Auditory specialisations for echo ranging

The mechanisms of attenuation outlined above are particularly important in facilitating accurate measurement of the echo delay, and hence of distance, because they prevent the auditory system from being overloaded by the outgoing sound pulses. In order to measure echo delay, the bat's auditory system must be capable of resolving very small time intervals. Most mammals are unsuited to this task because their auditory systems take too long to recover between successive sound stimuli - usually some tens of milliseconds need to elapse after a pulse of sound before full sensitivity is recovered. But in echolocation, most echoes will return within 30 ms of the outgoing pulse, given that bats track prey at distances of no more than 3 or 4 m and that sound travels at 334 ms"1.

It is therefore not surprising that the auditory neurons of bats are found to recover rapidly when stimulated with paired pulses of sound, roughly resembling the natural pairing of pulse and echo. For instance, summed potentials recorded from the midbrain of Myotis show some response to a second pulse that follows the first after only 0.5 ms and full recovery is achieved in 2 ms. Similar tests of hearing in fruit bats have shown that most species are slow to recover from the first pulse, but rapid recovery is found in one genus, Rousettus, which has evolved an echolocation capability independently of the insectivorous bats. Because this is the only respect in which the hearing of Rousettus differs conspicuously from that of its non-echolocating relatives, this rapid recovery may represent the most fundamental adaptation of the auditory system for echolocation (Grinnell & Hagiwara, 1972).

There is more to it than this, however, if a bat is to measure the arrival time of an echo rather than merely be aware of its existence. For echo delay to be coded by the auditory system as a time interval, some classes of inter-neurons must be able to act as accurate time markers. Interneurons that appear to be specialised as time markers for echolocation have been found in the inferior colliculus of bats belonging to several different genera. In fact, the majority of neurons in the inferior colliculus are found to have suitable properties, the most important of which is that their response is highly phasic. When one of these neurons is stimulated with a pair of FM pulses, it responds to each sound pulse by generating a single spike, or at the most two (Fig. 6.12a).

Repeated presentation of the stimulus shows that the delay between stimulus and response, the response latency, remains very consistent from one presentation to the next. The neuron's recovery is also sufficiently rapid that its response to the second pulse is essentially independent of that to the first pulse. As a result, the time interval between the two pulses is coded with remarkable precision in the pattern of spikes in the interneuron (Fig. 6.12a). Furthermore, the response latency remains almost constant regardless of the intensity of the sound stimulus, which is a crucial specialisation

20 kHz

2 ms

6 ms

Figure 6.12 Properties of time-marking interneurons in the inferior collicu-lus of bats. (a) An oscilloscope display showing a sound stimulus, consisting of a pair of identical, frequency-modulated pulses 6 ms apart (above), and the interneuron's response (below). Each dot represents the occurrence of a single spike in the responding neuron, and the pair of dot columns is generated by 16 consecutive presentations of the stimulus. (b) A similar display showing the response of a collicular neuron to a single frequency-modulated pulse, consisting of a downward sweep from 40 kHz to 20 kHz, stretched over three different durations (shown diagrammatically, above). The interneuron's response (dot columns, below) shifts in register with the occurrence of 25 kHz in the FM sweep (indicated by the arrowheads, above). This suggests that the neuron is responding to this particular frequency in the stimulus. (a from Pollak, 1980; b from Bodenhamer, Pollak & Marsh, 1979.)

for time marking in echolocation. The rate of recovery remains fast enough for the neurons to respond consistently to the second pulse, at short pulse intervals, even when the first pulse is as much as 30 dB louder than the second (Pollak et al., 1977; Pollak, 1980).

Each of these interneurons shows an exceptionally sharp tuning to a particular frequency within the FM sweep used for echolocation. This can be seen in their threshold curves, which have nearly vertical slopes on either side of the best frequency rather than the more usual V-shaped curve. As soon as the intensity of the appropriate frequency rises a few decibels above threshold, the neuron produces its single spike. Consequently, the neuron is fired promptly by the same frequency component in each sound pulse and so locks on to each event with consistent precision. This is shown nicely by stretching the sound pulse used as a stimulus over a longer time: for each stimulus duration, the neuron responds at the same relative position in the FM sweep (Fig. 6.12b).

Information about time of arrival provided by these neurons is passed on up the auditory pathway, particularly to the auditory region of the cerebral cortex. Neurons that are sensitive to time delays between the outgoing pulse and the returning echo have been found in the auditory cortex of all bat species tested. Such neurons are well suited to measuring the range of a target, based on the information provided by the time-marking neurons of the inferior colliculus. It is often the case that these cortical neurons are selective for different frequencies in the pulse and in the echo. This may reflect the fact that, during active flight, the echo will inevitably be Doppler shifted to a higher frequency than the outgoing pulse. The equivalent point in time for pulse and echo will therefore be indicated by separate time-marking neurons in the inferior colliculus.

The delay-sensitive neurons of the auditory cortex respond vigorously to paired FM pulses simulating natural pulse/echo pairs but hardly respond at all to FM pulses presented singly or to CF pulses. The response consists of a short train of spikes and occurs, in the majority of neurons, only if the time interval between pulse and echo is appropriate. If one of these neurons is presented with a sequence of pulse/echo pairs, in which the echo delay is progressively reduced to simulate the bat's approach to a target, it responds strongly only to a narrow range of delays (Fig. 6.13a). By systematically varying the echo delay and recording the neuron's responses, it can be seen that each of these neurons is sharply tuned to a particular echo delay, termed the best delay (Fig. 6.13b). Some of these delay-tuned neurons even respond preferentially to a narrow range of pulse-repetition rates, which correspond to the approach stage in intercepting prey (Wong, Maekawa & Tanaka, 1992).

