The electroreceptors of elasmobranchs are strongly modulated by the fish’s own ventilation but this source of potential interference is suppressed within the medulla. The mechanism for the suppression is thought to be based on the common mode nature of the ventilatory noise, i.e. it is of the same amplitude and phase for all of the electroreceptors, compared with environmental electric fields which affect the receptors differentially. Evidence for the common mode suppression hypothesis is provided here in skates by the observation that the response to an artificial common mode stimulus that is independent of ventilation and delivered through an electrode inserted into the animal’s gut is also suppressed by the medullary neurons; the extent to which a particular neuron suppresses the responses to the gut stimulus and to ventilation is similar. In addition, a potential modulation of 5–150 μV is measured between the skate’s interior and the sea water during ventilation and this appears to be responsible for the self-stimulation. By passing d.c. or sinusoidal currents through the gut electrode it is demonstrated that this ventilatory potential is due to the variable shunting of a standing d.c. potential across the fish’s skin by the opening and closing of the mouth and gill slits during ventilation. Osmoregulatory ion-pumping appears to contribute to the production of the d.c. potential.
Elasmobranch fishes possess an acutely sensitive electrosensory system with electoreceptor organs known as ampullae of Lorenzini. These ampullary organs are arranged in a series of 4–6 clusters on each side of the body. At each cluster, the receptor cells are located in subdermal alveoli at the blind endings of canals that radiate outwards and open to the exterior via skin pores. The receptor epithelium and the canal walls have a very high resistance in comparison with the conductivity along the canals themselves. With this arrangement, afferent firing rate effectively measures changes in potential difference between the sea water at the skin opening and the basal surface of the receptor cells in the interior of the animal. The best-documented behavioral role of the elasmobranch electroreceptors is in prey detection, where the system is used to detect the d.c. and low-frequency bioelectric fields of other animals in the aquatic environment. However, it is also possible that uniform electric fields such as those induced by the motion of the animal itself or of ocean currents within the earth’s magnetic field provide useful information for orientation and navigation. The anatomy, physiology and behavioral uses of the elasmobranch electrosense are reviewed in Bodznick and Boord (1986), Kalmijn (1988) and Montgomery (1988).
The d.c. bioelectric fields of prey animals appear to result from differences in the electrical properties among the various body surfaces the animals have in contact with the surrounding water (Kalmijn, 1988) and in many cases probably include a major contribution from the potentials produced across ion-exchange surfaces. In fishes these d.c. potentials are thought to be modulated at low frequencies by changes in resistance along the current pathway caused by the opening and closing of the mouth and gill slits during ventilation. One problem inherent in the electrosensory system is that the animal must detect extremely weak bioelectric fields produced by other animals on top of its own very similar bioelectric fields. The electric fields produced by elasmobranchs are typically of lower amplitude than those produced by bony fishes (Kalmijn, 1974) but, even so, recordings from primary afferent fibers of the electrosensory system in the thornback ray, Platyrhinoidis triseriata, show afferent input to be strongly modulated by the animal’s own ventilation (Montgomery, 1984). An interesting observation is that, in the secondary neurons of the electrosensory system in Platyrhinoidis, this ventilatory modulation is virtually absent. The animals are able to distinguish between afferent activity due to their own ventilation (reafference) and that due to extrinsic fields and they can effectively suppress the ventilatory reafference in the brain.
What is the mechanism for this noise suppression? Montgomery (1984) showed that ventilatory modulation is of similar amplitude and phase in all afferent fibers from one ampullary cluster; in other words, ventilatory modulation is common mode. Afferents from canals of opposite orientation respond differentially to extrinsic fields, but show a very similar response to ventilation. This indicates that the ventilatory reafference is primarily driven by a potential modulation between the animal’s interior and the surrounding water rather than by the small potential differences measured in the water around the fish during ventilation. These observations support an earlier suggestion by Kalmijn (1974) that the grouping of ampullae into clusters sharing a common internal reference may permit the elimination of unwanted noise by a mechanism of common mode suppression. New and Bodznick (1990) have recently shown that suppression of ventilatory reafference is also present in the little skate (Raja erinacea), though apparently to a lesser degree than that shown in Platyrhinoidis. They found that in Raja afferent modulation during ventilation is effectively common mode within all the electro-sensory afferents, even those from opposite sides of the body, and direct evidence was obtained for a contribution of contralateral input to the noise suppression mechanism.
