Reversible Inactivation of the Cerebellar Interpositus Nucleus Completely Prevents Acquisition of the Classically Conditioned Eye-Blink Response David J

Downloaded from learnmem.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press

Reversible Inactivation of the Cerebellar Interpositus Nucleus Completely Prevents Acquisition of the Classically Conditioned Eye-Blink Response David J. Krupa 1 and Richard F. Thompson Neur0sciences Program University of S0uthem Calif0mia Los Angeles, California 90089-2520

Abstract These results confirm and extend the original findings that appropriate lesions Numerous studies from several (either temporary or permanent) of the laboratories report that temporary interpositus nucleus completely prevent inactivation of the cerebellar interpositus acquisition of the conditioned eye-blink nucleus and regions of overlying cortex response. Other issues regarding reversible during eye-blink conditioning completely inactivation studies are also discussed. prevents acquisition of the conditioned eye-blink response (CR) without affecting Introduction the ability to learn the CR in subsequent training without inactivation. Recently, Numerous lines of evidence, ranging in diver- these results have been challenged by the sity from purely theoretical considerations (Marr suggestion that learning was not completely 1969; Albus 1971) to human functional imaging blocked in these studies. Instead, it has been studies (Molchan et al. 1994; Logan and Grafton suggested that low levels of responses on 1995; Blaxton et al. 1996), have consistently impli- test sessions might represent a retarded cated the cerebellum as being critically involved in form of learning caused by drug effects on a number of learning related tasks (for instance, see cerebellar cortex. The present study was Lisberger 1988; Thatch et al. 1992; Thompson and designed to address this issue directly. Very Krupa 1994). Among the many tasks in which the low doses of muscimol were used to cerebellum is critically involved, a number of stud- selectively inactivate the interpositus ies from several laboratories have demonstrated nucleus of rabbits during five conditioning that the cerebellum is essentially involved in acqui- sessions. Animals performed no significant sition and expression of classically conditioned dis- levels of CRs during those sessions. Training crete skeletal movements, in particular, the classi- was continued four more sessions without cally conditioned eye-blink response. any inactivations to test whether any Briefly, the evidence supporting cerebellar in- learning had occurred during the previous volvement in eye-blink conditioning includes the five sessions. Detailed analysis of responses following: (1) Appropriate lesions of a restricted during session six revealed that learning region of lateral cerebellar hemisphere, including was completely blocked by the low doses of lesions limited to the anterior interpositus nucleus, muscimol infused into the interpositus completely and permanently abolish both acquisi- during the first five sessions. Animals tion and expression of conditioned eye-blink re- subsequently acquired the CR normally. sponses (CRs) without affecting performance of the reflexive, unconditioned response (UR) (Lin- coln et al. 1982; McCormick and Thompson 1984a; Lavond et al. 1985; Yeo et al. 1985; Steinmetz et al. 1Corresponding author. Present address: Department of Neurobiology, Duke University Medical Center, Durham, 1992). (2) Electrophysiological recordings of neu- North Carolina 27710. ronal activity from within this region of cerebellum

LEARNING & MEMORY 3:545-556 91997 by Cold Spring Harbor Laboratory Press ISSN1072-0502/97 $5.00

L E A R N / N G & M E M O R Y 545 Downloaded from learnmem.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press Krupa and Thompson

reveal populations of neurons that respond not Subsequent studies have largely ruled out this only to conditioned and unconditioned stimuli but second possibility. Reversibly inactivating the also during performance of CRs in a manner that brain-stem motor nuclei and surrounding reticular precedes and predicts their occurrence, suggesting formation responsible for performance of the CR a causal role (McCormick and Thompson 1984b; has no effect at all on the ability to acquire the CR Berthier and Moore 1986, 1990; Foy et al. 1992). but does block CR expression, thus ruling out (3) Electrical microstimulation of brain stem pre- these structures as possible loci for the memory cerebellar nuclei or fibers immediately afferent to trace (Zhang and Lavond 1991; Krupa et al. 1996). the cerebellum serves as an effective conditioned Similarly, inactivation of the magnocellular red stimulus (CS) or unconditioned stimulus (US), de- nucleus, a structure that receives direct input from pending on stimulus location. CRs elicited by these the interpositus and is essentially involved in eye- stimuli are completely abolished by appropriate, blink conditioning, also prevents CR expression focal lesions of lateral cerebellum (Mauk et al. with no effect on CR acquisition (Clark and Lavond 1986; Steinmetz et al. 1986; Lavond et al. 1987). 1993; Krupa et al. 1993). Finally, inactivation of all Collectively, these results demonstrate that the cer- of the output fibers of the interpositus nucleus dur- ebellum is necessary for both acquisition and ex- ing conditioning by either injection of tetrodotoxin pression of the eye-blink CR. in the superior cerebellar peduncle or lidocaine Recently, several studies, each using reversible into the white matter ventral to the interpositus inactivation techniques, have provided very com- has the same effect: complete blockade of CR ex- pelling evidence that the essential locus of memory pression with no effect on CR acquisition (Nord- formation and storage (i.e., the memory trace) for holm et al. 1993; Krupa and Thompson 1995). Col- this type of learning is located within the restricted lectively, these results, along with the cerebellar region of lateral cerebellum encompassing the in- reversible inactivation results (above), provide terpositus nucleus and localized regions of overly- compelling data in support of a cerebellar memory ing cerebellar cortex that project to the interposi- trace: Inactivating this localized region of cerebel- tus nucleus. Clark et al. (1992) initially reported lum (including the interpositus nucleus and over- that temporary, reversible inactivation (by local lying cortex) during conditioning completely pre- cooling) of this localized region of cerebellum dur- vents learning from occurring, whereas inactivating eye-blink conditioning in rabbits completely ing essential structures in the eyeblink circuit that prevented acquisition of the eye-blink CR without are downstream from the cerebellum (including all affecting the ability to acquire the CR following of the output from this region of cerebellum) has removal of the cooling and without affecting the no effect on the ability to learn the CR but does ability to perform the UR. This initial result was prevent CR expression. subsequently confirmed and extended by several Recently, Bloedel and Bracha (1995) have sug- studies, each using different techniques such as gested that our previous reversible inactivation local microinjection into the anterior interpositus studies (in which the interpositus nucleus and lo- nucleus of muscimol, baclofen, or lidocaine to fo- calized regions of overlying cortex were reversibly cally and reversibly inactivate this region of cer- inactivated with either muscimol or lidocaine; ebellum (Krupa et al. 1993; Nordholm et al. 1993; Krupa et al. 1993; Nordholm et al. 1993) did not Hardiman et al. 1996; Ramirez et al. 1997). The completely prevent learning from occurring but, results of each of these studies were the same: instead, resulted in "...a low level of conditioned Appropriate inactivation of the critical region of responses during the retention testing [which] re- cerebellum completely prevented acquisition of flects not an absence of learning but rather a 're- the eye-blink CR with no effect on subsequent tarded' process of learning caused by the drug ef- learning following removal of inactivation and with fect on the cerebellar cortex"(p. 12). Although the no effect on the ability to perform the UR. These data from those previous studies directly and com- results indicate that the memory trace for this type pletely contradict this suggestion (see Discussion), of learning must be localized either (1) within the the present study was designed to test this possi- region of cerebellum that was inactivated in these bility in even more detail. Here, we used a very low studies or (2) in some other structure(s) down- dose of muscimol to selectively inactivate the in- stream from this region of cerebellum that receives terpositus nucleus during five sessions of eye-blink cerebellar input that is essential for both acquisi- conditioning. Retention tests (which would reveal tion and expression of the CR. any signs of learning even on the first trial of test-

