- Email: [email protected]
Activity of the crayfish caudal photoreceptor submitted to a conditioning paradigm
Brain Research, 124 (1977) 44%456 © Elsevier/North-Holland Biomedical Press, Amsterdam - Printed in The Netherlands
449
ACTIVITY OF THE CRAYFISH C A U D A L P H O T O R E C E P T O R SUBMITTED TO A C O N D I T I O N I N G P A R A D I G M
M. ROGER* and C. GALEANO Laboratoire de Neurophysiologie et Psyehophysiologie**, Faeultd des Sciences Fondamentales et Appliqudes, Poitiers (France) and Ddpartement de Physiologie et Pharmacologic, Universitd de Sherbrooke, Sherbrooke, Qudbec (Canada)
(Accepted July 28th, 1976)
SUMMARY The activity of the crayfish caudal photoreceptor (Ph) was studied during the application of a classical conditioning paradigm. This cell is a primary photoreceptor and, at the same time, a secondary neuron in the mechanoreceptive pathway. The electrical stimulation of afferent mechanoreceptor fibers, the analog of the conditioned stimulus (ACS), produced a slight increase in the Ph firing rate. The photic stimulation, the analog of the unconditioned stimulus (AUS), produced a marked increase in the Ph firing rate. After ACS-AUS pairing, a decrease in the Ph firing rate appeared during the ACS-AUS interval, and it was considered the analog of the conditioned response (ACR). This inhibitory response gradually vanished during extinction. The application of ACS or AUS series alone did not produce a response similar to the ACR, this excluding sensitization as a possible cause. The ACR meets several conventional criteria accepted in learning, supporting the idea that the crayfish caudal photoreceptor is part of a small neuronal network able to learn in the isolated nerve cord preparation.
INTRODUCTION The analysis of neural learning mechanisms is extremely complex in higher vertebrates; for this reason it is desirable to resort to simplified preparations in which histological, biochemical and electrophysiological studies can be carried out in parallel to the investigation of behavior. * Visiting professor in Canada, Coop6ration Franco-Qu6becoise, Sous-commission h la recherche scientifique et technologique. ** ERA, CNRS, no. 07-0624.
450 This article deals with the study of the plastic capabilities of the crayfish caudal ganglion photoreceptors, whose activity shows long-term changes as a result of precise modifications in their sensory input¢,, v.I ~. These two photoreceptor cells i1,~:' are at the same time, interneurons in the mechanosensitive pathway from the tailg, ~z: moreover, they have a pacemaker. Thus, the photoreceptor output represents the integrated activity of photosensitive, synaptic and pacemaker membranes on the spike-generator membrane. The photoreceptor axon activity in the dark is characteristic in different preparations4,6,s,9,~v: (l) the "isolated" photoreceptors (the isolated caudal ganglion in a saline solution bath with all mechanoreceptor afferents cut), driven by the pacemaker, display a symmetric last and rhythmic firing, and (12) the photoreceptors connected to all mechanoreceptors (the caudal ganglion with intact roots) are submitted to the simultaneous action of ipsi- and contralateral mechanoreceptors. Excitation of ipsilateral fibers elicits a brief photoreceptor discharge followed by transitory inhibition and excitation of contralateral fibers elicits only strong photoreceptor inhibition 9. In this case the firing of both photoreceptors is symmetric, slower and more irregular than that of the "isolated" photoreceptors. The photoreceptor output during single photic stimulation shows a |ransitory increase of frequency and rhythmicity whereas repetitive photic stimulations elicit progressive attenuated responses suggesting habituation 6. Thus, the caudal ganglion fulfill the minimum requirements to be submitted to a conditioning paradigm : there are two cells, each of them having two sensory inputs (photic and from mechanoreceptors) and one output. MATERIAL AND METHODS Eleven isolated ventral nerve cord preparations (Fig. 1) which showed good records throughout the experiments were studied; they were obtained by dissecting 7 0 r c o n e c t e s virilis and 4 0 r c o n e c t e s rusticus 5-10 cm in length. The animals were immobilized in cold water and the ventral nerve cord was isolated taking care of not injuring the third superficial root of the lVth ganglion and the second root of the Vlth ganglion. The cord was placed in a bath containing renewed Van Harreveld solution 13 at pH 7.5-7.6 in a dark Faraday cage at 18-22 '~C. A fine bundle containing the photoreceptor axon was dissected in one hemicord in the 3-4 connective and placed on a monopolar electrode for recording. The third superficial root with tile motoneuron axons was placed on another monopolar electrode for recording. The second root of the Vlth ganglion with mechanoreceptor afferent fibers was placed on a bipolar electrode for stimulation. We used hook stainless steel electrodes lifted into a layer of mineral oil. Signal amplification, monitoring and display were standard : data were stored on analog magnetic tape for further analysis in a PDP-9 computer (Digital Equipment Corp.). We used the programs of Feldman and Roberge 2,3 for the detection and classification of spikes and storage of data as interspike intervals. In some experiments we used a Fabri-Tek 1062 computer and a Franz-Caldwell's instant ratemeter.
