The comparator hypothesis of conditioned response generation: manifest conditioned excitation and inhibition as a function of relative excitatory strengths of CS and conditioning context at the time of testing

In the present research water-deprived rats were used in a conditioned lick suppression paradigm to test and further develop Rescorla's (1968) contingency theory, which posits that excitatory associations are formed when a conditioned stimulus (CS) signals an increase in unconditioned stimulus (US) likelihood and that inhibitory associations develop when the CS signals a decrease in US likelihood. In Experiment 1 we found that responding to a CS varied inversely with the associative status of the context in which the CS was trained and that this response was unaltered when testing occurred in a distinctively dissimilar context with a different conditioning history, provided associative summation with the test context was minimized. These results suggest that manifest excitatory and inhibitory conditioned responding is modulated by the associative value of the training context rather than that of the test context. In Experiment 2 it was demonstrated that postconditioning decreases in the associative value of the CS training context reduced the effective inhibitory value of the CS even when testing occurred outside of the training context. Moreover, this contextual deflation effect was specific to the CS training context as opposed to any other excitatory context. Collectively, these studies support the comparator hypothesis, which states that conditioned responding is determined by a comparison of the associative strengths of the CS and its training context that occurs at the time of testing rather than at the time of conditioning. This implies that all associations are excitatory and that responding indicative of conditioned inhibition reflects a CS-US association that is below (or near) the associative strength of its comparator stimulus. It is suggested that response rules which go beyond a monotonic relation between associative value and response strength can partially relieve learning theories of their explanatory burdens, thereby allowing for simpler models of acquisition.     

Within-compound associations between the context and the conditioned stimulus

Three experiments were conducted to test for the presence of associations between contextual cues and the nominal conditioned stimulus (CS) in fear conditioning. Rats were given tone-footshock pairings and were tested for their fear of the nominal CS, the tone, in a different context. Experiments 1 and 2 demonstrated that rats given nonreinforced exposure to the training context following conditioning were less fearful of the CS. Experiment 3 indicated that additional footshock presentations in the training context following conditioning produced greater fear of the CS. In both procedures postconditioning treatments designed to directly alter only the associative strength of the training context yielded parallel changes in the conditioned response to the CS. These data suggest that within-compound associations are formed between the context and the CS during classical conditioning.     

Mechanisms underlying retarded emergence of conditioned responding following inhibitory training: evidence for the comparator hypothesis

The comparator hypothesis posits that conditioned responding is determined by a comparison at the time of testing between the associative strengths of the conditioned stimulus (CS) and stimuli proximal to the CS at the time of conditioning. The hypothesis treats all associations as being excitatory and treats conditioned inhibition as the behavioral consequence of a CS that is less excitatory than its comparator stimuli. Conditioned lick suppression by rats was used to differentiate four possible sources of retarded responding to an inhibitory CS. These include habituation to the unconditioned stimulus (US), latent inhibition to the CS, blocking of the CS-US association by the conditioning context, and enhanced excitatory associations to the comparator stimuli. Prior research has demonstrated the first three phenomena. Therefore, we employed parameters expected to highlight the fourth one--the comparator process. In Experiment 1, our negative contingency training was shown to produce a conditioned inhibitor that passed inhibitory summation and retardation tests. In Experiment 2 we found transfer of retardation from an inhibitory CS to a novel stimulus when the location where retardation-test training occurred was excitatory, which is indicative of contextual blocking and/or comparator effects. In Experiment 3, extinction of the conditioning context was found to attenuate retardation regardless of whether extinction occurred before or after the CS-US pairings of the retardation test. This indicates that much of the present retardation was due to the comparator process rather than to contextual blocking. Experiment 4 demonstrated that habituation to the US did not contribute to retardation in the present case. Collectively, these studies suggest that retardation following inhibitory training can be explained without recourse to any of the traditional mechanisms of conditioned inhibition.     

