The International Journal of Comparative Psychology is sponsored by the International Society for Comparative Psychology. It is a peer-reviewed open-access digital journal that publishes studies on the evolution and development of behavior in all animal species. It accepts research articles and reviews, letters and audiovisual submissions.
Volume 28, 2015
Introduction to Timing and Time Perception
Interval Timing Behavior: Comparative and Integrative Approaches
Special Issue on Timing and Time Perception
Humans and rats can discriminate different fixed intervals (FIs) that are signaled by different stimuli. With only a few pairings of stimuli with intervals, temporal performance becomes a function of the stimuli, with responding increasing earlier for stimuli that signal shorter FIs compared to stimuli that signal longer FIs. As predicted by timing and conditioning models, the amount of training with the different stimuli and intervals determines the development of such stimulus control. This study reviews some earlier work from our group suggesting that the amount of training is necessary, but not sufficient, to account for the development of stimulus-controlled performance. Moreover, it describes an experiment in which participants were trained in a computerized shooting task with three FIs (target speeds) signaled by three stimuli (different background colors). In the first phase, the number of trials trained with each FI was held constant (60 trials each) across five experimental groups, but the order in which these trials were trained differed between groups, from a randomly determined FI in each trial (intermixed) to three consecutive blocks of 60 trials each (blocked). Intermediate groups had blocks of 10, 20, and 30 consecutive trials of each FI. Results showed that, although the amount of training was held constant across groups, the longer the training block the fewer the participants who demonstrated stimulus-controlled performance. In Phase 2, another 60 trials of each FI were trained, but intermixed for all groups. Results showed stimulus-controlled performance for all participants. These results represent another instance in which the amount of training is necessary, but not sufficient, for the development of stimulus control in temporal discriminations, and describe the effect of the number of consecutive trials within a block of training on temporal discriminations.
Temporal Averaging Across Stimuli Signaling the Same or Different Reinforcing Outcomes in the Peak Procedure
The present study examined factors that affect temporal averaging in rats when discriminative stimuli are compounded following separate training indicating the availability of reward after different fixed intervals (FI) on a peak procedure. One group of rats, Group Differential, learned that a flashing light stimulus signaled that one type of food pellet reward could be earned for lever pressing after an FI 5 s interval and that a second type of food pellet reward could be earned after an FI 20 s interval in the presence of a tone stimulus. A second group of rats, Group Non-Differential, was similarly trained except that both types of rewards were scheduled across flash and tone trials. When given non-reinforced flash + tone compound test trials the interval containing the maximal response rate was no different than on flash alone test trials, although some responding also appeared near the long FI time. After these FI contingencies were reversed (flash signaled FI 20 s and tone signaled FI 5 s), however, further compound test trials more clearly revealed a temporal averaging pattern in both groups. The peak interval was shifted to the right of the FI 5 stimulus. Moreover, Group Differential rats acquired the reversed discrimination somewhat more rapidly than Group Non-Differential rats, and in a final selective satiation test Group Differential rats responded less in later intervals after they had been sated on the FI 20 s reward. These data suggest that temporal averaging in stimulus compound tests occurs even when the stimuli being combined signal qualitatively different rewards, but that decreasing the value of one of those rewards can shift responding away from the relevant time interval in a selective satiation test. However, when an especially salient stimulus (e.g., flashing light) signals a short FI, rats tend to process the compound stimulus more in terms of its individual elements.
We examined whether temporal context influences how animals produce a time interval. Six pigeons pecked one key to start an interval and then another key to end the interval. Reinforcement followed whenever the interval duration fell within a range of values signaled by the keylight colors. During Phase 1, keylight colors S1 and L1, intermixed across trials, signaled the ranges (0.5-1.5 s) and (1.5- 4.5 s), respectively. During Phase 2, colors S2 and L2 signaled the ranges (1.5-4.5 s) and (4.5-13.5 s), respectively. We asked whether the intervals produced in the presence of L1 and S2, stimuli signalling the same range, varied with their temporal context, short in Phase 1, long in Phase 2. The results showed that a) the intervals produced in the presence of the different keylight colors accorded with the main properties of temporal differentiation, including Weber’s law, b) the L1 intervals had slightly higher means than the S2 intervals, a weak contrast effect, c) the L1 intervals also had higher variability than the S2 intervals. An extension of the learning-to-time model to temporal differentiation tasks reproduced some of the major features of the data but left unanswered how context might change the model parameters.