Neurons having similar best delays are grouped together in inwardly directed columns within a specific area of the auditory cortex (the location of the auditory cortex is shown in Fig. 6.13c). In the moustached bat, Pteronotus parnellii, these columns are arranged systematically with best delay increasing along the cortical surface from anterior to posterior (Fig. 6.13d). The delay-tuned neurons are thus arranged topographically according to the distance they encode and so provide a neural representation of target range. This delay-tuned area of the auditory cortex is quite distinct

Figure 6.13 Echo-ranging neurons in the auditory cortex of bats. (a) A histogram (above) summarising the response of an echo-ranging neuron from Myotis to a sequence of pulse/echo pairs (below) simulating the natural approach to a target. (b) The response of another neuron from Myotis, expressed as a percentage of the maximum number of spikes, as a function of the delay between a simulated pulse and echo of constant amplitude. (c) A diagram of the brain of Myotis, viewed from the left and above, showing the auditory region of the cerebral cortex. (d) The FM area of the auditory cortex in Pteronotus, showing the systematic distribution of echo-ranging neurons according to their best delay. The solid lines labelled with a number are contours of best delay in milliseconds. There are three horizontal clusters of neurons, indicated by broken lines, each tuned to only one of the three harmonics (H2, H3, H4) in the echo. (a from Wong etal., 1992; b and c modified after Sullivan, 1982; d modified after O'Neill & Suga, 1982.)

Figure 6.13 Echo-ranging neurons in the auditory cortex of bats. (a) A histogram (above) summarising the response of an echo-ranging neuron from Myotis to a sequence of pulse/echo pairs (below) simulating the natural approach to a target. (b) The response of another neuron from Myotis, expressed as a percentage of the maximum number of spikes, as a function of the delay between a simulated pulse and echo of constant amplitude. (c) A diagram of the brain of Myotis, viewed from the left and above, showing the auditory region of the cerebral cortex. (d) The FM area of the auditory cortex in Pteronotus, showing the systematic distribution of echo-ranging neurons according to their best delay. The solid lines labelled with a number are contours of best delay in milliseconds. There are three horizontal clusters of neurons, indicated by broken lines, each tuned to only one of the three harmonics (H2, H3, H4) in the echo. (a from Wong etal., 1992; b and c modified after Sullivan, 1982; d modified after O'Neill & Suga, 1982.)

from the tonotopic area. For reasons that are not yet understood, the auditory cortex of Myotis is less highly differentiated: there is a tendency for neurons with larger best delays to be located more posteriorly but this hardly constitutes a neural map of best delay. Nor is the delay-tuned area clearly separated from the neighbouring tonotopic area; in fact, there is a considerable overlap (Wong & Shannon, 1988).

The echolocation pulses of Pteronotus contain three higher harmonics in addition to the fundamental frequency, and each delay-tuned neuron responds to only one of these three harmonics in the echo (H2, H3 and H4 in Fig. 6.13d). By varying the intensity of the simulated echo when testing these neurons, it has proved possible to measure the threshold of the response at each delay tested and so to construct a threshold curve. A conspicuous result is that each of these curves has an upper threshold as well as a lower one, which means that the neuron fails to respond if the echo is too loud as well as if it is too quiet. Also, neurons tuned to shorter delays tend to have a higher threshold coupled with a narrower delay range. Each of the delay-tuned neurons is thus tuned to a particular combination of time delay and intensity appropriate to echoes returning from a certain distance (O'Neill & Suga, 1982).

The best delays in the FM area of the auditory cortex of Pteronotus cover a range from 0.4 ms at the anterior edge to 18 ms at the posterior edge (Fig. 6.13d), which corresponds to target ranges from 7 to 310 cm. This agrees closely with the range over which bats are observed to detect and react to targets (see section 6.6). An especially large number of neurons, reflected in cortical surface area, is devoted to delays from 3 to 8 ms (50 to 140 cm), corresponding roughly with the approach stage of target interception.

Especially with the larger values of echo delay, it is obvious that the neural response to the emitted pulse must be considerably delayed if it is to reach the cortex at the same time as the response to the echo. In fact, a large range of response latencies is found among the time-marking neurons of the inferior colliculus. It is therefore possible that the neural pathways through the colliculus are acting as delay lines and the cortical neurons are effectively coincidence detectors for particular combinations of pulse and echo latencies. The neural map of target range could thus be assembled in a similar manner to the barn owl's map of interaural time differences.

Because the FM area of the auditory cortex is sharply separated from the tonotopic area in Pteronotus, it is possible to inactivate them separately with drugs. As with the corresponding experiments in the barn owl's brain

(see section 6.4), this provides a test of the behavioural function of these regions. When the FM area is inactivated with the drug, fine discrimination of target range is impaired, though coarse discrimination is still possible, and frequency discrimination remains unaffected. This confirms that the FM area is involved in the perception of distance, as expected from the responses of its interneurons and their topographical arrangement, but it appears to have little to do with frequency discrimination (Riquimaroux, Gaioni & Suga, 1991). The opposite result is obtained when the tonotopic area is inactivated with the drug, and this area appears to be specialised for analysis of the Doppler-shifted CF signal.

Essentials of Human Physiology

Essentials of Human Physiology

This ebook provides an introductory explanation of the workings of the human body, with an effort to draw connections between the body systems and explain their interdependencies. A framework for the book is homeostasis and how the body maintains balance within each system. This is intended as a first introduction to physiology for a college-level course.

Get My Free Ebook


Post a comment