J o u n m l of

J Comp Physiol A (1989) 164: 459-474

~orlsoP/,

Comparal~e Neural, and P h y s i o l o g y A Behavioral ~,~, 9 Springer-Verlag 1989

Physiology of lateral line mechanoreceptive regions in the elasmobranch brain H. Bleckmann*, O. Weiss* *, and T.H. Bullock Neurobiology Unit, Scripps Institution of Oceanography and Department of Neurosciences, School of Medicine, University of California, San Diego, La Jolla, California 92093, USA Accepted September 16, 1988

Summary. The physiology of mechanoreceptive lateral line areas was investigated in the thornback guitarfish, Platyrhinoidis triseriata, from medulla to telencephalon, using averaged evoked potentials (AEPs) and unit responses as windows to brain functions. Responses were analysed with respect to frequency sensitivity, intensity functions, influence of stimulus repetition rate, response latency, receptive field (RF) organization and multimodal interaction. 1. Following a quasi-natural vibrating sphere stimulus, neural responses were recorded in the medullary medial octavolateralis nucleus (MON), the dorsal (DMN) and anterior (AN) nucleus of the mesencephalic nuclear complex, the diencephalic lateral tuberal nucleus (LTN), and a telencephalic area which may correspond to the medial pallium (Figs. 2, 3, 13, 14, 15, 16). 2. Within the test range of 6.5-200 Hz all lateral line areas investigated responded to minute water vibrations. Best frequencies (in terms of displacement) were between 75 and 200 Hz with threshold values for AEPs as low as 0.005 gm peak-to-peak (p-p) water displacement calculated at the skin surface (Fig. 6). 3. AEP-responses to a vibrating sphere stimulus recorded in the M O N are tonic or phasic-tonic, i.e., responses are strongest at stimulus onset but Abbreviations: A N anterior nucleus of the mesencephalic nuclear complex; A E P average evoked potential; D M N dorsomedial nucleus of the mesencephalic nuclear complex; L M N lateral nucleus of the mesencephalic nuclear complex; L T N lateral tuberal nucleus; M O N medial octavolateralis nucleus; M U A multiunit activity; p L L N posterior lateral line nerve; P S T H peristimulus time histogram; R F receptive field

Present address: * UniversitS.t Bielefeld, Fakult/it ffir Biologie II, Postfach 8640, D-4800 Bielefeld 1, Federal Republic of Germany 9* Abteilung Anatomie der Medizinischen Fakult/it Universit/it G6ttingen, D-3400 G6ttingen, Federal Republic of Germany

last for the whole stimulus duration in form of a frequency following response (Fig. 3). D M N and AN responses are phasic or phasic-tonic. Units recorded in the M O N are phase coupled to the stimulus, those recorded in the D M N , A N or L T N are usually not (Figs. 5, 8, 9). Diencephalic L T N and telencephalic lateral line responses (AEPs) often are purely phasic. However, in the diencephalic L T N tonic and/or off-responses can be recorded (Fig. 11). 4. For the frequencies 25, 50, and 100 Hz, the dynamic intensity range of lateral line areas varies from 12.8 to at least 91.6 dB (AEP) respectively 8.9 and 92 dB (few unit and single unit recordings) (Fig. 7). 5. Mesencephalic, diencephalic, and telencephalic RFs, based on the evaluation of AEPs or multiunit activity (MUA), are usually contralateral (AN and LTN) or ipsi- and contralateral (telencephalon) and often complex (Figs. 10, 12, 16). 6. In many cases no obvious interactions between different modalities (vibrating sphere, electric field stimulus, and/or a light flash) were seen. However, some recording sites in the mesencephalic A N and the diencephalic L T N showed bimodal interactions in that an electric field stimulus decreased or increased the amplitude of a lateral line response and vice versa (Fig. 13 B).

Introduction Fishes and aquatic amphibians use the mechanoreceptive lateral line to detect weak water movements (Dijkgraaf 1963; Bleckmann 1986). F r o m 5 up to about 200 Hz, primary lateral line afferents have vibration thresholds down to 0.0025 gm peak-topeak (p-p) displacement (Kuiper 1967; Harris and

460

H. Bleckmann et al. : Lateral line mechanoreception in elasmobranchs

van Bergeijk 1962; Bleckmann and Topp 1981; Sand 1981; Gray 1984). The pattern of impulses carried by primary lateral line afferents encodes information about the nature of the stimulus with respect to duration, amplitude, frequency, and phase (e.g., M/inz 1985; Elepfandt and Wiedemer 1987). If the activity of several neuromasts, which may differ with respect to the alignment of their most sensitive axis, is integrated over time and space, the additional information of stimulus direction and, perhaps, stimulus distance may be obtained (Hoin-Radkovski et al. 1984). Thus the peripheral lateral line provides the brain with all cues necessary to evaluate a complex wave stimulus. In contrast to the considerable attention attracted to the peripheral aspects of the lateral line, relatively little has been done with the central physiology (for reviews see Roberts 1981; Bleckmann and Bullock, in press). This, especially, holds true for cartilaginous fishes. The few papers which address this question report on the physiology of the first relay station of primary lateral line afferents, i.e., on the physiology of the medial octavolateralis nucleus (Hoagland 1935; Alnaes 1973; Paul and Roberts 1977a, b; Caird 1978) or the midbrain tectum and torus semicircularis (Callens et al. 1967; Knudsen 1977; Nederstigt and Schellart 1986; Zittlau et al. 1985, 1986). Platt et al. (1974) are the first to localize areas responsive to lateral line nerve shock in diencephalon and telencephalon. Bleckmann et al. (1987) used the guitarfish, Platyrhinoidis triseriata, to localize the areas responsive to a posterior lateral line nerve shock in mesencephalon, diencephalon, and telencephalon and to characterize the intensity functions and temporal dynamics in each region. The present paper extends the results on Platyrhinoidis, mainly by recording medullary, mesencephalic, diencephalic, and telencephalic lateral line responses to quasiphysiological stimuli. The responses recorded from higher brain centers are analysed with respect to frequency sensitivity, response latency, intensity response functions, influence of stimulus repetition rate, receptive field (RF) organization, and interaction with electric field stimuli and light flashes. For comparison some experiments were done with the hornshark, Heterodontusfrancisci. Materials and methods Male and female thornback guitarfish, Platyrhinoidis triseriata (Rhinobatoideae, Elasmobranchii, Chondrichthyes), weighing 190-900 g, were collected in shallow water off the coast of San Diego. Hornsharks, Heterodontus francisci (Heterodontidae, Chondrichthyes), weighing 300-640 g, were captured near Catalina Island. Surgical procedures to expose the brain, as well

as treatment of the fish prior to and during an experiment were similar with those reported elsewhere (Bleckmann et al. 1987). Stimulation. Local quasi-natural stimulation of the lateral line was produced by a vibrating plastic sphere (17.6 mm diameter), attached to a 2 mm diameter rod and driven in the axis of the rod by a vibrator (Ling Dynamic Systems, Model 102). The vibrator, which was placed so that the movement of the sphere was approximately perpendicular or parallel to the skin surface, rested on a separate steel platform ensuring that no vibrations were transmitted indirectly to the fish via the ground and experimental tank. Stimulation frequencies varied between 6.5 and 200 Hz. Varying peak-to-peak (p-p) displacements were obtained by means of an attenuator. For calibration the sphere movement was monitored in the absence of an animal for each stimulus frequency under a microscope (max. resolution 3.6 gm p-p displacement). For amplitudes < 3.6 lain, the displacement of the sphere was calculated by extrapolation. To avoid possible boundary layer effects (Kalmijn 1988) the distance between the surface of the fish and the sphere was no less than 4 mm. The actual p-p displacement (A) of the wave stimulus was calculated according to the equation

