Classical conditioning requires the existence of an unconditioned stimulus [UCS] that elicits an unconditioned response [UCR], that is, that reliably elicits an unlearned response, in the experimental subject. UCRs [unlearned responses] are also known as reflexes. The UCR is usually a physiological response that can reliably be elicited by a UCS, for example, salivation [the UCR] in response to the smell or sight of food [the UCS], particularly if one is hungry, or an eye blink [the UCR] in response to a puff of air [the UCS] blown into the eye. The classical conditioning procedure also requires a conditioned stimulus [CS], a stimulus of which the subject can be made aware but which initially does not cause the UCR, followed by a conditioned response, the same response as the UCR, but eventually in reaction to a different stimulus. For example, the CS in the puff of air example might be simply the sound of a buzzer, resulting, after conditioning is complete, in a blink [CR] caused by the CS alone.
Classical conditioning, then, would proceed as follows, using the four components and four steps.
CS: The CS [conditioned stimulus]—for example, the sound of a buzzer—is presented in several trials.
UCS: Each presentation of the CS is followed closely by presentation of the UCS [unconditioned stimulus]—for example, the puff of air.
UCR: Presentation of the UCS causes a UCR [an eye blink].
CR: After a sufficient number of presentations of the CS followed by the UCS, the experimenter presents the CS without the UCS. If a response, an eye blink, occurs, the UCR is now called a conditioned response [CR]. The eye blink response to the buzzer has been conditioned [learned].
Shown graphically, the sequence is
If the CS now produces a CR, with no presentation of the UCS, it can be said that conditioning [learning] has occurred and
Higher order conditioning. Higher order conditioning, that based upon previous learning, may also occur in the classical conditioning paradigm. In higher order conditioning, what was the CS comes to serve as a UCS. For example, if the experimenter always turned on a desk light before sounding the buzzer to begin classical conditioning [to produce an eye blink at the sound of the buzzer], the turning on of the light may eventually itself produce the eye blink, independent of the buzzer. In this case, the buzzer has become a UCS, and the turning on of the light has become a CS. Consequently, although initially [light] → CS[buzzer] → UCS[air puff] → UCR[eye blink] → CR[eye blink] higher order conditioning proceeds
And higher order conditioning [learning] occurs:
Classical conditioning terminology. Specific terminology is used to describe the classical conditioning procedure.
The process of learning a conditioned response is called acquisition. Usually, conditioning is faster if only a short time elapses between the presentation of the CS and the UCS.
The reverse process—that is, unlearning—can occur also and is called extinction. If the CS is presented for a time without the UCS, the CR will eventually cease [be extinguished].
If the CS is again presented later, however, the CR will sometimes return temporarily [this temporary return is called spontaneous recovery]. But the CR will disappear unless the UCS is at times reinstated
Andrew C.W. Weeks,1,4 Steve Connor,1 Richard Hinchcliff,2 Janelle C. LeBoutillier,3 Richard F. Thompson,2 and Ted L. Petit3
Andrew C.W. Weeks
1 Department of Psychology, Nipissing University, North Bay, Ontario P1B 8L7, Canada;
Find articles by Andrew C.W. Weeks
Steve Connor
1 Department of Psychology, Nipissing University, North Bay, Ontario P1B 8L7, Canada;
Find articles by Steve Connor
Richard Hinchcliff
2 Neuroscience Program, University of Southern California, Los Angeles, California 90089-2520, USA;
Find articles by Richard Hinchcliff
Janelle C. LeBoutillier
3 Department of Psychology, University of Toronto, Toronto, Ontario M1C 1A4, Canada
Find articles by Janelle C. LeBoutillier
Richard F. Thompson
2 Neuroscience Program, University of Southern California, Los Angeles, California 90089-2520, USA;
Find articles by Richard F. Thompson
Ted L. Petit
3 Department of Psychology, University of Toronto, Toronto, Ontario M1C 1A4, Canada
Find articles by Ted L. Petit
Author information Article notes Copyright and License information Disclaimer
1 Department of Psychology, Nipissing University, North Bay, Ontario P1B 8L7, Canada;
2 Neuroscience Program, University of Southern California, Los Angeles, California 90089-2520, USA;
3 Department of Psychology, University of Toronto, Toronto, Ontario M1C 1A4, Canada
4Corresponding author.E-mail ac.ugnissipin@wwerdna; fax [705] 474-1947.
