Robert D. Bodenhamer

Time and Frequency domain processing in the inferior colliculus of echolocating bats.

Tone bursts and frequency-modulated (FM) signals were presented to Mexican free-tailed bats and tuning curves, discharge patterns, and discharge latencies of single units in the inferior colliculus were recorded. Cells were broadly tuned to tone bursts, with most Q 10 values ranging from 3 to 20. However, in response to FM stimulation the discharges of neurons were closely synchronized to the time of occurrence of restricted frequency components within the FM sweep. These excitatory frequencies (EFs) were generally unaffected by changes in the starting frequency or intensity of the stimulus. Thus, in response to FM signals, the cells exhibited a much greater frequency selectivity than that observed following tone burst stimulation. Across the population of neurons sampled, EFs covering a wide frequency range were found, and the different EFs were represented in a systematic fashion within the colliculus. The frequencies in an FM biosonar signal or echo will thus be neurally represented both by the time of occurrence of neuronal discharges and by the location of the discharging cells within the nucleus. The potential role of this dual frequency coding in spectral and temporal processing of biosonar signals and echoes is discussed, with emphasis on the neural coding of target range.

Response characteristics of single units in the inferior colliculus of mustache bats to sinusoidally frequency modulated signals

Summary1.The discharge properties of single units in the inferior colliculus of the mustache bat,Pteronotus parnellii, were evaluated with tone bursts and sinusoidal frequency modulated (SFM) stimuli. The SFM signals were designed to mimic the modulation patterns imposed upon the echoes reflected from the beating wings of flying insects.2.Two groups of neurons were distinguished on the basis of their best frequencies (BFs). The first group was called the filter neurons and had BFs between 60–64 kHz. These neurons are involved in processing the constant frequency component of the bats echolocation calls. The second group, or non-filter, neurons, had BFs below 60 kHz, or above 64 kHz.3.Ninety percent of the filter neurons and 80% of the non-filter neurons responded to SFM signals with discharges that were phase locked to the modulation waveform. However, the two groups of neurons differed markedly in their sensitivity to very low modulation depths. The majority of filter neurons displayed locked discharges to SFM depths as low as ±50 Hz, and some even locked to depths of ±10 Hz, while only two non-filter units (out of 71 tested) locked to depths of ±50 Hz, and none locked to ±10 Hz.4.Most filter units displayed a marked selectivity, or tuning, for particular modulation depths, rates, and signal intensities. In these tuned or selective units, locking could be evoked only over a fairly narrow range of depths, rates, and/or intensities. When SFM parameters outside of the neurons tuned range were presented, only phasic-on or on-off discharges could be evoked.5.The influence of the SFM carrier frequency was also evaluated in filter units. In some units, the particular carrier frequency used was not critical for evoking locked discharges so long as the signal encroached upon the units tuning curve. In the majority of filter units, however, there was a marked asymmetry in that some carrier frequencies were much more effective in evoking locked discharges than were other frequencies. In some asymmetric units, frequencies below the BF were most effective and in other units carrier frequencies above the BF were most effective.6.We examined the inhibitory surround patterns in 15 filter units under the assumption that a predominant inhibitory flank on one or another side of the BF might account for the asymmetry in the effective SFM carrier frequencies. Detailed analyses in eight filter neurons yielded no relationship between the position of the inhibitory side bands and the asymmetry of SFM frequencies evoking locked discharges.