A = ~ where dr and do are the radial and angular displacement components of the wave stimulus (cf. Harris and van Bergeijk 1962). This calculation, which does not consider any boundary layer effects, was used to define the stimulus. If not otherwise stated the overall stimulation time was 500 ms with a rise and fall time of 20 and 100 ms, respectively. Every stimulus was followed by a pause of 1 to 55 s. For stimulating the ampullary electroreceptors an approximately uniform electric field of 5~500 ms duration was produced by passing the output of an optical stimulus isolation unit to two carbon rods positioned to the left and right of the fish. A field strength of 2-68 gV/cm was used. The electric field, whose isopotential lines were parallel to the longitudinal axis of the fish, was monitored with two electrodes 25 cm apart. Visual stimuli were provided by a stroboscopic flash directed to one or both eyes through a light pipe. Data collection. Averaged evoked potentials (AEPs), multiunit activity (MUA), few units, and single units were recorded with an electrode in the medulla, mesencephalon, diencephalon, or telencephalon using varnish-insulated tungsten microelectrodes (Frederick Haer, Brunswick, ME, USA), laboratory made indium-alloy-filled glass microelectrodes (Dowben and Rose 1953), or glass microelectrodes filled with 3 M NaC1. In all cases a silver wire inserted into the water bath served as a reference. Electrode impedances measured at 22 Hz varied from 1~4 Mg2 (tungsten electrodes), 0.%2 Ms (indium-alloy metal microelectrodes), and 2-17 M~2 (KC1 filled glass capillary micro-electrodes). Neural responses were band-pass filtered from 3-3000 Hz (AEPs) or 400-3000 Hz (unit responses) and 4 to 32 responses were summed on a digital signal processor (Nicolet 1170). Peri-stimulus time histograms (PSTHs) were generated using a level detector. Electrode penetrations were vertical to the plane of the table, hence to the frontal plane of the ray. The electrode was advanced with a microdrive (Burleigh Model PZ 550) which was attached to a micromanipulator and equipped with a digital depth readout. Determination of threshold values and RFs. When lateral line responses were encountered the effect of the vibration frequency of the sphere was measured in the range 6.5 to 200 Hz. For a given frequency the stimulus intensity was gradually reduced

H. Bleckmann et al. : Lateral line mechanoreception in elasmobranchs as increasingly sensitive body regions were found until no response could be elicited even at the region with the lowest threshold. The stimulus was then augmented to threshold at which an AEP could be recorded or at which a neuron responded in about 50% of the trials. In case of less well isolated units the threshold was judged with aid of an acoustic monitor and PSTHs calculated on-line. Usually the procedure was repeated several times until a given threshold was consistent within • 1 dB (for the forebrain • 2 dB). After completion of the threshold curve the function generator was adjusted to 100 Hz, a frequency to which all lateral line areas were very displacement sensitive. The sphere was then moved in 5 or 10 m m steps in the anterior, thereafter in the posterior direction until no or only a very weak neural response (judged on the amplitude of the AEP and/or a PSTH) could be recorded. Thereafter the sphere was moved I or 2 cm lateral or medial and the whole procedure was repeated. Because of the limited mobility of the vibrating sphere assembly, RFs could not be mapped on the ventral surface of the ray. In all cases the shortest distance between the surface of the sphere and the skin of the fish was adjusted to exactly 5 ram. Electrolytic lesions were made to mark recording sites for histological confirmation by passing 10 gA of constant cathodal current for 20 s. For the recovery of the lesions the animals some hours later were deeply reanesthetized with MS 222 (1:5000 in sea water), and sequentially perfused through an 18 gauge cannula in the conus arteriosus with 250-500 ml each of 1) elasmobranch Ringer's solution containing MS 222 (1:5000) and Heparin (150 Units/kg body weight), and 2) a fixative containing 1.5% glutaraldehyde and 1% paraformaldehyde in a 0.1 M, pH 7.4 phosphate buffer. The removed brains were stored for 1-2 days at 4 ~ in a 25% sucrose solution, then cut serially at 35 p.m on a cryostat (Histostat Microtome, American Optical), air dried and subsequently stained with eresyt violet. The number of experimental animals (N) and the number of recording sites or averaged responses (n) is given for every experiment.

Results Medulla

In cartilaginous fishes the medullary medial octavolateralis nucleus (MON) is the primary target of the mechanoreceptive lateral line (e.g., Bodznick and Northcutt 1980). If the brain of P. triseriata is penetrated with an electrode at the level 2 (Fig. 1), local AEPs to a weak vibrating sphere stimulus can be recorded at a depth which closely corresponds to the M O N (Fig. 2). Like acoustic frequency following responses recorded from the hindbrain in different vertebrate taxa (Bullock 1986), AEPs recorded in the M O N often show a delayed, tonic wave form that maintains a distorted first harmonic of the stimulus frequency and which may add, depending on recording depth and stimulus frequency, the second harmonic, sometimes only in the middle of a 500 ms response. In the examples shown (Figs. 2, 3) the eighth nerve was cut bilaterally. Consequently acoustic test

461 2

I VII

oils

VIII

IX

i Fig. 1. Dorsal view of the brain of Platyrhinoidis triseriata. Bars numbered 1 and 2 represent the rostro-caudal levels of transsections shown in Figs. 2 and 14. alll anterior lateral line lobe; alln anterior lateral line nerve; C cerebellum; O T optic tectum; plll posterior lateral line lobe; plln posterior lateral line nerve; T telencephalon. Cranial nerves are labeled by R o m a n numerals. Drawing of the brain provided by courtesy of R.G, Northcurt

stimuli did not evoke a neural response within the test range 40 to 1000 Hz, although the loudness of the stimuli was about 70 dB above the human hearing threshold. Depending on the vibration amplitude of the sphere both AEPs and units recorded in the M O N show a phasic or phasic-tonic response (Figs. 3, 4) which may be followed by a silent period (Fig. 5A, B). Thus AEPs and units recorded in the M O N clearly encode the stimulus duration if higher intensities are used. MON-units are phase coupled to the stimulus, at intermediate or high stimulus intensities firing up to one spike per stimulus cycle (Fig. 4). Phase-locking increases with stimulus strength and may reach a maximum (Fig. 5A, B). Without exception, responses recorded in the M O N had their 'best' frequency (in terms of displacement) in the range 100-150 Hz (Fig. 6A). Within this range a p-p displacement of 0.02 ~tm was sufficient to generate a neural response in the most sensitive preparation and recording site. On either side of the 'best' frequency the sensitivity of neural responses on average decreased by 1 2 . 5 d B _ 3 . 7 (from 6.5-100Hz) or 20.9 dB• (from 100-200 Hz) per octave ( N = 6 , n = 20). For all lateral line areas investigated the dynamic intensity range between threshold and maximal response, was determined for the frequencies