A need for individual data analyses for assessments of temporal control: Invertebrate fixed interval performance
We outline several experimental variables and addressed the inconsistencies of these variables within the invertebrate fixed interval literature. We posit previous inconsistencies within the invertebrate fixed interval literature may be due to the utilization of aggregate versus individual analyses and contend individual analyses are critical in order for conclusions to be made about species’ abilities to emit responses that can come under temporal control. To exemplify these statements, we exposed honey bees Apis mellifera lingustica (N=13) to either an FI 15-sec or FI 30-sec schedule of reinforcement and analyzed subjects’ cumulative response records, response bin levels, quarter lives, inter-response time patterns, response duration patterns, and trial durations. No measures clearly indicated individual subjects’ responding came under temporal control of the fixed interval schedules; however, pooled group analyses did produce seemingly clear evidence of temporal control.
- 1 supplemental PDF
The aim of the present study was to investigate sex-related variations in the perception of the duration of emotional stimuli (human faces). Twenty male and 20 female participants estimated the duration of angry, ashamed and neutral faces marking 0.4 to 1.6 s intervals. Female faces were used in one session, and male faces in the other. Compared to the angry faces condition, intervals were underestimated when ashamed faces were shown. However, the intervals in neither conditions were significantly overestimated or underestimated compared to the neutral condition. Even more critical is the fact that there was an underestimation by male participants of the duration of male faces compared to female faces; and female participants overestimated the duration in the anger condition, compared with the shame condition, only when male faces were presented. Moreover, the emotional effects on the participants’ performance were correlated to inter-individual differences in empathic abilities. The findings are discussed in terms of sex differences, of social context, and of how attention is solicited and arousal generated by emotions.
Previous studies showed that humans and mice can maximize their rewards in two alternative temporal discrimination tasks by incorporating exogenous probabilities and endogenous timing uncertainty into their decisions. The current study investigated if the probabilistic relations modulated the temporal discrimination performance in scenarios with more than two temporal options. In order to address this question, we tested humans (Experiment 1) and mice (Experiment 2) in the dual-switch task, which required subjects to discriminate three time intervals (short, medium, and long durations) in a sequential fashion. The latencies of switches from short to medium and from medium to long option were the main units of analysis. The results revealed that the timing of switches between the first two options (short-to-medium) were sensitive to probabilistic information in both humans and mice. However, mice but not humans adapted the timing of their subsequent switches between the last two options (medium-to-long) based on the probabilistic information associated with these latter options. These results point at a suboptimal tendency in the temporal decisions of humans with multiple options.
Models of interval timing typically include a response threshold to account for temporal production. The present study sought to evaluate the dependent concurrent fixed-interval fixed-interval schedule of reinforcement as a tool for selectively isolating the response threshold in rats, pigeons, and humans. In this task, reinforcement is available either at one location after a short delay or at another location at a longer delay. Because the reinforced location is not signaled, subjects normally respond on the first location and, if reinforcement is not delivered, then switch to the second location. The latency to switch between locations served as the primary dependent measure. After training rats, pigeons, and humans with equal reinforcement magnitudes in the short and long delays, the magnitude of reinforcement was increased threefold on the long-delay location. Consistent with model predictions, this biasing procedure decreased estimates of the response threshold of rats and pigeons, but also reduced temporal control in these species and increased response-threshold estimates in humans. Human and pigeon performance also suggested a magnitude-induced increase in the speed of the internal clock. Collectively, these results suggest that differences in reinforcement magnitude between response alternatives appear to modulate the response threshold, but not selectively, and may provide guidance for better isolating response-threshold effects in humans.
Interval timing is a complex cognitive process that involves the estimation of time within the seconds-to-minutes range. This temporal processing depends on cortico-striatal interactions, as well as an optimal dopaminergic function. On the other hand, the circadian system controls physiological and behavioral functions with periods close to 24 hr. We have previously reported that short-time perception in mice is influenced by the circadian pacemaker, with dopamine signaling as a link between both temporal systems. In this work we evaluated the involvement of melatonin in the circadian modulation of interval timing, as well as the interaction between this hormone and dopamine levels in the striatum. We report that melatonin-depleted rats, by pinealectomy, present an impairment in their ability to estimate a short (24 s) target duration in the peak-interval procedure. Moreover, melatonin administration in drinking water restores interval timing precision in pinealectomized rats. We also show that circadian desynchronization causes a transient impairment in the timing task. In addition, melatonin administration affects interval timing only when rats are trained and tested during the night. Furthermore, we report that melatonin depletion increases striatal dopamine availability, which is reverted by external melatonin administration. Taken together, our findings add further support to the notion that the circadian system modulates interval timing, probably by using melatonin as an output to regulate dopaminergic functions in brain areas that are important for interval timing mechanisms.