L E A R N / N G & M E M 0 R Y 546 Downloaded from learnmem.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press

INTERPOSITUS INACTIVATION BLOCKS CR ACQUISITION

ing) following removal of this inactivation revealed restraint in a Plexiglas restrainer and the sound no evidence at all of learning with no effect at all attenuating behavioral recording chamber for 1 on the rate of subsequent learning in the absence hour. Rabbits were randomly assigned to one of of inactivation. The present results, therefore, con- two groups. One group (n = 8; termed the "Mus- firm and extend the previous results of this and cimol" group) received five daily tone-air-puff con- other laboratories; namely, appropriate lesions (ei- ditioning sessions during which the interpositus ther temporary or permanent) of a restricted re- nucleus was reversibly inactivated with a microin- gion of lateral cerebellum, encompassing mini- jection of the ~/-amino-butyric acid (GABA) agonist mally the anterior interpositus nucleus, completely muscimol (see below). These rabbits then received prevent acquisition of the classically conditioned 2 days of rest followed by four more daily condi- eye-blink response and have no effect on perfor- tioning sessions without any inactivation to test mance of the reflex UR. whether any learning had occurred during the previous five sessions with inactivation. Finally, these rabbits received a last conditioning session in Materials and Methods which muscimol was again infused into the interpositus to test the effects of interpositus inactiva- SURGICAL PROCEDURES tion on retention of the CR and performance of the UR. Under aseptic surgical procedures, 16 New The second group (n = 8) of rabbits (termed Zealand albino rabbits (Oryctolagus cuniculus, the "Control" group) received an infusion of saline -2.5 kg at time of surgery) were each implanted vehicle (0.1 ~tl) into the interpositus nucleus be- with a chronic, stainless steel guide cannula (25 fore each of the first five control sessions. During gauge, 0.65 mm o.d.) fitted with an internal stain- each of these five control sessions, the rabbits less steel stylet (to ensure patency) that extended were restrained and placed in the behavioral re- 1.5 mm beyond the base of the guide cannulae. cording chamber but no stimuli were presented. Surgical procedures consisted of a midline incision Rabbits remained restrained in the chamber for the through the scalp, retraction of the periostium, and same length of time as the Muscimol group, -60 a small craniotomy (1.5 mm diam.) above the ste- min per session. Spontaneous eye-blink activity reotaxic coordinates of the interpositus nucleus. was recorded during these sessions by recording The stylet tip was aimed at the dorsal aspects of the nictitating membrane movement at intervals iden- anterior region of the left cerebellar interpositus tical to those in which stimuli were presented to nucleus, stereotaxically positioned 0.7 mm ante- the rabbits in the Muscimol group. Following these riot, 5.8 mm lateral, and 14.5 mm ventral to the five sessions, rabbits in the Control group also re- skull suture with h positioned 1.5 mm below ceived 2 days rest followed by four tone-air-puff bregma according to the stereotaxic atlas of Mc- conditioning sessions. As with the Muscimol Bride and Klemm (1968). The cannulae were an- group, no infusions were administered during chored to the skull using dental acrylic and three these sessions. Finally, Control rabbits also re- stainless steel skull screws. A small receptacle for ceived one last conditioning session in which mus- attaching a minitorque potentiometer and an air- cimol (1 nmole) was infused into the interpositus puff nozzle during the behavioral training ses- to test the effects on retention of the CR and per- sions was also cemented to the skull. A 1.O-ram formance of the UR. loop of 6-0 surgical suture (Ethilon) was placed in Each tone-air-puff conditioning session con- the apex of the left nictitating membrane (NM). sisted of 100 trials divided into 10 blocks of 10 Surgical anesthesia consisted of ketamine (60 mg/ trials. Each block of 10 trials consisted of 1-tone kg), xylazine (8 mg/kg), and halothane (1%-3% in alone trial followed by 4 paired tone-air-puff trials oxygen). All animals were treated in accordance followed by 1 air-puff alone trial followed by 4 with National Institutes of Health (NIH) guidelines. more paired tone-air-puff trials. Intertrial interval All rabbits received 7 days postoperative recovery. varied randomly between 20 sec and 40 sec [mean (3/) - 30 sec]. Paired tone-air-puff trials consisted of a tone CS (350 msec, 1 kHz, 85 riB) paired with BEHAVIORAL TRAINING PROCEDURES a coterminating corneal air-puff US (100 msec, 2.1 Following postoperative recovery, rabbits re- N/cm 2 pressure at the source). The tone was pre- ceived (on day 8) one session of habituation to sented through a small loudspeaker placed 30 cm