451 ANALOG
OF US ANALOG
/
~I. ~
OF CS
.~'
REC.
Fig. 1. Scheme of the 'isolated cord' preparation. The analog of the unconditioned stimulus was provided by a tungsten lamp focused on the Vlth ganglion with a light-guide. The analog of the conditioned stimulus was the electrical stimulation of the second root of the Vlth ganglion which had mechano-
receptor afferent fibers; only one was represented in thin line. The axon of one photoreceptor cell (in thick line) dissected in the hemicord of the 3~, connective, was placed on a hook electrode for recording. The third superficial root with the motoneuron axons was on another hook electrode. The experimental session was a replication of a classical conditioning schedule. The analog of the unconditioned stimulus (AUS) was a photic stimulus provided by a tungsten filament lamp focused on the VIth ganglion with a light guide, driven by a Grass stimulator, of 10-50 ft.-cd intensity and 1-5 sec duration. The photic stimulation was kept constant in each preparation throughout the experiment; it elicited a marked increase in the photoreceptor firing rate. The analog of the conditioned stimulus (ACS) was the electrical stimulus of the second root of the VIth ganglion, ipsilateral to the photoreceptor; the parameters were adjusted at the beginning of each experiment to cause a clear cut change in the activity of the motoneurons of the IIIrd superficial root, usually a brisk inhibition, and a minimal increase in the photoreceptor firing rate. The usual values were: train duration I-2.5 sec; pulse rate 20-50/sec; pulse duration 0.1-0.5 msec; intensity 3-11 V. These parameters were kept constant in each experiment and efficiency was controlled by the responses of motoneurons. During conditioning, the photic stimulation (AUS) was applied to the VIth ganglion 1-5 sec after the electrical stimulation of afferent fibers to the photoreceptor (ACS). The A C S - A U S interval was kept constant in each preparation. The set of both stimuli was a trial and the intertrial interval was 30 sec; a session consisted in a block of 90 or more trials. During extinction, the electrical stimulation of the second root was provided alone in blocks of 90-120 stimuli at the same intertrial interval. Sensitization was checked at the beginning and at the end of each experiment by the application of 20-100 isolated electrical stimuli at 30 sec interval.
452 A complete experiment including sensitization, conditioning, extinction, reconditioning, extinction and final sensitization, took about 8-10 h. The responsiveness of the ventral cord and the quality of records were optimized by a careful continuous perfusion with the saline solution and minimal manipulation of electrodes. RESU LTS
(1) Control of sensitization Photic stimulation of the Vlth ganglion. The repetitive stimulation elicited progressively attenuated responses during 10-20 trials, then the photoreceptor responses reached a stable level which was a fraction of that evoked by the first stimulus, as previously reported 6. There was never a clear cut evidence of post-stimulus inhibition (Fig. 2). Electrical stimulation of the second root of the Vlth ganglion. The repetitive stimulation of ipsilateral afferent fibers produced a minimum and transitory increase of the photoreceptor firing rate without evidence of post-stimulus inhibition (Fig. 3) even for hundreds of trials. The photoreceptor activity during the application of intermittent photic or electrical stimulation (section (1) above) was considered the 'baseline' activity.
(2) Conditioning The association of both stimuli produced progressive changes in the photoreceptor activity which was clearly different from the 'baseline' activity. At the beginning of the session, each stimulus elicited its characteristic response, without changes during the inter-stimuli interval (Fig. 4A). Then, just before the photic response, an inhibition of the photoreceptor appeared (analog of the conditioned response, ACR). The build-up of this inhibition increased and eventually stopped the firing for one second or more (Fig. 4B and C). The inter-stimuli firing rate showed a drop of about 75 o/.,oof the rate at the beginning of the session.
SENSITIZATION BASELINE ACTIVITY 100I i ~
:k'.
~
2"0
3b
4"0
-'J
.~4
1"o 5"0
30
SEC
Fig. 2. Instant firing rate of the photoreceptor responses to the analog of the unconditioned stimulus during the sensitization session. "Baseline" activity. Small numbers correspond to the photostimuli of one second duration. Note the attenuation of the responses at the beginning and the constance after the tenth stimulus. There was no inhibition previous to, or following the photoresponse, Abscissa: time in seconds. Ordinate: frequency per second.
453 SENSITIZATION BASELINE ACTIVITY 100"
7O
I . . . .
Ib
ACS
. . . .
2'o'
'
see
trials 1-10
GR5
Fig. 3. Post-stimulus histogram of 10 consecutive photoreceptor responses to the analog of the conditioned stimulus (ACS) during the sensitization session• 'Baseline' activity. Note the photoreceptor transitory response; after a few seconds the rate returned to control level• Abscissa: time in seconds; ordinate: number of spikes.
