- Email: [email protected]
Effect of electric organ discharge on ampullary receptors in a mormyrid
Braht Research, 145 (1978) 85-96 ~ Elsevier/North-Holland Biomedical Press
85
E F F E C T OF E L E C T R I C O R G A N D I S C H A R G E ON A M P U L L A R Y RECEPTORS IN A M O R M Y R I D
C. C. BELL and C. J. RUSSELL Neurological Sciences Institute, 1120 N. W. 20th Avenue, Portland, Ore. 97209 (U.S.A.)
(Accepted August 8th, 1977)
SUMMARY (1) Afferents from ampullary receptors were shown to be strongly affected by the electric organ discharge (EOD) in the mormyrid Gnathonemus petersii. (2) Over a broad range of resistivities (4-60 kf~cm) the response to the EOD was similar to the response to a brief (50-200,usec) outside positive pulse, i.e. an intitial acceleration of the discharge rate followed by a deceleration. (3) Brief biphasic positive-negative or negative-positive pulses where both phases were of equal amplitude and duration had no effect on ampullary afferents. Each phase had an effect, however, when given in isolation. These results suggest that a DC component in the EOD may cause the response in the ampullary afferent. (4) The response of ampullary afferents decreased sharply as resistivity was lowered below 10 k ~ c m . Responses to the EOD in mormyromast afferents also decreased. These effects were probably due to loading of the EOD at low resistivities and to a more rapid spatial decay of EOD voltage. (5) Responses of ampullary afferents to the EOD were much less affected by external non-conducting objects than were the responses of mormyromasts. These observations plus other considerations indicate that mormyromasts must still be held to play the major role in active electrolocation. Unresponsiveness of ampullary afferents to the EOD can not be taken as a reason for this, however.
INTRODUCTION Electroreceptors may be divided into two major classes, the ampullary or tonic receptors on the one hand and the tuberous or phasic on the otherS,e2, 23. 'Ampullary' and ~tuberous' refer to the distinct morphology of the two classes while 'tonic' and 'phasic' reflect the different functional properties. Primary afferents from ampullary receptors are spontaneously active. They respond to a voltage step across the skin
86 with a change in discharge rate. Eventually the rate returns to prestimulus levels, but strong stimuli can affect the rate for several minutes in mormyrids and a somewhat shorter time in gymnotoids 5. The peak sensitivity to sinusoidal stimuli is between 5 and l0 Hz ~,14A6. In contrast, afferents from tuberous or phasic receptors are usually silent in the absence of stimuli. They respond only at the onset of a voltage step, giving one or several spikes. Where sinusoidal stimuli have been used, peak sensitivities of 300-4000 Hz have been seen with low frequencies having little or no effect 1,11,18,~,~. In some gymnotoids the peak sensitivity of the phasic receptors is close to the frequency of maximum power in the electric organ discharge (EOD) of the fish 1,11,t3,1s,e~. Ampullary receptors are found in electric fish as well as in non-electric catfish and elasmobranchs. Tuberous receptors, however, are seen only in electric fish. This differential distribution, and the similarity between the power spectrum of the EOD and the spectral sensitivity of tuberous but not ampullary receptors, have led to the widely accepted idea that ampullary receptors are not affected by the EOD, and therefore could play no role in active electrolocation v. We were led to re-examine the question because of results obtained in another study (Bell and Russell, in preparation) on the mormyrid lateral line lobe. It is known that the cells in this structure which receive input from tuberous receptors are also strongly affected by a corollary discharge of the command to fire the electric organ 26. We found that the corollary discharge of the command signal affects cells receiving input from ampullary receptors, as well as those receiving from the tuberous type. Whatever the function of such a corollary discharge, its presence is most easily understood if the primary afferents which also affect these cells are normally influenced by the EOD. In the studies reported here, most ampullary afferents were strongly affected by the EOD. METHODS Ten mormyrid fish of the species Gnathonemus petersii were used. During recording, fish were restrained in two different ways. The first three animals were held by nylon netting stretched on a light plastic frame. The netting was placed on both sides of the body but was cut away in the region of the operculum to allow respiratory movements and to expose the dorsal branch of the posterior lateral line nerve. The second method of restraint proved to be better (see below) and was used in the last 7 experiments. A plastic rod 1 cm in diameter was attached to the skull with dental cement and then held firmly in .a clamp. The fish were further restrained in the lateral direction with 5 plastic dowels arranged vertically along each side of the animal. After exposing the nerve, a small rubber dam was glued around the wound to hold mineral oil. The restraining of the fish and the surgery were done under anesthesia (1:20,000 MS222), and on completion of the surgery the anesthetic was removed. The EOD returned within 15 min. After desheathing the nerve, small fiber bundles were hooked over an Ag-AgCI electrode for recording. Stimuli were delivered between an Ag-AgCI disk electrode in one corner of the chamber (about 15 cm from the head of the animal) and a movable silver ball electrode, which could be placed
87 at any point along the skin. With this arrangement of electrodes, making the nearby electrode positive causes a potential across the receptor which is positive outside and negative inside s . No attempt was made to record the absolute voltage at the pore and only relative stimulus amplitudes were studied. The currents were taken from the dial of a constant current stimulus isolation unit (Tektronix, 2620), the readings of which had been shown earlier to correspond to current as measured with a resistor. Most recordings were made at resistivities between 10 and 13 k~cm. Resistivity was varied in three experiments. It was lowered by adding normal saline (0.9 ~ ) and raised by successive dilutions with distilled water. The EOD was recorded with electrodes at either end of the electric organ. On occasion the areas under the two different phases were measured from filmed records. Ampullary receptor afferents can be readily distinguished from afferents arising from the two types of tuberous receptors found in mormyrids, the mormyromast and Knollenorgan a,15,23. Afferents from ampullary receptors discharge in the absence of stimulation; those from mormyromasts do not. Some Knollenorgan afferents are active without stimulation, but their spontaneous activity is quite different from that of ampullary afferents. It is generally more irregular and frequently contains intervals of less than 1 msec. Interspike intervals in ampullary afferents can also be close to 1 msec in duration, but only with very intense stimulation and never spontaneously. However, a more certain way of distinguishing ampullary from spontaneously active Knollenorgan afferents is by the responses to long-lasting and constant polarization. The Knollenorgan afferent will respond at a short latency to the on of outside positive stimuli with one or, more rarely, two spikes. The response latency is nearly constant, varying less than 0.5 msec, even with large variations in stimulus intensity. Most importantly the spontaneous discharge rate is not affected during long-lasting (e.g. 100 msec) stimuli of either polarity. In contrast, the discharge rate of ampullary afferents is very sensitive to the amplitude and polarity of a maintained polarization. The rate is increased by outside positive and decreased by outside negative stimuli (Figs. I and 3). In contrast to the Knollenorgan, response latencies are quite variable. Thus by examining spontaneous activity, the degree of variation in response latency, and the effect of long duration stimuli of different intensities and polarities, the three types of afferents can be easily identified. RESULTS Most ampullary afferents recorded in water of 10 k ~ c m or higher were obviously affected by each EOD (Fig. 1). Of the fibers recorded at resistivities of 10 kflcm or higher, 51 out of 63 were clearly related to the EOD. The fact that 5 out of 12 responded during the first three experiments, while 46 out of 51 responded during the last 7, suggest that the method of restraint used in the latter experiments was less damaging to the receptors. The responses were always of the same form: a broad acceleration lasting about 30 msec, followed by a still broader deceleration. In many cases the response had the appearance of a damped oscillation with successively smaller peaks and valleys being seen in the histogram.