462

H. Bleckmann et al. : Lateral line mechanoreception in elasmobranchs

/ /~ I/ ~V'~'~ / / ~ / /

.~ / / / EG /

~///~DN/

~EG

400pm 700 lOOO

1300

~ ~

16oo

~ ~

2200 23o0

50 pV I

13oo

2800

~

0100

50

pV I

3700

"\, " 1mm

J~ ~-'~ ~/~

4000 4300 4600

', ~

4000

"J~

51oo 200

pVL ~

150Hz, 15jum

100 me

Fig. 2. Line tracing of a transverse section through the medulla at level 2 (Fig. 1). In this preparation nerve eight was cut bilaterally. Here and in Fig. 14 the electrode track is indicated by a vertical line with crossbars marking recording depths, shown on the right. The intermediate dashed line points to the recording site where a lesion was made. Right: AEPs (n= 8) recorded in response to an ipsilateral 100 Hz vibrating sphere stimulus of 6 ~tm p-p displacement. Negativity of the brain electrode is down in all figures. The onset (phasic) response is a negativepositive peak followed by a tonic frequency following response which lasts for the whole stimulus duration. Here and in all other figures the stimulus marker (bottom) shows the voltage delivered to the Ling vibrator

Fig. 3. AEPs (n = 16) recorded in the ipsilateral MON. Note that the tonic response component follows and finally may double the stimulus frequency. Stimulus duration was 500 ms, with rise- and fall times of 100 ms, respectively. Stimulus amplitudes are given in the figure as p-p displacement

25, 50, and 100 Hz. Both sensitivity and dynamic intensity range varies widely between MON-units. For instance, at 100 Hz a unit may show saturation at 6 or at 130 gm p-p displacement. Responses to the 50 Hz and 25 Hz stimulus usually did not show saturation at the maximal possible stimulus amplitude which could be generated with the Ling vibrator (if the upper limit of the dynamic range due to technical reasons could not be determined this is indicated by a > sign). The attempt to determine the upper limit of the dynamic intensity range often required large stimulus amplitudes which may stimulate eighth nerve fibers. For this reason control experiments were done with animals whose eighth nerves were cut bilaterally. With respect to threshold, latency, and dynamic intensity range, these animals gave responses similar to those of intact preparations (see results marked by dashed lines in Figs. 6 and 7) thus making it unlikely that the responses recorded in intact preparations were contaminated by eighth nerve input.

For AEPs recorded in the M O N the dynamic intensity range at 100Hz was _>54dB, at 50Hz_>30dB and at 2 5 H z > 2 8 d B ( N = n = 2 ) . The corresponding values for unit responses varied from 22.4-59 dB, _>33 dB, and _>38.5 dB ( N = 5, n = 7). At a fixed stimulus amplitude (tested only for amplitudes larger than 100 gm p-p) responses to 100 Hz are stronger than to 50 Hz or 25 Hz (t-test, P<0.001) (Fig. 7). The latencies of ipsilateral AEP and unit responses significantly (t-test, P<0.01) decreased with increasing stimulus amplitude, e.g., at 100 Hz unit latencies varied between 11.8 ms_+ 2 (p-p displacement 137 gin) and 26.5 m s + 6 . 8 (3.6 gm) ( N = 6, n = 9). For a given stimulus amplitude latencies did not significantly increase with decreasing stimulus frequency, comparing 25, 50, and 100 Hz (t-test, P > 0.11). However, response latencies sharply increased with increasing stimulus rise time, e.g., at a rise time of 10 ms a unit may respond to a 100 Hz (p-p displacement 11 ~tm) stimu-

L a t e r a l line stimulus

>

H. Bleckmann et al. : Lateral line mechanoreception in elasmobranchs 4 ~Jm

7

20

34

6O

150

1 O0 ms

Fig. 4. Example of a single unit recorded in the ipsilateral M O N

in response to a vibrating sphere (50 Hz). Here and in Figs. 5, 8, and l l the stimulus amplitudes are given to the fight of each recording as p-p displacement. Note that at higher displacement amplitudes the unit is phase coupled to the stimulus

lus with a latency of 22 ms, while a rise time of 300 ms may cause a latency of 166 ms. M O N responses did not require a low stimulus repetition rate. Even at an interstimulus interval of I s responses still were 70.5% • 11 of those recorded at an interstimulus interval of 30 s ( N = n = 2).

Midbrain If the p L L N of Platyrhinoidis is stimulated electrically, local neural responses (AEPs and units) can be recorded from the dorsal nucleus of the mesencephalic nuclear complex (DMN), the anterior nucleus of the mesencephalic nuclear complex (AN), and the tectum mesencephali (Bleckmann et al. 1987). Neural responses can also be evoked within these areas with a vibrating sphere stimulus (Bleckmann et al. 1987). In the present study optimal stimulation sites for midbrain recordings often were just above the contralateral infraorbital lateral line canal. D M N and A N responses show displacement thresholds and frequency responses similar to those of MON-responses. Based on AEPs, the lowest p-p displacement thresholds of the mid-

463

brain D M N and A N were 0.005 ~tm (100 Hz) and 0.08 pm (150 Hz), respectively. At 25 Hz they were 2.3 ~tm and 8.3 ~tm, at 6.5 Hz 3.2 and 7.3 ~tm (Fig. 6). On either side of the 'best' frequency, displacement sensitivity of D M N responses on average decreased by 11.3 dB-+8.6 (6.5-100 Hz) and 1 8 . 4 d B + 9 . 1 (100-200Hz) per octave ( N = 3 , n > 12). The corresponding dB values for the A N are 9.0_+2.8 and 15.6-t-8.1, respectively ( N = 4 , n > 7). The 'tuning' of D M N (AEP and unit) responses was not significantly different from that of M O N responses (t-test, 0 . 4 3 < P < 0 . 5 4 ) . However, A N responses were significantly less sharply 'tuned' than M O N responses (t-test, 0 . 0 0 2 < P < 0.06). Depending on stimulus amplitude and recording site, D M N and A N responses may be phasic or phasic-tonic. In contrast to M O N units, D M N and A N units are not or only poorly phase coupled to the stimulus (Figs. 8, 9). The dynamic intensity ranges of D M N and A N responses are roughly similar to those of M O N responses, but A N may be significantly lower in range (Fig. 7). For AEPs recorded in the D M N (AN), the dynamic intensity ranges were >91.6 dB (31-46 dB) at 100 Hz, 56-65 dB (29 to > 3 2 dB) at 50 Hz, and _>52 dB ( > 2 0 dB) at 25 Hz. The corresponding values for units were 18.2-92dB (19.4 to > 6 0 d B ) , 19.1-66dB ( > 3 8 r i B ) , and 8.9 to >_52 dB (26 to > 35 dB). For a given stimulus amplitude DMN-responses on average do not decrease with decreasing stimulus frequency. At 50 Hz, the DMN-responses at the upper limit of the test range (150 ~tm p-p displacement) were equal ( 9 2 . 8 % + 2 8 ) to the maximal response to 100 Hz stimuli (130 ~tm p-p displacement). The 25 Hz stimulus at 105 ~m p-p even caused a response 138% _+62 of the 100 Hz maximum (t-test, P=0.012). The corresponding values for the A N are 91.5% _+22 (50 Hz) and 5 3 . 1 % _ 32 (25 Hz). In both the D M N and AN, response latencies for a given stimulus frequency and amplitude were significantly (t-test, P < 0 . 0 4 ) larger than for the M O N , determined in separate experiments. Response latencies increased (t-test, P < 0 . 0 0 3 ) with decreasing stimulus amplitude. To give numerical examples" the latencies of units recorded in the contralateral D M N increased from 22.6 ms_+4.8 (100 Hz, 130 gm p-p) to 38 ms_+4.7 (2.2 lam) ( N = 4, n = 9). The corresponding values for A N units are 2 3 . 3 m s + 1 . 2 and 61.5ms_+15 ( N = 4 , n = 8 ) , respectively. Stimulus rise time always strongly influenced the strength of neural responses. For instance, summing 16 trials at 100 Hz, 10 lain stimulus caused 338 spikes in D M N at a rise time of 10 ms but only 56 spikes at a rise time of 200 ms,