Mechanisms of blocking by contextual stimuli

The influence of contextual stimuli on the conditioning and performance of responding to a discrete stimulus was examined in the US preexposure paradigm using both context shift manipulations and a measure of context conditioning. Four groups of rats received both repeated exposure to an electric shock US in one context (Context 1), and repeated nonshocked exposure to a second context (Context 2). Two additional groups of rats received exposure to these contexts, but never received shock presentations. Rats exposed to shock learned to escape from the stimuli of Context 1, but did not escape from the stimuli provided by Context 2. Rats not exposed to shock failed to escape from either context. All rats then received a single CER conditioning session in which four pairings of a 3-min noise CS and shock US were presented. Half the rats received those CS-US pairings in the excitatory Context 1, while the remaining rats received those pairings in the neutral Context 2. Finally, half the rats in each of the CER conditioning treatments received extinction test trials of the noise CS in Context 1, while the remaining rats received those test trials in Context 2. Thus, this design factorially manipulated the presence of excitatory or neutral contextual stimuli during both conditioning and testing of a discrete CS. In comparison with the two groups of rats never preexposed to shock alone, attenuation in acquisition of conditioned suppression observed during test trials occurred only when CER conditioning had been administered in the excitatory Context 1, and this effect was manifested when testing occurred in either the excitatory Context 1 or the neutral Context 2. These results support the model of R. A. Rescorla and A. R. Wagner (1972) (in A. H. Black & W. F. Prokasy (Eds.) Classical Conditioning II, pp. 64–99, New York: Appleton-Century-Crofts) which asserts that contextual stimuli and sicrete CSs compete for limited associative strength supportable by a given US.     

Testing response generation rules

Robbins (1988) reported data that he viewed as inconsistent with Miller and Schachtman's (1985a) comparator hypothesis of conditioned response generation. Here we explain why we do not find his experiments a compelling test of the comparator hypothesis. We also briefly review other studies that tested the same predictions of the comparator hypothesis that Robbins examined. We conclude that there is considerable evidence that following excitatory or inhibitory conditioning with a target conditioned stimulus (CS) and unconditioned stimulus (US), extinction of other cues that were present during CS training ordinarily increases excitatory responding and decreases inhibitory responding to the CS. However, consistent with Robbins's conclusion, there is scant evidence that after CS-US training, enhancing the associative value of other cues that were present during CS training influences excitatory or inhibitory responding to the CS. The implications of these conclusions for the comparator hypothesis as an explanation of differences in acquired behavior and as a heuristic tool are considered.     

Human classical conditioning. The status of the CS

Recent reports of failure to obtain blocking in human galvanic skin response (GSR) conditioning, together with our own equivocal results with eyelid conditioning, have motivated us to re-examine the status of the conditioned stimulus (CS) in human conditioning studies. The issues raised by compound stimuli, by contextual cues and occasion setting stimuli, and by cross-modal transfer are considered in the light of data from our laboratories. These data include observations on the interchangeability of stimulus modalities during acquisition, the use of varying information loads embedded in occasion setting displays, the comparison of alternative blocking designs and the analysis of response topography in relation to stimulus variability. They suggested that an adequate account of the CS in human conditioning studies must recognize that it is dynamically processed and reprocessed both during and after acquisition.     

Conditioned inhibition of fear: Evidence for a competing response mechanism

In two experiments, inhibitory conditioning was attempted by presenting a discrete CS in a neutral stimulus environment shortly following the termination of either shock (Experiment 1) or a second discrete CS which had been paired in a forward manner with shock (Experiment 2). Evidence of successful inhibitory conditioning was mixed in Experiment 1, where the properties of the CS were assessed within an escape-from-fear procedure. Postresponse presentations of the CS enhanced performance, whereas the presentation of the CS prior to responding did not have the expected degrading effect on performance. In Experiment 2, the inhibitory properties of the CS were assessed by combining this stimulus with an excitatory CS and presenting the compound to rats engaged in a water-reinforced licking response. Less response suppression was found in reaction to this compound relative to three separate comparison conditions, thus witnessing the success of the inhibitory-conditioning procedure used. The common assumption that inhibitory conditioning results from the nonreinforcement of a CS in a situation where reinforcement is expected, i.e., one which contains previously reinforced cues, is not supported by these data, for no previously reinforced cues were simultaneously presented with the CS during inhibitory training. The data are in agreement with a conditioned antagonistic-response interpretation of inhibitory conditioning.     