88
A
E!
i m,
~
j
.
i
ii
I i
i i
ii
i
ii J
i
i
i
I
4 0 msee
Fig. 1. Effect of electric organ discharge (EOD) on ampullary afferent. A : response of afferent to outside positive (above) and outside negative (below) stimuli. In each case the lower trace indicates duration and polarity of the stimulus but not amplitude. As indicated by dots, EODs occurred during pause following outside positive and during burst following outside negative stimulii B: effect of EOD on same fiber. Effect is shown with superimposed traces (5), raster displays and post-EOD histogram. EOD shown in bottom record. Small spike seen immediately after EOD in the superimposed traces is simultaneously recorded mormyromast or Knoltenorgan afferent. Calibration marks on histogram indicate 100 counts/bin. In all ampullary afferents which were considered to be affected by the EOD, the influence could be seen with each EOD. However, there was no response if the E O D occurred when the spontaneous activity was silenced either by an outside negative stimulus or following an outside positive one. Such a lack of response to the EOD can be seen in Figs. 1A and in the top line of Fig. 2. Presumably, if polarization had been less intense the response would have been diminished rather than completely blocked. Fig. 2 also shows the effect of single EODs on a normally active ampullary afferent. Some afferents responded more strongly to the EOD than others. The weaklyor non-responding afferents appeared to arise from more anterior receptors, but this was not examined systematically. The acceleration-deceleration form of the response was essentially the same as the response to a brief (50-200 #sec) outside positive pulse (Fig. 2). The response to a brief pulse is a shortened form of the response to a long-lasting stimulus of the same polarity. A brief outside positive pulse causes acceleration-deceleration, while a brief negative pulse causes the reverse. Such responses to brief pulses were seen in all 14 fibers tested. Thresholds for a clear response to stimuli of 200 #sec duration were determined in 8 fibers. Thresholds for brief pulses were 2 0 4 0 times higher than for long ones. This gives an indication of the absolute stimulus amplitude across the skin since the threshold of these receptors to long pulses has been found to be 2-10 mV when measured between the pore opening and a distant electrode z. A biphasic stimulus produces no effect when the component pulses are brief (durations of 50-200 #sec were tested) and of equal amplitude and duration (Fig. 2, lower right). Given in isolation, however, each component can have a clear effect.
89 OUTSIDE
OUTSIDE
POSITIVE
ii]ll]llil+iilllil 'trilmii+ii,;i
T
i iil
+
i iii
iHltLtti ...... ....
150 msec
]III+~+
NEGATIVE
illliJtll tiiJ
' iPlIIt11111tlll[l rl]ll
I
iilliii
I00 msec
Lift ....... JJ.i ijl.t r rTr]l Wl . . . . . . . . 13 tlr+r
15 msec
.
r
]Till]Ill 1
[tl'l~
I
. . . . . . . . . . . . . .
i
,,l,!H!! !1]II I!!!+! Illl[r
200
I I I
psec
i
o,
!!]!!
J!!]!!]!!! !!!!J!!! !
yf[r!
rrrrrl]lr!imrr
L kn.L~.
1
~lu ~ ,i,i .Ib~J, Js.~.~.l,LlkJ"
A
BIPHASIC
m
40 mse¢
I~ULSES ,m
+
EOD
A
-20 I
~.I.,.1, ...... I,+~.L .....
-,-L~.~.L"
I Fig. 2. Effects of pulses of different durations and polarities on ampullary afferent. Effects of 150 msec, 15 msec and 200 #sec pulses of both polarities are shown. Stimulus amplitude is the same for 150 and 15 msec pulses but increased by factor of 3 for 200 #sec pulses. The absolute value of positive and negative pulses at any one duration is the same. Histograms below 200/+sec traces show the average effect of these stimuli. EOD occurrences are marked by dots. EOD has no effect if given when afferent is silenced following off of 150 msec outside positive stimulus or during 150 msec outside negative stimulus. As shown in the other traces a slight acceleration follows the EOD if the afferent is discharging at its resting rate. A histogram showing the effect of the EOD is shown at lower left. At lower right the effects of biphasic positive-negative and negative-positive pulses are shown. Each phase is of the same amplitude and duration as the monophasic pulses immediately above which produced clear effects on the discharge rate of the fiber. All histograms are based on 65-70 stimuli. Bars at right of each histogram indicate 20 counts/bin. As indicated by time calibration marks, the two top traces were taken at slower sweeps than the remaining records and histograms.
Mono- and biphasic pulses contain the same high frequencies, but the monophasic pulse also has a DC component.
This DC component
thus seems to be responsible
f o r t h e effect o f a b r i e f p u l s e o r o f t h e E O D o n t h e a m p u l l a r y a f f e r e n t . In
Gnathonemus petersii, t h e E O D is b i p h a s i c . D u r i n g t h e i n i t i a l p h a s e t h e a n -
t e r i o r e n d o f t h e e l e c t r i c o r g a n is p o s i t i v e , a n d c u r r e n t flows o u t o f t h e s k i n i n f r o n t o f t h e e l e c t r i c o r g a n w h e r e t h e r e c e p t o r s are. D u r i n g t h e s e c o n d p h a s e t h e a n t e r i o r
90
B
A II K ohm-eros
II K ohm-cms A 4 K ohm-cms
il
A
ilL
2 K ohm-cms
2 K ohm-cms
C
2 K ohm-cms
Fig. 3. Effects of resistivity on responses of ampullary and mormyromast afferents to the EOD. A : ampullary afferent recorded at the three resistivities indicated. OCcurrence of EOD indicated by arrowhead. B: upper two histograms: mormyromast afferent recorded at the two resisfivities indicated. Bottom histogram: responses of ampullary afferent to EOD shown for comparison at same time base as mormyromast response. C: one of two occasions on which slowing of ampullary afferent occurred following the EOD. Histograms in A and those of mormyromast in B are based on 205-215 EODs. Bottom histogram in B based on 600 and histogram in C is based on 700 EODs. The calibration mark in bottom histogram of B is at 25 counts/bin, other calibrations marks indicate 50 counts/ bin. Note expanded time calibration in B.
end is negative, and current flows in through the receptor area. Resistivity changes have a pronounced affect on the amplitude and waveform of the EOD2,1°. At high resistivities ( > 50 kl'lcm) the second phase is smaller in amplitude but longer in duration than the first. As resistivity is lowered, second phase amplitude becomes larger than that of the first phase and its duration becomes equally brief, in addition, at resistivities below 20 kflcm both phases show the effects of loading and the absolute amplitudes begin to fall steeply. One would expect that such changes in the EOD would affect a receptor's response. At resistivities between t0 and 13 k l l c m where most experiments were done the second, head negative, phase of the EOD is larger than the first, head positive, phase. The ratio of peak first phase to peak second phase voltage was 0.5-0.6. The net flow of current was therefore inward across the skin, as in the case of an outside positive pulse. The similarity between the effect of such a pulse and the EOD is therefore reasonable. Five ampullary afferents were recorded at a resistivity of 11 k~cm. Saline (0.9 ~ NaC1) was then added to lower the resistivity in one or two stages. In each case the response decreased with decreasing resistivity (Fig. 3A). At intermediate resistivities of about 4 kflcm, the responses were similar in form, though smaller in amplitude than at 11 kllcm. None of these fibers discharged in relation to the EOD
91 at resistivities between 1.7 and 2 kf~cm. Six additional ampullary afferents were recorded at 2 kf~cm only. Four of these gave no response, while two showed a moderate slowing of their discharge rate (Fig. 3). Such deceleration responses were not seen in any other ampullary afferents and seem paradoxical, since the second phase of the EOD is twice the amplitude of the first at this low resistivity, and the response should therefore be similar to the response to an outside positive pulse. Both afferents responded normally to long pulses of both polarities, and the receptors of both were located just in front of the electric organ at the posterior margin of the specialized electroreceptor epithelium. A possible explanation of the slowing is that local currents may have differed from the overall current. Responses of mormyromast afferents also fell in amplitude with decreasing resistivity. The number of spikes/EOD decreased and the latency increased (Fig. 3B, 2 upper histograms). This was seen in 8 mormyromast afferents. In 4 of these cases we succeeded in holding the fiber while decreasing the resistivity to the point where the response disappeared (2-4 kf~cm) and also while increasing it back to the original value. In each case the response returned and was the same as the initial response. Three ampullary afferents were recorded after increasing the resistivity to 60 k ~ c m . All were clearly related to the EOD, and the form of the response was the same as that seen at resistivities of 10-13 kf~cm, i.e. an acceleration-deceleration sequence. At the higher resistivity the second phase of the EOD was lower in amplitude than the first though broader in duration. The areas under the curves were nearly equal (first/second = 0.93) indicating only a small DC component in the outside positive direction. The fact that the response was nevertheless strong and similar to that following an outside positive pulse suggests that the frequencies contained in the two phases may be important as well as the DC components. The finding that ampullary afferents are affected by the EOD raises the question of involvement of these receptors in active electrolocation. We tested this possibility by recording EOD responses from 10 ampullary afferents before and after placing non-conducting objects near the receptor pores. We used rectangles of thin plastic, either 8 /, 20 × 0.3 mm or 12 ~< 35 >~ 0.3 ram. We tried centering the rectangle on the pore, and also placing it just anterior or just posterior. The response of ampullary afferents was unaffected by such objects even when placed within one millimeter of the skin. Lack of effect was judged by comparing histograms of the response before and after placing the object. In order to maximize the possibility of seeing an effect without damaging the receptor we also placed squares of plastic film (Saran Wrap) 20 /. 20 mm on the skin, centering them on the receptor. The plastic film was in contact with the skin at several points and at no point was it more than I m m away. In one of the three fibers in which this was done the plastic film seemed to have a slight effect (Fig. 4, left side). In the middle histogram, made while the film was over the receptor, there is a slight broadening of the acceleration portion of the response and a slight delay and reduction in the deceleration portion. The effect is admittedly marginal and was obtained under extreme conditions. It was, however, the only occasion on which an object appeared to have some influence and its inclusion illustrates the near insensitivity of the ampullary afferent response to nearby non-conducting
92
AMPULLARY
MORMY~ST
no object
J
object present L ___
no object
i
EOD
50 msec
EOD
m
5 msec
Fig. 4. Effects of objects on responses of ampullary and mormyromast efferents to the EOD. For both receptor types, the top histograms were made before introducing the object, the middle histograms while the object was present and the bottom histograms after removing the object. For the ampultary afferent the object was a 20 x 20 mm square of plastic film resting lightly on the skin and centered on the receptor. For the mormyromast it was an 8 x 20 x 0.3 mm plastic plate held 3-4 mm from the skin, also centered over the receptor. The few bins with reduced counts immediately after the EOD in histograms of the ampullary response are an artifact, due to rejecting some ampullary spikes along with the spikes of a simultaneously recorded mormyromast afferent. Note the different time bases of ampullary and mormyromast histograms. Ampullary histograms based on 225 to 235 EODs each; mormyromast histograms on 165 to 185 EODs. Vertical calibrabion marks indicate 50counts/ bin for ampuUary and 100 counts/bin for mormyromast histograms.