464

H. Bleckmann et al. : Lateral line mechanoreception in elasmobranchs 2.0 IJm

6.0

m

~

75

2 units [ _ _ 1 0 0 ms

A 19 IJm

~

[

a4

88

A

2 units

~

1 100 ms

13 Fig. 5A, B. PSTH (n - 8) of single units recorded in the ipsilateral M O N in response to a vibrating sphere stimulus. Stimulus frequencies are 100 Hz (A), and 50 Hz (B)

counting for the stimulus duration. A corresponding example for the A N is 150 vs 21 spikes. For full recovery, D M N and A N responses require a lower stimulus repetition rate than M O N responses. Response magnitude (amplitude of AEP and/or number of spikes evoked) at an interstimulus interval of 1 s was 40%_+24 (DMN), respectively 42% + 17 (AN) of that recorded at 30 s ( N = 3 or 4, 3
Diencephalon By directly shocking the pLLN, Bleckmann et al. (1987) have shown that in P. triseriata the diencephalic lateral tuberal nucleus (LTN) receives lateral line input. LTN-potentials evoked by a vibrating sphere are often, but not always, phasic, especially if low stimulus amplitudes are used. At higher stimulus intensities an additional tonic and/or offresponse component may become manifest (Fig. 11). In the case shown in Fig. 11 the nature of the response depended highly on stimulus repetition rate. If the interstimulus time interval was 55 s, both the tonic and the off-response components became manifest as soon as the p-p displacement of the stimulus surpassed 25 gm at 25 Hz, 12 gm at 50 Hz, and 10 ~tm at 100 Hz. At an interstimulus interval of 7 s the tonic response component appeared at 88 ~tm (50Hz) and 50 ~tm (100 Hz), respectively. At the 7 s repetition interval the 25 Hz stimulus caused no response at all, even if the stimulus amplitude was 150 gm p-p. For amplitudes >87 gm (stimulus frequency 50 and 100 Hz) a weak phase locking of the response could be observed.

H. Bteckmann et al. : Lateral line mechanoreception in elasmobranchs 200

200-

MON

100:

" ~ ~ . .

10

465

-

100

./S t

lo

1

.1

.1

.01

.01

200"

E

100

=L

v C

10

(]3

E (D 0

ffl a .01

"

i

i

=

~

i

i

~oo

.

i

J

i

i

200"~

100

10

/ .1

Te]enc.

LTN I

.01 0

I

g

l

l

l l 100

l

2;0

.

.

.

.

.

.

.

Fig. 6. Examples of threshold curves obtained by AEP ( o - - o ) , few unit ( A - - , ) , multi unit ( x - - x ), and single unit (ram--m) recordings from the ipsilateral MON, the contralateral D M N , AN, tectum, LTN, and the ipsilateral pallium (some of the curves shown for the D M N and A N have already been published; Bleckmann et at. 1987). Read omitted vertical axis label from the labeled vertical axis to the left. In all cases 8 or 16 responses were averaged. Experiments done with animals whose eighth nerve was cut bilaterally are shown here and in Fig. 7 by dashed lines

2;0

Frequency(Hz) The thresholds of L T N responses are highly dependent on stimulus repetition rate. At an interstimulus interval of 55 s minimal thresholds may be as low as 0.2 gm p-p displacement at 100 Hz, 1.9 gm at 50 Hz, and 11 pm at 6.5 Hz (Fig. 6). On either side of the 'best' frequency the sensitivity of L T N responses on average decreased by 6.9 dB_+ 1.4 (6.5-100 Hz) or 7.4 dB+_4.4 (100-200 Hz) per octave ( N = 3 , 4 5 5 dB at 100 Hz, 28.4-36.5 dB at 50 Hz, and > 16.4 dB at 25 Hz. The corresponding dynamic intensity ranges for unit responses were > 4 5 dB, _>38.7 dB, and _>7.8 dB (Fig. 7). Contrary to medullary responses, maximal LTN-re-

sponses often do not or only slightly decrease with decreasing stimulus frequency. At 50 Hz, the magnitude of LTN-responses at the upper limit of the test range (i.e., at 150 gm p-p) was 1 0 6 % + 4 3 of that caused by the 100 Hz stimulus at 130 gin. Even the 25 Hz stimulus at 105 gm caused a response which, with 67% 4-45, was not significantly (t-test, P = 0.284) different from that caused by the 100 Hz stimulus (Fig. 7). Peak latencies of AEPs recorded in the L T N were 4 2 m s _ 1 1 (100Hz, 137 gm p-p displacement), 48 m s + 1 9 . 8 (50 Hz, 150 gm), and 55 ms_+21.2 (25 Hz, 105 gm) ( N = 3 , n = 3). For a given stimulus frequency and amplitude, latencies were significantly (t-test, 0.001 < P < 0.049) larger than those determined for the M O N or D M N . In the LTN, response latencies did not significantly (t-test, P-> 0.17) increase with

466

H. Bleckmann et al. : Lateral line mechanoreception in elasmobranchs 100

100t ~

.

Hz

50

~

Hz

25

200

1501

?4' ....

//

(3.) 100

200~

n-

c~ F:

Hz

400

5~

o

1007

0

0

9 125

.

.

.

.

.

.

.

75

.

.

.

0

150 0

_

.

.

.

.

.

.

75

.

.

.

.

.

.

.

150 0

.

.

.

.

.

.

75

.

150

Fig. 7. Examples of intensity response curves obtained by AEP ( o - - o ) , few unit (A--A), multi unit ( x - - x ), and single unit ( s - - s ) recordings from the ipsilateral MON, the eontralateral DMN, AN, tectum, LTN, and the ipsilateral pallium. In all cases 8 or 16 responses were averaged. Read omitted vertical axis label from the first labeled vertical axis to the left. Note that the scale of the vertical axis may be different for different graphs

p-p Displacement (pro) decreasing stimulus amplitude, but this may be due to the small sample size. However, in the two preparations tested, an increase in rise time again led to a sharp decrease in response latency. For instance, at a rise time of 10 ms (stimulus frequency 100 Hz, p-p displacement 10.4 gm) the latency was 50 ms, while at a rise time of 300 ms the latency was 87 ms. In the same experiments the response amplitude decreased from 32 gV (rise time 10 ms) to 9.7 gV (rise time 300 ms). L T N responses did require a higher stimulus repetition interval than D M N and AN responses. Responses at an interstimulus interval of 1 s were only 9.8% ___8.9 of those recorded at an interstimulus interval of 30 s (N=2, n=4). The R F of a few unit L T N response was mapped in one case in 10 mm (rostro-caudal) and

5 mm (medio-lateral) steps. The RF of these units was large and complex and seemed to follow the course of the lateral line canals (Fig. 12). In some preparations and recording sites no obvious interactions between electro- and mechanoreceptive lateral line responses were seen (Fig. 13 A). However, in 3 out of 5 cases an electric field stimulus inhibited the response to a vibrating ball stimulus and vice versa. In the case shown in Fig. 13 B, the strength of inhibition depended on the polarity of the electric field.