Contextually Based Conditional Discrimination of the Rabbit Eyeblink Response

Rabbits received conditional discrimination training using contextual stimuli to set the occasion for stimulus pairings during eyelid conditioning. Specifically, animals were exposed to either the presence or the absence of an oscillating chamber light throughout the intertrial interval (50 ± 10 s). For half the animals, this light signaled paired presentations of a discrete tone conditioned stimulus (CS) and air puff unconditioned stimulus (US) while darkness signaled presentations of only the tone CS. The remaining animals experienced the opposite contextual relationship to the conditioning stimuli. These trial types occurred pseudo-randomly across a session, with all transitions between contextual settings (i.e., light or dark) taking place immediately at the CS–US offset. Under these conditions, animals successfully utilized the contextual stimuli as conditional cues for differential responding to the shared CS. Moreover, both light and dark were equally effective as discriminative stimuli. A subset of animals received further training in which the contextual contingency was removed by restricting all conditioning to the CS-alone context. Without the contingency in place, subsequent CS presentations (paired and CS-alone) evoked equivalent conditioned responding across three sessions of training. Following the reinstatement of the contextual contingencies, discriminatory responding was immediately observed and returned to previous levels within three sessions. Finally, animals appeared to use the static representation of the conditional cue, rather than the phasic transition between cues, for discriminatory responding. These findings are discussed in terms of current neurobiological models of eyelid conditioning.     

Conditioned suppression tests of the context-blocking hypothesis: testing in the absence of the preconditioned context

Two experiments are reported that use rats in a conditioned suppression situation. The experiments, designed to remove confounds that have complicated interpretations of prior research, tested the context-blocking hypothesis, the proposition that static apparatus cues or conditioning contexts can block conditioning to discrete conditioned stimuli (CSs). Experiment 1, like previous work, tested for conditioning to the target CS in the same context that had been preconditioned and in which target conditioning had occurred; the experiment demonstrated a context-blocking like effect. Experiment 2 tested for conditioning not only in the preconditioned context but also in a nonpreconditioned context. Evidence for context blocking appeared similar in the two test situations. This suggests that conditioned contexts block the acquisition of associative strength by discrete CSs at the time of target conditioning (e.g., Rescorla & Wagner, 1972) and not through performance factors at the time of testing (e.g., Gibbon & Balsam, 1981).     

Role of the hippocampus in blocking and conditioned inhibition of the rabbit's nictitating membrane response.

Bilateral aspiration of the dorsal hippocampus produced a disrupttion of blocking of the rabbit's nictitating membrane response in Kamin's two-stage paradigm (Experiment 1) but had no effect on the formation of a Pavlovian conditioned inhibitor (Experiment 2). The results of Experiment 1 indicated that normal animals and those with cortical lesions given conditioning to a light-plus-tone conditioned stimulus (CS) gave conditioned responses (CRs) to both the light and the tone during nonreinforced presentations of each (test phase). If, however, compound conditioning was preceded by tone acquisition, only the tone elicited a CR during testing; that is, blocking was observed. In rabbits with hippocampal lesions, however, CRs were given to both the light and the tone during testing whether or not compound conditioning was preceded by tone acquisition. The data from Experiment 2 showed that rabbits with hippocampal lesions could discriminate as well as normal rabbits and those with cortical lesions between a light (CS+) and light plus tone (CS-). In addition, when the inhibitory tone was subsequently paired with the unconditioned stimulus in retardation testing, animals in all three lesion conditions acquired the CR at the same rate. Thus, it appears that hippocampal lesions do not disrupt conditioned inhibition. The results of these experiments were taken as support for the view that the hippocampus is responsible for "tuning out" stimuli that have no adaptive value to the organism.     