objects. C a r b o n rods, which are highly conductive, were tried. Effects o f the rods o n the E O D response were also slight, b u t greater t h a n those o f n o n - c o n d u c t i n g objects. H o w e v e r , the c a r b o n rods a l t e r e d the s p o n t a n e o u s discharge rate, p r e s u m a b l y because o f the interface currents which t h e y g e n e r a t e l L Such rate changes could, in turn, affect the response, a n d the c a r b o n r o d results were therefore j u d g e d a m b i g u o u s with r e g a r d to the question o f active electrolocation. F o r c o m p a r i s o n , m o r m y r o m a s t afferents were also r e c o r d e d a n d e x a m i n e d for the effect o f n e a r b y plastic objects on their response to the E O D . The c o n t r a s t with a m p u l l a r y afferents was striking. Plastic rectangles smaller t h a n those tried with the a m p u l l a r y afferents (e.g. 5 x 13 mm) h a d a clear influence at distances f r o m the skin o f 3-5 mm. I f the rectangle was centered over the pore the response was decreased, if placed at either e n d o f the p o r e it was enhanced, as described by S z a b o a n d Hagiw a r a 24. Such b e h a v i o r was seen in all 12 o f the m o r m y r o m a s t afferents which were
93 examined in isolation (Fig. 4, on the right). On three occasions mormyromast afferents were recorded simultaneously with ampullary afferents. Plastic rectangles close to the ampullary pore had no effect on the ampullary afferent but strongly affected the mormyromast afferent, causing an increase in latency and a reduction in the number of spikes/response. DISCUSSION Ampullary afferents in Gnathonemuspetersii have been recorded in two different studies and found unresponsive to the EOD15, 91. These negative results may have been due to a low external resistivity. In addition, in one of the studies where mechanical sensitivity of ampullary receptors was the major interest, only afferents from the chin appendage were examined 21. Such anterior receptors are less likely to be affected by the EOD than those in the posterior part of the animal where we recorded. This is because at low resistivities the EOD-induced voltage across the skin probably decreases quite steeply in the posterior to anterior direction 2. In a study on Gymnotus carapo, Suga saw an ampullary receptor discharging in relation to the EOD 20. Although this was the exception, different recording conditions, such as higher resistivity, might increase the responsiveness. Afferents which appear to have been ampullary were recorded by Hagiwara et al. 9 in Electrophorus electricus. These afferents responded strongly to the low voltage EOD of that fish which is 20-40 V in amplitude, entirely monophasic and lasts 2-3 msec. On the basis of our results one would expect ampullary receptors to be influenced by such a discharge. One would also expect ampullary afferents to respond to the EOD in other fish with broad monophasic discharges, such as Mormyrus rume and some species of Hypopomus 4. The decrease in the responses to the EOD of both ampullary and mormyromast afferents at low resistivities can be attributed to a decrease in the voltage across the receptor during the discharge. This decrease has two causes: one is the decreased voltage produced by the electric organ at low resistivities, due to loading; the second is a more rapid spatial decay of transcutaneous voltage in front of the electric organ, due to the cable properties of the fish and the lower external resistivity 2. An alternative explanation is that the NaC1 added to reduce resistivity had a direct depressive effect on the receptor. Such depressive effects of NaCI have been seen by Roth in ampullary receptors of catfish~L In our experiments the final concentration of NaC1 at the lowest resistivity was about 4 mM/liter. At such concentrations Roth found the spontaneous discharge frequency to be reduced to less than one-half its original value. The response to formerly effective long duration (700 msec) stimuli disappeared even when intensity was increased by a factor of 10. In our experiments, however, there was no significant change in spontaneous activity. For example, the fiber in Fig. 3A discharged at rates of 92/sec, 99/sec and 88/sec in resistivities of 11, 4 and 2 k~, respectively. Furthermore, long duration stimuli remained effective, although higher intensities were necessary because of the lowered resistance between the two stimulating electrodes, both of which were in the bath (see Methods). It thus seems unlikely
94 that the loss of responsiveness to the EOD was due to a direct depressive effect of NaCI on the receptor. The fact that responses of mormyromast afferents were similarly reduced also argues against a direct effect. Epithelial tissue is present between mormyromast receptor cells and the external medium and this would probably reduce the diffusion of ions 22. Dependence on resistivity of the EOD-induced voltage across the receptor is important for active electrolocation. The mormyromast which is almost certainly the main receptor involved (see below) appears to have a rather narrow operating range. Szabo and Hagiwara 24 found that, once threshold was crossed, the response increased rapidly, reaching a maximum at less than two times threshold. The narrow operating range results in a high incremental sensitivity useful in active electrolocation. But it requires that the operating point be kept within relatively close limits. It is therefore of interest that the peak to peak voltage of the EOD, as recorded between a point on the anterior skin and a distant electrode, varied relatively little with resistivity changes between 20 and 180 kf~cm 2 and that Gosse s found that G. peterMi live in resistivities between t3 and 180 kD,cm. The EOD voltage decreases sharply as resistivity falls below 20 kllcm 2 and in this range the responses of both ampullary and mormyromast receptors also decreased with decreasing resistivity. At some resistivity below 2 kf~cm one would expect mormyromast as well as ampuUary afferents to stop responding and active electrolocation to become impossible. The question of which receptor class is primarily responsible for active electrolocation is important in the physiology of these fish. Knollenorgans are not believed to play a role 5,2a because their responses, once threshold is crossed, vary only slightly with stimulus intensity. Individual fibers could not signal the small variations in EOD-induced voltage upon which active electrolocation must depend. Furthermore their threshold is extremely low so they probably all discharge with each EOD, at least at normal resistivities. The proportion of the receptor population responding to each EOD doesn't vary, therefore, and could not be used to signal presence of an object. Finally, command associated inhibition of the central cells on which Knollenorgan afferents end is consistent with lack of any role in active electrolocation ~6. Several authors have suggested a communication role for Knollenorgan receptors r',~3. The responses of both ampullary and mormyromast afferents are graded with stimulus intensity. The markedly different effects of external objects on responses of the two types, however, indicates that mormyromast and not ampullary afferents are primarily responsible for active electrolocation. The reason for the different effects of objects is not clear, however. As discussed above, lowering resistivity lowers the amplitude of the EOD, and this in turn causes a reduced response in the afferents from both receptor types (Fig. 3). One concludes that ampullary and mormyromast afferents both go from strong to minimal responses over roughly similar ranges of EO D-induced transcutaneous voltages. But the striking differences in the effects of objects on responses of the two receptor types are difficult to reconcile with this conclusion (Fig. 4). A nearby non-conducting object reduces current flow through the skin immediately beneath it and this results in a lower EOD-induced voltage across the skin at that point. If responses of the two receptor types are equally affected by resistivity-induced
