection of IID. Outputs of IID sensitive neurons in both VLVps are projected bilaterally to ICc shell and the connections are inhibitory, whereas outputs of ITD sensitive neurons in ICc core projected to ICc shell are excitatory. It is shown using our neuronal models of VLVps, ICc core, and ICc shell that the neural map, in which ITD and IID are represented along the axes perpendicular mutually, can be generated in ICc shell by the excitatory inputs from ICc core and the inhibitory inputs from bilateral VLVps, as observed experimentally. We also show that the firing rate of ICc shell neuron are well represented by a suppressive multiplication of ITD and IID inputs.