These clear-cut results were obtained in 5 preparations; 4 preparations exhibited less obvious modifications of the response and two preparations did not show changes during the inter-stimuli interval.
(3) Extinction When clear inhibition was obtained during the inter-stimuli interval, the photic stimulation was suspended and extinction started without delay, with the electrical stimulation alone. The response to trial one, which was not modified by the photic stimulation, showed an inhibition of 10-15 sec duration, taking the place of the previous photic response (Fig. 5). The latency of this inhibition remained unchanged at the beginning of the session and after 15-25 trials it became irregular; the intensity diminished progressively and disappeared in 50-100 trials. At this stage, the electrical CONDITIONING B
A
C
80,
60-
401;
,
,
,
3
•
,
ACS trials 91-100 GR 5
,
,
i
,
,
i
,
sec
I"
,
,
i
i
1
,
i
i
,
i
i , . . m m
AUS trials 171-180
trials 201-210
Fig. 4. Post-stimulus histogram of 3 blocks of 10 successive photoreceptor responses during 3 different stages of the conditioning session. Horizontal full line: analog of the conditioned stimulus (ACS). Horizontal dotted line: analog of the unconditioned stimulus (AUS). Note the development of inhibition during the interval ACS-AUS. Abscissa: time in seconds• Ordinate: number of spikes.
454 EXTINCTION 1i
•
2
m
•
//////////// / /// // /// / //////////////////////ItlN/ 3 m
~
4atom
,re
~~~~~~~~~~/~~~~~i/l~/////////////////tlt//H///llllltitlmt/// ~~~~~~ l ////I////l///ltl// "124 i
•
GRIt
Fig. 5. Inter-spike interval during the extinction session. Horizontal full lines: analog of the conditioned stimulus; the asterisk indicates the moment when the analog of the unconditioned stimulus was delivered during the conditioning session. The vertical bars signal the time of arrival of the spikes. Note the intense inhibition at the beginning of the session; after 124 trials it disappeared. stimulation (ACS) elicited only a slight and transitory increase of the photoreceptor firing rate, as in the 'baseline' control (see second paragraph of section (I)).
(4) Reconditioning After extinction, we started a new conditioning session. The inhibition during the inter-stimuli interval (ACR) was prominent at a very early stage, reaching the strongest level in 10-15 trials and keeping this intensity throughout the session. A second extinction abolished the inhibition. In order to investigate possible remaing effects on the photoresponse after the extinction series, we applied without delay a series of 30 photostimuli whose responses were now similar to those of the corresponding 'baseline' control (see first paragraph section (1)). DISCUSSION The most striking finding of this research is that the crayfish caudal photoreceptors are able to change their output when they are submitted to a classical conditioning paradigm. Was the inhibitory response reported during the conditioning session a conditioned response, and if so, was the photoreceptor cell able to learn': Learning occurs when the association of two stimuli gives rise to a new response or increase a previous one, either to its full value or partially, the final response is a certain function of the input-output pattern. Since we are dealing with electrophysiological responses in a small network, it is difficult to speak of "learning" in the usual behavioral context. Nevertheless, by analogy, we can speak of 'behavior of a neuron'~ insofar as it is also influenced by the present and past events which are capable of eliciting plastic changes in its functional characteristics l~In our experiments we associated the electrical stimulation of mechanoreceptor afferents to the Vlth ganglion (ACS) with the photostimulation (AUS). We found that after pairing, the response elicited by the ACS was an inhibition of the photoreceptor cells (ACR). Was this inhibitory response initially obtained at the onset of the ACS alone'? In other words, was this a case of sensitization ~'?