95 voltage changes, they should also be affected equally by object induced changes (cf. Figs. 3 and 4). A possible explanation may be found in the difference in the effective stimulus which drives the two types of receptors. The DC component of a pulse or EOD results in a residual charge remaining on the capacitance of skin and receptor. This residual, relatively long-lasting voltage is probably the effective stimulus for the ampullary receptor. The anterior body of the fish can be seen as a passive cable with resistance and capacitance in parallel across the skin. In such a cable, low frequencies decrement more slowly with distance than do high frequencies. This means that the low frequency residual voltage will spread with little decrement from one skin region to another. At any one point the low frequency components of the transcutaneous voltage would represent the average of such voltages over a broad area. The result would be minimal local variations and minimal effects of local changes in external resistance. In contrast, the high frequency components to which the mormyromast is sensitive decrement rapidly, permitting sharp local differences and allowing small external objects to have effects. Other considerations also suggest that mormyromast rather than ampullary afferents play the major role in active electrolocation. (l) The mormyromast is insensitive to low frequencies and thus filters out non-EOD sources which could interact with EOD voltage and be noise to an active electrolocation system. The marked effect of extraneous stimuli on the response of an ampullary afferent to the EOD was noted in Figs. 1 and 2. (2) The lack of spontaneous activity in mormyromasts not only enhances the signal to noise ratio but also ensures that the receptor will be fully recovered when an EOD occurs. (3) Variability in the response to individual EODs in the absence of any environmental change would be undesirable for active electrolocation. The greater the variability, the larger the number of EODs and the longer the time required to detect a change in a receptor's response following an environmental change. The low variability of the mormyromast in comparison to the ampullary afferents suggests therefore that these receptors are better suited for electrolocation. (4) The response of the ampullary receptor does not reach a peak until 20 msec after the EOD, while the first spike of the mormyromast occurs at less than 5 msec and the entire response is usually finished by 15 msec. The shorter latency could well be important in reducing reaction time. ACKNOWLEDGEMENTS We thank Jaqueline Bolen for typing the manuscript and its drafts. The work was supported by Grants from N1H (NINCDS-06728) and NSF (BMS 73-06867.)
REFERENCES I Bastian, J., Frequency response characteristics of electroreceptors in weakly electric fish (Gymnotoidei) with a pulse discharge, J. comp. Physiol., 112 (1976) 165 180. 2 Bell, C. C., Bradbury, J. and Russell, C. J., The electric organ of a mormyrid as a current and voltage source, J. eomp. Physiol., 110 (1976) 65 88.
96 3 Bennett, M. V. L., Electroreceptors in Mormyrids, Cold. Spr. Harb. Symp. quant. Biol., 30 (1965) 245-262. 4 Bennett, M. V. L., Electric organs. In W. S. Hoar and D. S. Randall (Eds.), Fish Physiology, Academic Press, New York, 1971, pp. 347-491. 5 Bennett, M. V. L., Electroreception In W. S. Hoar and D. S. Randall (Eds.), l~sh Physiology, Academic Press. New York, 1971, pp. 493-574: 6 Dunning, B.D., A Quantitative and Comparative Analysis of the Tonic Electroreeeptors o]' Gnathonemus, Gymnotus and Kyrptopterus, Ph.D. Thesis, University of Minnesota, 1973. 7 Fessard, A., (Ed.), Electroreceptors and Other Specialized Receptors in Lower Vertebrates, Vol. Ilt/3, Handbook of Sensory Physiology, Springer-Verlag, Berlin, 1974. 8 Gosse, J. P., Le milieu aquatique et I'ecologie des poissons dons la region de Yangambi, Sci. Zool., 116 (1963) 113-249. 9 Hagiwara, S., Szabo, T. and Enger, P. S., Physiological properties of electroreceptors in the electric eel, Electrophorus eleetrieus, J. Neurophysiol., 28 (1965) 775-783. 10 Harder, W., Schief, A. und Uhlemann, H., Zur Funktion des elektrischen Organs yon Gnathonemus petersii (Gthr 1862) (Mormyriformes, Teleostei), Z. Vergl. Physiol., 48 (t964) 302-331. 11 Hopkins, C. D., Stimulus filtering and electroreception: tuberous electroreceptors in three species of gymnotoid fish, J. comp. Physiol., III (1976) 171-207. 12 Kalmijn, A. J., The detection of electric fields from inanimate and animate sources other than electric organs. In A. Fessard (Ed.), Handbook of Sensory Physiology, Vol. 111/3, Springer, New York, 1974, pp. 147-200. 13 Knudsen, E., Behavioral thresholds to electric signals in high frequency electric fish, J. comp. Physiol., 91 (1974) 333-353. 14 Peters, R. and Buwalda, R., Frequency responses of the electroreceptors ('small pit organs') of the catfish, lctalurus nebulosus, J. comp. PhysioL, 79 (1972) 29-38. 15 Roth, A., Propri6t6s functionelles et morphologiques des differents organes de la ligne laterale des Mormyrides, J. PhysioL (Paris), 59 (1967) 486. 16 Roth, A., Central neurons involved in the e[ectroreception of the catfish Kyrptopterus, J. comp. Physiol., 100 (1975) 135-146. 17 Roth, A., Zur Funktionweise der Elektrorezeptoren in der Haut von welsen (lctalurus): der Einflus der Ionen in susswasser, Z. Vergl. Physiol., 75 (1971) 303-322. 18 Scheich, H., Bullock, T. H. and Hamstra, R. H., Coding properties of two classes of afferent nerve fibers: high frequency electroreceptors in the electric fish, J. Neuroph~TioL, 36 (1973) 39-60. 19 Scheich, H. and Bullock, T. H., The detection of electric fields from electric organs. In A. Fessard (Ed.), Electroreceptors and Other Specialized Receptors in Lower Vertebrates, Handbook of Sensoo' Physiology, Vol. 111/3, Springer, Berlin, 1974, pp. 201-256. 20 Suga, N., Etectrosensitivity of specialized and ordinary lateral line organs of the electric fish, Gymnotus carapo. In P. Cahn (Ed.), LateratLine Detectors, Bloomington, Ind., 1967, pp. 395-409. 21 Szabo, T., l~ber eine bisher unbekannte Funktion der sog. ampullaren Organe bei Gnathonemus petersii, Z. Vergl. Physiol., 66 (1970) 164-175. 22 Szabo, T., Anatomy of the specialized lateral line organs of electroreception. In A. Fessard (Ed.),
Electroreceptors and Other Specialized Receptors in Lower Vertebrates, Handboo~ of Sensory Phsyiology, Vol. 111/3, Springer, Berlin, 1974, pp. 13-56. 23 Szabo, T. and Fessard, A., Physiology of electroreceptors. In A. Fessard (Ed.), Electroreceptors and Other Specialized Receptor in Lower Vertebrates, Handbook of Sensory Physiology, Vol. 11l/3, Springer, Berlin, 1974, pp. 59-124. 24 Szabo, T. and Hagiwara, S., A latency-change mechanism involved in sensory coding of electric fish (mormyrids), Physiol. Behav., 2 (1967) 331-335. 25 Viancour, T. A., Neuronal response autorhythmicity associated with non-mechanical frequency filtering in the electrosense system of Eigenmannia Virescens, Neurosci. Abstr., 2 (1976) Abstr. 335. 26 Zipser, B. and Bennett, M. V. L., Interaction of electrosensory and electromotor signals in lateral line lobe of a morrnyrid fish, J. Neurophysiol., 39 (1976) 713-721.