L E A R N / N G & M E M O R Y 547 Downloaded from learnmem.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press Krupa and Thompson in front of the rabbit. The air puff was delivered training trials (40 paired, 5 CS alone, and 5 US through a 3-mm i.d. tube positioned 1 cm from the alone) to ensure that the radiolabeled muscimol center of the rabbit's cornea. Behavioral responses effectively abolished the CR acquired previously were measured with a minitorque potentiometer (as did the previous, nonradioactive infusions). Im- attached to the suture loop in the rabbits' left NM. mediately after the fiftieth trial, each rabbit was CRs were defined as any 0.5-mm or greater exten- lightly anesthetized with halothane and decapi- sion of the NM occurring anytime between 35 tated. Their brains were removed rapidly and then msec and 250 msec (US onset) after CS onset. On frozen in liquid isopentane on dry ice. The frozen CS alone trials, a CR was counted as any response brains were sectioned on a cryostat (20 lam). The I>0.5 mm occurring anytime between 35 msec and frozen sections, along with 3H standards (Amer- 750 msec after CS onset. On both paired and tone sham), were apposed to X-ray film (Amersham Hy- alone trials, any response occurring between 0 perfflm) for 21 days. The exposed films were de- msec and 35 msec after CS onset was considered as veloped, and the brain sections were fixed in for- an a response or a spontaneous response and was malin fumes and stained with cresyl violet. The not counted as a CR. URs were defined as any extent of muscimol diffusion throughout the inter- movement of the NM within 500 msec following positus/dentate nuclei was then determined. US onset (the minimum resolvable movement was Following all training sessions, all rabbits in the 100 tam). Control group (and the two rabbits in the Musci- mol group not infused with 3Hqabeled muscimol) were injected intravenously with a lethal dose of MUSCIMOL INFUSIONS sodium pentobarbital, then perfused through the One hour prior to each of the first five training aorta with 0.9% saline followed by a 10% formalin sessions, rabbits in the Muscimol group received solution. The position of the stylet tips was marked an infusion of muscimol (1.0 nmole in 0.1 ~al of by passing 80 pA (for 8 sec) of anodal current isotonic saline vehicle; Sigma) into the left inter~ through a stainless steel lesioning electrode that positus nucleus. This low dose of muscimol was had been lowered through the guide cannula to chosen (based on previous work) to inactivate the the exact depth of the stylet tip. This position is interpositus nucleus without diffusing throughout the same as that of the inner injection cannula tip the overlying cerebellar cortex and inactivating when it was fully lowered during saline and mus- that region of cerebellum. Rabbits in the Control cimol infusions. The brains were then removed, group were infused with O. 1 pl of saline vehicle 1 embedded in an albumen gel, and stored in 10% hr prior to each of the first five sessions. Infusion formalin until they were sectioned (80 ~m) on a procedures for all rabbits involved removal of the freezing microtome. The sections were stained internal stylet from the guide cannula, insertion of with cresyl violet and Prussian blue, and the loca- a stainless steel injector cannula (31 ga, 0.21 mm tion of the stylet tip was determined. o.d.) that extended 1.5 mm below the base of the outer guide cannula, infusion of the drug at 0.3 /al/min, removal of the injector cannula 3 min after DATA ANALYSIS cessation of infusion, and, finally, reinsertion of the The mean percentage of CRs was calculated internal styler. for the paired trials for each block and each session. Data were analyzed with mixed analyses of HISTOLOGY variance with Group as the between subjects fac- tor and Session or Block as the repeated measure. To determine the extent of diffusion of the Where appropriate, post hoc Newman-Keuls tests 1.0-nmole dose of muscimol in cerebellum, 3H-la- were used to further analyze significant main ef- beled muscimol was infused into the interpositus fects or interactions. A significance level of 0.05 nucleus of six of the eight rabbits in the Muscimol was used for all statistical tests. group. Each rabbit was infused with 1.0 nmole (in O. 1 pl of saline) of 3H-labeled muscimol on the day after the tenth training session. In each infusion, Results tracer doses of 3H-labeled muscimol (2 vtCi/vtg; NEN) were mixed with the unlabeled muscimol. Cannula placements of two of the rabbits in One hour after infusion, each rabbit received 50 the Muscimol group were located outside the in-

L E A R N I N G & M E M O R Y 548 Downloaded from learnmem.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press INTERPOSITUS INACTIVATION BLOCKS CR ACQUISITION terpositus nucleus in the lateral ansiform lobule. Both of these rabbits showed significant levels of A learning during the first five conditioning sessions 100- in which muscimol was infused, each exceeding at least 41% CRs during at least one session and each 80- reaching a learning criterion of eight CRs in nine consecutive trials. In one rabbit, infusion of mus- 60- cimoI on session 10 had no effect at all on CR g- performance relative to performance during ses- o 40- sion 9 in which no infusions were administered (94% on session 9 vs. 100% on session 10). Exami- 20- nation of the spread of 3H-labeled muscimol in this rabbit revealed labeling predominantly located in i 1 2 3 4 5 Test ansiform cortex with very low concentrations of Session muscimoI in the dorsal aspects of the dentate B nucleus. No labeling was detected in the interposi- 100] tus nucleus. CRs of the other rabbit on session 10 (59%) were lower than on session 9 (96%) but were not completely abolished. Analysis of the spread of 3H-labeled muscimol in this rabbit revealed the highest concentration of muscimol in ansiform cortex as well as a very low concentration 0 40 of muscimol in the most dorsal-lateral aspects of the interpositus nucleus and dorsal aspects of the 20 dentate. Most of the interpositus nucleus, however, was not labeled. On the basis of these criteria, 0 1 2 3 4 5 6 7 8 9 10 it was concluded that the muscimol infusions during sessions 1-5 and 10 were not effectively inac- Block tivating the entire extent of the interpositus Figure 1 : (A) Mean (_+S.E.) percentage CRs for Musci- nucleus in these rabbits. These animals, therefore, mol and Control Groups for each training session. On were not included in further analyses. Similarly, sessions 1-5, rabbits received microinjections of 1.0- the cannula placement of one rabbit in the control nmole muscimol [Muscimol group (E3)] or saline vehicle group was outside the interpositus. Infusion of [Control group (0)] into the interpositus nucleus. The Muscimol group was presented with paired tone-air- muscimol on session 10 had no effect on CR per- puff training during inactivation, whereas the Control formance (relative to session 9). This rabbit was group received no stimuli. On sessions 6-9, no infusions also excluded from further analysis. One rabbit in were administered and both groups received paired the Control group dislodged its headstage prior to training. On the Test session, both groups received an completion of conditioning and was, therefore, ex- injection of muscimol (1 nmole) to assess its effects on cluded from further study. retention of the CR. There were no differences between Infusions of muscimol into the interpositus groups during sessions 6-9. Inactivation of the interposi- nucleus of the remaining six rabbits in the Musci- tus nucleus with this low dose of muscimol completely mol group before each of the first five sessions blocked acquisition of the eye-blink CR without affect- completely prevented any expression of eye-blink ing the ability to learn the CR during sessions 6-9 during CRs during those sessions (Fig. 1A). Responses dur- which no infusions were administered. (B) Mean (___s.E.) percentage CRs for Muscimol and Control groups as a ing those sessions did not differ significantly from function of training block on session 6 (first session with- the spontaneous levels recorded in the Control out infusion). Each block consisted of eight paired tone- group that was restrained and placed in the behav- air-puff trials. Again, there were no differences between ioral recording chamber during sessions 1-5 but groups. Responses of the Muscimol group at the start of presented no stimuli. An ANOVA with factors training were the same as Controls. Both Muscimol and Group and Session yielded no significant main ef- Control rabbits also showed the same rate of acquisition fects of Group, Session, or their interaction. during the session. There was no evidence at all that On session 6, the first session without any in- Muscimol animals had learned the CR during sessions activation, rabbits in the Muscimol group showed 1-5.

