- Email: [email protected]
State-dependent learning following electrical stimulation of the hippocampus: Intact and split-brain rats
Physiology & Behavior, Vol. 34, pp. 133-139. Copyright©Pergamon Press Ltd., 1985. Printed in the U.S.A.
0031-9384/84 $3.00 + .00
State-Dependent Learning Following Electrical Stimulation of the Hippocampus: Intact and Split-Brain Rats 1
D. C. MclNTYRE, R. J. STENSTROM, D. TAYLOR, K. A. S T O K E S A N D N. E D S O N D e p a r t m e n t o f Psychology, Carleton University, Ottawa, Canada, K I S 5B6 R e c e i v e d 27 A p r i l 1984 McINTYRE, D. C., R. J. STENSTROM, D. TAYLOR, K. A. STOKES AND N. EDSON. State-dependent learning following electrical stimulation of the hippocampus: Intact and split-brain rats. PHYSIOL BEHAV 34(1) 133-139, 1985.--In experiment 1, electrical stimulation of the posterior hippocampus was shown to produce state-dependent learning (SDL) for a step-out inhibitory avoidance task in rats. Stimulation sites in either the right or left hippocampus were equally effective in producing this effect. Similarly, the presence or absence of afterdischarge (AD) following the stimulation did not differentially affect performance on the task. In experiment 2, forebrain bisection ameliorated the behavioral deficits in the animals receiving stimulation before testing but not before training (N/S group), while those stimulated before training but not before testing (S/N group) remained impaired; thus, providing a demonstration of asymmetrical SDL. Variations in extent of the commissurotomy differentially affected the laterality of the afterdischarge but not the performance in the SDL task. Speculation as to the mechanisms of this SDL effect was presented. State-dependent learning
Memory
Retrieval
Hippocampus
L O W intensity electrical brain stimulation (EBS) has been used in many studies to provoke specific brain loci shortly following a training experience. It is presumed that if the locus is important for memory, such post-trial stimulation will interfere with the consolidation of the experience and degrade recall on the subsequent test trial the following day. Such an effect has been observed on the performance of an inhibitory avoidance response following post-trial stimulation o f the caudate nucleus [6, 9, 32, 33], the amygdala [1, 9, 13, 15] and the hippocampus [2, 14, 15, 30]. Consequently, EBS has been considered to be a valuable tool in the neuroanatomical, neurochemical dissection of the memory process [11]. An alternative to the consolidation-disruption hypothesis of post-trial treatment is that in some of these brain structures EBS may be producing state-dependent learning (SDL). The latter refers to the phenomenon where retrieval of information acquired under normal or altered-state conditions is most likely when the conditions imposed at the time of training are reinstated during testing [22]. EBS of the candate nucleus has been shown to produce SDL of an inhibitory avoidance response [18,25], a fact which could account for its amnestic potential. Such a S D L test of the amygdaia with EBS proved negative [18,25], suggesting the memory disruptive effects of post-trial anygdala EBS may represent true consolidation failure. The hippocampus also has been implicated in the memory process, primarily from amnesia rather than SDL studies. Recently, however, we have shown [29] that provocation of a kindled epileptic focus in the hippocampus will produce SDL. It was of theoretical interest,
Split-brain
Brain stimulation
therefore, to determine whether such stimulation in normal, non-kindled hippocampal sites is also able to provide the conditions capable of supporting SDL. EXPERIMENT 1 In this experiment, we examined EBS of a single locus in the right versus left posteroventral hippocampus of nonkindled, normal rats for its ability to support S D L in a stepout inhibitory avoidance paradigm. METHOD
Animals The subjects were 60 male Wistar rats (Woodlyn Farms, Guelph, Ont.), weighing 280-420 g at the time of surgery. All rats were housed singly with food and water available ad lib throughout the experiment.
Inhibitory Avoidance Apparatus Inhibitory avoidance (IA) training and testing occurred in a large wooden box (2' x 2' x 13.5" high), painted an even gray, with a brass grid floor, in which a smaller wooden box ( 8 x 6 x 1 3 . 5 inches high) was placed in one corner. The smaller box had a 3" semicircular opening in one wall which faced out into the larger box. A guillotine door restricted passage of an animal through the opening either out of or back into the smaller box. A 100 W bulb was located over the smaller box to endow it with a slight aversive property compared to the darker large box. The grid floor of the large box
XThis work was supported by an NSERC grant to D. C. Mclntyre.
133
McINTYRE El" AL.
134 was connected to a scrambler which delivered a 1.0 mA d.c. shock for five seconds when activated. PROCEDURE
Handling All rats were handled three times during the week preceding surgery for three minutes per session. During the week following surgery, all rats were handled twice during the baseline E E G recording sessions in the EBS box.
Surgery Animals were anesthetized with Nembutal (60 mg/kg, IP). Bipolar electrodes were implanted bilaterally into the hippocampi using the coordinates: 3.7 mm posterior to bregma, 5.1 mm lateral to the midline and 7.0 mm below the skull [24]. The electrodes were constructed from two twisted strands of 0.127 mm diameter Nichrome wire, Diamel insulated, and crimped into male Amphenol pins. The pins and a ground pin were inserted into a plastic headplug [20] and secured to the skull with jeweller's screws and dental acrylic.
Groups Ten days following surgery, the animals were assigned to one of seven groups. Three training-test groups were duplicated for both the right and left hippocampal EBS loci. These training-testing conditions were as follows: training and testing preceded by EBS (S/S); EBS before training but not before testing (S/N); EBS before testing but not before training (N/S); and a single control, to be compared to both the right and left hippocampal groups, receiving no stimulation before training or testing (N/N).
Stimulation On both training and test days, all rats were connected to the stimulation/recording equipment and placed in the small glass EBS box (12x 12x 12 inches) for a minimum of 45 seconds immediately prior to the beginning of an IA trial. If an animal was stimulated, it received a 60 Hz sine wave of 100 ~ A (peak-to-peak) intensity to the appropriate electrode (right or left hippocampus) for five seconds. If hippocampal afterdischarge (AD) was not provoked, the rat was removed 15 seconds later and placed in the IA chamber for immediate training or testing; i f A D was observed, the rat was removed 15 seconds after the termination of the AD and then placed in the IA chamber. On the non-stimulation trials, the animal received an additional 45 seconds in which only E E G recordings were taken before being placed in the IA chamber.
FIG. 1. Electrode placements for the left and right hippocampal stimulation sites (excluding the N/N group). Circles represent S/S animals; triangles, N/S animals; and squares, S/N animals.