455 In behavioral studies the best control for sensitization is a group given both the CS and the US in unpaired random sequence. In our electrophysiological experiments we used, instead, a sequence of ACS, then a sequence of AUS (or in reverse order), before and after the pairing, in the same preparation. This schedule was chosen because the photoreceptor activity is submitted to changes according to the dark adaptational level of the cell. Therefore, with irregular timing in the application of the photic stimulus, the photoresponses became irregular. We considered it important to keep constant the adaptational level of the photoreceptor, in order to identify clearly any possible response due to training. It was easy to recognize a change in the firing rate beyond the 'baseline' activity, because we always used the same preparation as control. Sensitization cannot account for the results obtained, because the 'conditioned' response did not appear either before training (when both ACS and AUS were presented an equal number of times but never paired)l,10,14 or after extinction, when we repeated the isolated AUS and ACS series. Moreover, the inhibitory response (ACR) appeared when the sequence of pairing was ACS-AUS and not in the reverse case; the length of the ACS-AUS interval was also important, and no ACR was obtained when it was longer than 10 sec. The inhibitory response meets several of the conventional criteria accepted in learning: it appeared by precise pairing, disappeared by extinction, reappeared by reconditioning and disappeared again by new extinction. At this point, it could not be elicited either by a long series of ACS alone, or by an AUS series alone. The selected response to the ACS was a slight increase in the photoreceptor rate and the response obtained after training was an inhibition. We may now wonder whether it was a new response or one that existed in the previous repertoire of the photoreceptors. Kennedy 9 and Wilkens and Larimer 17 showed that the isolated stimulation of the second ipsilateral root of the Vlth ganglion elicited a dual effect on the photoreceptor: a short discharge followed by inhibition. Likewise, Galeano and Bdliveau 4 showed that the average influence of mechanoreceptors on ipsilateral photoreceptors during a period of several seconds was weakly inhibitory. The train of stimulation we used in this experiment did not show inhibitory effects during sensitization (as in section (1) second paragraph) or after extinction. It is difficult to understand how an AUS giving high rate responses can associate with the ACS and give, finally, an inhibitory ACR. Galeano and B61iveau 5 reported that the photoreceptors are not isolated elements into the ventral cord, but part of a network connected to motoneurons, interneurons and mechanoreceptor fibers. It is possible that connections from photoreceptor efferents to interneurons and then to mechano- and/or photoreceptors again are involved in the build-up of the inhibitory response. These experiments show that the application of the ACS and AUS in a strict temporal relationship creates a situation with the characteristics of a conditioned response, thus supporting the idea that the crayfish caudal photoreceptor cells are part of a small neuronal network able to learn. ACKNOWLEDGEMENTS We wish to thank Dr. E. Ramon Moliner (D6partement d'Anatomie, Universit6 de Sherbrooke) for comments on this manuscript.
456 REFERENCES I Eisenstein, E. M., The use of invertebrate systems for studies on the bases oflearning and memory. In G. C. Quarton, T. Melnechuk and F. O. Schmitt (Eds.), The Neurosciences: A Study Program, Rockefeller University Press, New York, 1967, pp. 653-665. 2 Feldman, J . F . , M~thodes d'Analyse des Impulsions Nerveuses, M. Sc. These, Universit6 de Montr6al, 1969. 3 Feldman, J. F. and Roberge, F. A., Computer detection, classification and analysis of neuronal spike sequences, InJbrm., 9 (1971 ) 185-~197. 4 Galeano, C. and Beliveau, S., Mechanoreceptor and photoreceptor tonic integration in the crayfish, Canad. J. PhysioL Pharmacol., 51 (1973) 949-958. 5 Galeano, C. and Beliveau, S., Functional connections of the crayfish caudal photorcceptor, Acf~ neurol. Lat.-amer., 20 (1974) 15-21. 6 Galeano, C. and Chow, K. L., Response of caudal photoreceptor of crayfish to continuous and intermittent photic stimulation, Canad. J. Physiol. Pharmacol., 49 (1971) 699-706. 7 Galeano, C., Zamfirescu, F. and Beliveau, S., Plasticity in the crayfish caudal ganglion, IRCS Med. Sci., 4 (1976) 103. 8 Herman, H. T. and Olsen, R. E., Afferent stochastic modulation of crayfish caudal photoreceptov units, J. gen. Physiol., 147 (1968) 209-217. 9 Kennedy, D., Physiology of photoreceptor neurons in the abdominal nerve cord of the crayfish, J. gen. Physiol., 46 0963) 551-572. l0 Miller, N. E., Certain facts of learning relevant to the search for its physical basis. In G. C. Quarton, T. Melnechuk and F. O. Schmitt (Eds.), The Neurosciences: A Study Program. Rockefeller University Press, New York, 1967, pp. 643-652. 11 Prosser, C. L., Action potentials in the nervous system of the crayfish. II. Response to illumination of the eye and caudal ganglion, J. cell. comp. Physiol., 4 (1934) 363-377. 12 Roger, M. et Galeano, C., Modification de la r6ponse du photor~epteur caudal de 1'6crevis~ durant une situation de conditionnement classique, J. Physiol. (Paris), 71 (1975) 308- 309A. 13 Van Harreveld, A., A physiological solution for fresh water crustacean, Proc Soc. e.~'p. Bh~l. (N. Y.), 34 (1936) 428~,32. 14 Wells, M. J., Evolution and associative learning. In P. N. R. Usherwood and D. R. Newth (Eds.). Simple Nervous Systems, Arnold, London, 1975, 445-473. 15 Welsh, J. H., The caudal photoreceptor and response of the crayfish to light, J. ceil. ~ , , p . Physiol., 4 (1934) 379-388. 16 Wiersrna, C. A.G., Behavior of neurons. In F. O. Schmitt and F. G. Worden (Eds.i, l'he Nelo'osciences: Third Study Program, MIT Press, Cambridge, Mass., 1974, 419-431. 17 Wilkens, L. A. and Larimer, J. L., The CNS photoreceptor of crayfish : morphology and synaptic activity, J. comp. Physiol., 80 (1972) 389 407.