L E A R N ! N G & M E M O R Y 549 Downloaded from learnmem.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press Krupa and Thompson

no signs of any learning from the previous five cant acquisition was seen across the 10 tone alone conditioning sessions (in which the interpositus trials of session 6 [F(9,90) = 2.63, P < 0.01]. This was inactivated) and subsequently learned the CR tone alone test trims measure is particularly sensi- as if naive. Their performance on session 6 and tive because it includes any responses for a period subsequent rate of acquisition on sessions 7-9 was up to 750 msec after CS onset. Therefore, if rabbits indistinguishable from control rabbits that had in the Muscimol group were performing any long been infused with saline on sessions 1-5 but pre- latency CRs, these would be detected during the sented with no stimuli on those sessions. A t-test tone alone test trials. However, no difference was on trials to criterion (eight CRs in nine consecutive seen between groups on this or any other measure trials) for the two groups revealed no significant of acquisition. In summary, inactivation of the in- difference in learning rate It(10)--0.29, P > 0.05; terpositus nucleus with a very low dose of musci- mean _+ s.n. for Muscimol group = 153 -+ 74 trials, mol completely prevented learning from occurring Control group = 140 _+ 80). Again, an ANOVA with without any effects at all on the ability to acquire factors Group and Session revealed no significant the CR in subsequent training without inactivation. effect of Group or interaction (Fvalues < 1), but in Consistent with previous results, infusions of this case, a strong effect of Session was fotmd muscimol into the interpositus nucleus had no ef- [F(3,30) = 52.4, P < 0.0001], indicating that signifi- fect on the performance of the air puff-evoked UR cant acquisition had occurred over the 4 days of (Fig. 2). There was no significant difference be- training. Infusion of muscimol into the interpositus tween UR amplitudes of the Muscimol group (re- of both Muscimol rabbits and Control rabbits on corded on air-puff alone test trials) on session 5 session 10 (Test) completely abolished the CR during which muscimol infusions inactivated the learned during sessions 6-9 (see Fig. 1A). interpositus and session 6 in which no infusions To ensure that the low level of CRs performed were administered It(5) - -0.04, P > 0.05]. (Be- by the Muscimol group on session 6 (16%) did not cause Control animals were not presented with represent a "retarded" process of learning (as suggested by Bloedel and Bracha 1995; see Introduc- tion) but, instead, simply reflected the normal CR acquisition that would be expected of naive ani- 57 T mals during the course of a single training session, 4 we compared the performance of the Muscimol E group with the Control group on a block-by-block g3 and a trial-by-trial basis. If any learning had oc- curled during sessions 1-5, it would be manifest, ~a < on session 6, in the form of a higher rate of re- m sponding by the Muscimol group during the initial trials of session 6 and/or a higher rate of acquisition over the course of that (and/or subsequent) sessions. The results clearly demonstrate that nei- 0 ther occurred. A block-by-block (each training 5 6 9 Test block consisted of eight paired tone-air-puff trials) Session analysis of the mean percentage of CRs for each Figure 2: Mean (_+S.E.) UR amplitudes measured on air- group revealed no differences at all between puff alone trials for the Muscimol group on session 5 groups (Fig. 1B). A Group x Block ANOVA yielded (last session with inactivation) and session 6 (first train- no significant effect of Group or interaction ing session without inactivation). Mean UR amplitudes [F(1,10) < 1 and F(9,90) = 1], but a significant ef- for session 9 (last training session without inactivation) fect of Block was found [F(9,90) - 4.98, and Test session (with inactivation) are collapsed across P < 0.0001 ], indicating that significant acquisition group, because both groups were treated identically in this phase of the experiment and there were no signifi- occurred over blocks on the first day of training cant differences between groups. Inactivation of the in- without inactivation. Furthermore, trial by trial terpositus nucleus with 1 nmole of muscimol had no analysis of responses on tone alone test trials dur- effect on the ability to perform the reflexive air puff- ing session 6 also revealed no differences at all evoked UR. Solid bars indicate sessions with muscimol between groups and no significant interaction of infusion; open bars indicate sessions without muscimol Group and Trial (both F values < 1). Again, signifi- infusion.

L E A R N ! N G & M E M O R Y 550 Downloaded from learnmem.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press INTERPOSITUS INACTIVATION BLOCKS CR ACQUISITION