Timing is a ubiquitous process that underlies a great variety of human activities and depends on highly conserved neuronal circuitry, the cortico-striatal loops. The peak interval (PI) task is an operant task that conditions subjects to initiate and terminate behavioral responses bracketing a fixed interval associated with reinforcement. Performance in this task depends on the efficacy of temporal control processes that coordinate interval encoding and decoding, instrumental response innitiation, cessation and maintenance, and motor control. Here, we used the PI procedure to characterize temporal control in zQ175 knockin (KI) and BAC HD transgenic (Tg) mice generated to model Huntington's Disease (HD), and contrast the result with previously published R6/2 Tg PI data. HD is a progressive neurodegenerative disorder that involves degeneration of the same neural circuits underlying temporal information processing and control of motor output. Our results indicate that temporal control is disrupted in R6/2 Tg and zQ175 KI mice but intact in BAC HD Tg mice. Trial-by-trial analysis of break-run patterns in response rates indicated that shifts in zQ175 KI response curves were driven by significant delays in response initiation and cessation. Similar temporal control deficits were previously reported in HD patients and R6/2 transgenic HD mice. These findings support the use of zQ175 mice in preclinical studies of HD-related cognitive deficits. They provide evidence of a strong homology between the human and rodent neural bases of temporal information processing, temporal response control, and their pathology in neurodegeneration.
The accuracy of time judgments depends upon many factors, including the sensory properties of the to-be-timed stimulus. In auditory Pavlovian fear conditioning, an initially neutral tone (conditioned stimulus, CS) predicts the arrival of an aversive event (unconditioned stimulus, US) at a fixed time interval. The temporal relation between the CS and US events is encoded, leading to the development of a temporal pattern of responding. Little attention has been paid to the potential impact of the characteristics of the CS tone on the development of this temporal pattern. Here we compared the acquisition of the temporal pattern of conditioned responses of rats to different CS tone frequencies. Rats were first conditioned to lever press for food. Then, while lever pressing for food, they were presented with 60-s tones of two very different frequencies 1kHz or 11kHz, each paired with a foot-shock given 30s after tone onset. This fear conditioning led to the appearance of conditioned suppression of the lever pressing. On probe trials the tone duration was 60 s, and the reinforcer was omitted. With training, a pattern of suppression evolved during the probe trials, showing a maximum of suppression near the programmed time of the shock US, however the 11kHz CS tone yielded better temporal control than did the 1kHz tone. A second experiment investigated rats’ abilities to discriminate between two times of shock arrival (10s or 30s) predicted by the different tone frequencies (1kHz or 11kHz), In this experiment, rats showed poorer discriminative timing performance when the lower frequency (1kHz) was associated with the longer duration (30s). Our results suggest a strong impact of the CS sensory properties on the expression of temporal learning within the context of auditory fear conditioning in rats.
Anticipatory timing plays a critical role in many aspects of human and non-human animal behavior. Timing has been consistently observed in the range of milliseconds to hours, and demonstrates a powerful influence on the organization of behavior. Anticipatory timing is acquired early in associative learning and appears to guide association formation in important ways. Importantly, timing participates in regulating goal-directed behaviors in many schedules of reinforcements, and plays a critical role in value-based decision making under concurrent schedules. In addition to playing a key role in fundamental learning processes, timing often dominates when temporal cues are available concurrently with other stimulus dimensions. Such control by the passage of time has even been observed when other cues provide more accurate information and can lead to sub-optimal behaviors. The dominance of temporal cues in governing anticipatory behavior suggests that time may be inherently more salient than many other stimulus dimensions. Discussions of the interface of the timing system with other cognitive processes are provided to demonstrate the powerful and primitive nature of time as a stimulus dimension.