449
ACTIVITY OF THE CRAYFISH C A U D A L P H O T O R E C E P T O R SUBMITTED TO A C O N D I T I O N I N G P A R A D I G M
M. ROGER* and C. GALEANO Laboratoire de Neurophysiologie et Psyehophysiologie**, Faeultd des Sciences Fondamentales et Appliqudes, Poitiers (France) and Ddpartement de Physiologie et Pharmacologic, Universitd de Sherbrooke, Sherbrooke, Qudbec (Canada)
(Accepted July 28th, 1976)
SUMMARY The activity of the crayfish caudal photoreceptor (Ph) was studied during the application of a classical conditioning paradigm. This cell is a primary photoreceptor and, at the same time, a secondary neuron in the mechanoreceptive pathway. The electrical stimulation of afferent mechanoreceptor fibers, the analog of the conditioned stimulus (ACS), produced a slight increase in the Ph firing rate. The photic stimulation, the analog of the unconditioned stimulus (AUS), produced a marked increase in the Ph firing rate. After ACS-AUS pairing, a decrease in the Ph firing rate appeared during the ACS-AUS interval, and it was considered the analog of the conditioned response (ACR). This inhibitory response gradually vanished during extinction. The application of ACS or AUS series alone did not produce a response similar to the ACR, this excluding sensitization as a possible cause. The ACR meets several conventional criteria accepted in learning, supporting the idea that the crayfish caudal photoreceptor is part of a small neuronal network able to learn in the isolated nerve cord preparation.
INTRODUCTION The analysis of neural learning mechanisms is extremely complex in higher vertebrates; for this reason it is desirable to resort to simplified preparations in which histological, biochemical and electrophysiological studies can be carried out in parallel to the investigation of behavior. * Visiting professor in Canada, Coop6ration Franco-Qu6becoise, Sous-commission h la recherche scientifique et technologique. ** ERA, CNRS, no. 07-0624.
450 This article deals with the study of the plastic capabilities of the crayfish caudal ganglion photoreceptors, whose activity shows long-term changes as a result of precise modifications in their sensory input¢,, v.I ~. These two photoreceptor cells i1,~:' are at the same time, interneurons in the mechanosensitive pathway from the tailg, ~z: moreover, they have a pacemaker. Thus, the photoreceptor output represents the integrated activity of photosensitive, synaptic and pacemaker membranes on the spike-generator membrane. The photoreceptor axon activity in the dark is characteristic in different preparations4,6,s,9,~v: (l) the "isolated" photoreceptors (the isolated caudal ganglion in a saline solution bath with all mechanoreceptor afferents cut), driven by the pacemaker, display a symmetric last and rhythmic firing, and (12) the photoreceptors connected to all mechanoreceptors (the caudal ganglion with intact roots) are submitted to the simultaneous action of ipsi- and contralateral mechanoreceptors. Excitation of ipsilateral fibers elicits a brief photoreceptor discharge followed by transitory inhibition and excitation of contralateral fibers elicits only strong photoreceptor inhibition 9. In this case the firing of both photoreceptors is symmetric, slower and more irregular than that of the "isolated" photoreceptors. The photoreceptor output during single photic stimulation shows a |ransitory increase of frequency and rhythmicity whereas repetitive photic stimulations elicit progressive attenuated responses suggesting habituation 6. Thus, the caudal ganglion fulfill the minimum requirements to be submitted to a conditioning paradigm : there are two cells, each of them having two sensory inputs (photic and from mechanoreceptors) and one output. MATERIAL AND METHODS Eleven isolated ventral nerve cord preparations (Fig. 1) which showed good records throughout the experiments were studied; they were obtained by dissecting 7 0 r c o n e c t e s virilis and 4 0 r c o n e c t e s rusticus 5-10 cm in length. The animals were immobilized in cold water and the ventral nerve cord was isolated taking care of not injuring the third superficial root of the lVth ganglion and the second root of the Vlth ganglion. The cord was placed in a bath containing renewed Van Harreveld solution 13 at pH 7.5-7.6 in a dark Faraday cage at 18-22 '~C. A fine bundle containing the photoreceptor axon was dissected in one hemicord in the 3-4 connective and placed on a monopolar electrode for recording. The third superficial root with tile motoneuron axons was placed on another monopolar electrode for recording. The second root of the Vlth ganglion with mechanoreceptor afferent fibers was placed on a bipolar electrode for stimulation. We used hook stainless steel electrodes lifted into a layer of mineral oil. Signal amplification, monitoring and display were standard : data were stored on analog magnetic tape for further analysis in a PDP-9 computer (Digital Equipment Corp.). We used the programs of Feldman and Roberge 2,3 for the detection and classification of spikes and storage of data as interspike intervals. In some experiments we used a Fabri-Tek 1062 computer and a Franz-Caldwell's instant ratemeter.