Histology and Data Analysis IA data were analysed non-parametrically using the Mann-Whitney U Test, while AD durations were assessed with analysis of variance. The electrode placements were confirmed in 40 /z thick, frozen sections, mounted and stained with cresyl violet. RESULTS
IA Training and Testing Animals were placed singly in the small step-out box facing away from the guillotine door. When the animal turned around to face the door, the latter was opened and a stop watch started. The dependent measure was the length of time taken from the raising of the door until the rat placed all four paws on the grid floor of the larger box. At this point, during the training trial, five seconds of footshock was applied, following which the rat was returned to its home cage. No shock was administered 24 hours later after the test trial. On the test trial, if a rat remained in the small box for 500 seconds without fully stepping out, the trial was ended and a score of 500 seconds recorded.
The electrode placements for the right and left stimulation sites in the hippocampus are indicated in Fig. 1. There were no obvious differences in the locations of the electrode tips between the different experimental groups; therefore, the behavioral differences between groups likely reflected differences in group treatment. Examination of the IA training trial scores indicated that there were no reliable differences between the groups in their initial step-out latencies. Regarding the test trial, animals which received the same treatment prior to training as before testing (same-state groups, N/N and S/S) showed good recall of the IA response by exhibiting high test trial scores (see Fig. 2). This was true for both the right and left hemisphere
SDL AND HIPPOCAMPAL S T I M U L A T I O N
500
W
400 >-
¢J Z
500
•
F I
135
o
o o
I.-I
200
I00
i
N/N
I N/S
, S/N
S/S
LEFT HIPPO C A M P A L
N/S
S/N
S/S
the hippocampal commissure, which presumably accounts for the immediate appearance of AD contralateral to the stimulating electrode, as observed in experiment 1. We previously reported [28] that interference with the communication between the hemispheres via forebrain commissurotomy greatly changed the nature of the SDL resulting from right or left hippocampal kindled convulsions. We observed that the convulsions emanating from the left hemisphere continued to provide symmetrical SDL after commissurotomy, like intact animals [28,29], while those from the right hippocampus showed a complete recovery of all retrieval impairments in the altered-state groups. In the present experiment, we determined the effect of forebrain commissurotomy on the IA responses of left versus right hippocampal stimulated rats, following the previous SDL paradigm of experiment 1.
RIGHT LOCUS
FIG. 2. Median step-out latencies during training (black) and testing trials for all groups. Individual animals are indicated on the test trial bars. Those with AD are shown as filled circles and those without AD as open circles.
EBS loci. Those groups receiving a different treatment before training compared to testing (altered-state groups, S/N and N/S) showed significantly deleterious test trial performance compared to either the N/N or S/S groups. Considering those receiving EBS in the left hippocampus, the N/S group was inferior to the N/N group (p <0.003) and the S/S group (p<0.014), but not the S/N group. Similarly, the latter was different from the N/N group (p<0.001) and the S/S group (p<0.004). A roughly similar profile was exhibited by those receiving EBS of the right hippocampus, where the two altered-state groups, N/S and S/N, were different from the N/N group (p,<0.001) but not each other, while only the S/N group differed from the S/S group (p<0.05). This statistical tendency towards asymmetrical SDL in the right hippocampus and symmetrical SDL in the left hippocampus has been observed previously in rats with a kindled epileptic focus in the hippocampus [29]. Typically, asymmetrical SDL is indicated when the retention test scores of the two altered-state groups (N/S and S/N) is significantly different. In most stimulated rats, a brief duration bilateral AD was provoked which was similar from the right and left sites (mean seconds of AD---standard error in the right hippocampus=31.7_+5.1 seconds; left hippocampus=34.9_+5.0 seconds). The animals without AD, however, showed performance which was typical of their group, suggesting that stimulation per se was adequate to provide the state changes reported above (see Fig. 2). This same conclusion was reached when stimulating the caudate nucleus in a SDL paradigm [18]. These results indicate that stimulation of the posterior hippocampus is capable of providing the conditions necessary to support SDL, similar to the production of secondarily generalized kindled convulsions from this same area [28,29]. EXPERIMENT 2 The two hippocampi are richly interconnected through
METHOD
Animals
Sixty male Wistar rats were used in this experiment. All animals were housed singly and given food and water ad lib. All handling procedures were identical to experiment 1. Surgery
Surgery was performed in two stages. All rats were anesthetized, as before, and subjected to commissure bisection via an anterior approach [16]. To achieve this goal, a 2 mm wide groove was drilled in the exposed cranium on top of the midline suture from 8 to 13 mm anterior to bregma, until only a thin sheet of skull remained. A number 22 scalpel blade, its back ground down to the shape of a scythe, was inserted through the sheet of bone between the hemispheres, with the tip pointed caudalward. The blade tip was moved posterior approximately 12 mm, then ventral 4-5 mm to section the corpus callosum and hippocampal commissures, and withdrawn slowly, with a downward arc at the exit point to fully bisect the genu of the corpus callosum. The incision area was disinfected with penicillin G and closed with wound clips. Three weeks of recovery was allowed before animals were operated upon again for implantation of bilateral electrodes, as in experiment 1. RESULTS
Eleven rats were culled from experiment due to misplaced electrodes, poor health or off-midline bisections. The electrode tips of those remaining are presented in Fig. 3. There was no consistent difference between groups in terms of stimulation sites. The commissure transections were classified as belonging to one of three categories: total splits (n= 16), which included complete bisection of the corpus callosum and hippocampal commissure; extensive splits (n=24), which involved complete bisection of the hippocampal commissure but not all of the corpus callosum; and partial splits (n=9), where a portion of both the corpus callosum and hippocampal commissure remained intact. It would seem intuitive that the likelihood of producing animals in the "extensive" subgroup was very low, since separation of the corpus callosum from the hippocampal commissure seems improbable, yet this was our most common outcome (49% of the rats). Perhaps when the blade was inserted through the genu of the corpus callosum and intercepted the anterior hippocampal commissure, it was able to draw the commissure ventrally away
136
MclNTYRE El AL. As in experiment 1, the presence of AD tbllowing the EBS did not consistently affect the test trial performance of an animal, compared to those with no AD, again indicating that stimulation p e r s e is the critical variable in inducing the SDL deficits (see Fig. 5). Assessment of the AD durations revealed no differences between the right and left hippocampal sites (right hippocampus=62.0-+7.8 seconds; left hippocampus=69.9___7.1 seconds), however, the bisection did result in an exacerbation of the discharge compared to the intact rats of experiment l (p<0.001). The AD in the total splits, and most of the extensive splits, was unilateral, while the partial subgroup exhibited bilateral discharge. G E N E R A L DISCUSSION
FIG. 3. Electrode placements for the left and right hippocampal stimulation sites (excluding the N/N group) in the commissurotomized rats. Circles represent S/S animals; triangles, N/S animals; and squares, S/N animals.