Organisms are constantly extracting information from the temporal structure of the environment, which allows them to select appropriate actions and predict impending changes. Several lines of research have suggested that interval timing is modulated by the dopaminergic system. It has been proposed that higher levels of dopamine cause an internal clock to speed up, whereas less dopamine causes a deceleration of the clock. In most experiments the subjects are first trained to perform a timing task while drug free. Consequently, most of what is known about the influence of dopaminergic modulation of timing is on well-established timing performance. In the current study the impact of altered DA on the acquisition of temporal control was the focal question. Thirty male Sprague-Dawley rats were distributed randomly into three different groups (haloperidol, d-amphetamine or vehicle). Each animal received an injection 15 min prior to the start of every session from the beginning of interval training. The subjects were trained in a Fixed Interval (FI) 16s schedule followed by training on a peak procedure in which 64s non-reinforced peak trials were intermixed with FI trials. In a final test session all subjects were given vehicle injections and 10 consecutive non-reinforced peak trials to see if training under drug conditions altered the encoding of time. The current study suggests that administration of drugs that modulate dopamine do not alter the encoding temporal durations but do acutely affect the initiation of responding.
Long-interval timing fills the gap between the traditional range of short-interval timing (i.e., seconds to minutes) and the limited range of circadian entrainment (i.e., approximately a day). A number of reports suggest that rats time long intervals. However, a recent report proposed that anticipation of long, but noncircadian, intervals is highly constrained. We tested the hypothesis that long-interval timing is highly constrained by examining a number of cases: 7, 8, 9, 11, 12, and 13 hour intermeal intervals. We found evidence for long interval timing in each case. Long interval timing appears to be robust.
Individual Differences in Novelty-Seeking are Associated with Different Patterns of Preference in a Risk-Sensitivity Procedure in Rats
The preferences of organisms faced with changing conditions in food delivery situations have been studied under the rubric of risk-sensitivity. Optimal foraging theory often applies the energy budget model to explain the preferences shown by organisms, but in this paper we suggest a different approach, one based on the study of individual differences. A sample of rats was classified as high and low novelty-seeking. Afterwards, they were maintained at 75% or 90% of their body weight and exposed to a risk-sensitivity procedure. The results show that the novelty-seeking model is associated with different patterns of preference under a risk-sensitivity procedure, but that these patterns do not correlate with the level of food deprivation employed. Furthermore, we found that the spontaneous alternation between options in a choice situation correlates with the organism’s preference during a risk procedure. Considering recent findings in the area of animal and human decision-making, our results are explained in terms of altered behavioral processes.
The Development of Juvenile-Typical Patterns of Play Fighting in Juvenile Rats does not Depend on Peer-Peer Play Experience in the Peri-Weaning Period
Play fighting in rats involves attack and defense of the nape. To protect the nape, rats use a variety of defensive tactics, with different strains having specific preferences. Targeting of the nape is established before weaning and defense matures over the course of the week preceding and the week proceeding weaning. Thus, it is possible that experience from engaging in immature forms of play is needed to consolidate the nape as the playful target and for the development of the juvenile-typical pattern of defense. Two experiments were conducted to evaluate this possibility. For the first experiment, male rats were reared over the week post-weaning in either pairs or alone, and their play tested with unfamiliar partners when juveniles (31-34 days). For the second experiment, during the week preceding weaning, male and female rats were placed into one of three conditions: (1) with the mother and no peers, (2) with same-sex siblings but no mother or (3) with both the mother and same-sex siblings. The subjects were tested in same-sex, same-condition pairs when juveniles (31-34 days). Rats from all conditions, in both experiments, attacked the nape during play fighting and developed the same juvenile-typical patterns of playful defense. This suggests that the experience of peer-peer play in the peri-weaning period is not necessary for the development of the attack and defense components of juvenile-typical play.
Free-running spontaneous alternation refers to the animal’s tendency to prefer the least recently visited locations in successive spatial choices, which is attributed to the animals’ choice between stimuli based on prior experience. Turn alternation, which is observed in directional choices preceded by a forced turn in one direction, also reflects the animals’ tendency to alternate between directional choices but this tendency has been assumed to rely on other cues (e.g., proprioceptive cues) derived from the prior responses (e.g., forced turn in one direction). Based on previous studies, the turn alternation appears to rely on more primitive (lower-form) information features and to be a more frequently observed empirical phenomenon than the spontaneous alternation. We investigated these two behavioral alternation tendencies in Artemia sp. Experiment 1 tested the continuous spontaneous alternation (cSAB) performance of Artemia sp. in two different mazes: t-maze (three options) and plus maze (four options). Experiment 2 tested the turn alternation performance of Artemia sp. counter-balancing the direction of initial forced-turn between subjects. Our results showed that Artemia sp. had nearly chance level spontaneous alternation performance in the t-maze and plus maze whereas a higher than chance level turn alternation performance. These results support the ubiquity of turn alternation tendency across species and point at the lack of spontaneous alternation in Artemia sp.