any stimuli during sessions 1-5, comparisons of the anterior interpositus nucleus. In two rabbits UR amplitudes between groups cannot be made (infused with muscimol during sessions 1-5), inef- for these sessions.) There were also no significant fective placements were located in ansiform cor- differences between UR amplitudes recorded tex, dorsal and lateral to the interpositus nucleus. on session 9 (no infusion) and session 10 in which Infusion of muscimol into the more dorsal of these muscimol was infused into the interpositus locations had no effect at all on acquisition or re- nucleus of both groups to test retention [no sig- tention of the eye-blink CR. Muscimol infusion into nificant effect of Group, F(1,10)= 2.2; Session, the other location did not prevent acquisition or F(1,10) - 1.7; or their interaction, F(1,10) = 2.3; all expression of the CR. However, muscimol infusion P values > 0.05]. into this region did partially affect CR performance Histological reconstructions of all cannula on session 10 (above). In one rabbit (infused with placements are shown in Figure 3A. Effective saline during sessions 1-5), the cannula was lo- placements are located within or just adjacent to cated just rostral and ventral to the anterior interpositus. Infusion of muscimol (on session 10) in this rabbit had no effect on performance of the CR. To determine the maximal extent of muscimol diffusion, 3H-labeled muscimol was infused into the interpositus nuclei of four rabbits with effective cannula placements. 3Hqabeled muscimol was also infused into the two Muscimol rabbits with ineffective placements (see above). Figure 3B (left) shows the autoradiograph of the largest extent of diffusion of 3H-labeled muscimol. Muscimol diffused throughout the anterior interpositus nucleus as well as regions of the dentate nucleus. In this particular rabbit, there was a very low concentration of muscimol in the most ventral aspects of ansiform cortex (also see Fig. 4, top row). There was no diffusion outside of the cerebellum. The maximal extent of muscimol diffusion for each of the four rabbits with effective cannula placements that were infused with 3H-labeled muscimol is shown in Figure 4. In each case, diffusion of 3H- labeled muscimol was restricted almost entirely to the interpositus/dentate nuclei. There was no evidence at all of muscimol diffusion outside of the cerebellum. These very localized patterns of diffusion were further confirmed by the ineffective Figure 3: (A) Cannula locations for each of the rabbits. (11) Effective muscimol infusion sites; (O) controls; (V) placements. Infusion of muscimol through cannu- ineffective placements. Abbreviations: (ANS) ansiform lae located just outside the interpositus had no ef- cortex; (DE) dentate nucleus; (FA) fastigial nucleus; (HVl) fect on CR acquisition or expression. hemispheric Iobule VI; (icp) inferior cerebellar peduncle; (IN) interpositus nucleus; (IO) inferior olive; (PF) paraflocculus. Numerals at bottom of standard sections represent distance (mm) rostral from the ~. skull suture. Discussion (B) Autoradiograph (left) showing the largest extent of The results of the present study are clear: In- ~H-labeled muscimol diffusion. Superimposed on the activation of a very localized region of cerebellum, autoradiograph is an outline drawing of the Nissl-stained including dorsal aspects of the anterior interposi- section from which it was exposed (right). Diffusion of tus nucleus, with a very low dose of muscimol muscimol is restricted to the interpositus/dentate nuclei with very low levels of muscimol diffusion into the most completely prevented acquisition of the condi- ventral aspects of overlying ansiform cortex. There is no tioned eye-blink response. Subsequent learning in evidence at all of diffusion outside of the cerebellum. the absence of muscimol was completely unaf- Calibration bar, 3.0 mm. fected, ruling out the possibility of any lingering

L E A R N I N G I M E M 0 R Y Downloaded from learnmem.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press Krupa and Thompson

cerebellar cortex. There was no spread of muscimol outside the cerebellum. These results confirm and extend the results of previous studies from several laboratories in which reversible inactivation techniques were used to temporarily block cerebellar activity during eye- blink conditioning. Each of these studies used entirely different methods of inactivation: local cooling (Clark et al. 1992), lidocaine infusion (Nord- holm et al. 1993), baclofen microinjection (Ramirez et al. 1997), and muscimol microinjection with air-puff US (Krupa et al. 1993) or muscimol microinjection with periorbital shock US (Hardi- man et al. 1996). Despite these different techniques, the results of each study were the same: Cerebellar inactivation prevented acquisition of the eye-blink CR with no effect on the ability to learn the CR in subsequent training without inactivation. The present data directly contradict the recent challenges by Bloedel and Bracha (1995) regarding the results of our (and other) previous studies. Those authors contend that we used "unusually high amounts of muscimol", thereby creating the possibility that the drug diffused to and inhibited neurons at extracerebellar sites. Furthermore, they suggest that because we inactivated regions of cerebellar cortex, "it is possible that a low level of conditioned responses during the retention testing Figure 4: Maximal extent of ZH-labeled muscimol dif- reflects not an absence of learning but rather a fusion (shown on standard sections) in four rabbits with 'retarded' process of learning caused by the drug effective cannula placements. Muscimol diffused throughout the interpositus nucleus and regions of den- effect on the cerebellar cortex." The authors, how- tate nucleus. In some rabbits, very low concentrations of ever, offer no evidence at all to support these muscimol diffused into the most ventral aspects of ansi- claims. As described below, the results of the pre- form cortex. There was no diffusion of muscimol outside sent study decisively rule out these arguments. of the cerebellum. In our previous study (Krupa et al. 1993), the intent was to test whether eye-blink conditioning could occur while both the interpositus nucleus as effects of muscimol. A block-by-block comparison well as regions of overlying cerebellar cortex were of percent CRs over the first conditioning session inactivated. To inactivate this region of cerebel- without inactivation revealed no differences at all lum, we used a muscimol dose of 14 nmoles in 1.0 between the Muscimol and Control groups. There jal. Although Bloedel and Bracha (1995) character- were also no differences on subsequent condition- ize this dose as "unusually high" and suggest that ing sessions. Examination of percent CRs on tone muscimol might have diffused out of the cerebel- alone test trials during session 6 also revealed no lum and inactivated extracerebellar sites, all of the differences between groups. This test is particu- available data (including data from their own labo- larly sensitive because it would reveal any long la- ratory) argue against this possibility. tency CRs if they had developed. Muscimol inacti- For instance, prior to our study, a number of vations also had no effect on the ability to perform other laboratories had previously used doses of the air puff-evoked UR. Quantitative autoradiogra- muscimol similar to or greater than ours to selec- phy confirmed that the spread of muscimol was tively inactivate localized regions within the brain localized to the dentate/interpositus nuclei and, in including similar regions of cerebellum (Hikosaka some rabbits, the most ventral aspects of lateral and Wurtz 1985a,b; Martin and Ghez 1988; Van

L E A R N I N G I M E M 0 R Y Downloaded from learnmem.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press