from the corpus callosum and bisect it during the blade's posterior movement. Many of these rats (79%) additionally experienced bisection of either the anterior commissure and/or dorsal portion of the massa intermedia (thalamus) and the posterior commissure. An example of typical total and extensive splits is presented in Figs. 4A and B, respectively. Examination of the IA data indicated that there were no differences between the seven groups on their initial steprOUt latencies on the training trial. Inspection of the test trial scores revealed that forebrain bisection generally ameliorated the deficits of the altered-state groups, particularly the two N/S groups, compared to experiment 1. In the left hemisphere, the only impairment was in the S/N group (compared to the N/N, p <0.05), while in the right hemisphere, the S/N group was deficient compared to the N/N (p<0.003), S/S (p<0.006) and N/S (p<0.027) groups. No other comparisons were reliably different. These data are presented in Fig. 5. The extent of the bisection was not consistently associated with a unique behavioral outcome within groups, suggesting that partial interhemispheric dysfunction is as effective, in this context, as complete dysfunction (see Fig. 5).
In experiment 1, posterior hippocampal stimulation was clearly capable of providing the conditions necessary to support SDL in non-kindled, commissurally-intact rats. The absence of AD during the stimulation seemed to make little difference in the animal's test trial score, compared to others in its group showing AD, thus indicating that stimulation p e r s e is sufficient to produce the performance decrements in the altered-state groups. These observations have been shown similarly for EBS of the caudate nucleus [18]. There were no reliable performance differences between the same groups in the right versus left hemispheres, although the right hemisphere groups tended to show slightly inferior performance in all stimulated conditions. Clearly, the lateralized SDL seen in the kindled hippocampai experiments reported previously [28,29] was not observed in the present non-kindled preparation. This issue has been discussed elsewhere [28]. In experiment 2, forebrain bisection resulted in a general amelioration of the altered-state performance deficits seen in the intact rats of experiment 1. This abatement o f interference was virtually complete in the two N/S groups, while the two S/N groups continued to show inferior recall. Therefore, forebrain bisection resulted in a demonstration of asymmetrical SDL. Since EBS before the training trial only (S/N) had a more severe effect on recall than before the test trial only (N/S), this might suggest that retrieval of a memory rendered state dependent at the time of consolidation is harder to retrieve from storage than one consolidated in the normal state and retrieved during an altered state. On the other hand, the EBS before training simply might have degraded the memory such that, in a marginally different state (normal), retrieval was not good enough to show adequate performance on the test trial, while those experiencing training in the normal state had a more robust memory, and the marginally altered state produced by EBS was not sufficiently different to hinder the test trial recall. This proposal assumes, therefore, that EBS of the bisected hippocampus produces both a weak SDL effect and a weak consolidation-disruption effect, both of which act synergistically to degrade test trial performance in the S/N group (and in post-trial amnesia studies?), but only the weakened SDL component would be operative in the N/S group. The mechanism of the SDL or its asymmetry is unknown. One possibility, however, might relate to biochemical and/or hormonal interactions in the brain following the release of both peripheral and central catecholamines. For example, there has been considerable recent evidence that manipulation of adrenal epinephrine during training or testing can dramatically influence the probability of successful recall
SDL AND HIPPOCAMPAL STIMULATION
FIG. 4. Photomarcographs of representative split-brain subjects. (A) A total split; (B) An extensive split. See the text for descriptions.
137
138
MclNI'YRE E7 ALo
~u
500
bJ
>.c.D Z w i-
l
400
300
_J 200 Z C3 tO
ioo
~E
I N/N
N/S
S/N
SIS
LE F T H IPPOCANIPAL
N/S
SIN
SIS
RIGHT LOCUS
FIG. 5. Median step-out latencies during training (black) and testing trials for all groups. Individual animals are indicated on the test trial bars. Circles represent total splits; triangles, extensive splits; and squares, partial splits. See the text for descriptions. Filled characters indicate elicited AD, while open characters denote no AD.
during subsequent testing, providing both amnesia [3, 8, 23] and SDL [10] depending upon the paradigm. In fact, injection of epinephrine IP following training seems to produce its amnestic effect by a SDL mechanism, since subsequent pretrial injection at testing eliminated the performance deficit [10]. In addition, we have shown that adrenalectomy will attenuate both the amnestic [17] and SDL [19] properties of an amygdala kindled convulsion, even though the adrenalectomy had little influence on the brain seizure per se. It is interesting that the adrenalectomy did not completely abolish the SDL, but rather changed it from a symmetrical to asymmetrical form, where only those convulsed before training but not testing showed a recall deficit--like the bisected S/N subjects in the present report. The central adrenergic receptors seem to be involved also in a considerable portion of this memory modulation by providing both amnesia [4, 5, 12.31] and SDL [34]. Since both the central and peripheral
adrenergic mechanisms are likely involved in the memoryrecall process, perhaps in the present experiment, bisecting the forebrain dampened the magnitude of the inter-hemispheric (adrenergic?) response to the unilateral hippocampal EBS sufficiently to attenuate most of the recall deficit directly in the brain, or indirectly by changing the brain output to the adrenal system, or both. Such an indirect mechanism might be considered based on the following data: a chronic reduction of forebrain NE or adrenalectomy alone does not consistently influence acquisition or retention of the IA response, but when combined as a single treatment, the two manipulations result in a severe performance decrement [27]. This decrement has been shown to be substantially related to levels of corticosterone released from the adrenal glands during training [21], and that one of the target organs for this release is the hippocampus [261. Therefore, in our experiments, possibly unilateral EBS of the hippocampus altered adrenal corticosterone output which, subsequently, had a more iateralized influence upon the hippocampus in the rats with forebrain bisection compared to the intact rats. Whatever the precise mechanism of this SDL demonstration, these results suggest a degree of caution in the interpretation of amnesia in post-trial EBS studies as manifestations of a destroyed memory trace, rather than a retrieval deficit or conditioning effect. Although Gold and Zornetzer 161 have suggested recently that "administering a treatment prior to learning in an attempt to discover something fundamentally unambiguous regarding memory itself is at best complicated and at worst u s e l e s s , " such a pretrial treatment can tell you something about what the treatment does not do to the memory. In the present experiment, the pre-trial EBS treatrneut does not destroy the memory of the training experience (S/S groups), but renders it unavailable for recall when the training and testing conditions are not the same with regard to that treatment (N/S and S/N groups). Interestingly, Gold and Zornetzer [6] concluded their review by suggesting "qn normal memory formation the specific pattern of arousal present in the brain at the time of training may become an integral component of the stored information . . . and might need to be reproduced, or at least approximated, at the time of retrieval for stored information to be elaborated/" We think this statement, in its general lbrm, has been shown convincingly in the pre-trial hippocampal EBS experiments presented here, and for the caudate nucleus [18,25], both structures which show - a m n e s i a " to post-trial stimulation.