451 ANALOG
OF US ANALOG
/
~I. ~
OF CS
.~'
REC.
Fig. 1. Scheme of the 'isolated cord' preparation. The analog of the unconditioned stimulus was provided by a tungsten lamp focused on the Vlth ganglion with a light-guide. The analog of the conditioned stimulus was the electrical stimulation of the second root of the Vlth ganglion which had mechano-
receptor afferent fibers; only one was represented in thin line. The axon of one photoreceptor cell (in thick line) dissected in the hemicord of the 3~, connective, was placed on a hook electrode for recording. The third superficial root with the motoneuron axons was on another hook electrode. The experimental session was a replication of a classical conditioning schedule. The analog of the unconditioned stimulus (AUS) was a photic stimulus provided by a tungsten filament lamp focused on the VIth ganglion with a light guide, driven by a Grass stimulator, of 10-50 ft.-cd intensity and 1-5 sec duration. The photic stimulation was kept constant in each preparation throughout the experiment; it elicited a marked increase in the photoreceptor firing rate. The analog of the conditioned stimulus (ACS) was the electrical stimulus of the second root of the VIth ganglion, ipsilateral to the photoreceptor; the parameters were adjusted at the beginning of each experiment to cause a clear cut change in the activity of the motoneurons of the IIIrd superficial root, usually a brisk inhibition, and a minimal increase in the photoreceptor firing rate. The usual values were: train duration I-2.5 sec; pulse rate 20-50/sec; pulse duration 0.1-0.5 msec; intensity 3-11 V. These parameters were kept constant in each experiment and efficiency was controlled by the responses of motoneurons. During conditioning, the photic stimulation (AUS) was applied to the VIth ganglion 1-5 sec after the electrical stimulation of afferent fibers to the photoreceptor (ACS). The A C S - A U S interval was kept constant in each preparation. The set of both stimuli was a trial and the intertrial interval was 30 sec; a session consisted in a block of 90 or more trials. During extinction, the electrical stimulation of the second root was provided alone in blocks of 90-120 stimuli at the same intertrial interval. Sensitization was checked at the beginning and at the end of each experiment by the application of 20-100 isolated electrical stimuli at 30 sec interval.
452 A complete experiment including sensitization, conditioning, extinction, reconditioning, extinction and final sensitization, took about 8-10 h. The responsiveness of the ventral cord and the quality of records were optimized by a careful continuous perfusion with the saline solution and minimal manipulation of electrodes. RESU LTS
(1) Control of sensitization Photic stimulation of the Vlth ganglion. The repetitive stimulation elicited progressively attenuated responses during 10-20 trials, then the photoreceptor responses reached a stable level which was a fraction of that evoked by the first stimulus, as previously reported 6. There was never a clear cut evidence of post-stimulus inhibition (Fig. 2). Electrical stimulation of the second root of the Vlth ganglion. The repetitive stimulation of ipsilateral afferent fibers produced a minimum and transitory increase of the photoreceptor firing rate without evidence of post-stimulus inhibition (Fig. 3) even for hundreds of trials. The photoreceptor activity during the application of intermittent photic or electrical stimulation (section (1) above) was considered the 'baseline' activity.
(2) Conditioning The association of both stimuli produced progressive changes in the photoreceptor activity which was clearly different from the 'baseline' activity. At the beginning of the session, each stimulus elicited its characteristic response, without changes during the inter-stimuli interval (Fig. 4A). Then, just before the photic response, an inhibition of the photoreceptor appeared (analog of the conditioned response, ACR). The build-up of this inhibition increased and eventually stopped the firing for one second or more (Fig. 4B and C). The inter-stimuli firing rate showed a drop of about 75 o/.,oof the rate at the beginning of the session.
SENSITIZATION BASELINE ACTIVITY 100I i ~
:k'.
~
2"0
3b
4"0
-'J
.~4
1"o 5"0
30
SEC
Fig. 2. Instant firing rate of the photoreceptor responses to the analog of the unconditioned stimulus during the sensitization session. "Baseline" activity. Small numbers correspond to the photostimuli of one second duration. Note the attenuation of the responses at the beginning and the constance after the tenth stimulus. There was no inhibition previous to, or following the photoresponse, Abscissa: time in seconds. Ordinate: frequency per second.
453 SENSITIZATION BASELINE ACTIVITY 100"
7O
I . . . .
Ib
ACS
. . . .
2'o'
'
see
trials 1-10
GR5
Fig. 3. Post-stimulus histogram of 10 consecutive photoreceptor responses to the analog of the conditioned stimulus (ACS) during the sensitization session• 'Baseline' activity. Note the photoreceptor transitory response; after a few seconds the rate returned to control level• Abscissa: time in seconds; ordinate: number of spikes.