INTERPOSITUS INACTIVATION BLOCKS CR ACQUISITION

Neerven et al. 1989; Keating and Thach 1991; Mink fused out of the cerebellum and inactivated extra- and Thach 1991), contradicting the suggestion that cerebellar sites. our dose was unusually high. Martin (1991), using In light of these results, the data from the pre- a combination of both 3Hqabeled muscimol auto- sent study conclusively rule out the possibility that radiography and L-[14C]glucose uptake reported inadvertent and undetected inactivation of extra- that the maximal radial spread of a similar dose of cerebellar sites by muscimol might be the reason muscimol in cerebral cortex was -2.5 mm, a result for the complete prevention of eye-blink condition- similar to the spread of muscimol we reported ing during cerebellar inactivation. Here, we used a (Krupa et al. 1993, p. 990; Fig. 2). Furthermore, very low dose of muscimol (1.0 mole) to selec- Martin (1991) demonstrated that the radial spread tively inactivate the interpositus nucleus. In addi- of muscimol is very stable for at least 2 hr after tion to using a very low dose, the infusion volume infusion, arguing against the possibility of musci- was also reduced to 0.1 pl, one-tenth the volume mol spreading outside the cerebellum during the typically used in previous studies. As pointed out time course of our experiments. by Bracha et al. (1994), muscimol spread appears Consistent with these results, our own autora- to be mainly attributable to replacement of extra- diographic analysis of the spread of 3H-labeled cerebellar fluid induced by injection pressure. Re- muscimol found no evidence at all of diffusion of sults of our autoradiographs are entirely consistent muscimol outside of the cerebellum. Previous and with this point. As such, the much lower volume of subsequent work in our laboratory, combining infusate used in the present study, coupled with electrophysiological recordings, behavioral mea- the lower dose of muscimol, would result in a very sures, and quantitative autoradiography confirmed restricted sphere of diffusion. This was confirmed that infusion of 14 moles of muscimol into the by our autoradiography that demonstrated that the interpositus nucleus inactivated this structure as radial spread of muscimol was -1.5 mm. In none of well as regions of cerebellar cortex overlying this the autoradiographs was there any evidence at all nucleus without diffusing outside the cerebellum of muscimol diffusion into any extracerebellar (Krupa et al. 1992; Krupa 1993). structures. In summary, these data argue decisively Finally, evidence from their own laboratory ar- against the suggestion by Bloedel and Bracha gues against the suggestion by Bloedel and Bracha (1995) that the complete and total prevention of that our dose of muscimol was unusually high. In eye-blink conditioning following appropriate cer- their recent study, Bracha et al. (1994) used a dose ebellar inactivation might be the result of musci- of 3.5 nmoles to inactivate the interpositus mol diffusing to (and inactivating) extracerebellar nucleus. The authors estimated the effective radial sites during the course of the eye-blink condition- spread of drug to be -1.6 mm. In our study, the ing sessions. There is absolutely no evidence of volume of cerebellum inactivated included both extracerebellar actions in any of the studies using the interpositus nucleus as well as regions of over- reversible inactivation of the interpositus nucleus lying cerebellar cortex, a volume estimated to be during training (Clark et al. 1992; Krupa et al. 3.7 times the volume inactivated by Bracha et al. As 1993; Nordholm et al. 1993; Hardiman et al. 1996; such, the required dose of muscimol to inactivate Ramirez et al. 1997). Finally, we note that Bracha et this greater volume would be correspondingly al. (1994) did not infuse muscimol during acquisi- larger. Therefore, the dose of 3.5 nmoles used by tion and is therefore irrelevant to the present issue Bracha et al. to inactivate just the interpositus of the role of the cerebellum in learning the CR. nucleus corresponds very well with our propor- The further suggestion by Bloedel and Bracha tionally larger dose of 14 nmoles used to inactivate (1995) that the low level of CRs performed on the both the interpositus as well as overlying cortex. first session without inactivation might somehow Thus, although Bloedel and Bracha characterize represent a retarded form of learning is simply not our dose of 14 nmoles as "unusually high", their supported by the data. In our previous study, the own data demonstrate that it was not. In short, all animals that had been infused with muscimol dur- of the available evidence, including the data of Bra- ing the first six conditioning sessions performed an cha et al., indicates that the dose of muscimol (14 average of 19% CRs on session 7, the first condi- nmoles) used in our original study was precisely tioning session without inactivation. However, the dose necessary to inactivate the region of cer- these CRs merely represented the normal rate of ebellum intended. More importantly, there is sim- responses that would be expected by naive ani- ply no evidence whatsoever that muscimol dif- mals over the first session of conditioning. This

L E A R N / N G & M E M O R Y 553 Downloaded from learnmem.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press Krupa and Thompson was confirmed by a block-by-block and trial-by-trial tion and expression of the CR. However, the pre- comparison of the performance of the muscimol cise functional roles of cerebellar cortex and the group with control animals during that first session interpositus nucleus in eye-blink conditioning re- without inactivation. This comparison revealed no main largely unknown. Further study will be re- differences at all between groups. quired to elucidate the role of each of these struc- The present study reinforces this result. On tures in the process of eye-blink conditioning. session 6, the first session without inactivation, the Muscimol group performed 16% CRs. However, examination of those CRs clearly demonstrates that MUSCIMOL-INDUCED DEPRESSION IN THE those responses represented a normal rate of ac- INTERPOSITUS NUCLEUS? quisition expected during the first conditioning session. If any learning had occurred during ses- Recently, Perrett and Mauk (1995) suggested sions 1-5, this would be expressed on sessions 6-9 that inactivation of the interpositus nucleus with in the form of either performance of a higher num- muscimol during eye-blink conditioning did not ber of CR~s early in session 6 and/or a faster rate of block learning per se but instead resulted in a de- acquisition on session 6 and subsequent sessions. pression of mossy fiber synaptic efficacy within However, the data confirm that this did not occur. this nucleus that apparently masked plasticity that There were no differences at all between the Mus- had occurred in cerebellar cortex during training. cimol and Control groups in either the number of However, these authors offer no evidence to sup- CRs performed early in training (including the very port this idea. Numerous lines of evidence rule out first trials) or in the rate of acquisition during ses- this possibility. sion 6 and subsequent sessions. Comparison of First, in our previous study, both the interposi- performance on tone alone test trials revealed no tus nucleus as well as cerebellar cortex were inac- differences between groups, ruling out the possi- tivated with muscimol during conditioning. In the bility that the Muscimol rabbits were performing present study, using a much lower dose of musci- long latency responses that would not be detected mol, only the interpositus nucleus was inactivated. on paired tone-air-puff trials. The very low dose of However, percent CRs as well as the rate of CR muscimol used in this study was such that inacti- acquisition on the first session without muscimol vation was limited to the interpositus nucleus. Mus- inactivation were identical for both groups. These cimol did not inactivate overlying regions of cer- measures were also identical to controls infused ebellar cortex shown previously to be involved in with saline. If muscimol infusions were somehow eyeblink conditioning. In summary, the present re- causing a synaptic depression within the interposi- sults decisively rule out the suggestions by Bloedel tus, it would seem likely that there would be dif- and Bracha (above) and confirm all of the previous ferences in CR performance by these different results of this and other laboratories: Temporary groups of rabbits, but there were none. inactivation of a restricted region of cerebellum, Second, other studies have used local cooling minimally encompassing the anterior interpositus (Clark et al. 1992) and lidocaine (Nordholm et al. nucleus, completely prevents acquisition of the 1993) to inactivate the interpositus nucleus during eyeblink CR, with no effect on the UR and with no eye-blink conditioning. These methods of inactiva- effect at all on the ability to learn the CR in subse- tion would abolish both pre- as well as postsynap- quent training without inactivation. tic activity within the interpositus, which would Although the muscimol dose used here did not prevent any form of synaptic plasticity from occur- inactivate significant regions of cerebellar cortex, ring. Use of these methods of inactivation also the results of the present study do not rule out a completely prevented eye-blink conditioning from possible role for cerebellar cortex in acquisition or occurring (above). The performance of animals on expression of the eyeblink CR. Several lines of evi- the first session without inactivation in these stud- dencc support the hypothesis that, during eye- ies, however, was the same as performance of rab- blink conditioning, plasticity occurs in both the bits infused with muscimol. If muscimol infusions cerebellar cortex as well as the interpositus into the interpositus had resulted in a long-term nucleus. For instance, permanent lesions of cer- depression, the rates of acquisition in each of these ebellar cortex (which spare the interpositus studies should be different, but they are not. nucleus) have been shown to significantly impair Third, Yeo and colleagues (Hardiman et al. (and in some cases, completely prevent) acquisi- 1996; Ramnani and Yeo 1996) recently tested the