Received 2006 Jun 12; Accepted 2007 Mar 27.
Copyright © 2006, Cold Spring Harbor Laboratory Press
Abstract
Eye-blink conditioning involves the pairing of a conditioned stimulus [usually a tone] to an unconditioned stimulus [air puff], and it is well established that an intact cerebellum and interpositus nucleus, in particular, are required for this form of classical conditioning. Changes in synaptic number or structure have long been proposed as a mechanism that may underlie learning and memory, but localizing these changes has been difficult. Thus, the current experiment took advantage of the large amount of research conducted on the neural circuitry that supports eye-blink conditioning by examining synaptic changes in the rabbit interpositus nucleus. Synaptic quantifications included total number of synapses per neuron, numbers of excitatory versus inhibitory synapses, synaptic curvature, synaptic perforations, and the maximum length of the synapses. No overall changes in synaptic number, shape, or perforations were observed. There was, however, a significant increase in the length of excitatory synapses in the conditioned animals. This increase in synaptic length was particularly evident in the concave-shaped synapses. These results, together with previous findings, begin to describe a sequence of synaptic change in the interpositus nuclei following eye-blink conditioning that would appear to begin with structural change and end with an increase in synaptic number.
The classically conditioned eye-blink response is known to be supported by well defined neural circuits. The pairing of the conditioned stimulus [CS, tone] to the unconditioned stimulus [US, air puff] requires the cerebellum and the interpositus nucleus in particular [Thompson et al. 1998; Christian and Thompson 2003]. The importance of the interpositus nucleus was confirmed by the finding that temporary inactivation of this structure prevents both the acquisition and performance of the conditioned response [Krupa et al. 1993]. Further, reversible lesions to brain regions efferent to the interpositus nuclei have failed to block both the formation and expression of the conditioned response [Kim and Thompson 1997].
While the neural circuit responsible for eye-blink conditioning is now defined, our understanding of the underlying cellular mechanisms is less complete. Research to date has suggested synaptic involvement since the disruption of synaptic enzymes and neurotransmitter receptors leads to impaired acquisition of the conditioned response [Bracha et al. 1998; Gomi et al. 1999; Chen and Steinmetz 2000a, b]. Kleim et al. [2002] quantified synaptic number in the interpositus nuclei of the rat following eye-blink conditioning and found an increase in the number of excitatory synapses [synaptogenesis] but no increase in number of inhibitory synapses following 5 d of training.
The current experiment was conducted to not only examine changes in synaptic number but also consider synaptic structural change in the rabbit interpositus nuclei. Further, synapses were analyzed from tissue harvested 1 h after the animals reached a criteria level for acquisition of the learning task. This differed from Kleim et al. [2002], where 5 d of training were completed regardless of the rate of task acquisition. The change in observation timing in the current experiment was seen as an opportunity to view synaptic changes evident directly after the acquisition of the conditioned response. Based on the earlier findings in the rat, we hypothesized that synaptic changes would be evident in the rabbit interpositus nucleus following eye-blink conditioning.
Results
Behavioral results
Animals in the paired group began to show consistent conditioned response [CR] eye-blinks after as little as 80 and as many as 320 trials [Fig. 1]. Since animals did not continue daily sessions after they reached the learning criteria [eight out of nine CRs], the total number of trials required to reach the criteria differed between animals. Importantly, animals in the unpaired group were matched to the paired animals so that the number of CS and US presentation was identical between the groups.
Open in a separate window
Figure 1.
The percentage of conditioned responses [mean ± SE] across daily sessions. Paired animals [n = 8] show a large increase in conditioned responses while unpaired animals [n = 7] do not.
Synaptic counts
A mean of 56 [+3.5] synapses per animal was counted though 53 [+2.9] dissector pairs and a mean volume of 127 [+6.1] μm3. The neuronal sampling resulted in a mean of 122 [+12] neurons counted through 60 dissector pairs and a volume of 2.4 × 104 μm3 per animal. Importantly, the number of neurons per unit volume did not change significantly between the groups. This satisfied the condition of equal neurons and volume as stated in the Materials and Methods section for using “synapses per neuron” as an appropriate measurement. A MANOVA consideration of the unbiased estimate of the number of synapses per neuron revealed no difference in total synapses, excitatory, or inhibitory contacts [Table 1], and no differences in synapses with different curvatures or perforations.