These clear-cut results were obtained in 5 preparations; 4 preparations exhibited less obvious modifications of the response and two preparations did not show changes during the inter-stimuli interval.
(3) Extinction When clear inhibition was obtained during the inter-stimuli interval, the photic stimulation was suspended and extinction started without delay, with the electrical stimulation alone. The response to trial one, which was not modified by the photic stimulation, showed an inhibition of 10-15 sec duration, taking the place of the previous photic response (Fig. 5). The latency of this inhibition remained unchanged at the beginning of the session and after 15-25 trials it became irregular; the intensity diminished progressively and disappeared in 50-100 trials. At this stage, the electrical CONDITIONING B
A
C
80,
60-
401;
,
,
,
3
•
,
ACS trials 91-100 GR 5
,
,
i
,
,
i
,
sec
I"
,
,
i
i
1
,
i
i
,
i
i , . . m m
AUS trials 171-180
trials 201-210
Fig. 4. Post-stimulus histogram of 3 blocks of 10 successive photoreceptor responses during 3 different stages of the conditioning session. Horizontal full line: analog of the conditioned stimulus (ACS). Horizontal dotted line: analog of the unconditioned stimulus (AUS). Note the development of inhibition during the interval ACS-AUS. Abscissa: time in seconds• Ordinate: number of spikes.
454 EXTINCTION 1i
•
2
m
•
//////////// / /// // /// / //////////////////////ItlN/ 3 m
~
4atom
,re
~~~~~~~~~~/~~~~~i/l~/////////////////tlt//H///llllltitlmt/// ~~~~~~ l ////I////l///ltl// "124 i
•
GRIt
Fig. 5. Inter-spike interval during the extinction session. Horizontal full lines: analog of the conditioned stimulus; the asterisk indicates the moment when the analog of the unconditioned stimulus was delivered during the conditioning session. The vertical bars signal the time of arrival of the spikes. Note the intense inhibition at the beginning of the session; after 124 trials it disappeared. stimulation (ACS) elicited only a slight and transitory increase of the photoreceptor firing rate, as in the 'baseline' control (see second paragraph of section (I)).
(4) Reconditioning After extinction, we started a new conditioning session. The inhibition during the inter-stimuli interval (ACR) was prominent at a very early stage, reaching the strongest level in 10-15 trials and keeping this intensity throughout the session. A second extinction abolished the inhibition. In order to investigate possible remaing effects on the photoresponse after the extinction series, we applied without delay a series of 30 photostimuli whose responses were now similar to those of the corresponding 'baseline' control (see first paragraph section (1)). DISCUSSION The most striking finding of this research is that the crayfish caudal photoreceptors are able to change their output when they are submitted to a classical conditioning paradigm. Was the inhibitory response reported during the conditioning session a conditioned response, and if so, was the photoreceptor cell able to learn': Learning occurs when the association of two stimuli gives rise to a new response or increase a previous one, either to its full value or partially, the final response is a certain function of the input-output pattern. Since we are dealing with electrophysiological responses in a small network, it is difficult to speak of "learning" in the usual behavioral context. Nevertheless, by analogy, we can speak of 'behavior of a neuron'~ insofar as it is also influenced by the present and past events which are capable of eliciting plastic changes in its functional characteristics l~In our experiments we associated the electrical stimulation of mechanoreceptor afferents to the Vlth ganglion (ACS) with the photostimulation (AUS). We found that after pairing, the response elicited by the ACS was an inhibition of the photoreceptor cells (ACR). Was this inhibitory response initially obtained at the onset of the ACS alone'? In other words, was this a case of sensitization ~'?