The demonstration that ventilatory modulation is common mode shows that noise suppression could operate via a common mode suppression mechanism, and this is further supported by the importance of the contralateral input for the suppression (New and Bodznick, 1990). However, other possibilities exist. For example, in the weakly electric fish Gnathonemuspetersii, ampullary electroreceptors respond to the animal’s own electric organ discharge. This unwanted reafference is removed by an elegant mechanism of modifiable efference copy (for a review, see Bell, 1986). Although no evidence was found for this mechanism in Raja (New and Bodznick, 1990), one aim of the present study was to provide a more direct test of the common mode suppression hypothesis versus other mechanisms by determining whether an experimentally imposed common mode electrosensory input, independent of ventilation, would also be suppressed within the central nervous system. In addition, we provide new information on the characteristics and origins of the bioelectric fields associated with ventilation in elasmobranchs.
MATERIALS AND METHODS
Experiments were performed on 20 specimens of the little skate Raja erinacea Mitchill at the Marine Biological Laboratory in Woods Hole, MA. Animals were caught in short trawl tows in Vineyard Sound and kept in cooled sea water until their return to the laboratory, where they were maintained in holding tanks at a temperature of 14°C. For experimentation, the animals were anesthetized by immersion in tricaine methanesulfonate (approximately 0.02%), the cranium was opened to expose the brain, which was then decerebrated by a transection at the optic chiasm. Some animals were then paralyzed by intravenous injection of tubocurarine chloride (3 mg kg−1). In others, the spinal cord was transected about 1 cm behind the brain to permit normal ventilatory movements while immobilizing the animal’s trunk and tail. A salt-bridge electrode of PE90 tubing filled with 1.5% agar in sea water was inserted into the gut via the anus. An additional Ag/AgCl electrode, made from 0.2mm diameter silver wire coated with Teflon except near the tip, was implanted through a small skin incision into the interior of the animal in the region between the hyoid and buccal ampullary clusters on the head. The skin incision was then sealed with tissue adhesive (Histoacryl, Trihawk). The animals were positioned on a Plexiglas head holder to stabilize the brain for microelectrode recording. A Plexiglas plate inserted through the mouth provided support for the cranium, but also meant that the mouth was held partly open. In the case of paralyzed animals, a stream of oxygenated sea water directed into the mouth provided for respiration. The rays were immersed in sea water up to the level of the cranial opening, and the water temperature in the experimental bath was regulated at 8–10°C. These procedures followed NIH guidelines for care and use of experimental animals and were approved by the Institutional Animal Care and Use Committees of the Marine Biological Laboratory and Wesleyan University.
Electrosensory afferent activity was recorded with glass micropipettes (4 mol l−1 NaCl; 25 MΩ) in the anterior lateral line nerve within the cranium. Platinum-black-tipped indium electrodes (2–5 μm tip diameter, 2–7 MΩ) were used to make recordings from neurons within the dorsal octavolateralis nucleus, where ascending efferent neurons (AENs), the principal output neurons, were identified by antidromic stimulation from the lateral mesencephalic nucleus. Single unit activity was recorded during ventilation, during stimulation from the gut electrode and during stimulation by uniform extrinsic electric fields. Spike data were analyzed from poststimulus time histograms of 25–50 ventilatory cycles or stimulus presentations. The response was measured as the peak-to-peak change in firing rate produced by the stimulus or, in the case of a zero-firing rate nonlinearity, the increase above spontaneous firing rate was measured and this valued was doubled to obtain the effective peak-to-peak modulation.
Ventilation was monitored with the piezoelectric crystal of a phonograph cartridge coupled to movements of the branchial chamber. This signal was used to trigger the histogram program at the beginning of expiration. The stimulus delivered through the gut electrode was a continuous 1 or 2 Hz sine wave centered about zero and applied between the gut electrode and salt bridges (1.5% agar in sea water) located along all four sides of the experimental aquarium. The amplitude was adjusted to give a 20 μV peak-to-peak signal measured between the interior of the animal and an indifferent electrode placed in the seawater bath. These amplitude and frequency values were chosen to approximate normal ventilatory potential modulations recorded from skates. Longitudinal and transverse uniform fields were applied through salt bridges as a continuous 2 Hz sine wave, typically at a peak-to-peak amplitude of 2 μV cm−1. This intensity is within the linear range of the intensity response functions of afferents and dorsal nucleus neurons. The response of an electrosensory unit to a uniform field depends on the orientation of its canal with respect to the field. For this reason, the overall response of units to uniform fields was taken as the vector addition of their responses to the two fields presented at right angles. The receptive field of units was located by a small (5 mm) roving dipole electrode.