Blockade of GABAA receptors in the interpositus nucleus modulates expression of conditioned excitation but not conditioned inhibition of the eyeblink response

The cerebellum and related brainstem structures are essential for excitatory eyeblink conditioning. Recent evidence indicates that the cerebellar interpositus and lateral pontine nuclei may also play critical roles in conditioned inhibition (CI) of the eyeblink response. The current study examined the role of GABAergic inhibition of the interpositus nucleus in retention of CI. Male Long-Evans rats were implanted with a cannula positioned just above or in the anterior interpositus nucleus before training. The rats were trained with two different tones and a light as conditioned stimuli, and a periorbital shock as the unconditioned stimulus. CI training consisted of four phases: 1) excitatory conditioning (8 kHz tone paired with shock); 2) feature-negative discrimination (2 kHz tone paired with shock or 2 kHz tone concurrent with light); 3) summation test (8 kHz tone or 8 kHz tone concurrent with light); and 4) retardation test (light paired with shock) After reaching a criterion level of performance on the feature-negative discrimination (40% discrimination), 0.5 μl picrotoxin (a GABA_A receptor antagonist) was infused at one of four concentrations, each concentration infused during separte test sessions. Picrotoxin transiently impaired conditioned responses during trials with the excitatory stimulus (tone) in a dose-dependent manner, but did not significantly impact responding to the inhibitory compound stimulus (tone-light). The results suggest that expression of conditioned inhibition of the eyeblink conditioned response does not require GABAergic inhibition of neurons in the anterior interpositus nucleus.     

Blockade of GABAA Receptors in the Interpositus Nucleus Modulates Expression of Conditioned Excitation but not Conditioned Inhibition of the Eyeblink Response

The cerebellum and related brainstem structures are essential for excitatory eyeblink conditioning. Recent evidence indicates that the cerebellar interpositus and lateral pontine nuclei may also play critical roles in conditioned inhibition (CI) of the eyeblink response. The current study examined the role of GABAergic inhibition of the interpositus nucleus in retention of CI. Male Long-Evans rats were implanted with a cannula positioned just above or in the anterior interpositus nucleus before training. The rats were trained with two different tones and a light as conditioned stimuli, and a periorbital shock as the unconditioned stimulus. CI training consisted of four phases: 1) excitatory conditioning (8 kHz tone paired with shock); 2) feature-negative discrimination (2 kHz tone paired with shock or 2 kHz tone concurrent with light); 3) summation test (8 kHz tone or 8 kHz tone concurrent with light); and 4) retardation test (light paired with shock). After reaching a criterion level of performance on the feature-negative discrimination (40% discrimination), 0.5 microl picrotoxin (a GABAA receptor antagonist) was infused at one of four concentrations, each concentration infused during separate test sessions. Picrotoxin transiently impaired conditioned responses during trials with the excitatory stimulus (tone) in a dose-dependent manner, but did not significantly impact responding to the inhibitory compound stimulus (tone-light). The results suggest that expression of conditioned inhibition of the eyeblink conditioned response does not require GABAergic inhibition of neurons in the anterior interpositus nucleus.     