Telencephalon In Platyrhinoidis a

telencephalic area which may correspond to the medial pallium, receives input from the mechanoreceptive lateral line nerve

H. Bleckmann et al. : Lateral line mechanoreception in elasmobranchs

467

0.3 p m

1,5

3.8 2 units 200

ms

14

Fig. 9. Few-unit response recorded in the contralateral A N during a vibrating ball stimulus (12.5 Hz, 105 gm p-p displacement). Note the weak phase coupling and that the phasic-tonic response is followed by a silent period

were made at 100 to 400 gm depth intervals. To find out whether visual and electric field stimuli are processed in different telencephalic areas, a stroboscopic light flash (delivered ipsi-, contra- or bilaterally), an electric field stimulus (50 ms duration, 7-14 gV/cm), and a 100 Hz vibrating ball stimulus of 30 gm p-p displacement (applied ipsior contralaterally) were given in succession. To minimize a possible interaction of neural responses caused by different stimulus modalities, stimuli were separated by 700 ms (light flash and electric field stimulus) or 500 ms (electric field and vibrating ball stimulus), respectively. One example of a vibrating ball response obtained through a vertical track through the telencephalon is shown in Fig. 14. In this 450 g animal the horizontal extension of the telencephalic lateral line area was about 0.8 mm in the medial-lateral and 0.8-1 mm in the anterior-posterior direction (Fig. 15). Good responses were found at a depth of about

105

~

2 units I 100 ms

Fig. 8. Few-unit responses recorded in the contralateral D M N during a 25 Hz vibrating ball stimulus. 16 responses were averaged

(Bleckmann et al. 1987). The same authors could also evoke telencephalic responses with stroboscopic light flashes or with weak uniform electric fields. To learn whether the telencephalon of Platyrhinoidis responds to a vibrating sphere stimulus as well, we mapped the forebrain at 400 gm mediolateral and rostro-caudal intervals. Recordings

i....7 i i

( I

"i\\ t

\

~ A ~

~

_

I

Fig. 10. Top view of P. triseriata showing the position of the vibrating sphere (dots). Irregular objects posteriorly are 'thorns' of the skin. Depending on the site of stimulation, the A N responses (n = 8), made in succession from the same recording site, differ with respect to the number of spikes evoked (expressed as PSTH shown to the right). RF was mapped using a stimulus displacement of 0.7 ~tm p-p. Dot size represents 0 8 spikes (o), 9-16 spikes (e), and 16-42 spikes (e) within the first 200 ms of stimulation. Responses _<8 spikes are believed to be not distinguishable from spontaneous activity. Good stimulation sites approximately follow the route of the lateral line canal system. Stimulus repetition rate in all cases was 0.01 Hz

468

H. Bleckmann et al. : Lateral line mechanoreceptiou in elasmobranchs

A

7~m

L

B

2,3

pm

2

6

1 unit

I 1 O0 ms

1 unit [ _ _ 100 m s

Fig. l l A , B. PSTH ( n = 8 ) of few-unit responses recorded in the contralateral LTN to a vibrating sphere stimulus. Stimulus frequencies are 50 Hz (A), and 100 Hz (B)

l lti /

",/

,,'"\

Fig. 12. Complex RF in the diencephalon. Top view of P. triseriata showing the positions of the vibrating sphere (dots). Depending on the site of stimulation, recordings, made in succession from the same recording site within the contralateral LTN, differ with respect to the number of spikes evoked (expressed as PSTH shown to the right). For the rostral stimulation area (borders of stimulation areas are drawn by dotted lines), R F was mapped using a stimulus frequency of 100 Hz and a stimulus of 2 ~m p-p displacement. For the more caudal area, R F was mapped using again a stimulus frequency of 100 Hz but a stimulus displacement of 7.5 Ixm p-p. Weaker stimuli failed to give any responses in this body area. Dot size represents 0-15 spikes (o), 16-30 spikes (e), and 30-65 spikes (O) within the first 150 ms of stimulation. Responses < 15 spikes are believed to be not distinguishable from spontaneous activity. Good stimulation sites follow the route of the lateral line canal system. All PSTH are based on the average of 8 responses

H. Bleckmann et al. : Lateral line mechanoreception in elasmobranchs

A

469

B ~Lhl LI~

L__

I|

I iIILII

II

I~lllt II I

I~l~l

EhLILLLL~L_

JJ

I

/ III li IIII II

Illil

Irll|~l~l~llld I IL I

IlilU~

!

/ 200 pv l _ _

loo.v 1 _ _

300 ms

300 ms

Electric field

Lateral line stimulus

Fig. 13. A From top to bottom: AEPs (n=4) recorded after stimulating the infraorbital lateral line canal with a vibrating ball (100 Hz, 2 gm p-p water displacement, 500 ms duration), an electric field stimulus delivered to the water bath (44 gV/cm, 300 ms duration), and both (lowest record). When both stimuli were applied they were given in succession, separated by 300 ms (see stimulus trace). Insets shown to the right: PSTHs (n = 8) of multiple unit responses following the same stimulus regime. Arrows point to artifacts due to the electric field stimulus. The two insets shown at the left are examples of multiple unit responses following a lateral line stimulus (top left) (100 Hz, 17.5 I~m p-p displacement, 500 ms duration), or an electric field (mid left) stimulus (44 gV/cm, 500 ms duration). Note that the high-amplitude lateral line stimulus in addition to the phasic response also causes a late response component. Time bars 300 ms, vertical bars 100 pV. B Example of interaction between the electroreceptive- and mechano-receptive lateral line. From top to bottom: AEPs (n = 8) recorded after stimulating the infraorbital latera! line canal with a vibrating ball (100 Hz, 30 gm p-p water displacement, 500 ms duration), an electric field stimulus delivered to the water bath (68 gV/cm, 300 ms duration), and both (lowest record). When both stimuli were applied they were given in succession, separated by 300 ms (see stimulus trace). Note that the lateral line response shows a decrease in amplitude (number of spikes) when both stimuli were given in succession. Insets: Corresponding PSTHs (n=8) (few-unit response). Arrows point to artifacts caused by the electric field stimulus. Time bar 300 ms, vertical bar 200 pV

2.8-3.2 ram. Despite intensive efforts, none of our preparations gave telencephalic responses to light flashes either ipsi- or contralateral and/or to weak electric field stimuli, although such responses have