Potential modulations measured between the Ag/AgCl electrodes inside the body and those in the water during ventilation were amplified by a PAR model 113 d.c. differential preamplifier with half-amplitude low-frequency cut-off at 0.03 Hz. In four animals, this so-called ventilatory potential was monitored after the intravenous injection of 1–4 mmol of NaCl (1 or 2 ml injections into the caudal vein) to salt-load the animal. This was done in order to stimulate increased osmoregulatory ion transport and thereby to assess its contribution to the potential. In two other cases, the internal electrode was implanted in an otherwise intact animal to record the size of the ventilatory potential and to monitor changes induced by placing the animal in diluted sea water.
With a Ag/AgCI electrode inserted beneath the skin of a skate, an electrical potential modulation is measured between the fish’s interior and the surrounding sea water coincident with ventilatory movements. This ventilatory potential lasts 1–3 s but can vary greatly in both waveform and amplitude among fish (Fig. 1A) and in a single fish at different times. Most often it takes the form of a biphasic wave, which is inside-negative during expiration and then inside-positive during inspiration, but in other cases the biphasic potential is reversed in polarity. More rarely the ventilatory potential is monophasic, triphasic or even more complex.
In two unrestrained and unoperated animals the ventilatory potential amplitude was 5–27 μV. The ventilatory potential in operated animals set up for physiological recordings ranged in peak-to-peak amplitude from less than 5 μV to more than 150 μV, but was usually 15–25 μV. Gradual changes in amplitude were often seen in the hour immediately after surgery. Cycle by cycle changes in the intensity of the normal breathing movements affected the ventilatory potential amplitude (see bottom record, Fig. 1A), and occasional ‘coughs’, when water was vigorously ejected through the spiracle, generally caused brief, large inside-positive ventilatory potentials.
The pattern of firing of afferent electrosensory fibers during ventilation was as one would predict given their known excitatory response to a cathodal stimulus located at the skin pore. That is, when the internal potential became positive relative to the sea water, the firing rate in the afferents was accelerated; negative inside potential suppressed afferent firing (Fig. 1A). As noted previously (Montgomery, 1984; New and Bodznick, 1990), this is true regardless of the location of the skin pore of a particular receptor on the body surface. The degree of afferent modulation was also directly proportional to the amplitude of the co-occurring ventilatory potential (Fig. IB).
The ventilatory potential was investigated experimentally. Placing a skate in 70% sea water resulted in a threefold increase in the ventilatory potential over 30 min; it returned to normal soon after the fish was returned to full sea water. Salt-loading animals by intravenous injection of 1–4 mmol of NaCl in order to stimulate enhanced osmoregulatory ion transport caused a relatively small or no immediate change in the potential, but after a lag of 5–20 h the potential increased dramatically to a peak of 230–350 μV and then returned to more normal levels after 72–96 h (Fig. 2).
The origin of the ventilatory potential was studied further with the use of a seawater-agar bridge electrode inserted through the anus into the fish’s gut. Current from a constant-current isolator, passed between the gut electrode and salt-bridge electrodes on the four sides of the tank, created an electrical potential between the inside of the fish near the ampullary clusters in the head and the surrounding sea water that, as detailed below, appeared to be nearly the same over all areas of the skin surface that were under water. A 3–15 Hz sinusoidal potential produced by a current stimulus (200–300 μA) through the gut electrode and monitored between the internal Ag/AgCl electrode and the indifferent electrode in the sea water showed an amplitude modulation during ventilation due to apparent changes in the resistive pathway between the animal’s interior and the sea water outside. The peak-to-peak amplitude of the potential created across the body surface decreased during expiration and increased during inspiration (Fig. 3A), coinciding with the separate phases of the usual bipolar ventilatory potential. In cases where the waveform of the ventilatory potential was nearly monophasic, the resistance changes measured in this way were similarly unidirectional and coincident with the ventilatory potential (Fig. 3C).