85
E F F E C T OF E L E C T R I C O R G A N D I S C H A R G E ON A M P U L L A R Y RECEPTORS IN A M O R M Y R I D
C. C. BELL and C. J. RUSSELL Neurological Sciences Institute, 1120 N. W. 20th Avenue, Portland, Ore. 97209 (U.S.A.)
(Accepted August 8th, 1977)
SUMMARY (1) Afferents from ampullary receptors were shown to be strongly affected by the electric organ discharge (EOD) in the mormyrid Gnathonemus petersii. (2) Over a broad range of resistivities (4-60 kf~cm) the response to the EOD was similar to the response to a brief (50-200,usec) outside positive pulse, i.e. an intitial acceleration of the discharge rate followed by a deceleration. (3) Brief biphasic positive-negative or negative-positive pulses where both phases were of equal amplitude and duration had no effect on ampullary afferents. Each phase had an effect, however, when given in isolation. These results suggest that a DC component in the EOD may cause the response in the ampullary afferent. (4) The response of ampullary afferents decreased sharply as resistivity was lowered below 10 k ~ c m . Responses to the EOD in mormyromast afferents also decreased. These effects were probably due to loading of the EOD at low resistivities and to a more rapid spatial decay of EOD voltage. (5) Responses of ampullary afferents to the EOD were much less affected by external non-conducting objects than were the responses of mormyromasts. These observations plus other considerations indicate that mormyromasts must still be held to play the major role in active electrolocation. Unresponsiveness of ampullary afferents to the EOD can not be taken as a reason for this, however.
INTRODUCTION Electroreceptors may be divided into two major classes, the ampullary or tonic receptors on the one hand and the tuberous or phasic on the otherS,e2, 23. 'Ampullary' and ~tuberous' refer to the distinct morphology of the two classes while 'tonic' and 'phasic' reflect the different functional properties. Primary afferents from ampullary receptors are spontaneously active. They respond to a voltage step across the skin
86 with a change in discharge rate. Eventually the rate returns to prestimulus levels, but strong stimuli can affect the rate for several minutes in mormyrids and a somewhat shorter time in gymnotoids 5. The peak sensitivity to sinusoidal stimuli is between 5 and l0 Hz ~,14A6. In contrast, afferents from tuberous or phasic receptors are usually silent in the absence of stimuli. They respond only at the onset of a voltage step, giving one or several spikes. Where sinusoidal stimuli have been used, peak sensitivities of 300-4000 Hz have been seen with low frequencies having little or no effect 1,11,18,~,~. In some gymnotoids the peak sensitivity of the phasic receptors is close to the frequency of maximum power in the electric organ discharge (EOD) of the fish 1,11,t3,1s,e~. Ampullary receptors are found in electric fish as well as in non-electric catfish and elasmobranchs. Tuberous receptors, however, are seen only in electric fish. This differential distribution, and the similarity between the power spectrum of the EOD and the spectral sensitivity of tuberous but not ampullary receptors, have led to the widely accepted idea that ampullary receptors are not affected by the EOD, and therefore could play no role in active electrolocation v. We were led to re-examine the question because of results obtained in another study (Bell and Russell, in preparation) on the mormyrid lateral line lobe. It is known that the cells in this structure which receive input from tuberous receptors are also strongly affected by a corollary discharge of the command to fire the electric organ 26. We found that the corollary discharge of the command signal affects cells receiving input from ampullary receptors, as well as those receiving from the tuberous type. Whatever the function of such a corollary discharge, its presence is most easily understood if the primary afferents which also affect these cells are normally influenced by the EOD. In the studies reported here, most ampullary afferents were strongly affected by the EOD. METHODS Ten mormyrid fish of the species Gnathonemus petersii were used. During recording, fish were restrained in two different ways. The first three animals were held by nylon netting stretched on a light plastic frame. The netting was placed on both sides of the body but was cut away in the region of the operculum to allow respiratory movements and to expose the dorsal branch of the posterior lateral line nerve. The second method of restraint proved to be better (see below) and was used in the last 7 experiments. A plastic rod 1 cm in diameter was attached to the skull with dental cement and then held firmly in .a clamp. The fish were further restrained in the lateral direction with 5 plastic dowels arranged vertically along each side of the animal. After exposing the nerve, a small rubber dam was glued around the wound to hold mineral oil. The restraining of the fish and the surgery were done under anesthesia (1:20,000 MS222), and on completion of the surgery the anesthetic was removed. The EOD returned within 15 min. After desheathing the nerve, small fiber bundles were hooked over an Ag-AgCI electrode for recording. Stimuli were delivered between an Ag-AgCI disk electrode in one corner of the chamber (about 15 cm from the head of the animal) and a movable silver ball electrode, which could be placed
87 at any point along the skin. With this arrangement of electrodes, making the nearby electrode positive causes a potential across the receptor which is positive outside and negative inside s . No attempt was made to record the absolute voltage at the pore and only relative stimulus amplitudes were studied. The currents were taken from the dial of a constant current stimulus isolation unit (Tektronix, 2620), the readings of which had been shown earlier to correspond to current as measured with a resistor. Most recordings were made at resistivities between 10 and 13 k~cm. Resistivity was varied in three experiments. It was lowered by adding normal saline (0.9 ~ ) and raised by successive dilutions with distilled water. The EOD was recorded with electrodes at either end of the electric organ. On occasion the areas under the two different phases were measured from filmed records. Ampullary receptor afferents can be readily distinguished from afferents arising from the two types of tuberous receptors found in mormyrids, the mormyromast and Knollenorgan a,15,23. Afferents from ampullary receptors discharge in the absence of stimulation; those from mormyromasts do not. Some Knollenorgan afferents are active without stimulation, but their spontaneous activity is quite different from that of ampullary afferents. It is generally more irregular and frequently contains intervals of less than 1 msec. Interspike intervals in ampullary afferents can also be close to 1 msec in duration, but only with very intense stimulation and never spontaneously. However, a more certain way of distinguishing ampullary from spontaneously active Knollenorgan afferents is by the responses to long-lasting and constant polarization. The Knollenorgan afferent will respond at a short latency to the on of outside positive stimuli with one or, more rarely, two spikes. The response latency is nearly constant, varying less than 0.5 msec, even with large variations in stimulus intensity. Most importantly the spontaneous discharge rate is not affected during long-lasting (e.g. 100 msec) stimuli of either polarity. In contrast, the discharge rate of ampullary afferents is very sensitive to the amplitude and polarity of a maintained polarization. The rate is increased by outside positive and decreased by outside negative stimuli (Figs. I and 3). In contrast to the Knollenorgan, response latencies are quite variable. Thus by examining spontaneous activity, the degree of variation in response latency, and the effect of long duration stimuli of different intensities and polarities, the three types of afferents can be easily identified. RESULTS Most ampullary afferents recorded in water of 10 k ~ c m or higher were obviously affected by each EOD (Fig. 1). Of the fibers recorded at resistivities of 10 kflcm or higher, 51 out of 63 were clearly related to the EOD. The fact that 5 out of 12 responded during the first three experiments, while 46 out of 51 responded during the last 7, suggest that the method of restraint used in the latter experiments was less damaging to the receptors. The responses were always of the same form: a broad acceleration lasting about 30 msec, followed by a still broader deceleration. In many cases the response had the appearance of a damped oscillation with successively smaller peaks and valleys being seen in the histogram.