REFERENCES 1. Bresnahan, E. and A. Routtenberg. Memory disruption by unilateral low level sub-seizure stimulation of the medial amygdaloid nucleus. Physiol Behav 9: 513-525, 1972. 2. Brunner, R. L. and R. R. Rossi. Hippocampal disruption and passive avoidance behaviour. Psychonom Sci 15: 228--229, 1969. 3. de Almeida, M. A. M. K., F. P. Kapczinski and I. Izquierdo. Memory modulation by post-training intraperitoneal, but not introcerebroventricular, administration of ACTH or epinephrine. Behav Neural Biol 39: 277-283, 1983. 4. Ellis, M. E. and R. P. Kesner. The noradrenergic system of the amygdala and aversive information processing. Behav Neurosei 97: 399-415, 1983. 5. Gallagher, M. and B. S. Kapp. Effect of phentolamine administration into the amygdala complex of rats on time-dependent memory processes. Behav Neural Biol 31: 90-95, 1981. 6. Gold, P. E. and S. F. Zornetzer. The mnemon and its juices: neuromodulation of memory processes. Behav Neural Biol 38: 151-189, 1983.
7. Gold, P. E. and R. A. King. Caudate stimulation and retrograde amnesia: a threshold and gradient. Behav Biol 7: 709-715, 1972. 8. Gold, P. E., R. McCarty and D. B. Sternberg. Peripheral catecholamines and memory modulation. In: Neurobiology ~f Learning and Memory, edited by H. Matthies. New York: Raven Press, 1981. 9. Gold, P. E., J. Macri and J. L. McGaugh. Retrograde amnesia produced by subseizure amygdala stimulation. Behav Biol 9: 671-680, 1973. 10. Izquierdo, I. and R. D. Dias. Memory as a state-dependent phenomenon: role of ACTH and epinephrine. Behav Neural Biol 38: 144-149, 1983. I1. Kesner, R. P. Brain stimulation: effects on memory. Behav Neural Biol 36: 315-367, 1982. 12. Kesner, R. P. and M. E. Ellis. Memory consolidation: brain region and neurotransmitter specificity. Neurosei Lett 39: 295300, 1983.
SDL AND HIPPOCAMPAL STIMULATION 13. LePiane, F. G. and A. G. Phillips. Differential effects of electrical stimulation of amygdala, caudate-putamen or substantia nigra pars compacta on taste aversion and passive avoidance in rats. Physiol Behav 21: 979-985, 1978. 14. Lidsky, A. and B. M. Slotnik. Effects of post-trial limbic stimulation on retention of a one-trial passive avoidance response. J Comp Physiol Psychol 76: 337-348, 1971. 15. McDonough, J. H. and R. P. Kessner. Amnesia produced by brief electrical stimulation of the amygdala or dorsal hippocampus in cats. J Comp Physiol Psychol 77: 171-178, 1971. 16. McIntyre, D. C. Split-brain rat: transfer and interference of kindled amygdala convulsions. Can J Neurol Sci 2: 429--437, 1975. 17. Mclntyre, D. C. Adrenalectomy: protection from kindled convulsion induced amnesia in rats. Physiol Behav 17: 789-795, 1976. 18. McIntyre, D. C. and J. L. Gunter. State-dependent learning induced by low intensity electrical stimulation of the caudate or amygdala nuclei in rats. Physiol Behav 23: 449--454, 1979. 19. McIntyre, D. C. and P. D. Warm. Adrenalectomy: Protection from kindled cofivulsion-induced dissociation in rats. Physiol Behav 20: 469--474, 1978. 20. Molino, A. and D. C. McIntyre. Another inexpensive headplug for the electrical recording and/or stimulation of rats. Physiol Behav 9: 273-275, 1972. 21. Ogren, S.-O. and K. Fuxe. On the role of brain noradrenaline and the pituitary-adrenal axis in avoidance learning 1: Studies with corticosterone. Neurosci Lett 5: 291-296, 1977. 22. Overton, D. A. Experimental methods for the study of statedependent learning. Fed Proc 33: 1800-1813, 1974. 23. Palfai, T. and T. J. Walsh. The role of peripheral catecholamines in reserpine-induced amnesia. Behav Neural Biol 27: 423-432, 1979.
139 24. Pellegrino, L. J. and A. J. Cushman. A Stereotaxic Atlas of the Rat Brain. New York: Appleton-Century-Crofts, 1967. 25. Phillips, A. G. and F. G. LePaine. Differential effects of electrical stimulation of amygdala or caudate on inhibitory shock avoidance: a role for state-dependent learning. Behav Brain Res 2: 103-111, 1981. 26. Roberts, D. C. S. and F. E. Bloom. Adrenal steroid-induced changes in fl-adrenergic receptor binding in rat hippocampus. Eur J Pharmacol 74: 37-41, 1981. 27. Roberts, D. C. S. and H. C. Fibiger. Evidence for interaction between central noradrenergic neurons and adrenal hormones in learning and memory. Pharmacol Biochem Behav 7: 191-194, 1977. 28. Stokes, K. A. and D. C. McIntyre. Lateralized state-dependent learning produced by hippocampal kindled convulsions: effect of slit-brain. (Submitted). 29. Stokes, K. A. and D. C. McIntyre. Lateralized asymmetrical state-dependent learning produced by kindled convulsions from the rat hippocampus. Physiol Behav 26: 163-169, 1981. 30. Vardaris, R. M. and K. E. Schwartz. Retrograde amnesia for passive avoidance produced by stimulation of dorsal hippocampus. Physiol Behav 6: 131-135, 1971. 31. Walsh, T. J. and T. Palfai. Memory storage impairment or retrieval failure: pharmacologically distinguishable processes. Pharmacol Biochem Behav 11: 453-456, 1979. 32. Wyers, E. J. and S. A. Deadwyler. Duration and nature of retrograde amnesia produced by stimulation of caudate nucleus. Physiol Behav 6: 97-103, 1971. 33. Wyers, E. J., H. V. S. Peeke, J. S. Williston and M. J. Herz. Retroactive impairment of passive avoidance by stimulation of the caudate nucleus. Exp Neurol 22: 350/66, 1968. 34. Zornetzer, S. F., M. S. Gold and J. Hendrickson. Alphamethyi-p-tyrosine and memory: state-dependency and memory failure. Behav Biol 12: 135-141, 1974.