Contextual conditioning in infants, but not older animals, is facilitated by CS conditioning

Context conditioning in infant Sprague-Dawley rats (postnatal days [PD] 15, 17, and 19), juveniles (PD 25), adolescents (PD 35), and adults (PD 71-89) was compared when CS conditioning did or did not occur in the context. Degree of CS conditioning within that context was equated across age, and separate groups at each age were given unpaired presentations of the CS and US within that context. Infants conditioned more effectively to context when CS-US pairings occurred in that context than when they did not, juveniles conditioned to context about equally with and without CS-US pairings in the context, and adolescents and adults conditioned less effectively to context when CS and US were paired than when unpaired. Adolescents had significant context conditioning despite CS-US pairings in the context but adults did not, and overall, context conditioning was strongest for adolescents. Supplementary experiments indicated that with more extensive conditioning experience, the infants' pattern of context conditioning became more similar to that of older animals, and with less conditioning experience the pattern of context conditioning by adults became more similar to that of younger animals, but infants never attained the adult pattern of context conditioning nor did adults attain the infant pattern. The potentiation of context conditioning by CS conditioning observed in infants is consistent with previous evidence derived from compound conditioning. Alternative explanations place common emphasis on infant-specific amodal processing. One views potentiation as a result of the greater perceived intensity of the stimulus compound (CS and context, in this case) during conditioning and the lesser generalization decrement in infants than adults when tested with a single element after conditioning with a compound. The other explanation emphasizes consequences of the redundancy inherent in intersensory compounds, within the theory of Bahrick and Lickliter.     

The Influence of Retention Interval on the US Preexposure Effect: Changes in Contextual Blocking over Time

A number of studies manipulating the length of the interval between conditioning and testing indicate spontaneous recovery from overshadowing, suggesting that certain instances of overshadowing represent a deficit in memory retrieval rather than a failure of animals to form an association between the overshadowed stimulus and the US. The present series of experiments examined the influence of lengthening the retention interval on blocking, another stimulus selection phenomenon that is typically interpreted as an acquisition deficit. The results indicated that when subjects were tested shortly (3 days) after training conditioning to a taste blocked subsequent conditioning to an odor conditioned in compound with that taste (Experiment 1), whereas prior conditioning to an odor did not block subsequent conditioning to a taste conditioned in compound with that odor (Experiment 2). This pattern of results was essentially unchanged when testing occurred at a longer (21-day) retention interval. However, there was evidence of a US preexposure effect in Experiment 2 when subjects in the US ONLY control condition were tested at the 3-day retention interval, but not when testing occurred 21 days after conditioning. Experiments 3 and 4 examined whether this loss of the US preexposure effect over time might actually represent a change in the degree of contextual blocking as the retention interval is lengthened. Exposure to the conditioning context either during the interval between Phase 1 and Phase 2 of conditioning (Experiment 3) or prior to Phase 1 of conditioning (Experiment 4) alleviated this US preexposure effect suggesting that the loss of the US preexposure effect as the retention interval is lengthened observed in Experiment 2 is due to changes in the degree of blocking by contextual stimuli over time. The results are discussed in terms of differential susceptibility of forgetting of two functional roles played by a contextual stimuli in the current situation-context as a CS and context as a retrieval cue for other CS-US associations.     

Learned Irrelevance Is Not the Sum of Exposure to CS and US

In three experiments hungry rats received appetitive conditioning trials with a light that signalled the delivery of sucrose solution. In Experiment 1, prior exposure to uncorrelated presentations of the conditioned stimulus (CS) and unconditioned stimulus (US) retarded conditioning significantly more than did prior exposure to the CS alone. In Experiments 2 and 3, groups exposed to uncorrelated presentations of the CS and US within the same session conditioned significantly more slowly than groups given separate sessions of exposure to the CS followed by sessions of exposure to the US (or vice versa). Some part of the learned irrelevance effect depends on exposure to a zero correlation between the CS and US, perhaps because this promotes learning that the CS predicts no change in the probability of the US.     

Acquisition of the classically conditioned eyeblink response in humans over the life span

Human subjects ranging in age from 18 to 85 years underwent classical conditioning of the eyeblink response to a tone conditioned stimulus (CS) and an air-puff unconditioned stimulus (UCS). There was a decline in percentage of conditioned responses with age. This decline was most noticeable in subjects over age 50. These conditioning deficits were not due to age-related changes in sensitivity to the tone CS or the air-puff UCS, nor could the conditioning deficits be attributed to an age-related decline in general cognitive abilities or to changes in spontaneous blink rates. The results are discussed in terms of using the classically conditioned eyeblink in humans in conjunction with the classically conditioned nictitating membrane response in rabbits as a model system for studying the neurobiology of age-related conditioning deficits.     