L E A R N / N G & M E M O R Y 554 Downloaded from learnmem.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press

INTERPOSITUS INACTIVATION BLOCKS CR ACQUISITION possibility of muscimol-induced depression di- Bloedel, J.R. and V. Bracha. 1995. On the cerebellum, rectly. They first trained rabbits to perform the CR. cutaneomuscular reflexes, movement control and the elusive engrams of memory. Behav. Brain Res. 68: 1-44. They then infused muscimol into the cerebellum and presented the animals with tone alone extinc- Bracha, V., M.L. Webster, N.K. Winters, K.B. Irwin, and J.R. tion training. They then tested the animals without Bloedel. 1994. Effects of muscimol inactivation of the muscimol to determine whether any extinction cerebellar interposed-dentate nuclear-complex on the performance of the nictitating-membrane response in the had occurred. They found that the rabbits immedi- rabbit. Exp. Brain Res. 100" 453-468. ately performed CRs at preinfusion rates: No dec- rement in responding had occurred while the in- Clark, R.E. and D.G. Lavond. 1993. Reversible lesions of the terpositus was inactivated with muscimol. This re- red nucleus during acquisition and retention of a classically conditioned behavior in rabbits. Behav. Neurosci. sult alone would appear to rule out the possibility 107" 264-270. of a synaptic depression in the interpositus during these studies. Clark, R.E., A.A. Zhang, and D.G. Lavond. 1992. Reversible lesions of the cerebellar interpositus nucleus during In summary, the present study and all other acquisition and retention of a classically conditioned studies in which a localized region of the cerebel- behavior. Behav. Neurosci. 106" 879-888. lum including the anterior interpositus nucleus is inactivated during initial training of the condi- Hardiman, M.J., N. Ramnani, and C.H. Yeo. 1996. Reversible inactivations of the cerebellum with muscimol prevent the tioned eye-blink response agree that no learning acquisition and extinction of conditioned nictitating occurs. All evidence to date strongly supports the membrane responses in the rabbit. Exp. Brain Res. hypothesis that the cerebellum is the site of 110" 235-247. memory storage for this form of learning, and there Hikosaka, O. and R.H. Wurtz. 1985a. Modification of is no convincing evidence to the contrary (see Yeo saccadic eye movements by GABA-related substances. I. 1991; Lavond et al. 1993; Thompson and Krupa Effect of muscimol and bicuculline in monkey superior 1994; Thompson and Kim 1996). colliculus. J. Neurophysiol. 53: 266-291.

--. 1985b. Modification of saccadic eye movements by GABA-related substances. II. Effects of muscimol in monkey substantia nigra pars reticulata. J. Neurophysiol. 53" 292-308. Acknowledgments We thank Christine G. Logan for critical comments and Krupa, D.J. 1993. Localization of the essential memory trace Anne C. Krupa for invaluable assistance during preparation of for a classically conditioned behavior. Ph.D. thesis, this manuscript. This research was supported by National University of Southern California, Los Angeles, CA. Science Foundation grant IBN-9215069, Office of Naval Krupa, D.J. and R.F. Thompson. 1995. Inactivation of the Research grant N00014-95-1152, National Institute of Mental superior cerebellar peduncle blocks expression but not Health grant 5P01-MH52194, and a grant from the Sankyo acquisition of the rabbit's classically conditioned eye-blink Company, LTD. response. Proc. Natl. Acad. Sci. 92" 5097-5101. The publication costs of this article were defrayed in part by payment of page charges. This article must therefore Krupa, D.J., J.K. Thompson, and R.F. Thompson. 1993. be hereby marked "advertisement" in accordance with 18 Localization of a memory trace in the mammalian brain. USC section 1734 solely to indicate this fact. Science 260: 989-991.

Krupa, D.J., J. Weng, and R.F. Thompson. 1996. Inactivation References of brainstem motor nuclei blocks expression but not Albus, J 9 1971. A theory of cerebellar function. Math. acquisition of the rabbit's classically conditioned eyeblink Biosci. 10: 25-61. response. Behav. Neurosci. 110" 219-227.

Berthier, N.E. and J.W. Moore. 1986. Cerebellar Purkinje cell Lavond, D.G., T.L. Hembree, and R.F. Thompson. 1985. activity related to the classically conditioned nictitating Effect of kainic acid lesions of the cerebellar interpositus membrane response. Exp. Brain Res. 63: 341-350. nucleus on eyelid conditioning in the rabbit. Brain Res. 326:179-182. 91990. Activity of deep cerebellar nuclear cells during Lavond, D.G., B.J. Knowlton, J.E. Steinmetz, and R.F. classical conditioning of nictitating membrane extension in Thompson. 1987. Classical conditioning of the rabbit eyelid rabbits. Exp. Brain Res. 83" 44-54. response with a mossy-fiber stimulation CS: II. Lateral reticular nucleus stimulation. Behav. Neurosci. Blaxton, T.A., T.A. Zeffiro, J.D.E. Gabrieli, S.Y. Bookheimer, 101 9676-682. M.C. Carrillo, W.H. Theodore, and J.F. Disterhoft. 1996. Functional mapping of human learning--a positron emission Lavond, D.G., J.J. Kim, and R.F. Thompson. 1993. tomography activation study of eyeblink conditioning. J. Mammalian brain substrates of aversive classical Neurosci. 16: 4032-4040. conditioning. Annu. Rev. Psychol. 44:31 7-342.