Although California sea lions (Zalophus californianus) are capable of forming complex mental concepts, they have failed to demonstrate mirror self-recognition, a skill that requires both a mental representation of one’s physical features and knowledge of a reflective surface. Many non-human species that do not recognize themselves in mirrors can nonetheless learn to use mirror reflections to locate and retrieve objects. A total of 7 sea lions housed at 2 separate facilities were tested on their ability to detect an object using a mirror. The results of a preliminary detection task in which sea lions were reinforced for looking at a mirror to locate an object suggested that 4 sea lions reliably learned to locate an object positioned below a mirror in one of three locations. A follow-up study was conducted to determine if 3 different sea lions could learn the task without training the animals to use the mirrors. Two of the 3 sea lions located a single object in 1 of the 3 locations statistically above chance when the mirrors were added to the task for the first time. With additional mirror exposure, 1 sea lion successfully achieved 100% accuracy in detecting familiar objects placed in 1 of 3 familiar locations. This sea lion also demonstrated her ability to detect an object via a mirror located in a novel, fourth position with 100% accuracy. When two novel objects were tested with four locations, the sea lion again performed well, detecting the objects 87.5%. The results suggest that sea lions have the ability to use mirrors to locate an object with minimal exposure to a mirror, but likely need additional experience with mirrors to efficiently use the properties of these reflective surfaces and understand that the image is a two-dimensional representation of a three-dimensional object.
- 2 supplemental images
- 2 supplemental files
Partial reinforcement reduces vulnerability to anti-anxiety self-medication during appetitive extinction
Inbred rats from the Roman low-avoidance strain (RLA-I), but not from the Roman high-avoidance strain (RHA-I) increased preference for ethanol after being exposed to sessions of appetitive extinction (Manzo et al. Physiol Behav 2014 123:86-92). RLA-I rats have shown greater sensitivity than RHA-I rats to a variety of anxiogenic situations, including those involving reward loss. Such increased fluid preference did not occur after acquisition (reinforced) sessions or in control groups with postsession access to water, rather than ethanol. Because ethanol has anxiolytic properties in tasks involving reward loss, oral consumption after extinction sessions was interpreted as anti-anxiety or emotional self-medication (ESM). The present research was an attempt to reduce or eliminate the ESM effect in RLA-I rats by giving them 50% partial reinforcement training during the acquisition of an instrumental response, a treatment known to induce resilience to loss-induced anxiety. As expected, partially reinforced RLA-I rats showed a higher resistance to extinction in comparison to continuously reinforced animals, displaying lower ethanol consumption than continuously reinforced rats during the postsession preference test. Partial and continuous control groups receiving water during the preference tests showed no changes in preference. These results suggest that exposure to reward uncertainty typical of partial reinforcement training can reduce ESM in rats genetically selected for high levels of anxiety.
Assessment of Demand for Food under Concurrent PR and FR Schedules in the Brushtail Possum (Trichosurus Vulpecula)
The aim of this study was to compare the demand by possums for foods under different arrangements of concurrent progressive-ratio and fixed-ratio schedules of reinforcement. In Experiment 1, every possible food pair made up of berries, chicken, egg, foliage, insects and mushroom was presented (30 pairs in total). The requirement on the progressive-ratio schedule increased within a session and the fixed-ratio was kept constant at 30. In Experiment 2, a subset of the foods from Experiment 1 were used (chicken, mushroom, egg and berries) and in separate conditions the fixed-ratio was either 30 or 10 responses. In Experiment 3, the foods were the same as used in Experiment 2 and the progressive-ratio schedule increased every five sessions and the fixed-ratio schedule was 30 responses. Exponential models of demand were applied to consumption rates to compare the parameters of initial demand, essential value and P max, and break point and cross point across foods. The models described the data well and consumption rates were similar when the incrementing schedules increased within- and across sessions. Demand was highest for berries, egg and locust in Experiment 1 and egg and chicken in Experiments 2 and 3. This finding has practical implications for understanding possum food preferences in the wild as a function of other available food sources.
