Interaural delay-dependent changes in the binaural difference potential of the human auditory brain stem response

Abstract

Binaural difference potentials (BDs) are thought to be generated by neural units in the brain stem responding specifically to binaural stimulation. They are computed by subtracting the sum of monaural responses from the binaural response, BD = B − (L + R). BDs in dependency on the interaural time difference (ITD) have been measured and compared to the Jeffress model in a number of studies with conflicting results. The classical Jeffress model assuming binaural coincidence detector cells innervated by bilateral excitatory cells via two delay lines predicts a BD latency increase of ITD/2. A modification of the model using only a single delay line as found in birds yields a BD latency increase of ITD. The objective of this study is to measure BDs with a high signal-to-noise ratio for a large range of ITDs and to compare the data with the predictions of some models in the literature including that of Jeffress. Chirp evoked BDs were recorded for 17 ITDs in the range from 0 to 2 ms at a level of 40 dB nHL for four channels (A1, A2, PO9, PO10) from 11 normal hearing subjects. For each binaural condition 10,000 epochs were collected while 40,000 epochs were recorded for each of the two monaural conditions. Significant BD components are observed for ITDs up to 2 ms. The peak-to-peak amplitude of the first components of the BD, DP1-DN1, is monotonically decreasing with ITD. This is in contrast with click studies which reported a constant BD-amplitude for ITDs up to 1 ms. The latency of the BD-component DN1 is monotonically, but nonlinearly increasing with ITD. In the current study, DN1 latency is found to increase faster than ITD/2 but slower than ITD incompatible with either variant of the Jeffress model. To describe BD waveforms, the computational model proposed by Ungan et al. [Hearing Research 106, 66–82, 1997] using ipsilateral excitatory and contralateral inhibitory inputs to the binaural cells was implemented with only four parameters and successfully fitted to the BD data. Despite its simplicity the model predicts features which can be physiologically tested: the inhibitory input must arrive slightly before the excitatory input, and the duration of the inhibition must be considerably longer than the standard deviations of excitatory and inhibitory arrival times to the binaural cells. With these characteristics, the model can accurately describe BD amplitude and latency as a function of the ITD.