88
A
E!
i m,
~
j
.
i
ii
I i
i i
ii
i
ii J
i
i
i
I
4 0 msee
Fig. 1. Effect of electric organ discharge (EOD) on ampullary afferent. A : response of afferent to outside positive (above) and outside negative (below) stimuli. In each case the lower trace indicates duration and polarity of the stimulus but not amplitude. As indicated by dots, EODs occurred during pause following outside positive and during burst following outside negative stimulii B: effect of EOD on same fiber. Effect is shown with superimposed traces (5), raster displays and post-EOD histogram. EOD shown in bottom record. Small spike seen immediately after EOD in the superimposed traces is simultaneously recorded mormyromast or Knoltenorgan afferent. Calibration marks on histogram indicate 100 counts/bin. In all ampullary afferents which were considered to be affected by the EOD, the influence could be seen with each EOD. However, there was no response if the E O D occurred when the spontaneous activity was silenced either by an outside negative stimulus or following an outside positive one. Such a lack of response to the EOD can be seen in Figs. 1A and in the top line of Fig. 2. Presumably, if polarization had been less intense the response would have been diminished rather than completely blocked. Fig. 2 also shows the effect of single EODs on a normally active ampullary afferent. Some afferents responded more strongly to the EOD than others. The weaklyor non-responding afferents appeared to arise from more anterior receptors, but this was not examined systematically. The acceleration-deceleration form of the response was essentially the same as the response to a brief (50-200 #sec) outside positive pulse (Fig. 2). The response to a brief pulse is a shortened form of the response to a long-lasting stimulus of the same polarity. A brief outside positive pulse causes acceleration-deceleration, while a brief negative pulse causes the reverse. Such responses to brief pulses were seen in all 14 fibers tested. Thresholds for a clear response to stimuli of 200 #sec duration were determined in 8 fibers. Thresholds for brief pulses were 2 0 4 0 times higher than for long ones. This gives an indication of the absolute stimulus amplitude across the skin since the threshold of these receptors to long pulses has been found to be 2-10 mV when measured between the pore opening and a distant electrode z. A biphasic stimulus produces no effect when the component pulses are brief (durations of 50-200 #sec were tested) and of equal amplitude and duration (Fig. 2, lower right). Given in isolation, however, each component can have a clear effect.
89 OUTSIDE
OUTSIDE
POSITIVE
ii]ll]llil+iilllil 'trilmii+ii,;i
T
i iil
+
i iii
iHltLtti ...... ....
150 msec
]III+~+
NEGATIVE
illliJtll tiiJ
' iPlIIt11111tlll[l rl]ll
I
iilliii
I00 msec
Lift ....... JJ.i ijl.t r rTr]l Wl . . . . . . . . 13 tlr+r
15 msec
.
r
]Till]Ill 1
[tl'l~
I
. . . . . . . . . . . . . .
i
,,l,!H!! !1]II I!!!+! Illl[r
200
I I I
psec
i
o,
!!]!!
J!!]!!]!!! !!!!J!!! !
yf[r!
rrrrrl]lr!imrr
L kn.L~.
1
~lu ~ ,i,i .Ib~J, Js.~.~.l,LlkJ"
A
BIPHASIC
m
40 mse¢
I~ULSES ,m
+
EOD
A
-20 I
~.I.,.1, ...... I,+~.L .....
-,-L~.~.L"
I Fig. 2. Effects of pulses of different durations and polarities on ampullary afferent. Effects of 150 msec, 15 msec and 200 #sec pulses of both polarities are shown. Stimulus amplitude is the same for 150 and 15 msec pulses but increased by factor of 3 for 200 #sec pulses. The absolute value of positive and negative pulses at any one duration is the same. Histograms below 200/+sec traces show the average effect of these stimuli. EOD occurrences are marked by dots. EOD has no effect if given when afferent is silenced following off of 150 msec outside positive stimulus or during 150 msec outside negative stimulus. As shown in the other traces a slight acceleration follows the EOD if the afferent is discharging at its resting rate. A histogram showing the effect of the EOD is shown at lower left. At lower right the effects of biphasic positive-negative and negative-positive pulses are shown. Each phase is of the same amplitude and duration as the monophasic pulses immediately above which produced clear effects on the discharge rate of the fiber. All histograms are based on 65-70 stimuli. Bars at right of each histogram indicate 20 counts/bin. As indicated by time calibration marks, the two top traces were taken at slower sweeps than the remaining records and histograms.
Mono- and biphasic pulses contain the same high frequencies, but the monophasic pulse also has a DC component.
This DC component
thus seems to be responsible
f o r t h e effect o f a b r i e f p u l s e o r o f t h e E O D o n t h e a m p u l l a r y a f f e r e n t . In
Gnathonemus petersii, t h e E O D is b i p h a s i c . D u r i n g t h e i n i t i a l p h a s e t h e a n -
t e r i o r e n d o f t h e e l e c t r i c o r g a n is p o s i t i v e , a n d c u r r e n t flows o u t o f t h e s k i n i n f r o n t o f t h e e l e c t r i c o r g a n w h e r e t h e r e c e p t o r s are. D u r i n g t h e s e c o n d p h a s e t h e a n t e r i o r
90
B
A II K ohm-eros
II K ohm-cms A 4 K ohm-cms
il
A
ilL
2 K ohm-cms
2 K ohm-cms
C
2 K ohm-cms
Fig. 3. Effects of resistivity on responses of ampullary and mormyromast afferents to the EOD. A : ampullary afferent recorded at the three resistivities indicated. OCcurrence of EOD indicated by arrowhead. B: upper two histograms: mormyromast afferent recorded at the two resisfivities indicated. Bottom histogram: responses of ampullary afferent to EOD shown for comparison at same time base as mormyromast response. C: one of two occasions on which slowing of ampullary afferent occurred following the EOD. Histograms in A and those of mormyromast in B are based on 205-215 EODs. Bottom histogram in B based on 600 and histogram in C is based on 700 EODs. The calibration mark in bottom histogram of B is at 25 counts/bin, other calibrations marks indicate 50 counts/ bin. Note expanded time calibration in B.
end is negative, and current flows in through the receptor area. Resistivity changes have a pronounced affect on the amplitude and waveform of the EOD2,1°. At high resistivities ( > 50 kl'lcm) the second phase is smaller in amplitude but longer in duration than the first. As resistivity is lowered, second phase amplitude becomes larger than that of the first phase and its duration becomes equally brief, in addition, at resistivities below 20 kflcm both phases show the effects of loading and the absolute amplitudes begin to fall steeply. One would expect that such changes in the EOD would affect a receptor's response. At resistivities between t0 and 13 k l l c m where most experiments were done the second, head negative, phase of the EOD is larger than the first, head positive, phase. The ratio of peak first phase to peak second phase voltage was 0.5-0.6. The net flow of current was therefore inward across the skin, as in the case of an outside positive pulse. The similarity between the effect of such a pulse and the EOD is therefore reasonable. Five ampullary afferents were recorded at a resistivity of 11 k~cm. Saline (0.9 ~ NaC1) was then added to lower the resistivity in one or two stages. In each case the response decreased with decreasing resistivity (Fig. 3A). At intermediate resistivities of about 4 kflcm, the responses were similar in form, though smaller in amplitude than at 11 kllcm. None of these fibers discharged in relation to the EOD
91 at resistivities between 1.7 and 2 kf~cm. Six additional ampullary afferents were recorded at 2 kf~cm only. Four of these gave no response, while two showed a moderate slowing of their discharge rate (Fig. 3). Such deceleration responses were not seen in any other ampullary afferents and seem paradoxical, since the second phase of the EOD is twice the amplitude of the first at this low resistivity, and the response should therefore be similar to the response to an outside positive pulse. Both afferents responded normally to long pulses of both polarities, and the receptors of both were located just in front of the electric organ at the posterior margin of the specialized electroreceptor epithelium. A possible explanation of the slowing is that local currents may have differed from the overall current. Responses of mormyromast afferents also fell in amplitude with decreasing resistivity. The number of spikes/EOD decreased and the latency increased (Fig. 3B, 2 upper histograms). This was seen in 8 mormyromast afferents. In 4 of these cases we succeeded in holding the fiber while decreasing the resistivity to the point where the response disappeared (2-4 kf~cm) and also while increasing it back to the original value. In each case the response returned and was the same as the initial response. Three ampullary afferents were recorded after increasing the resistivity to 60 k ~ c m . All were clearly related to the EOD, and the form of the response was the same as that seen at resistivities of 10-13 kf~cm, i.e. an acceleration-deceleration sequence. At the higher resistivity the second phase of the EOD was lower in amplitude than the first though broader in duration. The areas under the curves were nearly equal (first/second = 0.93) indicating only a small DC component in the outside positive direction. The fact that the response was nevertheless strong and similar to that following an outside positive pulse suggests that the frequencies contained in the two phases may be important as well as the DC components. The finding that ampullary afferents are affected by the EOD raises the question of involvement of these receptors in active electrolocation. We tested this possibility by recording EOD responses from 10 ampullary afferents before and after placing non-conducting objects near the receptor pores. We used rectangles of thin plastic, either 8 /, 20 × 0.3 mm or 12 ~< 35 >~ 0.3 ram. We tried centering the rectangle on the pore, and also placing it just anterior or just posterior. The response of ampullary afferents was unaffected by such objects even when placed within one millimeter of the skin. Lack of effect was judged by comparing histograms of the response before and after placing the object. In order to maximize the possibility of seeing an effect without damaging the receptor we also placed squares of plastic film (Saran Wrap) 20 /. 20 mm on the skin, centering them on the receptor. The plastic film was in contact with the skin at several points and at no point was it more than I m m away. In one of the three fibers in which this was done the plastic film seemed to have a slight effect (Fig. 4, left side). In the middle histogram, made while the film was over the receptor, there is a slight broadening of the acceleration portion of the response and a slight delay and reduction in the deceleration portion. The effect is admittedly marginal and was obtained under extreme conditions. It was, however, the only occasion on which an object appeared to have some influence and its inclusion illustrates the near insensitivity of the ampullary afferent response to nearby non-conducting
92
AMPULLARY
MORMY~ST
no object
J
object present L ___
no object
i
EOD
50 msec
EOD
m
5 msec
Fig. 4. Effects of objects on responses of ampullary and mormyromast efferents to the EOD. For both receptor types, the top histograms were made before introducing the object, the middle histograms while the object was present and the bottom histograms after removing the object. For the ampultary afferent the object was a 20 x 20 mm square of plastic film resting lightly on the skin and centered on the receptor. For the mormyromast it was an 8 x 20 x 0.3 mm plastic plate held 3-4 mm from the skin, also centered over the receptor. The few bins with reduced counts immediately after the EOD in histograms of the ampullary response are an artifact, due to rejecting some ampullary spikes along with the spikes of a simultaneously recorded mormyromast afferent. Note the different time bases of ampullary and mormyromast histograms. Ampullary histograms based on 225 to 235 EODs each; mormyromast histograms on 165 to 185 EODs. Vertical calibrabion marks indicate 50counts/ bin for ampuUary and 100 counts/bin for mormyromast histograms.