Recovery Cycles of Single Neurons in the Inferior Colliculus of Unanesthetized Bats Obtained with Frequency-modulated and Constant-Frequency Sounds

Summary1.Recovery cycles were recorded from individual units in the inferior colliculus of unanesthetized Mexican free-tailed bats,Tadarida brasiliensis mexicana. With most units, two brief FM signals which mimicked the natural biosonar signals of this species were used but with a few units two constant-frequency (CF) tone bursts were utilized. Up to 110 recovery cyles were generated by each unit in which the intensities of the first and second signals, the inter-pulse-intervals (IPIs), and the signal durations were systematically varied.2.In general, four recovery patterns were observed. The first is poor recovery where the units were typically unresponsive to the second stimulus (Fig. 2). The second is independent-like recovery where the units fired to both signals as if the initial had little or no effect upon the response to the second signal (Figs. 3–9). The third variety is selective recovery typified by more or less poor recovery except at one particular combination of stimulus parameters where recovery improved significantly (Figs. 10, 11). The fourth, and most dramatic, is pulse-enhanced recovery characterized by a facilitated firing to the second signal having a subthreshold intensity, responding contingent upon the presentation of a loud initial signal (Figs. 12, 13).3.The particular recovery pattern was highly dependent upon stimulus duration. The great majority of units exhibited a poor recovery pattern with signal durations less than 2 ms. Many of these neurons, however, also exhibited an independent-like recovery pattern with signals having longer durations. One unit exhibited poor recovery with 1.7 ms FM bursts, selective recovery with 3.4 ms FM signals and independent-like recovery with 6.9 ms FM signals (Figs. 2, 10 and 11).4.70% of the units sampled had independent-like recovery patterns with 3.4–8.0 ms signals. Of particular importance, each of these neurons recovered completely when the first signals were 30–50 dB more intense than the second signals and this commonly occurred at IPIs as short as 3.9 ms. In all of these cells, the initial signals were always 72–82 dB and the second signals ranged from 32–52 dB, intensities comparable to those of the orientation cries and returning echoes experienced during the goal-directed phase of echolocation.5.Three units had pulse-enhanced recovery patterns where the units responded to second signals having intensities of from 5–20 dB. At these low intensities the second signals failed to elicit responses when presented alone but evoked firings when preceded by initial signals of 70–80 dB. These intensities correspond to the pulse intensities reaching the ears upon the emission of loud orientation cries and the echo intensities expected at the point of initial detection of a target.6.Pulse-enhanced units were only seen with 26 kHz tone bursts and were never observed with FM signals. SinceTadarida have been observed to emit 26 kHz constant-frequency pulses during the search portion of echolocation, it is suggested that the pulse-enhanced responders are specialized for echodetection.

Cochleotopic Organization of the Mustache Bat’s Inferior Colliculus

In recent years considerable attention has been given to the neural correlates of echolocation in bats (Busnel and Fish, 1980). Favorite subjects for these studies have been the mustache bat, a New World species studied by American investigators, and the horseshoe bat, an Old World bat used by European scientists. Both of these bats are of particular interest because of their extraordinarily good frequency discrimination, coupled with one of the most highly specialized auditory systems of any known animal. An appreciation of these animals is best gained by first considering some of their behavioral characteristics. Both the mustache and horseshoe bats belong to a nontaxonomic group called the long CF/FM bats, a name deriving from the types of echolocation signals that they emit. The most prominent feature of the calls is a long constant frequency (CF) component, a tone burst in effect, followed by a less conspicuous, brief frequency modulated (FM) portion (Schnitzler and Henson, 1980). A point of importance is that the pattern of pulse emission has a highly dynamic quality. Most species of long CF/FM bats that have been studied compensate for Doppler-shifts in the echo CF component (Schnitzler and Henson, 1980; Suga and O’Neill, 1980; Henson et al., 1980). Doppler-shifts occur whenever there is a difference in flight speed between the bat and its target.

Frequency Selectivity of Constant Latency Neurons in the Inferior Colliculus of Mexican Free-Tailed Bats

There are neurons in the inferior colliculus of the Mexican free-tailed bat, Tadarida brasiliensis mexicana, which are characterized by a highly stable discharge latency to frequency modulated (FM) signals. These phasic constant latency responders (pCLRs) are capable of accurately encoding the time interval between simulated orientation cries and echoes, and thus appear to be well suited for encoding target range.