L E A R N / N G & M E M O R Y 555 Downloaded from learnmem.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press Krupa and Thompson

Lincoln, J.S., D.A. McCormick, and R.F. Thompson. 1982. Ramirez, O.A., A.F. Nordholm, D. Gellerman, J.K. Ipsilateral cerebellar lesions prevent learning of the classically Thompson, and R.F. Thompson. 1997. The conditioned conditioned nictitating membrane/eyelid response. Brain Res. eyeblink response: A role for the GABA-B receptor. 242:190-193. Pharmacol. Biochem. Behav. (in press).

Lisberger, S.G. 1988. The neural basis for learning of simple Ramnani, N. and C.H. Yeo. 1996. Reversible inactivations of motor skills. Science 242: 728-735. the cerebellum prevent the extinction of conditioned nictitating membrane responses in rabbits. J. Physiol. Logan, C.G. and S.T. Grafton. 1995. Functional anatomy of 495.1 : 159-168. human eyeblink conditioning determined with regional cerebral glucose metabolism and positron-emission Steinmetz, J.E., D.J. Rosen, P.F. Chapman, D.G. Lavond, and tomography. Proc. Natl. Acad. Sci. 92" 7500-7504. R.F. Thompson. 1986. Classical conditioning of the rabbit eyelid response with a mossy-fiber stimulation CS: I. Pontine Marr, D. 1969. A theory of cerebellar cortex. J. Physiol. nuclei and middle cerebellar peduncle stimulation. Behav. 202: 437-470. Neurosci. 100: 878-887.

Martin, J.H. 1991. Autoradiographic estimation of the extent Steinmetz, J.E., D.G. Lavond, and R.F. Thompson. 1989. of reversible inactivation produced by microinjection of Classical conditioning in rabbits using pontine nucleus lidocaine and muscimol in the rat. Neurosci. Lett. stimulation as a conditioned stimulus and inferior olive 127:160-164. stimulation as an unconditioned stimulus. Synapse 3" 225-233. Martin, J.H. and C. Ghez. 1988. Red nucleus and motor Steinmetz, J.E., D.G. Lavond, D. Ivkovich, C.G. Logan, and cortex: Parallel motor systems for the initiation and control of R.F. Thompson. 1992. Disruption of classical eyelid skilled movement. Behav. Brain Res. 28:217-223. conditioning after cerebellar lesions: Damage to a memory trace system or a simple performance deficit? J. Neurosci. Mauk, M.D., J.E. Steinmetz, and R.F. Thompson. 1986. 12: 4403-4426. Classical conditioning using stimulation of the inferior olive as the unconditioned stimulus. Proc. Natl. Acad. Sci. Thach, W.T., H.P. Goodkin, and J.G. Keating. 1992. The 83: 5349-5353. cerebellum and the adaptive coordination of movement. Annu. Rev. Neurosci. 15" 403--442. McBride, R.L. and W.R. Klemm. 1968. Stereotaxic atlas of the rabbit brain, based on the rapid method of photography Thompson, R.F. and D.J. Krupa. 1994. Organization of of frozen, unstained sections. Commun. Behav. Biol. memory traces in the mammalian brain. Annu. Rev. 2" 179-215. Neurosci. 17" 519-549.

McCormick, D.A. and R.F. Thompson. 1984a. Cerebellum: Thompson, R.F. and J.J. Kim. 1996. Memory-systems in the Essential involvement in the classically conditioned eyelid brain and localization of a memory. Proc. Natl. Acad. Sci. response. Science 223" 296-299. 93:13438-13444.

~. 1984b. Neuronal responses of the rabbit cerebellum Van Neerven, J., O. Pompeiano, and H. Collewijn. 1989. during acquisition and performance of a classically Depression of the vestibulo-ocular and optokinetic responses conditioned nictitating membrane-eyelid response. J. by intrafloccular microinjection of GABA-A and GABA-B Neurosci. 4:2811-2822. agonists in the rabbit. Arch. Italiennes de Biol. 127" 243-263.

Mink, J.W. and W.T. Thach. 1991. Basal ganglia motor Yeo, C.H. 1991. Cerebellum and classical conditioning of control. III. Pallidal ablation: Normal reaction time, muscle motor responses. Ann. N.Y. Acad. Sci. 627: 292-304. cocontraction, and slow movement. J. Neurophysiol. 65: 330-351. Yeo, C.H., M.J. Hardiman, and M. Glickstein. 1985. Classical conditioning of the nictitating membrane response of the Molchan, S.E., T. Sunderland, A.R. Mclntosh, P. Herscovitch, rabbit. I. Lesions of the cerebellar nuclei. Exp. Brain Res. and B.G. Schreurs. 1994. A functional anatomical study of 60: 87-98. associative learning in humans. Proc. Natl. Acad. Sci. 91 : 8122-8126. Received February 18, 1997; accepted in revised form April Nordholm, A.F., J.K. Thompson, C. Dersarkissian, and R.F. 19, 1997. Thompson. 1993. Lidocaine infusion in a critical region of cerebellum completely prevents learning of the conditioned eyeblink response. Behav. Neurosci. 107: 882-886.

Perrett, S.P. and M.D. Mauk. 1995. Extinction of conditioned eyelid responses requires the anterior lobe of cerebellar cortex. J. Neurosci. 15" 2074-2080.

L E A R N / N G & M E M O R Y 556 Downloaded from learnmem.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press

Reversible inactivation of the cerebellar interpositus nucleus completely prevents acquisition of the classically conditioned eye-blink response.

D J Krupa and R F Thompson

Learn. Mem. 1997, 3: Access the most recent version at doi:10.1101/lm.3.6.545

References This article cites 41 articles, 12 of which can be accessed free at: http://learnmem.cshlp.org/content/3/6/545.full.html#ref-list-1

License

Email Alerting Receive free email alerts when new articles cite this article - sign up in the box at the Service top right corner of the article or click here.