Table 1.
Synapses per neuron
Open in a separate window
Synaptic length
MANOVA and Tukey’s post-hoc analysis of the length of various synaptic types revealed a significant increase in excitatory synapses in the paired group [F[2,15] = 13.88, P < 0.05, Fig. 2]. When the synaptic shape of these excitatory synapses was considered, a significant increase in the size of concave and flat-shaped synapses was also observed in the paired animals [Fig. 2].
Open in a separate window
Figure 2.
The number of excitatory and inhibitory synapses per neuron [mean ± SE] and various curvature subtypes [excitatory only] between the three groups. [*] The paired group is significantly different from the other two groups [P < 0.05].
Discussion
Although synaptic change has long been a candidate mechanism underlying the maintenance of a conditioned response, it has been difficult to localize the small and distributed changes likely associated with this form of learning. The extensive research conducted on the neural circuitry that supports the eye-blink response provides a unique opportunity to know exactly where to look for synaptic changes. The current results indicate that synaptic change is associated with the acquisition of eye-blink classical conditioning. Although synaptogenesis was not observed, excitatory synapses were found to increase in length in the conditioned group. Further, within these excitatory contacts, learning was associated with larger concave and flat-shaped synapses. The following discussion will attempt to place the current results within the relevant literature, but it is important to mention that few studies have examined synaptic ultrastructure in the interpositus nuclei.
Synaptic number
As stated, eye-blink conditioning was not associated with an overall change in the number of synapses per neuron or a change in the number of any of the synaptic subtypes. These results differ from those of Kleim et al. [2002], who found a significant increase in the number of excitatory contacts in the interpositus nucleus following eye-blink conditioning in rats. Importantly, Kleim et al. conducted 5 d of training, where conditioning to the criteria used here was achieved on day three. The added training significantly lengthened the time frame for synaptic examination compared with the current study, where the interpositus synapses were examined from animals perfused 1 h after achieving the learning criterion. In some cases, this criterion was met on the first day of training. Thus, while appearing contradictory, the current finding may simply represent an earlier time point in a sequence of synaptic changes associated with this form of conditioning.
Regarding synaptic counts, it is important to acknowledge that the current study counted an average of ∼60 synapses per animal. Since counts of 100 or more are ideal for unbiased stereological techniques, it is possible that a larger synaptic sampling may have revealed differences in synapse per neuron ratios. The nonsignificant trend, however, does not indicate an increase in excitatory synapses in the interpositus nucleus [see Table 1].
Synaptic length
An increase in the length of the excitatory synapse in the interpositus nucleus was the central finding of this experiment. Although Kleim et al. [2002] did not examine synaptic size, Geinisman et al. [2000] found that trace eye-blink conditioning was associated with an increased size of synaptic contact area in the hippocampus. Synapses in the hippocampus have also been extensively studied following the induction of long-term potentiation [LTP], and increased synaptic size has been a consistent result during the maintenance phase of this phenomenon [Fifkova 1985; Weeks et al. 2000].
The functional relevance of synaptic size may be that larger synapses allow for more transmitter release and larger receptor beds [Petit 1995]. Mackenzie et al. [1999] measured NMDA receptor-mediated miniature calcium transients attributed to the release of single transmitter quanta. Subsequent ultrastructure analysis showed that synapse size correlated positively with the amplitude of the postsynaptic response. Mackenzie et al. suggested that larger synapses may express a greater number of NMDA receptors and therefore have greater connective strength.
The current results are similar to our previous finding where a significant increase in synaptic length was observed in the hippocampus 1 h after the induction of LTP [Weeks et al. 2000]. These results appear to indicate that the expression of both learning and LTP yield similar initial changes in synaptic ultrastructure. There are, however, important differences between the synaptic changes observed in the hippocampus and cerebellum.
It is also important to mention that the increased length reported here represents changes in individual PSDs or directly adjacent PSD segments that make up a perforated synapse. This analysis did not sum the larger group of synaptic segments that likely form on the large presynaptic terminals known to exist in the interpositus nuclei. That said, changes in the individual synapses likely translate to the larger terminal region, although future three-dimensional analysis is required to confirm this hypothesis.
Synaptic ultrastructure
One interesting finding was the very small proportion of perforated synapses evident in the interpositus nucleus [