In animals in which the normal ventilatory potential was otherwise quite small, a d.c. potential created by passing d.c. current of 100–250 μ A through the gut electrode resulted in a large normal-looking ventilatory potential modulation measured with the subdermal Ag/AgCl electrode (Fig. 3B). Reversing the polarity of the gut current reversed the polarity of each phase of the ventilatory potential. In animals with larger ventilatory potentials, a d.c. gut stimulus of the correct polarity and amplitude canceled and even reversed the polarity of the normal ventilatory potential (Fig. 3C). These results support the hypothesis that the ventilatory potential is a modulation of a standing d.c. potential that exists across the fish’s body surface. The d.c. gut stimulus required to cancel the normal ventilatory potential provides an estimate of this standing d.c. potential. The d.c. potential change created by a gut stimulus just sufficient to suppress the normal ventilatory potential ranged from 20–450 μV, and in almost all cases the required polarity of the gut stimulus was negative. In two animals exhibiting a ventilatory potential of reversed phase, i.e. inside-positive on expiration, negative on inspiration, a positive d.c. gut stimulus suppressed the ventilatory potential and a negative potential enhanced it.
Central suppression of common mode signals
The gut electrode was used to create a common mode stimulus, that is one of nearly the same amplitude and phase, for all of the electroreceptors as a means of testing the hypothesis of a central common mode rejection mechanism for suppressing ventilatory reafference. The potentials created by the gut electrode stimulus were measured with the subdermal Ag/AgCl electrode in the region of the ampullary clusters and an indifferent electrode, which was moved to a series of locations around the skate, in the bath. The records in Fig. 4 show that the gut electrode produced a good common mode stimulus with a nearly identical potential modulation for virtually all of the electrosensory canals with skin pores beneath the water surface, though the field in the region of the most caudal hyoid pores in some animals was slightly phase-shifted and smaller in amplitude than at other locations (Fig. 4, position E). The significant exception to this was found in the case of the dorsal and medial hyoid pore group. The skin pores of these receptors were located above the water surface in the experimental situation, and the potential on the surface of the skin in this region was phase-reversed in comparison with the potentials recorded in the sea water.
The common mode nature of the gut stimulus was confirmed in recordings from more than 200 primary electroreceptive afferents. Afferents from canals of different orientations and locations on the body responded differentially to uniform electric field stimuli, but gave nearly identical responses to the gut stimulus (Fig. 5). The ampullary organs of skates in a uniform electric field are sensitive to the potential drop across the skin plus the drop within the animal’s body along the length of the ampullary canal. Thus, as illustrated in Fig. 5, afferents from electroreceptor organs with longer canals appear to be more sensitive to uniform fields than those from organs with shorter canals. However, the amplitude of the receptors’ responses to the gut stimulus was independent of canal length, indicating that the skate’s interior, at least in the region of the ampullary clusters in the head, is nearly isopotential with respect to the gut stimulus and that the effective voltage drop is across the skin. The amplitude of the response to the gut stimulus was also similar to the afferent modulation produced by a ventilatory potential of the same magnitude (Fig. 1B). The common mode nature of the response to the gut stimulus was evident in all afferents with a receptive field located in the water. In several instances afferents were found whose receptive fields could not be located. These were presumed to be in the dorsomedial hyoid group and their responses were phase-reversed, consistent with the polarity of the field produced by the gut stimulus in that location.
The responses of 62 AENs and 85 primary afferents to the gut stimulus and to uniform electric fields in the water were compared. A signal to noise ratio (S/N) for each electroreceptive unit studied was defined as the overall response of the unit to uniform fields of 2 μV cm−1 (signal) divided by its response to the 20 μV gut stimulus (noise). The S/N in primary electroreceptor afferents was relatively homogeneous and averaged 0.96 (S.D. 0.7, N=85) (Fig. 6). In contrast, the S/N for AENs was quite variable, even within a single animal. Approximately one-third had S/N values similar to that of primary afferents and the remaining two-thirds were larger, indicating suppression of the response of these AENs to the gut stimulus compared with that to extrinsic fields (Figs 6, 7). The mean S/N for all AENs recorded was 4.6 (S.D. 9.1, N=62), a value 4.8 times higher than the average for primary afferents.