objects. C a r b o n rods, which are highly conductive, were tried. Effects o f the rods o n the E O D response were also slight, b u t greater t h a n those o f n o n - c o n d u c t i n g objects. H o w e v e r , the c a r b o n rods a l t e r e d the s p o n t a n e o u s discharge rate, p r e s u m a b l y because o f the interface currents which t h e y g e n e r a t e l L Such rate changes could, in turn, affect the response, a n d the c a r b o n r o d results were therefore j u d g e d a m b i g u o u s with r e g a r d to the question o f active electrolocation. F o r c o m p a r i s o n , m o r m y r o m a s t afferents were also r e c o r d e d a n d e x a m i n e d for the effect o f n e a r b y plastic objects on their response to the E O D . The c o n t r a s t with a m p u l l a r y afferents was striking. Plastic rectangles smaller t h a n those tried with the a m p u l l a r y afferents (e.g. 5 x 13 mm) h a d a clear influence at distances f r o m the skin o f 3-5 mm. I f the rectangle was centered over the pore the response was decreased, if placed at either e n d o f the p o r e it was enhanced, as described by S z a b o a n d Hagiw a r a 24. Such b e h a v i o r was seen in all 12 o f the m o r m y r o m a s t afferents which were
93 examined in isolation (Fig. 4, on the right). On three occasions mormyromast afferents were recorded simultaneously with ampullary afferents. Plastic rectangles close to the ampullary pore had no effect on the ampullary afferent but strongly affected the mormyromast afferent, causing an increase in latency and a reduction in the number of spikes/response. DISCUSSION Ampullary afferents in Gnathonemuspetersii have been recorded in two different studies and found unresponsive to the EOD15, 91. These negative results may have been due to a low external resistivity. In addition, in one of the studies where mechanical sensitivity of ampullary receptors was the major interest, only afferents from the chin appendage were examined 21. Such anterior receptors are less likely to be affected by the EOD than those in the posterior part of the animal where we recorded. This is because at low resistivities the EOD-induced voltage across the skin probably decreases quite steeply in the posterior to anterior direction 2. In a study on Gymnotus carapo, Suga saw an ampullary receptor discharging in relation to the EOD 20. Although this was the exception, different recording conditions, such as higher resistivity, might increase the responsiveness. Afferents which appear to have been ampullary were recorded by Hagiwara et al. 9 in Electrophorus electricus. These afferents responded strongly to the low voltage EOD of that fish which is 20-40 V in amplitude, entirely monophasic and lasts 2-3 msec. On the basis of our results one would expect ampullary receptors to be influenced by such a discharge. One would also expect ampullary afferents to respond to the EOD in other fish with broad monophasic discharges, such as Mormyrus rume and some species of Hypopomus 4. The decrease in the responses to the EOD of both ampullary and mormyromast afferents at low resistivities can be attributed to a decrease in the voltage across the receptor during the discharge. This decrease has two causes: one is the decreased voltage produced by the electric organ at low resistivities, due to loading; the second is a more rapid spatial decay of transcutaneous voltage in front of the electric organ, due to the cable properties of the fish and the lower external resistivity 2. An alternative explanation is that the NaC1 added to reduce resistivity had a direct depressive effect on the receptor. Such depressive effects of NaCI have been seen by Roth in ampullary receptors of catfish~L In our experiments the final concentration of NaC1 at the lowest resistivity was about 4 mM/liter. At such concentrations Roth found the spontaneous discharge frequency to be reduced to less than one-half its original value. The response to formerly effective long duration (700 msec) stimuli disappeared even when intensity was increased by a factor of 10. In our experiments, however, there was no significant change in spontaneous activity. For example, the fiber in Fig. 3A discharged at rates of 92/sec, 99/sec and 88/sec in resistivities of 11, 4 and 2 k~, respectively. Furthermore, long duration stimuli remained effective, although higher intensities were necessary because of the lowered resistance between the two stimulating electrodes, both of which were in the bath (see Methods). It thus seems unlikely
94 that the loss of responsiveness to the EOD was due to a direct depressive effect of NaCI on the receptor. The fact that responses of mormyromast afferents were similarly reduced also argues against a direct effect. Epithelial tissue is present between mormyromast receptor cells and the external medium and this would probably reduce the diffusion of ions 22. Dependence on resistivity of the EOD-induced voltage across the receptor is important for active electrolocation. The mormyromast which is almost certainly the main receptor involved (see below) appears to have a rather narrow operating range. Szabo and Hagiwara 24 found that, once threshold was crossed, the response increased rapidly, reaching a maximum at less than two times threshold. The narrow operating range results in a high incremental sensitivity useful in active electrolocation. But it requires that the operating point be kept within relatively close limits. It is therefore of interest that the peak to peak voltage of the EOD, as recorded between a point on the anterior skin and a distant electrode, varied relatively little with resistivity changes between 20 and 180 kf~cm 2 and that Gosse s found that G. peterMi live in resistivities between t3 and 180 kD,cm. The EOD voltage decreases sharply as resistivity falls below 20 kllcm 2 and in this range the responses of both ampullary and mormyromast receptors also decreased with decreasing resistivity. At some resistivity below 2 kf~cm one would expect mormyromast as well as ampuUary afferents to stop responding and active electrolocation to become impossible. The question of which receptor class is primarily responsible for active electrolocation is important in the physiology of these fish. Knollenorgans are not believed to play a role 5,2a because their responses, once threshold is crossed, vary only slightly with stimulus intensity. Individual fibers could not signal the small variations in EOD-induced voltage upon which active electrolocation must depend. Furthermore their threshold is extremely low so they probably all discharge with each EOD, at least at normal resistivities. The proportion of the receptor population responding to each EOD doesn't vary, therefore, and could not be used to signal presence of an object. Finally, command associated inhibition of the central cells on which Knollenorgan afferents end is consistent with lack of any role in active electrolocation ~6. Several authors have suggested a communication role for Knollenorgan receptors r',~3. The responses of both ampullary and mormyromast afferents are graded with stimulus intensity. The markedly different effects of external objects on responses of the two types, however, indicates that mormyromast and not ampullary afferents are primarily responsible for active electrolocation. The reason for the different effects of objects is not clear, however. As discussed above, lowering resistivity lowers the amplitude of the EOD, and this in turn causes a reduced response in the afferents from both receptor types (Fig. 3). One concludes that ampullary and mormyromast afferents both go from strong to minimal responses over roughly similar ranges of EO D-induced transcutaneous voltages. But the striking differences in the effects of objects on responses of the two receptor types are difficult to reconcile with this conclusion (Fig. 4). A nearby non-conducting object reduces current flow through the skin immediately beneath it and this results in a lower EOD-induced voltage across the skin at that point. If responses of the two receptor types are equally affected by resistivity-induced