455 In behavioral studies the best control for sensitization is a group given both the CS and the US in unpaired random sequence. In our electrophysiological experiments we used, instead, a sequence of ACS, then a sequence of AUS (or in reverse order), before and after the pairing, in the same preparation. This schedule was chosen because the photoreceptor activity is submitted to changes according to the dark adaptational level of the cell. Therefore, with irregular timing in the application of the photic stimulus, the photoresponses became irregular. We considered it important to keep constant the adaptational level of the photoreceptor, in order to identify clearly any possible response due to training. It was easy to recognize a change in the firing rate beyond the 'baseline' activity, because we always used the same preparation as control. Sensitization cannot account for the results obtained, because the 'conditioned' response did not appear either before training (when both ACS and AUS were presented an equal number of times but never paired)l,10,14 or after extinction, when we repeated the isolated AUS and ACS series. Moreover, the inhibitory response (ACR) appeared when the sequence of pairing was ACS-AUS and not in the reverse case; the length of the ACS-AUS interval was also important, and no ACR was obtained when it was longer than 10 sec. The inhibitory response meets several of the conventional criteria accepted in learning: it appeared by precise pairing, disappeared by extinction, reappeared by reconditioning and disappeared again by new extinction. At this point, it could not be elicited either by a long series of ACS alone, or by an AUS series alone. The selected response to the ACS was a slight increase in the photoreceptor rate and the response obtained after training was an inhibition. We may now wonder whether it was a new response or one that existed in the previous repertoire of the photoreceptors. Kennedy 9 and Wilkens and Larimer 17 showed that the isolated stimulation of the second ipsilateral root of the Vlth ganglion elicited a dual effect on the photoreceptor: a short discharge followed by inhibition. Likewise, Galeano and Bdliveau 4 showed that the average influence of mechanoreceptors on ipsilateral photoreceptors during a period of several seconds was weakly inhibitory. The train of stimulation we used in this experiment did not show inhibitory effects during sensitization (as in section (1) second paragraph) or after extinction. It is difficult to understand how an AUS giving high rate responses can associate with the ACS and give, finally, an inhibitory ACR. Galeano and B61iveau 5 reported that the photoreceptors are not isolated elements into the ventral cord, but part of a network connected to motoneurons, interneurons and mechanoreceptor fibers. It is possible that connections from photoreceptor efferents to interneurons and then to mechano- and/or photoreceptors again are involved in the build-up of the inhibitory response. These experiments show that the application of the ACS and AUS in a strict temporal relationship creates a situation with the characteristics of a conditioned response, thus supporting the idea that the crayfish caudal photoreceptor cells are part of a small neuronal network able to learn. ACKNOWLEDGEMENTS We wish to thank Dr. E. Ramon Moliner (D6partement d'Anatomie, Universit6 de Sherbrooke) for comments on this manuscript.
456 REFERENCES I Eisenstein, E. M., The use of invertebrate systems for studies on the bases oflearning and memory. In G. C. Quarton, T. Melnechuk and F. O. Schmitt (Eds.), The Neurosciences: A Study Program, Rockefeller University Press, New York, 1967, pp. 653-665. 2 Feldman, J . F . , M~thodes d'Analyse des Impulsions Nerveuses, M. Sc. These, Universit6 de Montr6al, 1969. 3 Feldman, J. F. and Roberge, F. A., Computer detection, classification and analysis of neuronal spike sequences, InJbrm., 9 (1971 ) 185-~197. 4 Galeano, C. and Beliveau, S., Mechanoreceptor and photoreceptor tonic integration in the crayfish, Canad. J. PhysioL Pharmacol., 51 (1973) 949-958. 5 Galeano, C. and Beliveau, S., Functional connections of the crayfish caudal photorcceptor, Acf~ neurol. Lat.-amer., 20 (1974) 15-21. 6 Galeano, C. and Chow, K. L., Response of caudal photoreceptor of crayfish to continuous and intermittent photic stimulation, Canad. J. Physiol. Pharmacol., 49 (1971) 699-706. 7 Galeano, C., Zamfirescu, F. and Beliveau, S., Plasticity in the crayfish caudal ganglion, IRCS Med. Sci., 4 (1976) 103. 8 Herman, H. T. and Olsen, R. E., Afferent stochastic modulation of crayfish caudal photoreceptov units, J. gen. Physiol., 147 (1968) 209-217. 9 Kennedy, D., Physiology of photoreceptor neurons in the abdominal nerve cord of the crayfish, J. gen. Physiol., 46 0963) 551-572. l0 Miller, N. E., Certain facts of learning relevant to the search for its physical basis. In G. C. Quarton, T. Melnechuk and F. O. Schmitt (Eds.), The Neurosciences: A Study Program. Rockefeller University Press, New York, 1967, pp. 643-652. 11 Prosser, C. L., Action potentials in the nervous system of the crayfish. II. Response to illumination of the eye and caudal ganglion, J. cell. comp. Physiol., 4 (1934) 363-377. 12 Roger, M. et Galeano, C., Modification de la r6ponse du photor~epteur caudal de 1'6crevis~ durant une situation de conditionnement classique, J. Physiol. (Paris), 71 (1975) 308- 309A. 13 Van Harreveld, A., A physiological solution for fresh water crustacean, Proc Soc. e.~'p. Bh~l. (N. Y.), 34 (1936) 428~,32. 14 Wells, M. J., Evolution and associative learning. In P. N. R. Usherwood and D. R. Newth (Eds.). Simple Nervous Systems, Arnold, London, 1975, 445-473. 15 Welsh, J. H., The caudal photoreceptor and response of the crayfish to light, J. ceil. ~ , , p . Physiol., 4 (1934) 379-388. 16 Wiersrna, C. A.G., Behavior of neurons. In F. O. Schmitt and F. G. Worden (Eds.i, l'he Nelo'osciences: Third Study Program, MIT Press, Cambridge, Mass., 1974, 419-431. 17 Wilkens, L. A. and Larimer, J. L., The CNS photoreceptor of crayfish : morphology and synaptic activity, J. comp. Physiol., 80 (1972) 389 407.