A subset of cells in Fig. 6 was recorded from five animals that were not paralyzed by curare injection but instead had spinal transections and were ventilating naturally. For both primary afferents and AENs in these animals the ventilatory modulation was also measured and an S/N value was determined where the signal was again the response to uniform fields (2 Hz sinusoidal stimuli; 2 μVcm−1) but the noise was the peak-to-peak ventilatory modulation. The uniform field stimuli were presented without a fixed phase relationship to the ventilatory cycle, which had a variable 1.5–3 s period. For primary afferents the mean S/N was 1.2 (S.D. 0.7, N=47) and, as with the gut stimulus, the S/N for AENs was quite variable but averaged 4.8 (S.D. 6.2, N=28), four times higher than for the afferents (Fig. 8). For these AENs the extent to which their responses to the gut stimulus and to the ventilatory potential were suppressed compared to uniform fields was positively correlated (Fig. 9, r=0.66).
The feeble, low-frequency electrical potentials recorded in the water near fishes during their ventilatory movements were among the first bioelectric fields identified as natural stimuli for the electric sense of elasmobranchs (Kalmijn, 1974). Such electric fields guide the well-aimed approach of elasmobranch predators to their concealed prey (Kalmijn, 1971). Through physiological recordings from the electroreceptors of freely ventilating elasmobranchs it later became evident that potentials associated with the animal’s own ventilation were also a potent source of electrosensory self-stimulation (Montgomery, 1984). From the common mode nature of the ventilatory reafference, it was inferred that the effective ventilatory potential responsible for self-stimulation was that existing between the animal’s interior and the surrounding water rather than the weak potential distribution measured in the sea water. In the current study we have directly measured a ventilatory potential modulation between the animal’s interior and the sea water with a Ag/AgCl electrode placed beneath the skin of the head, and it is evident from the correlation of electroreceptor firing with the waveform and amplitude of the potential that it is responsible for the ventilatory reafference. Although in our physiological set-up the animal was restrained by a head clamp and a mouthpiece that holds the mouth partly open, the ventilatory potential is not simply an artifact of this arrangement as it was recorded with a similar waveform and amplitude in two unoperated, unrestrained animals breathing naturally. Control measures made while probing the body surface over the recording electrode demonstrated that the ventilatory potential is also not a movement artifact.
The ventilatory potential is probably due to modulation of a standing d.c. potential between the animal’s interior and the water by the variable shunting of the potential across the gills during ventilation. This model is supported by experiments in which a current was passed between the gut electrode and the sea water. Altering the standing d.c. potential by passing a d.c. gut current in this manner predictably altered the ventilatory potential modulation. Thus, a sizable ventilatory potential of the usual waveform can be created in an animal in which none is otherwise apparent solely by changing the d.c. potential, and in other cases a normal ventilatory potential can be canceled and reversed with the appropriate polarity of d.c. gut stimulus. It is difficult to measure the actual size of the small d.c. potential between the fish’s interior and the water because of the unknown contribution to the measurements of the galvanic half-cell potential of the implanted electrode. Measurements of the abrupt change in potential at the onset of the d.c. gut current required to cancel exactly the normal ventilatory potential modulation do permit a measure of this d.c. potential, which in most cases turned out to be 200–300 μV but ranged from near 0 to 450 μV. This method may underestimate the actual potential by a small amount because of the capacitance across the animal’s body wall. In most cases the measured d.c. potential was inside-positive relative to the water, but in a few cases it was inside-negative, accounting for the reversed polarity of the potential fluctuations seen during ventilation in these animals.
Evidence of the impedance changes between the interior of the animal and the surrounding water is also provided by the gut stimulus. In this case, a sinusoidal current stimulus created a potential modulation across the skin that was reduced in amplitude during expiration when the gill and spiracles were open and enhanced during inspiration when the gill clefts were tightly shut. Thus, the partial shunting of the inside-positive d.c. potential found in most fish can account for the shift to a more negative potential during expiration and, similarly, a positive potential shift accompanies the resistance increase during inspiration. The reverse is true when the d.c. potential is inside-negative.