Electric field

Lateral line stimulus

been found in earlier experiments (Bleckmann et al. 1987). Nevertheless in several instances we examined whether the lateral line responses were affected by visual and/or electric field stimuli, but this was not the case. Sometimes we encountered units which were spontaneously active. However, these units could never be driven by any of the stimulus modalities and stimulus regimes tested. Following a 100 Hz, 30 ~tm p-p displacement, stimulus, in one out of three hornsharks a weak negative-positive going AEP was recorded 3.7 mm below the surface of the posterior third of the telencephalon. This response had a peak latency of 188 ms and did not survive an interstimulus interval of less than 5 s. Hence we assume it to be of telencephalic origin. In Platyrhinoidis, telencephalic AEPs following a vibrating ball stimulus may differ with respect to the polarity of the response and the number and latency of the peaks. In different preparations, responses from 1 up to 4 prominent peaks were performed. Positive peaks were P78, P196, and P360, i.e., at the latencies indicated in ms. In another case the telencephalic response consisted of two distinct waves, P175 and P320. The amplitude of these two peaks could be varied independently by changing the position of the vibrating ball stimulus (Fig. 16). This indicates that telencephalic RFs, like A N and L T N RFs, can be complex, a

470

H. Bleckmann et al. : Lateral line mechanoreception in elasmobranchs t

/e

~

~

,

*

~

~

300 pm 600 900 1200 1500

-' "-~

r

1800 2100 2400

~

~

2700

F~-~"~'~-"~ ~ - - ~ , ~ - ~ ....... ~

3000 3300 3600 39oo

~

"

~

,

~ ~

.

J

-

~

,

~ /

~

~

4200

47004500 ~

4900 5100

\

5300 5700 500 ms

Fig. 14. Line tracing of a transverse section through the telencephalon at level 1 (Fig. 1). The intermediate dashed line points to the recording site where a lesion was made. Right: AEPs (n = 4) recorded in the ipsilateral telencephalon in response to a vibrating sphere (100 Hz, 30 pm p-p displacement) positioned 5 mm above the infraorbital lateral line canal. Note that prominent AEPs appear only at a depth of 4.2-5.1 mm

finding which was confirmed in other preparations as well. The latency of telencephalic responses did not decrease with decreasing stimulus frequency. In the two preparations tested, an increase in rise time led to a sharp increase in response latency. 3600

For instance, a rise time of 10 ms (100 Hz, 25 gm) caused a latency of 161 ms, while at a rise time of 300 ms the latency was 343 ms. In the same experiments the response amplitude decreased from 37 gV (rise time 10 ms) to 17 I~V (rise time 300 ms). Different peaks of a telencephalic response could differ with respect to their dynamic properties. In one animal the late wave P184 did not survive a repetition rate > 0.1 Hz, whereas the extinction of the early wave P66 required a repetition rate > 4 Hz. Telencephalic lateral line responses, like LTN responses, show a remarkable response decrement. Responses at an interstimulus interval of 5 s were only 23%_+22 of those recorded at an interval of 30 s ( N = 2, n = 3). Telencephalic responses are less sensitive than hindbrain and midbrain responses to our standard form of stimulation. Minimal thresholds were 2.1 gm (100 Hz), 4.0 gm (50 Hz), and 45 gm (6.5 Hz). On each side of the 'best' frequency the sensitivity decreased by 7.6 dB + 1.4 (6.5-100Hz) and 5.7dB_+2.3 (100-200Hz) per octave ( N = 4 , 4_ 3 1 . 4 d B (50Hz), and > 1 6 . 9 d B (25 Hz). Different parts of telencephalic responses could differ with respect to frequency sensitivity. For instance, the maximal first negative component of the response may decrease from 100% (130 pm, 100 Hz) to 33% (105 gm, 25 Hz) whereas

2400

~ ~ 4~ o o/ o/ / ~ ~

,ooo

"

50 pV [

~ 400

ms

Fig. 15. AEPs (n=4) recorded in the ipsilateral telencephalon in response to a vibrating sphere (100 Hz, 30 gm p-p displacement) positioned 5 mm above the infraorbital lateral line canal. The recording locations are shown on a drawing of the dorsal view of the telencephalon. The number with each record is the recording depth in gm, i.e. the depth at which a maximum response was recorded

H. Bleckmann et al. : Lateral line mechanoreception in elasmobranchs

471

Fig. 16. AEPs (n=4) recorded at a given site in the pallium of the contralateral telencephalon in response to a vibrating sphere (100 Hz, 30 gm p-p amplitude). Recording depth was 4.6 mm. Note that the AEPs consist of two long latency low frequency waves which independently can change size depending on the exact sphere location (black dots on fish surface)

the second component of the response may increase from 67% (130gin, 100Hz) to 115% (105 ~tm, 25 Hz). Other cases, where the first and second wave responded in just the opposite way, were also observed. The magnitude of telencephalic responses only slightly decreased with decreasing stimulus frequency for similar stimulus intensities, as in the LTN. On average, the 25 Hz stimulus at the maximal 105 ~tm p-p displacement caused a neural response which still was 66.5% +_42 of that observed at 100 Hz and 150 ~tm. In the test range 30-300 ms telencephalic responses were not systematically affected by stimulus duration. If the peak amplitude of the AEPs caused by a 30 ms stimulus is defined as 100%, the response magnitude at 300 ms is 116% ( N = 3 , n = 3 ) . Peak latencies were 181 m s + 4.4 at 3 0 m s stimulus duration compared to 191 ms_+18.3 at 300 ms (t-test, P=0.41). In some instances frequency and amplitude modulated wave stimuli (65-135 Hz within 500 ms) were presented. However, in no case did these more complex stimuli lead to better responses. Discussion

This study confirms by use of quasi-natural lateral line stimuli, that cartilaginous fishes have a well defined lemniscal pathway for lateral line information processing (see also Bleckmann et al. 1987). Areas which respond to a vibrating sphere stimulus are so far confined to the medullary M O N , the midbrain D M N , tectum, and AN, the diencephalic LTN, and an unidentified telencephalic area which

may correspond to the medial pallium (Bleckmann et al. 1987). In the same areas as well as in the diencephalic posterior central thalamic nucleus neural responses have been reported after a p L L N shock (Bleckmann et al. 1987). As far as this study has gone, submodalities or clear subsets of lateral line mechanosensory responses have not turned up. However, the possibility remains that a wider assortment of quasi-natural stimuli with combinations of frequencies and with amplitude modulation or motion, such as looming, might reveal specialized properties of higher lateral line centers. Nevertheless there are some properties in which successive lateral line centers differ.

Frequency and intensity functions Within the test range 6.5-200 Hz, all lateral line areas were sensitive to water displacements. Up to the level of the midbrain, the range of lowest displacement threshold is narrow. The lack of a bimodal distribution of best frequencies, in terms of displacement, discourages a recognition of different functional subsystems based on frequency. Forebrain lateral line areas are less sharply 'tuned' and are less sensitive to our standard test stimuli than medullary and midbrain lateral line areas. In addition forebrain areas often show a reduced dynamic range of intensity. Both the higher threshold and the reduced dynamic range could result from suboptimal stimulation. The neurophysiology of forebrain lateral line areas indicates that both diencephalon and telencephalon are less concerned than medullary and midbrain areas with high-fre-

472

H. Bleckmannet al. : Lateral line mechanoreceptionin elasmobranchs

quency sinusoidal stimuli. This may support Kalmijn's (1988) argument that the so called 'best' frequencies represent only the upper frequency limit of the lateral line system. Hence, low-frequency sinusoidal waves, generated - for instance - by a swimming fish, may be of larger behavioral significance than expected from hindbrain and midbrain recordings. Our results show that stimulus rise time is one of the most crucial parameters in determining the kind of midbrain and forebrain response. Irrespective of frequency, a stimulus of a given amplitude will cause a sharp increase in response latency and a decrease in response magnitude if the rise time is extended. Hence fast transients with high acceleration are the kind of hydrodynamic events well suited to stimulate midbrain and forebrain lateral line areas.