In a widely used animal-metacognition paradigm, monkeys are positively reinforced with food for correct classifications of stimuli as sparse or dense and punished with timeouts for incorrect responses, but they also have access to an “uncertainty” response that moves them to the next trial without either of these forms of feedback. Rhesus monkeys use this uncertainty response most often for trials on which they are at greatest risk for making an error, suggesting that they are monitoring their ability to make these classifications. Capuchin monkeys do not succeed to the same degree on these tasks—conceivably as a result of differential contingencies in place between the sparse/dense responses (food delivery or timeout) and the uncertainty response (avoidance of a timeout but also no chance for food reward). Here, we used a novel variation of this task in which the outcomes of the three response classes (sparse, dense, uncertain) were functionally equivalent. All responses simply determined the delay interval before presentation of a second task (matching-to-sample), and that task yielded potential food rewards. Overall, capuchin monkeys used the dense and sparse responses appropriately, including some animals that had no prior experience in performing this classification task. However, none used the uncertainty response appropriately even when it was placed on the same contingency plane as the dense and sparse responses. This suggests that the failure of capuchin monkeys to use an uncertainty response is not the result of that response producing a qualitatively different outcome compared to the dense and sparse responses.
While relatively little is known about the psychology of domestic pigs, what is known suggests that pigs are cognitively complex and share many traits with animals whom we consider intelligent. This paper reviews the scientific evidence for cognitive complexity in domestic pigs and, when appropriate, compares this literature with similar findings in other animals, focusing on some of the more compelling and cutting-edge research results. The goals of this paper are to: 1) frame pig cognition and psychology in a basic comparative context independent of the livestock production and management setting; and 2) identify areas of research with pigs that are particularly compelling and in need of further investigation. We summarize and discuss several areas of comparative psychology, including nonsocial and social cognition, self-awareness, emotion, and personality. We conclude that there are several areas of research in which the findings are suggestive of complex psychology in pigs. We conclude by calling for more noninvasive cognitive and behavioral research with domestic pigs in non-laboratory settings that allow them to express their natural abilities.
Age Influences Male's Mating Preferences for Multiparous and Nulliparous Females in the Laboratory-bred Macaca Fascicularis
Most studies of partner choice and preferences in nonhuman primates have been concerned with male social rank. Females select high-ranking males, and high-ranking females can more readily gain access to males. Although researchers have mentioned males’ choices and their preferences for females, papers that focus on male preferences have been few. Past studies suggested that male primates prefer older females. We analyzed data collected by the every-other-day mating system in which a male alternately lived with each of two females. This allowed us to compare the biological and/or physical characteristics of pregnant females with those of nonpregnant females and to minimize social factors. Multiparous and/or older females were more frequently pregnant. In the nulliparous group, females that became pregnant for the first time tended to be younger than the other nonpregnant females paired in the mating set. While males generally prefer parous females, males prefer to mate with younger females among females that have never been pregnant. A female that gets pregnant at a young age could become pregnant again later in life. Such a mating preference would guarantee higher reproductive success for females in the wild as well.
This article consists of two experiments reporting conditioned flavor avoidance (or taste aversion) in laboratory rats that swam in the flavored water. A statistically reliable effect was demonstrated in Experiment 1 by using a simple conditioning procedure with sweet (sodium saccharin) water. Compared with control rats that had no swimming experience or those that swam in tap water, experimental rats showed avoidance of the sweet water in the choice test between it and tap water, if they had swum in the sweet water for 20 min over four days. Rinsing the rats off with tap water after the swimming had no effect on this flavor avoidance learning. This finding suggests that tasting the sweet water during swimming was critical. Experiment 2 confirmed the flavor avoidance learning in swimming rats by a differential conditioning procedure with sour (citric acid) and bitter (denatonium benzoate) solutions. Although the effect size was relatively small in the two experiments reported here, this new procedure may contribute to future research concerning Pavlovian conditioning due to its procedural simplicity.
Introduction to the Special Issue on Biological Constraints on Learning
This special section stemmed from a symposium on biological constraints on learning that was organized for the XVII Biennial Meeting of the International Society for Comparative Psychology in Bogotá, Colombia, September 2014. The symposium sought to revisit the topic of biological constraints on learning. Such is the aim of this special section as well, guided by the conviction that the topic demands further study. Some of the most important researchers in the area were invited to contribute. Their acceptance indicates the topic’s enduring importance.
Special Issue on Biological Constraints on Learning
The first and most prominent study of selective associations was the so-called bright-noisy-water experiment by Garcia and Koelling (1966). This study was a landmark in the development of thinking about biological constraints on learning and remains the most highly cited study of selective associations, even though it lacked important controls. I first describe the original experiment and initial criticisms of it. I then discuss the various control issues that were ignored in the original experiment but addressed in subsequent research. In this account, I rely primarily on research conducted in my laboratory, because the problems were not addressed by any other investigators. Along the way, I discuss the discovery of a selective sensitization effect related to the Garcia-Koelling findings, ways to rule out selective sensitization, and studies of selective associations in pre-weanling rats. I conclude with a look back at the impact of the Garcia-Koelling experiment and recommendations for new generations of students in the field.