The natural origin of the d.c. potential modulated during ventilation probably includes diffusional potentials across permeable surfaces, but a major contribution must also come from active ion pumping. The change in ventilatory potential observed within minutes of transferring a skate to dilute sea water and also the relatively small but short-latency changes seen in some animals after NaCl injection are probably due to diffusional potentials from permeable epithelia such as that of the gills. However, the much larger increases in ventilatory potential with salt-loading were only observed after a delay of several hours, which is more consistent with active osmoregulatory ion transport having been turned on. Burger (1962) found that after intravenous NaCl injection in the dogfish, Squalus acanthias, the rectal gland exhibited a rise in NaCl secretion only after a lag of 30–75 min. The rectal gland, which is the chief osmoregulatory organ in elasmobranchs (for a review, see Shuttleworth, 1988), creates substantial transepithelial potentials in secreting sodium and chloride. Studies on an isolated rectal gland preparation from the dogfish indicate that during active ion secretion the lumen of the gland, which opens into the gut, may be several millivolts negative compared with the plasma (Siegel et al. 1976). The gill is also known to be involved in active ion pumping in elasmobranchs and may make an additional but unknown contribution to the d.c. potential. Osmoregulatory physiologists who have measured d.c. potentials between the interior of fishes and the surrounding water using intraperitoneal salt-bridge electrodes have generally attributed these to branchial sources because of their similarity to transbranchial potentials measured in vitro (Bentley et al. 1976). It is notable that both gill ion permeability and the d.c. potentials measured in vivo are an order of magnitude smaller in elasmobranchs than in marine teleosts (Shuttleworth, 1988; Evans, 1980). As stated earlier, the bioelectric fields measured in the water during ventilation are also much smaller in elasmobranchs than in marine teleosts (Kalmijn, 1974).
It follows from the above considerations that the ventilatory potential and thus electrosensory reafference could be influenced by any factor that affects either the standing d.c. potential or the relative resistive changes during ventilation and, therefore, might be expected to be quite changeable under natural circumstances. The standing d.c. potential could be readily influenced by an intermittent dietary salt load, by swimming into the hyposaline water of an estuary, or by pH changes in the water or blood (Bentley et al. 1976). Resistive changes during ventilation are affected by variations in ventilatory movements as indicated above, but could also be affected by other factors such as the changes in gill circulation and permeability that accompany stress or strenuous activity. Circulating catecholamines may affect branchial circulation (Taylor, 1985) and also suppress rectal gland activity by restricting its blood supply (Shuttleworth, 1988).
The ampullary electroreceptors of elasmobranchs are responsive to extremely low frequencies, faithfully following potential modulations below 0.05 Hz (D. Bodznick, unpublished observations). However, at least over a reasonable range they are not sensitive to the absolute d.c. level and, by a process of accommodation, can apparently maintain high sensitivity to small potential modulations against a range of d.c. reference levels (Murray, 1965). The accommodation mechanism is not completely understood (Bennett and Clusin, 1976), but its importance for ampullary electroreceptor function is clear given the d.c. potentials measured here between the fish’s interior and the sea water.
Suppression of common mode signals
A central mechanism for rejecting common mode signals is particularly well suited for suppressing electrosensory reafference caused by the ventilatory potential because, although the potential can vary greatly in intensity and waveform at different times and in different animals, it is, nevertheless, always a common mode stimulus for the electroreceptors. Direct evidence for a common mode suppression mechanism in the medulla is provided here by experiments using the common mode stimulus created with the gut electrode. In our physiological set-up the gut electrode provides an effective means of creating a stimulus of common amplitude and phase for all the electroreceptors, with the exception of the dorsomedial hyoid group in which the skin pores are out of the water. This was verified by measuring the potential distribution in the water around the skate and, perhaps more convincingly, by monitoring the responses of a large population of primary afferent fibers. The major potential drop caused by this stimulus appears to be across the skin of the skate since receptor canal length had no apparent effect on the sensitivity of the response of the afferent fibers to the gut stimulus. Although there was considerable variability among AENs, most units exhibited significant suppression of the response to the gut stimulus compared with responses to extrinsic uniform fields. The uniform fields were similar in frequency and intensity to the gut stimulus but, unlike the gut stimulus, they affected the electroreceptors differentially. That the response to a common mode stimulus with no temporal relationship to ventilation is suppressed within the brain is clear evidence for the existence of a common mode rejection mechanism. This result does not rule out the possibility that other mechanisms, such as efference copy, also contribute to suppressing ventilatory reafference. However, in unparalyzed animals most individual AENs appeared to suppress the gut response and ventilatory reafference to about the same degree, despite considerable variation in the suppression among the AENs. Suppression of ventilatory modulation and of the response to the gut stimulus had a positive correlation coefficient (r) of 0.66 in the 29 cells measured and omitting one large outlying value, results in an r value of 0.76. This indicates that other mechanisms may not play a major role in suppressing the reafference due to ventilation at this level of the brain. The results are also important because they establish the gut electrode as a means of creating a well-controlled common mode stimulus that may be useful for further characterizing the medullary circuits responsible for suppressing ventilatory noise.