Temporal processing AEPs and units recorded within the M O N of Platyrhinoidis are tonic or phasic-tonic and follow a stimulus repetition rate up to at least one stimulus per second. In contrast, responses recorded from higher order nuclei are strikingly depressed at this repetition rate and at successively higher levels of the neuraxis the minimum time required to avoid response decrement during repetitive stimulation increases. This also holds true if the pLLN is stimulated electrically (Bleckmann et al. 1987). A striking refractoriness to repetitive or paired stimuli has also been observed in the higher levels of the electroreceptive lateral line of this species (Schweitzer 1983), the acoustic system of the carp (Echteler 1985a), and in the acoustic and lateral line system of the catfish (Finger and Bullock 1982). The circumscribed responsive areas in the diencephalon and telencephalon of Platyrhinoidis make these forebrain regions resemble the discrete sensory representations in mammals. However, an important difference remains. This is the apparent lack of pallial responses to quasi-natural lateral line stimuli presented at moderate repetition rates. Only further studies will show whether more biological stimuli, perhaps moving objects or spheres which vibrate with mixtures of frequencies, will elicit more robust responses. The possibility must be considered, however, that sensory pallial areas in fishes are more equivalent to secondary or tertiary sensory cortex than to primary sensory cortex in mammals.

Frequency encoding Tonic, phase locked responses to sinusoidal stimuli are commonly observed in primary lateral line af-

ferents (Caird 1978; Bleckmann and Topp 1981; Topp 1983; Mfinz 1985; Elepfandt and Wiedemer 1987). Tonic, phase locked responses could also be recorded from the MON, and, in rare instances and with weak phase locking, from the D M N or A N of Platyrhinoidis. Midbrain and forebrain lateral line responses are in general purely phasic. This suggests that stimulus frequency is less faithfully preserved in the temporal discharge pattern of neurons at higher brain levels, a finding valid for the acoustic system of the carp as well (Echteler 1985a). One means by which frequency information is encoded is by spatial segregation of neurons within a given nucleus according to their preferred frequencies. This neuronal arrangement, known as tonotopy, is commonly observed in the acoustic system of amphibians and terrestrial vertebrates (e.g., Pettigrew et al. 1981; Fuzessery and Feng 1981 ; Manley 1971 ; Scheich et al. 1979; Merzenich and Reid 1974). A tonotopic organization may also be present within the acoustic medial torus semicircularis of the carp (Echteler 1985 b) but has not been reported for the mechanoreceptive lateral line. In this study the threshold curves of units recorded at different locations in the hindbrain, midbrain- or forebrain were surprisingly similar in shape. Consequently a tonotopic organization of higher order lateral line nuclei was not observed.

Somatotopy Representation of sensory space in somatotopic or computed maps is commonly observed in sensory systems (e.g., Heiligenberg 1988). However, it is still not clear whether the preservation of spatial information is an important aspect of lateral line processing. In contrast to the electroreceptive lateral line (e.g., Carr et al. 1981), no simple somatotopic representation has been recorded from secondary medullary lateral line neurons. However, the defined - although fractured - mesencephalic, diencephalic, and telencephalic RFs found in the present study indicate that at least some spatial information is preserved up to the highest brain levels (see also Bleckmann et al. 1987). Complex somatotopic maps, determined with aid of falling water drops, have also been found in the cerebellum of Platyrhinoidis (Fiebig 1988).

Multimodality The elasmobranch lateral mesencephalic complex receives not only lateral line, but also auditory and electroreceptive input (Boord and Northcutt 1982;

H. Bleckmann et al. : Lateral line mechanoreception in elasmobranchs

Corwin and Northcutt 1982; Bodznick and Northcutt 1984; Schweitzer 1983, 1986). The organization of the mesencephalic nuclear complex is similar to that of the torus semicircularis, its possible homolog in electroreceptive teleosts (Northcutt 1978), in that different modalities maintain largely separate representations in distinct nuclei (Knudsen 1976, 1977, 1978; Bullock 1981; Cart etal. 1981; Schweitzer 1986). Despite this segregation of input, some multimodal units responding to certain combinations of tactile, lateral line, auditory, or visual inputs have been found in the torus and tectum of teleosts (Knudsen 1977; Nederstigt and Schellart 1986) and aquatic amphibians (Zittlau et al. 1985) as well as in the AN of Platyrhinoidis (Schweitzer 1986; Lowe 1987). Likewise, bimodal units responding to electric field and lateral line stimuli were found in the present study in the LTN. Units were either 'electric-field-depressed mechanosensory units' or 'lateral-line-depressed electrosensory units'. More refined stimulus devices may reveal more sophisticated types of muttimodal interaction not only at the level of the diencephalon but also at the level of the telencephalon. Acknowledgements. Supported by grants to H.B. from DFG (B1 242 1-2, 2-1), to T.H.B. from NSF and NIH and to O.W. from Education Abroad Program of UCSD and University of G6ttingen. We thank Dr. R.G. Northcutt for his generous offer to use some of his sharks for the physiological experiments. Drs. E. Fiebig and R.G. Northcutt in addition deserve thank for their help in identifying the brain areas in which diencephalic and telencephalic lesions were recovered.

References Alnaes E (1973) Unit activity of ganglionic and medullary second order neurons in the eel lateral line system. Acta Physiol Scand 88:49 61 Bleckmann H (1986) Role of the lateral line in fish behavior. In: Pitcher TJ (ed) The behaviour of teleost fishes. Croom Helm, London Sydney, pp 177 202 Bleckmann H, Bullock TH (in press) Central physiology of the lateral line system, with special reference to elasmobranchs. In: Coombs S, G6rner P, Mfinz H (eds) Neurobiology and evolution of the lateral line system. Springer, New York Berlin Tokyo Bleckmann H, Bullock TH, Jorgensen JM (1987) The lateral line mechanoreceptive mesencephalic, diencephalic, and telencephalic regions in the thornback ray, Platyrhinoidis triseriata (Elasmobranchii). J Comp Physiol A 161:67 84 Bleckmann H, Topp G (1981) Surface wave sensitivity of the lateral line system of the topminnow Aploeheilus lineatus. Naturwissenschaften 68 : 62~625 Bodznick DA, Northcutt RG (1980) Segregation of electroand mechanoreceptive inputs to the elasmobranch medulla. Brain Res 195:313-322 Bodznick DA, Northcutt RG (1984) An electrosensory area in the telencephalon of the little skate, Raja erinacea. Brain Res 298 : 117-124 Boord RL, Northcutt RG (1982) Ascending lateral line path-