95 voltage changes, they should also be affected equally by object induced changes (cf. Figs. 3 and 4). A possible explanation may be found in the difference in the effective stimulus which drives the two types of receptors. The DC component of a pulse or EOD results in a residual charge remaining on the capacitance of skin and receptor. This residual, relatively long-lasting voltage is probably the effective stimulus for the ampullary receptor. The anterior body of the fish can be seen as a passive cable with resistance and capacitance in parallel across the skin. In such a cable, low frequencies decrement more slowly with distance than do high frequencies. This means that the low frequency residual voltage will spread with little decrement from one skin region to another. At any one point the low frequency components of the transcutaneous voltage would represent the average of such voltages over a broad area. The result would be minimal local variations and minimal effects of local changes in external resistance. In contrast, the high frequency components to which the mormyromast is sensitive decrement rapidly, permitting sharp local differences and allowing small external objects to have effects. Other considerations also suggest that mormyromast rather than ampullary afferents play the major role in active electrolocation. (l) The mormyromast is insensitive to low frequencies and thus filters out non-EOD sources which could interact with EOD voltage and be noise to an active electrolocation system. The marked effect of extraneous stimuli on the response of an ampullary afferent to the EOD was noted in Figs. 1 and 2. (2) The lack of spontaneous activity in mormyromasts not only enhances the signal to noise ratio but also ensures that the receptor will be fully recovered when an EOD occurs. (3) Variability in the response to individual EODs in the absence of any environmental change would be undesirable for active electrolocation. The greater the variability, the larger the number of EODs and the longer the time required to detect a change in a receptor's response following an environmental change. The low variability of the mormyromast in comparison to the ampullary afferents suggests therefore that these receptors are better suited for electrolocation. (4) The response of the ampullary receptor does not reach a peak until 20 msec after the EOD, while the first spike of the mormyromast occurs at less than 5 msec and the entire response is usually finished by 15 msec. The shorter latency could well be important in reducing reaction time. ACKNOWLEDGEMENTS We thank Jaqueline Bolen for typing the manuscript and its drafts. The work was supported by Grants from N1H (NINCDS-06728) and NSF (BMS 73-06867.)
REFERENCES I Bastian, J., Frequency response characteristics of electroreceptors in weakly electric fish (Gymnotoidei) with a pulse discharge, J. comp. Physiol., 112 (1976) 165 180. 2 Bell, C. C., Bradbury, J. and Russell, C. J., The electric organ of a mormyrid as a current and voltage source, J. eomp. Physiol., 110 (1976) 65 88.
96 3 Bennett, M. V. L., Electroreceptors in Mormyrids, Cold. Spr. Harb. Symp. quant. Biol., 30 (1965) 245-262. 4 Bennett, M. V. L., Electric organs. In W. S. Hoar and D. S. Randall (Eds.), Fish Physiology, Academic Press, New York, 1971, pp. 347-491. 5 Bennett, M. V. L., Electroreception In W. S. Hoar and D. S. Randall (Eds.), l~sh Physiology, Academic Press. New York, 1971, pp. 493-574: 6 Dunning, B.D., A Quantitative and Comparative Analysis of the Tonic Electroreeeptors o]' Gnathonemus, Gymnotus and Kyrptopterus, Ph.D. Thesis, University of Minnesota, 1973. 7 Fessard, A., (Ed.), Electroreceptors and Other Specialized Receptors in Lower Vertebrates, Vol. Ilt/3, Handbook of Sensory Physiology, Springer-Verlag, Berlin, 1974. 8 Gosse, J. P., Le milieu aquatique et I'ecologie des poissons dons la region de Yangambi, Sci. Zool., 116 (1963) 113-249. 9 Hagiwara, S., Szabo, T. and Enger, P. S., Physiological properties of electroreceptors in the electric eel, Electrophorus eleetrieus, J. Neurophysiol., 28 (1965) 775-783. 10 Harder, W., Schief, A. und Uhlemann, H., Zur Funktion des elektrischen Organs yon Gnathonemus petersii (Gthr 1862) (Mormyriformes, Teleostei), Z. Vergl. Physiol., 48 (t964) 302-331. 11 Hopkins, C. D., Stimulus filtering and electroreception: tuberous electroreceptors in three species of gymnotoid fish, J. comp. Physiol., III (1976) 171-207. 12 Kalmijn, A. J., The detection of electric fields from inanimate and animate sources other than electric organs. In A. Fessard (Ed.), Handbook of Sensory Physiology, Vol. 111/3, Springer, New York, 1974, pp. 147-200. 13 Knudsen, E., Behavioral thresholds to electric signals in high frequency electric fish, J. comp. Physiol., 91 (1974) 333-353. 14 Peters, R. and Buwalda, R., Frequency responses of the electroreceptors ('small pit organs') of the catfish, lctalurus nebulosus, J. comp. PhysioL, 79 (1972) 29-38. 15 Roth, A., Propri6t6s functionelles et morphologiques des differents organes de la ligne laterale des Mormyrides, J. PhysioL (Paris), 59 (1967) 486. 16 Roth, A., Central neurons involved in the e[ectroreception of the catfish Kyrptopterus, J. comp. Physiol., 100 (1975) 135-146. 17 Roth, A., Zur Funktionweise der Elektrorezeptoren in der Haut von welsen (lctalurus): der Einflus der Ionen in susswasser, Z. Vergl. Physiol., 75 (1971) 303-322. 18 Scheich, H., Bullock, T. H. and Hamstra, R. H., Coding properties of two classes of afferent nerve fibers: high frequency electroreceptors in the electric fish, J. Neuroph~TioL, 36 (1973) 39-60. 19 Scheich, H. and Bullock, T. H., The detection of electric fields from electric organs. In A. Fessard (Ed.), Electroreceptors and Other Specialized Receptors in Lower Vertebrates, Handbook of Sensoo' Physiology, Vol. 111/3, Springer, Berlin, 1974, pp. 201-256. 20 Suga, N., Etectrosensitivity of specialized and ordinary lateral line organs of the electric fish, Gymnotus carapo. In P. Cahn (Ed.), LateratLine Detectors, Bloomington, Ind., 1967, pp. 395-409. 21 Szabo, T., l~ber eine bisher unbekannte Funktion der sog. ampullaren Organe bei Gnathonemus petersii, Z. Vergl. Physiol., 66 (1970) 164-175. 22 Szabo, T., Anatomy of the specialized lateral line organs of electroreception. In A. Fessard (Ed.),
Electroreceptors and Other Specialized Receptors in Lower Vertebrates, Handboo~ of Sensory Phsyiology, Vol. 111/3, Springer, Berlin, 1974, pp. 13-56. 23 Szabo, T. and Fessard, A., Physiology of electroreceptors. In A. Fessard (Ed.), Electroreceptors and Other Specialized Receptor in Lower Vertebrates, Handbook of Sensory Physiology, Vol. 11l/3, Springer, Berlin, 1974, pp. 59-124. 24 Szabo, T. and Hagiwara, S., A latency-change mechanism involved in sensory coding of electric fish (mormyrids), Physiol. Behav., 2 (1967) 331-335. 25 Viancour, T. A., Neuronal response autorhythmicity associated with non-mechanical frequency filtering in the electrosense system of Eigenmannia Virescens, Neurosci. Abstr., 2 (1976) Abstr. 335. 26 Zipser, B. and Bennett, M. V. L., Interaction of electrosensory and electromotor signals in lateral line lobe of a morrnyrid fish, J. Neurophysiol., 39 (1976) 713-721.