Contemporary models for the evolution of learning suggest that environmental predictability plays a critical role in whether learning is expected to evolve in a particular species, a claim originally made over 50 years ago. However, amongst many behavioral scientists who study insect learning, as well as amongst neuroscientists who study the brain architecture of insects, a very different view is emerging, namely that all animals possessing a nervous system should be able to learn. More specifically, the capacity for associative learning may be an emergent property of nervous systems such that, whenever selection pressures favor the evolution of nervous systems, for whatever reason, the capacity for associative learning follows ipso facto. One way to reconcile these disparate views of learning is to suggest that the assumed default in these evolutionary models, namely the non-learning phenotype, is incorrect: The ability to learn is, in fact, the default but, under certain conditions, selection pressures can override that ability, resulting in hard-wired, or considerably less plastic, responses. Thus, models for the evolution of learning actually may be models for the conditions under which inherent plasticity is overridden. Moreover, what have been revealed as the costs of learning in insects may, instead, be costs associated with far more complex cognitive feats than simple associative learning – cognitive skills that researchers are just now beginning to reveal.
Testing hypotheses about evolved psychological adaptations is the purview of human evolutionary psychology (HEP). A basic tenet of HEP is that the brain is comprised of specialized modules that evolved in response to selection pressures present in ancestral environments, and these modules support domain specific behavioral and cognitive processes that promoted survival and reproductive fitness during human evolutionary history. One set of cognitive domains involves learning and memory, and HEP has attempted to account for how evolutionary processes have shaped the design features supporting how humans acquire, store and retrieve information. Similarly, comparative psychology recognizes that cognitive traits of humans and animals are specialized to meet specific environmental challenges. However, these specializations are not regarded as species-specific, but rather reflect either adaptive modifications of general memory processes (e.g., episodic), or are processes that support a specific type of learning (e.g., taste aversions, imprinting, song learning). These alternatives to HEP emphasize the presence of quantitative rather than qualitative differences in learning and memory abilities. The goal of this paper is to examine these contrasting approaches of HEP and comparative psychology, and, using the survival processing effect (Nairne, Thompson, & Pandeirada, 2007, 2008) as an example, evaluate the plausibility of domain-specific adaptive hypotheses of human memory.
From Biological Constraints to Flexible Behavior Systems: Extending Our Knowledge of Sexual Conditioning in Japanese Quail
The traditional learning view involves the general process theory of learning that focuses onidentifying universal principles that apply to all species capable of learning from experience, and that operate across a wide variety of situations. Examples of behavior that contradictgeneral-process conceptions of learning have been in the past referred to as “biologicalconstraints”. Traditional learning theorists choose to consider these examplesas exceptions to otherwise universal principles of learning. On the contrary, thetypical ethologist is more likely to be concerned with how specific behaviorsmay have evolved and in an animal’s species typical responses to stimuli theyare likely to encounter in their natural environment. However, they also fail to embrace animallearning phenomena that occurs in the laboratory into their theoreticalframework. Behavior systems represent analternative to this view by providing a link between traditional views oflearning and ethology. Theyconceptualize experiential learning not as a set of universal principles, butas species typical processes that reflect the specific demands of the ecological niche in which the species evolved. The current paper reviews andbrings-to-date Domjan’s formulation of a sexual behavior system in maleJapanese quail. The system includes astimulus dimension consisting of species typical cues, local cues, andcontextual cues, and a response dimension consisting of general search, focalsearch, and copulatory behavior. Domjan’sformulation includes two diagrams that include symbols that representunconditioned and conditioned effects within the system. Our modification of the system focuses onadditional and up-to-date conditioned effects. In general, adding conditioning to the system increases potentialstimulus and response variation, thereby increasing the flexibility of thesystem as it has evolved as a result of continued observation andexperimentation.