0031-9384/84 $3.00 + .00
State-Dependent Learning Following Electrical Stimulation of the Hippocampus: Intact and Split-Brain Rats 1
D. C. MclNTYRE, R. J. STENSTROM, D. TAYLOR, K. A. S T O K E S A N D N. E D S O N D e p a r t m e n t o f Psychology, Carleton University, Ottawa, Canada, K I S 5B6 R e c e i v e d 27 A p r i l 1984 McINTYRE, D. C., R. J. STENSTROM, D. TAYLOR, K. A. STOKES AND N. EDSON. State-dependent learning following electrical stimulation of the hippocampus: Intact and split-brain rats. PHYSIOL BEHAV 34(1) 133-139, 1985.--In experiment 1, electrical stimulation of the posterior hippocampus was shown to produce state-dependent learning (SDL) for a step-out inhibitory avoidance task in rats. Stimulation sites in either the right or left hippocampus were equally effective in producing this effect. Similarly, the presence or absence of afterdischarge (AD) following the stimulation did not differentially affect performance on the task. In experiment 2, forebrain bisection ameliorated the behavioral deficits in the animals receiving stimulation before testing but not before training (N/S group), while those stimulated before training but not before testing (S/N group) remained impaired; thus, providing a demonstration of asymmetrical SDL. Variations in extent of the commissurotomy differentially affected the laterality of the afterdischarge but not the performance in the SDL task. Speculation as to the mechanisms of this SDL effect was presented. State-dependent learning
Memory
Retrieval
Hippocampus
L O W intensity electrical brain stimulation (EBS) has been used in many studies to provoke specific brain loci shortly following a training experience. It is presumed that if the locus is important for memory, such post-trial stimulation will interfere with the consolidation of the experience and degrade recall on the subsequent test trial the following day. Such an effect has been observed on the performance of an inhibitory avoidance response following post-trial stimulation o f the caudate nucleus [6, 9, 32, 33], the amygdala [1, 9, 13, 15] and the hippocampus [2, 14, 15, 30]. Consequently, EBS has been considered to be a valuable tool in the neuroanatomical, neurochemical dissection of the memory process [11]. An alternative to the consolidation-disruption hypothesis of post-trial treatment is that in some of these brain structures EBS may be producing state-dependent learning (SDL). The latter refers to the phenomenon where retrieval of information acquired under normal or altered-state conditions is most likely when the conditions imposed at the time of training are reinstated during testing [22]. EBS of the candate nucleus has been shown to produce SDL of an inhibitory avoidance response [18,25], a fact which could account for its amnestic potential. Such a S D L test of the amygdaia with EBS proved negative [18,25], suggesting the memory disruptive effects of post-trial anygdala EBS may represent true consolidation failure. The hippocampus also has been implicated in the memory process, primarily from amnesia rather than SDL studies. Recently, however, we have shown [29] that provocation of a kindled epileptic focus in the hippocampus will produce SDL. It was of theoretical interest,
Split-brain
Brain stimulation
therefore, to determine whether such stimulation in normal, non-kindled hippocampal sites is also able to provide the conditions capable of supporting SDL. EXPERIMENT 1 In this experiment, we examined EBS of a single locus in the right versus left posteroventral hippocampus of nonkindled, normal rats for its ability to support S D L in a stepout inhibitory avoidance paradigm. METHOD
Animals The subjects were 60 male Wistar rats (Woodlyn Farms, Guelph, Ont.), weighing 280-420 g at the time of surgery. All rats were housed singly with food and water available ad lib throughout the experiment.
Inhibitory Avoidance Apparatus Inhibitory avoidance (IA) training and testing occurred in a large wooden box (2' x 2' x 13.5" high), painted an even gray, with a brass grid floor, in which a smaller wooden box ( 8 x 6 x 1 3 . 5 inches high) was placed in one corner. The smaller box had a 3" semicircular opening in one wall which faced out into the larger box. A guillotine door restricted passage of an animal through the opening either out of or back into the smaller box. A 100 W bulb was located over the smaller box to endow it with a slight aversive property compared to the darker large box. The grid floor of the large box
XThis work was supported by an NSERC grant to D. C. Mclntyre.
133
McINTYRE El" AL.
134 was connected to a scrambler which delivered a 1.0 mA d.c. shock for five seconds when activated. PROCEDURE
Handling All rats were handled three times during the week preceding surgery for three minutes per session. During the week following surgery, all rats were handled twice during the baseline E E G recording sessions in the EBS box.
Surgery Animals were anesthetized with Nembutal (60 mg/kg, IP). Bipolar electrodes were implanted bilaterally into the hippocampi using the coordinates: 3.7 mm posterior to bregma, 5.1 mm lateral to the midline and 7.0 mm below the skull [24]. The electrodes were constructed from two twisted strands of 0.127 mm diameter Nichrome wire, Diamel insulated, and crimped into male Amphenol pins. The pins and a ground pin were inserted into a plastic headplug [20] and secured to the skull with jeweller's screws and dental acrylic.
Groups Ten days following surgery, the animals were assigned to one of seven groups. Three training-test groups were duplicated for both the right and left hippocampal EBS loci. These training-testing conditions were as follows: training and testing preceded by EBS (S/S); EBS before training but not before testing (S/N); EBS before testing but not before training (N/S); and a single control, to be compared to both the right and left hippocampal groups, receiving no stimulation before training or testing (N/N).
Stimulation On both training and test days, all rats were connected to the stimulation/recording equipment and placed in the small glass EBS box (12x 12x 12 inches) for a minimum of 45 seconds immediately prior to the beginning of an IA trial. If an animal was stimulated, it received a 60 Hz sine wave of 100 ~ A (peak-to-peak) intensity to the appropriate electrode (right or left hippocampus) for five seconds. If hippocampal afterdischarge (AD) was not provoked, the rat was removed 15 seconds later and placed in the IA chamber for immediate training or testing; i f A D was observed, the rat was removed 15 seconds after the termination of the AD and then placed in the IA chamber. On the non-stimulation trials, the animal received an additional 45 seconds in which only E E G recordings were taken before being placed in the IA chamber.
FIG. 1. Electrode placements for the left and right hippocampal stimulation sites (excluding the N/N group). Circles represent S/S animals; triangles, N/S animals; and squares, S/N animals.