473

ways to the midbrain of the clearnose skate Raja eglanteria. J Comp Neurol 207:274--282 Bullock TH (1981) Comparisons of the electric and acoustic senses and their central processing. In: Tavolga WN, Popper AN, Fay RR (eds) Hearing and sound communication in fishes. Springer, Berlin Heidelberg New York, pp 525 572 Bullock TH (1986) Interspecific comparison of brainstem auditory evoked potentials and frequency following responses among vertebrate classes. In: Cracco RQ, Bodis-Wollner I (eds) Evoked potentials. Alan R Liss, New York, pp 155164 Caird DM (1978) A simple cerebellar system: the lateral line lobe of the goldfish. J Comp Physiol 127:61-74 Callens M, Vandenbusche E, Greenway Ph (1967) Convergence of retinal and lateral line stimulation on tectum opticum and cerebellar neurones. Arch Int Physiol Biochem 75 : 148-150 Carr CE, Maler L, Heiligenberg W, Sas E (1981) Laminar organization of the afferent and efferent system of the torus semicircularis of gymnotid fish: morphological substrates for parallel processing in the electrosensory system. J Comp Neurol 203 : 649-670 Corwin JT, Northcutt RG (1982) Auditory centers in the elasmobranch brain: deoxyglucose localization and evoked potential recording. Brain Res 236: 261-273 Dijkgraaf S (1963) The functioning and significance of the lateral line organs. Biol Rev 38:51-106 Dowben RM, Rose JE (1953) A metal-filled microelectrode. Science 118 : 22 Echteler SM (1985 a) Organization of central auditory pathways in a teleost fish, Cyprinus carpio. J Comp Physiol A 156:267-280 Eehteler SM (1985b) Tonotopic organization in the midbrain of a teleost fish. Brain Res 338:387-391 Elepfandt A, Wiedemer L (1987) Lateral-line response to water surface waves in the clawed frog, Xenopus laevis. J Comp Physiol A 160:667-682 Fiebig E (1988) Connections of the corpus cerebelli in the thornback guitarfish, Platyrhinoidis triseriata (Elasmobranchii). Study with WGA-HRP and extracellular granule cell recording. J Comp Neurol 268 : 567 583 Finger TE, Bullock TH (1982) Thalamic center for the lateral line system in the catfish Ictalurus nebulosus: evoked potential evidence. J Neurobiol 13:39-47 Fuzessery ZM, Feng AS (198/) Frequency representation in the dorsal medullary nucleus of the leopard frog, Rana p. pipiens. J Comp Physiol 143:339-347 Gray SJ (1984) Interaction of sound pressure and particle acceleration in the excitation of the lateral-line neuromasts of sprats. Proc R Soc Lond B 220:299-325 Harris G, Bergeijk A van (1962) Evidence that the lateral-line organ responds to near-field displacements of sound sources in water. J Acoust Soc Am 34:1831-1841 Heiligenberg W (1988) Electrosensory maps form a substrate for the distributed and parallel control of behavioral responses in weakly electric fish. Brain Behav Evol 31: 6-16 Hoagland H (1935) Electrical responses from the lateral line nerves of fishes. V. Responses in the central nervous system. J Gen Physiol 18:89-91 Hoin-Radkovski I, Bleckmann H, Schwartz E (1984) Determination of source distance in the surface-feeding fish Pantodon buchholzi (Pantodontidae). Anim Behav 32: 840-851 Kalmijn A (1988) Hydrodynamic and acoustic field detection in elasmobranch and teleost fishes. In: Atema J, Fay RR,

474

H. Bleckmann et al. : Lateral line mechanoreception in elasmobranchs

Popper A, Tavolga W (eds) Sensory biology of aquatic animals. Springer, New York Berlin Tokyo, pp 83-130 Knudsen EI (1976) Midbrain responses to electroreceptive input in catfish. Evidence of orientation preferences and somatotopic organization. J Comp Physiol 106:51-67 Knudsen EI (1977) Distinct auditory and lateral line nuclei in the midbrain of catfishes. J Comp Neurol 173:417-432 Knudsen EI (1978) Functional organization in electroreceptive midbrain of the catfish. J Neurophysiol 41:350-364 Kuiper JW (1967) Frequency characteristics and functional significance of the lateral line organ. In: Cahn Phil (ed) Lateral line detectors. Indiana University Press, Bloomington London, pp 105-121 Lowe DA (1987) Single unit study of lateral line cells in the optic tectum of Xenopus laevis: evidence for bimodal lateral line/optic units. J Comp Neurol 257 : 396404 Manley JA (1971) Single unit studies in the midbrain auditory area of caiman. Z Vergl Physiol 71 : 255 261 Merzenich MM, Reid MD (1974) Representation of the cochlea within the inferior colliculus of the cat. Brain Res 77: 397-415 Mfinz H (1985) Single unit activity in the peripheral lateral line system of the cichlid fish Sarotherodon nilotieus L. J Comp Physiol A 157:555-568 Nederstigt LJA, Schellart NAM (1986) Aconsticolateral processing in the torus semicircularis of the trout Salmo gairdneri. Pfliigers Arch 406:151-157 Northcutt RG (1978) Brain organization in the cartilaginous fishes. In: Hodgson ES, Mathewson RF (eds) Sensory biology of sharks, skates, and rays. Office of Naval Research, Arlington, Va, pp 117-193 Paul DH, Roberts BL (1977a) Studies on a primitive cerebellar cortex. II. The projection of the posterior lateral line nerve to the lateral line lobes of the dogfish brain. Proc R Soc Lond B 195:467-478 Paul DH, Roberts BL (1977b) Studies on a primitive cerebellar cortex. II1. The projection of the anterior lateral line nerve

to the lateral line lobes of the dogfish brain. Proc R Soc Lond B 195 :47%496 Pettigrew AG, Anson M, Chung SH (1981) Hearing in the frog: a neurophysiological study of the auditory response in the midbrain. Proc R Soc Lond B 212:433457 Platt CJ, Bullock TH, Cz6h G, Kova~evi6 N, Konjevi6 Dj, Gojkovi6 M (1974) Comparison of electroreceptor, mechanoreceptor, and optic evoked potentials in the brain of some rays and sharks. J Comp Physiol 95:323-355 Roberts BL (1981) Central processing of acousticolateralis signals in elasmobranchs. In: Tavolga WN, Popper AN, Fay RR (eds) Hearing and sound communication in fishes. Springer, Berlin Heidelberg New York, pp 357-372 Sand O (1981) The lateral line and sound reception. In: Tavolga WN, Popper AN, Fay RR (eds) Hearing and sound communication in fishes. Springer, Berlin Heidelberg New York, pp 459-480 Scheich H, Bonke BA, Langner G (1979) Functional organization of some auditory nuclei in the Guinea fowl demonstrated by the 2-deoxyglucose technique. Cell Tissue Res 204:17027 Schweitzer J (1983) The physiological and anatomical localization of two electroreceptive diencephalic nuclei in the thornback ray, Platyrhinoidis triseriata. J Comp Physiol 153:331-341 Schweitzer J (1986) Functional organization of the electroreceptive midbrain in an elasmobranch (Platyrhinoidis triseriata): a single unit study. J Comp Physiol A 158:43-58 Topp G (1983) Primary lateral line responses to water surface waves in the topminnow A. lineatus (Pisces, Cyprinodontidae). Pfl/igers Arch 397:62-67 Zittlau KE, Claas B, M~nz H, G6rner P (1985) Multisensory interaction in the torus semicircularis of the clawed toad Xenopus laevis. Neurosci Lett 60:77-81 Zittlau KE, Claas B, Mfinz H (1986) Directional sensitivity of lateral line units in the clawed toad Xenopus laevis Daudin. J Comp Physiol A 158:469477