Taste aversion learning (aka conditioned taste aversions or CTA) can occur even when there is delay of some hours between experience of the taste and the subsequent onset of illness. This property of CTA is quite distinct from other forms of associative learning, where typically no association between two events is acquired if they are separated by more than a minute. This paper provides an overview of a series of recent experiments based on the assumption that long-delay CTA is possible only when no potentially overshadowing – or ‘concurrently interfering’ (Revusky, 1971) – events occur during the delay. The general method is one in which in a single conditioning session the rats are first given 8% sucrose, providing the sweet target taste, and 65 min later are injected with lithium chloride. What vary across experiments are the potentially interfering events occurring during the 65-min delay period. When the interfering event is a second, and quite different, taste, namely sour-tasting hydrochloric acid solution (HCl), this produces 1-trial overshadowing of the sucrose aversion, to a degree that is greater when HCl is given late in the delay period, greater when HCl is given in the same context as sucrose and greater when HCl has not been pre-exposed. Other intervening events can also overshadow sucrose aversion learning. These include placement into a novel context, as long as this occurs immediately before injection, and even stimuli that evoke memories of food-related experiences. These results can be accounted for by adding to the Rescorla-Wagner model (Rescorla & Wagner, 1972) the assumption that sickness is comprised of a succession of mini-bouts and the assumption that context-event associations (Wagner, 1981) are important in long-delay CTA.
Research concerned with visual dominance in appetitive and auditory dominance in aversive learning situations (selective associations) is reviewed. The present analysis stresses that the dominant sensory modality of stimulus control is determined by the relative affective valence acquired by a compound auditory-visual stimulus through reinforcement contingencies, rather than by whether the primary reinforcer is appetitive or aversive. For example, take two groups of rats or pigeons on exactly the same shock-avoidance contingency in a tone-light compound (TL), but with different contingencies when the compound is absent (TL). Responding came predominantly under (1) auditory control when conditions in TL were hedonically negative relative to those in (TL), and (2) visual control when conditions in (TL) made TL relatively positive. Selective associations here are a product of the relative hedonic state, positive or negative, established to the auditory-visual compound. Therefore, this constraint reflects a high level of functioning by a hedonic comparator -- with TL’s hedonic value contextually determined by the totality of the events encountered, and reinforcement contingencies, operating in its world. The physical particulars of the reinforcer in TL here, shock avoidance, clearly were not responsible for the hedonic psychological state TL produced. Weiss, Panlilio, and Schindler (1993a, 1993b) went on to show that these proclivities can be (1) reversed, and (2) overcome by a blocking design when the biologically-contingency-disadvantaged stimulus is first pretrained on its own. Relating the “hedonic model” to evolution is speculative. But, the hedonic model is scientifically integrative by relating this biological constraint to a variety of phenomena that involve incentive-motivational states. These include choice behavior, conditioned preference, behavioral contrast and appetitive-aversive interactions.
This paper describes a neural network account of misbehavior with an extant neural network model of conditioning. The model makes no distinction between learning (weight-change mechanisms) in operant and Pavlovian conditioning, but preserves the standard behavioral distinctions between types of stimuli, responses, and contingencies, with connectionist interpretations of some possible neuroanatomical substrates. Misbehavior has been traditionally conceived as a species-specific response R* that is unnecessary for a biologically significant reward S* but interferes with another response R that is necessary for S*. Misbehavior thus conceived has been explained as interfering Pavlovian conditioned responding. Three four-layer feedforward neural networks were designed to differ only in their output layers, as a connectionist interpretation of three hypothetical operant-Pavlovian relations in misbehavior, namely, interference (Pavlovian output to operant output lateral inhibitory connection), compatibility (Pavlovian output to operant output lateral excitatory connection), and independence (no lateral connection in the output layer). These relations are proposed as neural-network interpretations of neuroanatomical substrates of conditioning with three biologically significant stimuli, namely, food, water, and sexual mate, respectively. Each network first received pairings of contextual cues with its respective S*, to simulate pretraining with such stimuli. Then, networks received operant contingencies where S* was paired with the same contextual cues, as well as cues from a token dependently on R responding, defined as a minimal R activation of 0.5. Networks showed substantial misbehavior (qua conditioned R* responding) that interfered with R to different extents, food causing the most, sexual mate the least interference. Limitations, future directions, and implications for biological constraints and the generality of learning are discussed.
Many researchers have reported differential rates of learning and inferred selective associations between events reflecting biological constraints on learning that have evolved for each given species. Although we do not doubt that there are such biological constraints on learning, we suggest that some of the many claims may actually be spurious due to use of less than optimal research designs. We propose six methodological and inferential concerns that current researchers and reviewers of past research may find useful.
This is an obituary for Ethel Tobach, founder of this journal and of the International Society for Comparative Psychology.
- 1 supplemental image