In this study, as in a previous one on Raja erinacea (New and Bodznick, 1990), most AENs exhibited at least some ventilatory modulation. Some cells were strongly affected. This is in marked contrast to the case in Platyrhinoidis triseriata (Montgomery, 1984), in which virtually none of the dorsal nucleus neurons were affected by ventilation. The reason for this difference is not known. In the studies on Raja the device that stabilized the head also restricted movements of the mouth during ventilation; no such restraint was used with Platyrhinoidis. It is possible that there is a proprioceptive component to the normal suppression of ventilatory reafference that has been altered in Raja. In other ways the methodologies of the experiments with Platyrhinoidis and Raja were very similar and it may be that the results simply reflect a species difference. This would suggest that electrosensory sensitivity might be more adversely affected by ventilatory reafference in Raja than in Platyrhinoidis, but this can only be determined through behavioral studies.
The reasons for the variation observed among AENs in this study on Raja with respect to their cancellation of common mode stimuli are also not known. In the case of ventilatory reafference it is possible that some of the apparent variation in the suppression is actually due to variation in the amplitude of the ventilatory potential itself among animals or during a single experiment. We attempted to minimize this by periodically sampling primary afferents throughout the experiment to ensure that the ventilatory reafference was consistently strong, but some variation was unavoidable. However, the suppression of responses to the gut stimulus was as variable among AENs as suppression of responses to ventilation and the positive correlation between the suppression of the gut stimulus and ventilatory reafference indicates that other factors are more important. Differences in the general health and condition of the fish might account for some variation. However, units in even a single electrode track sometimes showed quite different degrees of common mode suppression.
In the recording set-up used in these studies the top of the head is above the water level and the skin openings of the dorsomedial hyoid ampullae are out of the water. As a result, this small group of 10–12 electroreceptors on each side appeared to respond 180° out of phase with the other receptors during stimulation by the gut electrode and presumably also during ventilation. This should significantly degrade the common mode signal suppression in any units for which these ampullae provide inhibitory inputs. Similarly, AENs with excitatory receptive fields amongst these dorsomedial receptors should show poor common mode suppression, though none was apparently included in our sample of units. This artifact of the set-up seems to be a probable source of some of the variation observed among AENs. Lastly, since the extent of common mode signal suppression was measured in each case in comparison to the cell’s response to uniform electric fields, some of the apparent variation among AENs in their suppression of common mode stimuli may actually be due to variations in their uniform field sensitivity. Recent studies on the carpet shark (Bodznick and Montgomery, 1992) indicate that differences in AEN receptive field configurations might result in just such differences in uniform field sensitivity among the AENs. This variation would be beyond that expected as a result of varying ampullary canal lengths, which affect uniform field sensitivity in AENs and primary afferents alike.
Despite the variation, the effective suppression of an imposed artificial common mode stimulus by many AENs is the best evidence to date for a central common mode suppression mechanism. Direct evidence of the mechanism should also be found in the receptive fields of AENs, which must have inhibitory areas in addition to their small, ipsilateral excitatory fields. Several factors have made a convincing demonstration of these inhibitory areas difficult. However, this has recently been accomplished in experiments on the carpet shark, the results of which are reported in the following paper.
The authors thank Greg Hjelmstad for helpful discussion and technical assistance during a part of the study. The work was funded by an NSF grant to D.B. and Fullbright Fellowship to J.C.M.