Histology and Data Analysis IA data were analysed non-parametrically using the Mann-Whitney U Test, while AD durations were assessed with analysis of variance. The electrode placements were confirmed in 40 /z thick, frozen sections, mounted and stained with cresyl violet. RESULTS
IA Training and Testing Animals were placed singly in the small step-out box facing away from the guillotine door. When the animal turned around to face the door, the latter was opened and a stop watch started. The dependent measure was the length of time taken from the raising of the door until the rat placed all four paws on the grid floor of the larger box. At this point, during the training trial, five seconds of footshock was applied, following which the rat was returned to its home cage. No shock was administered 24 hours later after the test trial. On the test trial, if a rat remained in the small box for 500 seconds without fully stepping out, the trial was ended and a score of 500 seconds recorded.
The electrode placements for the right and left stimulation sites in the hippocampus are indicated in Fig. 1. There were no obvious differences in the locations of the electrode tips between the different experimental groups; therefore, the behavioral differences between groups likely reflected differences in group treatment. Examination of the IA training trial scores indicated that there were no reliable differences between the groups in their initial step-out latencies. Regarding the test trial, animals which received the same treatment prior to training as before testing (same-state groups, N/N and S/S) showed good recall of the IA response by exhibiting high test trial scores (see Fig. 2). This was true for both the right and left hemisphere
SDL AND HIPPOCAMPAL S T I M U L A T I O N
500
W
400 >-
¢J Z
500
•
F I
135
o
o o
I.-I
200
I00
i
N/N
I N/S
, S/N
S/S
LEFT HIPPO C A M P A L
N/S
S/N
S/S
the hippocampal commissure, which presumably accounts for the immediate appearance of AD contralateral to the stimulating electrode, as observed in experiment 1. We previously reported [28] that interference with the communication between the hemispheres via forebrain commissurotomy greatly changed the nature of the SDL resulting from right or left hippocampal kindled convulsions. We observed that the convulsions emanating from the left hemisphere continued to provide symmetrical SDL after commissurotomy, like intact animals [28,29], while those from the right hippocampus showed a complete recovery of all retrieval impairments in the altered-state groups. In the present experiment, we determined the effect of forebrain commissurotomy on the IA responses of left versus right hippocampal stimulated rats, following the previous SDL paradigm of experiment 1.
RIGHT LOCUS
FIG. 2. Median step-out latencies during training (black) and testing trials for all groups. Individual animals are indicated on the test trial bars. Those with AD are shown as filled circles and those without AD as open circles.
EBS loci. Those groups receiving a different treatment before training compared to testing (altered-state groups, S/N and N/S) showed significantly deleterious test trial performance compared to either the N/N or S/S groups. Considering those receiving EBS in the left hippocampus, the N/S group was inferior to the N/N group (p <0.003) and the S/S group (p<0.014), but not the S/N group. Similarly, the latter was different from the N/N group (p<0.001) and the S/S group (p<0.004). A roughly similar profile was exhibited by those receiving EBS of the right hippocampus, where the two altered-state groups, N/S and S/N, were different from the N/N group (p,<0.001) but not each other, while only the S/N group differed from the S/S group (p<0.05). This statistical tendency towards asymmetrical SDL in the right hippocampus and symmetrical SDL in the left hippocampus has been observed previously in rats with a kindled epileptic focus in the hippocampus [29]. Typically, asymmetrical SDL is indicated when the retention test scores of the two altered-state groups (N/S and S/N) is significantly different. In most stimulated rats, a brief duration bilateral AD was provoked which was similar from the right and left sites (mean seconds of AD---standard error in the right hippocampus=31.7_+5.1 seconds; left hippocampus=34.9_+5.0 seconds). The animals without AD, however, showed performance which was typical of their group, suggesting that stimulation per se was adequate to provide the state changes reported above (see Fig. 2). This same conclusion was reached when stimulating the caudate nucleus in a SDL paradigm [18]. These results indicate that stimulation of the posterior hippocampus is capable of providing the conditions necessary to support SDL, similar to the production of secondarily generalized kindled convulsions from this same area [28,29]. EXPERIMENT 2 The two hippocampi are richly interconnected through
METHOD
Animals
Sixty male Wistar rats were used in this experiment. All animals were housed singly and given food and water ad lib. All handling procedures were identical to experiment 1. Surgery
Surgery was performed in two stages. All rats were anesthetized, as before, and subjected to commissure bisection via an anterior approach [16]. To achieve this goal, a 2 mm wide groove was drilled in the exposed cranium on top of the midline suture from 8 to 13 mm anterior to bregma, until only a thin sheet of skull remained. A number 22 scalpel blade, its back ground down to the shape of a scythe, was inserted through the sheet of bone between the hemispheres, with the tip pointed caudalward. The blade tip was moved posterior approximately 12 mm, then ventral 4-5 mm to section the corpus callosum and hippocampal commissures, and withdrawn slowly, with a downward arc at the exit point to fully bisect the genu of the corpus callosum. The incision area was disinfected with penicillin G and closed with wound clips. Three weeks of recovery was allowed before animals were operated upon again for implantation of bilateral electrodes, as in experiment 1. RESULTS
Eleven rats were culled from experiment due to misplaced electrodes, poor health or off-midline bisections. The electrode tips of those remaining are presented in Fig. 3. There was no consistent difference between groups in terms of stimulation sites. The commissure transections were classified as belonging to one of three categories: total splits (n= 16), which included complete bisection of the corpus callosum and hippocampal commissure; extensive splits (n=24), which involved complete bisection of the hippocampal commissure but not all of the corpus callosum; and partial splits (n=9), where a portion of both the corpus callosum and hippocampal commissure remained intact. It would seem intuitive that the likelihood of producing animals in the "extensive" subgroup was very low, since separation of the corpus callosum from the hippocampal commissure seems improbable, yet this was our most common outcome (49% of the rats). Perhaps when the blade was inserted through the genu of the corpus callosum and intercepted the anterior hippocampal commissure, it was able to draw the commissure ventrally away
136
MclNTYRE El AL. As in experiment 1, the presence of AD tbllowing the EBS did not consistently affect the test trial performance of an animal, compared to those with no AD, again indicating that stimulation p e r s e is the critical variable in inducing the SDL deficits (see Fig. 5). Assessment of the AD durations revealed no differences between the right and left hippocampal sites (right hippocampus=62.0-+7.8 seconds; left hippocampus=69.9___7.1 seconds), however, the bisection did result in an exacerbation of the discharge compared to the intact rats of experiment l (p<0.001). The AD in the total splits, and most of the extensive splits, was unilateral, while the partial subgroup exhibited bilateral discharge. G E N E R A L DISCUSSION
FIG. 3. Electrode placements for the left and right hippocampal stimulation sites (excluding the N/N group) in the commissurotomized rats. Circles represent S/S animals; triangles, N/S animals; and squares, S/N animals.