Attenuation of age-related conditioning deficits in humans by extension of the interstimulus interval

Young (17-22 years) and older (61-86 years) persons underwent classical conditioning of the eye-blink response to a tone conditioned stimulus (CS) and an airpuff unconditioned stimulus (UCS) at 1 of 3 interstimulus intervals (ISIs; 400, 650, and 900 ms). As in a previous study, older subjects conditioned more slowly and emitted fewer conditioned responses at the optimal 400-ms ISI. At longer ISIs, however, this age-related disruption of classical conditioning was attenuated. These differences in conditioning were not due to nonassociative factors, such as sensitivity to the tone CS or airpuff UCS or to differences in spontaneous blink rate, nor were they due to differences in general cognitive abilities. The results are discussed in terms of the role of temporal relationships between stimuli in conditioning across the life span.     

Time course of cardiac conditioned responses in restrained rats as a function of the trace CS-US interval

Two experiments aimed at understanding the temporal characteristics of trace-conditioned heart-rate responses to a 0.5-s tone (conditioned stimulus [CS]) in restrained rats. A CS paired with a tail-shock (unconditioned stimulus [US]) elicited lasting bradycardiac responses. The magnitude and extinction rate of conditioned responses (CR) were independent of the CS-US interval (interstimulus interval [ISI], 3 s to 20 s). Unreinforced test trials were analyzed for CR topography. Responding was delayed in groups with longer ISIs, but CR latencies, peak and decay times were not proportional to the ISI. Response peaks tended to cluster either about 6 s after CS onset, or about 10 s with a slow decay, depending on the ISI. The authors postulated 2 components of auditory stimulus traces involved in cardiac conditioning, maximally active 6 s and 10 s respectively after CS onset. The topography of the CR could be constrained to combinations of associative strength and instantaneous activity of these 2 components.     

Initial excitatory conditioning with the truly random control procedure in rats: The effects of density of the conditioned stimulus

The present experiments were designed to test whether initial excitatory conditioning to a stimulus was obtained, using a within-subject design of the truly random control (TRC) procedure, and to examine the effects of the density of the conditioned stimulus on it. In Experiment 1, the effect of the three levels of CS density (50%, 25%, 20%) was assessed at six different stages of training, using the conditioned suppression of licking in rats. Although all of the groups given the TRC procedure showed evidence of initial excitatory conditioning, it was more marked in the group with the highest CS density than in the other groups. Experiment 2 confirmed the results of Experiment 1, using a between-subject design. The findings are in line with predictions from the Rescorla-Wagner model, and are discussed with reference to other contemporary theories of associative learning.     

Recent and Remote Extinction Cues Reduce Spontaneous Recovery

Five appetitive conditioning experiments with rats examined the ability of extinction cues (ECs) to reduce spontaneous recovery after extinction procedures that varied the temporal relation between the ECs and the conditioned stimulus (CS) and included presentations of additional events (e.g. other stimuli correlated with extinction, CSs, and the US). In extinction, two different ECs were presented either closely in time before (i.e. recent to), or more distant (i.e. remote) from a nonreinforced CS. Both recent and remote ECs reduced spontaneous recovery to the CS when present during testing (Experiments 1-4). Each EC reduced recovery despite the addition during extinction of a second EC and CS (Experiments 1, 3, and 4), and the US (Experiment 3). Experiments 4 and 5 investigated the recent and remote ECs' tendency to control a serial occasion setting discrimination involving the target CS under explicit training conditions. Neither EC gained such discriminative control. Possible explanations of the results are discussed, including configural learning, occasion setting, and contextual cue control.