Dynamic temporal signal processing in the inferior colliculus of echolocating bats

Abstract

In nature, communication sounds among animal species including humans are typical complex sounds that occur in sequence and vary with time in several parameters including amplitude, frequency, duration as well as separation, and order of individual sounds. Among these multiple parameters, sound duration is a simple but important one that contributes to the distinct spectral and temporal attributes of individual biological sounds. Likewise, the separation of individual sounds is an important temporal attribute that determines an animal's ability in distinguishing individual sounds. Whereas duration selectivity of auditory neurons underlies an animal's ability in recognition of sound duration, the recovery cycle of auditory neurons determines a neuron's ability in responding to closely spaced sound pulses and therefore, it underlies the animal's ability in analyzing the order of individual sounds. Since the multiple parameters of naturally occurring communication sounds vary with time, the analysis of a specific sound parameter by an animal would be inevitably affected by other co-varying sound parameters. This is particularly obvious in insectivorous bats, which rely on analysis of returning echoes for prey capture when they systematically vary the multiple pulse parameters throughout a target approach sequence. In this review article, we present our studies of dynamic variation of duration selectivity and recovery cycle of neurons in the central nucleus of the inferior colliculus of the frequency-modulated bats to highlight the dynamic temporal signal processing of central auditory neurons. These studies use single pulses and three biologically relevant pulse-echo (P-E) pairs with varied duration, gap, and amplitude difference similar to that occurring during search, approach, and terminal phases of hunting by bats. These studies show that most collicular neurons respond maximally to a best tuned sound duration (BD). The sound duration to which these neurons are tuned correspond closely to the behaviorally relevant sounds occurring at different phases of hunting. The duration selectivity of these collicular neurons progressively increases with decrease in the duration of pulse and echo, P-E gap, and P-E amplitude difference. GABAergic inhibition plays an important role in shaping the duration selectivity of these collicular neurons. The duration selectivity of these neurons is systematically organized along the tonotopic axis of the inferior colliculus and is closely correlated with the graded spatial distribution of GABA(A) receptors. Duration-selective collicular neurons have a wide range of recovery cycle covering the P-E intervals occurring throughout the entire target approaching sequences. Collicular neurons with low best frequency and short BD recover rapidly when stimulated with P-E pairs with short duration and small P-E amplitude difference, whereas neurons with high best frequency and long BD recover rapidly when stimulated with P-E pairs with long duration and large P-E amplitude difference. This dynamic variation of echo duration selectivity and recovery cycle of collicular neurons may serve as the neural basis underlying successful hunting by bats. Conceivably, high best frequency neurons with long BD would be suitable for echo recognition during search and approach phases of hunting when the returning echoes are high in frequency, large in P-E amplitude difference, long in duration but low in repetition rate. Conversely, low best frequency neurons with shorter BD and sharper duration selectivity would be suitable for echo recognition during the terminal phase of hunting when the highly repetitive echoes are low in frequency, small in P-E amplitude difference, and short in duration. Furthermore, the tonotopically organized duration selectivity would make it possible to facilitate the recruitment of different groups of collicular neurons along the tonotopic axis for effective processing of the returning echoes throughout the entire course of hunting.

Keywords: duration selectivity; echolocation; inferior colliculus; pulse-echo pairs; recovery cycle; temporal signal processing.

Figure 1

(A) band-pass, (B) short-pass, (C) long-pass, and (D) all-pass duration tuning curves of collicular neurons of the FM bat, Eptesicus fuscus. Each horizontal dashed line indicates the 50% maximal response. The selectivity of each echo duration curve is expressed with a best duration (BD, indicated with an arrowhead) and a normalized duration width (nDW indicated with a double-arrowheads). NA: a BD is not available [Adapted from Wu and Jen (2006)].

Figure 2

(A–H) The echo duration tuning curves of two collicular neurons determined with the echo pulses of three pulse-echo (P-E) pairs and with single pulses. The type of duration tuning curve, the BD and the nDW are shown within each plot. Note that the nDW progressively decreases with lengthening of PD and P-E gap of three P-E pairs [Adapted from Wu and Jen (2006)].

Figure 3

(A,B) The echo duration tuning curves of a collicular neuron determined with three P-E pairs with varied pulse duration and P-E gap at 20-dB (unfilled) and 10-dB (filled) P-E amplitude differences [Adapted from Jen and Wu (2008)].

Figure 4

(A,B) Echo duration tuning curves of two collicular neurons determined with three P-E pairs of varied pulse duration and P-E gap before (predrug, A,C) and during bicuculline (B) or GABA (D) application [Adapted from Wu and Jen (2006)].

Figure 5

(Aa,Ba) The distribution of nDW of collicular neurons obtained before (filled circles and triangles) and during (unfilled circles and triangles) bicuculline (Aa) and GABA (Ba) application in relation to the recording depth. (Ab,Bb) The distribution of percent change in nDW in relation to the recording depth. The linear regression line and correlation coefficient for each plot are shown by a solid line and r [From Jen and Wu (2006)].

Figure 6

(A) The recovery cycle of two duration-selective collicular neurons measured with P-E pairs that vary in P-E gap, duration, and amplitude differences (PA-EA = 0, 10, and 20 dB). The BDs of these two neurons are 1.5 ms (A) and 10 ms (B). Note that neuron A has the shortest 50% recovery time (dashed line) when stimulated with 1.5 ms of P and E at 10-dB P-E amplitude difference (A-2, arrow) while neuron B recovers rapidly when stimulated with 10 ms of P and E at 20-dB P-E amplitude difference (B-3, arrow) [Adapted from Wang et al. (2008)].