from the corpus callosum and bisect it during the blade's posterior movement. Many of these rats (79%) additionally experienced bisection of either the anterior commissure and/or dorsal portion of the massa intermedia (thalamus) and the posterior commissure. An example of typical total and extensive splits is presented in Figs. 4A and B, respectively. Examination of the IA data indicated that there were no differences between the seven groups on their initial steprOUt latencies on the training trial. Inspection of the test trial scores revealed that forebrain bisection generally ameliorated the deficits of the altered-state groups, particularly the two N/S groups, compared to experiment 1. In the left hemisphere, the only impairment was in the S/N group (compared to the N/N, p <0.05), while in the right hemisphere, the S/N group was deficient compared to the N/N (p<0.003), S/S (p<0.006) and N/S (p<0.027) groups. No other comparisons were reliably different. These data are presented in Fig. 5. The extent of the bisection was not consistently associated with a unique behavioral outcome within groups, suggesting that partial interhemispheric dysfunction is as effective, in this context, as complete dysfunction (see Fig. 5).
In experiment 1, posterior hippocampal stimulation was clearly capable of providing the conditions necessary to support SDL in non-kindled, commissurally-intact rats. The absence of AD during the stimulation seemed to make little difference in the animal's test trial score, compared to others in its group showing AD, thus indicating that stimulation p e r s e is sufficient to produce the performance decrements in the altered-state groups. These observations have been shown similarly for EBS of the caudate nucleus [18]. There were no reliable performance differences between the same groups in the right versus left hemispheres, although the right hemisphere groups tended to show slightly inferior performance in all stimulated conditions. Clearly, the lateralized SDL seen in the kindled hippocampai experiments reported previously [28,29] was not observed in the present non-kindled preparation. This issue has been discussed elsewhere [28]. In experiment 2, forebrain bisection resulted in a general amelioration of the altered-state performance deficits seen in the intact rats of experiment 1. This abatement o f interference was virtually complete in the two N/S groups, while the two S/N groups continued to show inferior recall. Therefore, forebrain bisection resulted in a demonstration of asymmetrical SDL. Since EBS before the training trial only (S/N) had a more severe effect on recall than before the test trial only (N/S), this might suggest that retrieval of a memory rendered state dependent at the time of consolidation is harder to retrieve from storage than one consolidated in the normal state and retrieved during an altered state. On the other hand, the EBS before training simply might have degraded the memory such that, in a marginally different state (normal), retrieval was not good enough to show adequate performance on the test trial, while those experiencing training in the normal state had a more robust memory, and the marginally altered state produced by EBS was not sufficiently different to hinder the test trial recall. This proposal assumes, therefore, that EBS of the bisected hippocampus produces both a weak SDL effect and a weak consolidation-disruption effect, both of which act synergistically to degrade test trial performance in the S/N group (and in post-trial amnesia studies?), but only the weakened SDL component would be operative in the N/S group. The mechanism of the SDL or its asymmetry is unknown. One possibility, however, might relate to biochemical and/or hormonal interactions in the brain following the release of both peripheral and central catecholamines. For example, there has been considerable recent evidence that manipulation of adrenal epinephrine during training or testing can dramatically influence the probability of successful recall
SDL AND HIPPOCAMPAL STIMULATION
FIG. 4. Photomarcographs of representative split-brain subjects. (A) A total split; (B) An extensive split. See the text for descriptions.
137
138
MclNI'YRE E7 ALo
~u
500
bJ
>.c.D Z w i-
l
400
300
_J 200 Z C3 tO
ioo
~E
I N/N
N/S
S/N
SIS
LE F T H IPPOCANIPAL
N/S
SIN
SIS
RIGHT LOCUS
FIG. 5. Median step-out latencies during training (black) and testing trials for all groups. Individual animals are indicated on the test trial bars. Circles represent total splits; triangles, extensive splits; and squares, partial splits. See the text for descriptions. Filled characters indicate elicited AD, while open characters denote no AD.
during subsequent testing, providing both amnesia [3, 8, 23] and SDL [10] depending upon the paradigm. In fact, injection of epinephrine IP following training seems to produce its amnestic effect by a SDL mechanism, since subsequent pretrial injection at testing eliminated the performance deficit [10]. In addition, we have shown that adrenalectomy will attenuate both the amnestic [17] and SDL [19] properties of an amygdala kindled convulsion, even though the adrenalectomy had little influence on the brain seizure per se. It is interesting that the adrenalectomy did not completely abolish the SDL, but rather changed it from a symmetrical to asymmetrical form, where only those convulsed before training but not testing showed a recall deficit--like the bisected S/N subjects in the present report. The central adrenergic receptors seem to be involved also in a considerable portion of this memory modulation by providing both amnesia [4, 5, 12.31] and SDL [34]. Since both the central and peripheral
adrenergic mechanisms are likely involved in the memoryrecall process, perhaps in the present experiment, bisecting the forebrain dampened the magnitude of the inter-hemispheric (adrenergic?) response to the unilateral hippocampal EBS sufficiently to attenuate most of the recall deficit directly in the brain, or indirectly by changing the brain output to the adrenal system, or both. Such an indirect mechanism might be considered based on the following data: a chronic reduction of forebrain NE or adrenalectomy alone does not consistently influence acquisition or retention of the IA response, but when combined as a single treatment, the two manipulations result in a severe performance decrement [27]. This decrement has been shown to be substantially related to levels of corticosterone released from the adrenal glands during training [21], and that one of the target organs for this release is the hippocampus [261. Therefore, in our experiments, possibly unilateral EBS of the hippocampus altered adrenal corticosterone output which, subsequently, had a more iateralized influence upon the hippocampus in the rats with forebrain bisection compared to the intact rats. Whatever the precise mechanism of this SDL demonstration, these results suggest a degree of caution in the interpretation of amnesia in post-trial EBS studies as manifestations of a destroyed memory trace, rather than a retrieval deficit or conditioning effect. Although Gold and Zornetzer 161 have suggested recently that "administering a treatment prior to learning in an attempt to discover something fundamentally unambiguous regarding memory itself is at best complicated and at worst u s e l e s s , " such a pretrial treatment can tell you something about what the treatment does not do to the memory. In the present experiment, the pre-trial EBS treatrneut does not destroy the memory of the training experience (S/S groups), but renders it unavailable for recall when the training and testing conditions are not the same with regard to that treatment (N/S and S/N groups). Interestingly, Gold and Zornetzer [6] concluded their review by suggesting "qn normal memory formation the specific pattern of arousal present in the brain at the time of training may become an integral component of the stored information . . . and might need to be reproduced, or at least approximated, at the time of retrieval for stored information to be elaborated/" We think this statement, in its general lbrm, has been shown convincingly in the pre-trial hippocampal EBS experiments presented here, and for the caudate nucleus [18,25], both structures which show - a m n e s i a " to post-trial stimulation.
REFERENCES 1. Bresnahan, E. and A. Routtenberg. Memory disruption by unilateral low level sub-seizure stimulation of the medial amygdaloid nucleus. Physiol Behav 9: 513-525, 1972. 2. Brunner, R. L. and R. R. Rossi. Hippocampal disruption and passive avoidance behaviour. Psychonom Sci 15: 228--229, 1969. 3. de Almeida, M. A. M. K., F. P. Kapczinski and I. Izquierdo. Memory modulation by post-training intraperitoneal, but not introcerebroventricular, administration of ACTH or epinephrine. Behav Neural Biol 39: 277-283, 1983. 4. Ellis, M. E. and R. P. Kesner. The noradrenergic system of the amygdala and aversive information processing. Behav Neurosei 97: 399-415, 1983. 5. Gallagher, M. and B. S. Kapp. Effect of phentolamine administration into the amygdala complex of rats on time-dependent memory processes. Behav Neural Biol 31: 90-95, 1981. 6. Gold, P. E. and S. F. Zornetzer. The mnemon and its juices: neuromodulation of memory processes. Behav Neural Biol 38: 151-189, 1983.
7. Gold, P. E. and R. A. King. Caudate stimulation and retrograde amnesia: a threshold and gradient. Behav Biol 7: 709-715, 1972. 8. Gold, P. E., R. McCarty and D. B. Sternberg. Peripheral catecholamines and memory modulation. In: Neurobiology ~f Learning and Memory, edited by H. Matthies. New York: Raven Press, 1981. 9. Gold, P. E., J. Macri and J. L. McGaugh. Retrograde amnesia produced by subseizure amygdala stimulation. Behav Biol 9: 671-680, 1973. 10. Izquierdo, I. and R. D. Dias. Memory as a state-dependent phenomenon: role of ACTH and epinephrine. Behav Neural Biol 38: 144-149, 1983. I1. Kesner, R. P. Brain stimulation: effects on memory. Behav Neural Biol 36: 315-367, 1982. 12. Kesner, R. P. and M. E. Ellis. Memory consolidation: brain region and neurotransmitter specificity. Neurosei Lett 39: 295300, 1983.
SDL AND HIPPOCAMPAL STIMULATION 13. LePiane, F. G. and A. G. Phillips. Differential effects of electrical stimulation of amygdala, caudate-putamen or substantia nigra pars compacta on taste aversion and passive avoidance in rats. Physiol Behav 21: 979-985, 1978. 14. Lidsky, A. and B. M. Slotnik. Effects of post-trial limbic stimulation on retention of a one-trial passive avoidance response. J Comp Physiol Psychol 76: 337-348, 1971. 15. McDonough, J. H. and R. P. Kessner. Amnesia produced by brief electrical stimulation of the amygdala or dorsal hippocampus in cats. J Comp Physiol Psychol 77: 171-178, 1971. 16. McIntyre, D. C. Split-brain rat: transfer and interference of kindled amygdala convulsions. Can J Neurol Sci 2: 429--437, 1975. 17. Mclntyre, D. C. Adrenalectomy: protection from kindled convulsion induced amnesia in rats. Physiol Behav 17: 789-795, 1976. 18. McIntyre, D. C. and J. L. Gunter. State-dependent learning induced by low intensity electrical stimulation of the caudate or amygdala nuclei in rats. Physiol Behav 23: 449--454, 1979. 19. McIntyre, D. C. and P. D. Warm. Adrenalectomy: Protection from kindled cofivulsion-induced dissociation in rats. Physiol Behav 20: 469--474, 1978. 20. Molino, A. and D. C. McIntyre. Another inexpensive headplug for the electrical recording and/or stimulation of rats. Physiol Behav 9: 273-275, 1972. 21. Ogren, S.-O. and K. Fuxe. On the role of brain noradrenaline and the pituitary-adrenal axis in avoidance learning 1: Studies with corticosterone. Neurosci Lett 5: 291-296, 1977. 22. Overton, D. A. Experimental methods for the study of statedependent learning. Fed Proc 33: 1800-1813, 1974. 23. Palfai, T. and T. J. Walsh. The role of peripheral catecholamines in reserpine-induced amnesia. Behav Neural Biol 27: 423-432, 1979.
139 24. Pellegrino, L. J. and A. J. Cushman. A Stereotaxic Atlas of the Rat Brain. New York: Appleton-Century-Crofts, 1967. 25. Phillips, A. G. and F. G. LePaine. Differential effects of electrical stimulation of amygdala or caudate on inhibitory shock avoidance: a role for state-dependent learning. Behav Brain Res 2: 103-111, 1981. 26. Roberts, D. C. S. and F. E. Bloom. Adrenal steroid-induced changes in fl-adrenergic receptor binding in rat hippocampus. Eur J Pharmacol 74: 37-41, 1981. 27. Roberts, D. C. S. and H. C. Fibiger. Evidence for interaction between central noradrenergic neurons and adrenal hormones in learning and memory. Pharmacol Biochem Behav 7: 191-194, 1977. 28. Stokes, K. A. and D. C. McIntyre. Lateralized state-dependent learning produced by hippocampal kindled convulsions: effect of slit-brain. (Submitted). 29. Stokes, K. A. and D. C. McIntyre. Lateralized asymmetrical state-dependent learning produced by kindled convulsions from the rat hippocampus. Physiol Behav 26: 163-169, 1981. 30. Vardaris, R. M. and K. E. Schwartz. Retrograde amnesia for passive avoidance produced by stimulation of dorsal hippocampus. Physiol Behav 6: 131-135, 1971. 31. Walsh, T. J. and T. Palfai. Memory storage impairment or retrieval failure: pharmacologically distinguishable processes. Pharmacol Biochem Behav 11: 453-456, 1979. 32. Wyers, E. J. and S. A. Deadwyler. Duration and nature of retrograde amnesia produced by stimulation of caudate nucleus. Physiol Behav 6: 97-103, 1971. 33. Wyers, E. J., H. V. S. Peeke, J. S. Williston and M. J. Herz. Retroactive impairment of passive avoidance by stimulation of the caudate nucleus. Exp Neurol 22: 350/66, 1968. 34. Zornetzer, S. F., M. S. Gold and J. Hendrickson. Alphamethyi-p-tyrosine and memory: state-dependency and memory failure. Behav Biol 12: 135-141, 1974.