Serial discrimination-reversal learning with and without errors by the California sea lion

A sea lion under water performed virtually without errors on a series of three form-discrimination reversals. Significant training requirements were the combining of a previously well-established size cue preference with the nonpreferred form cue, followed by the gradual reduction of the size cue until it was completely eliminated. Orienting responses reached a peak and then decreased during progressive-dimensional-change training, suggesting critical stages in the transition of attention from the size dimension to the form dimension. Further experimentation revealed that intensive training during these critical stages obviated the need to reduce very gradually the size cue. Without special training sea lions make perseverative errors on a series of form-discrimination reversals. "Emotional" or nontest-oriented behavior was associated only with the occurrence of successive errors.     

Influence of two types of information trials on the formation of a discrimination-reversal learning set in Japanese monkeys

Abstract: Three groups of male adult Japanese monkeys were trained in 30 discrimination-reversal learning-set (LS) problems using the Wisconsin General Test Apparatus. A control group ( N = 3) was trained in the standard procedure: acquisition of up to 12 consecutive correct responses, reversal to the same criterion, and then shift to a new problem. To the other two groups, three information trials signaling the stimulus value in reversal were given after they reached the criterion in acquisition. Subjects in the positive-stimulus information (PSI) group ( N = 4) were exposed only to the reversal positive stimulus, and a response to it was rewarded. Subjects in the negative-stimulus information (NSI) group ( N = 4) were exposed only to the reversal negative stimulus, and any response to it was not rewarded. Information trials reduced reversal errors and NSI trials were more effective than PSI trials in error reduction ( p < .05). Information trials also aided in acquisition: the NSI and PSI groups learned the acquisition phase of the problems with significantly fewer errors than control group ( p < .01) in the last stage of training. The implications of these results are discussed in relation to strategy learning and attention learning in LS formation.     

Sequential learning using temporal context

The temporal context model (TCM) has been extensively applied to recall phenomena from episodic memory. Here we use the same formulation of temporal context to construct a sequential learning model called the predictive temporal context model (pTCM) to extract the generating function of a language from sequentially-presented words. In pTCM, temporal context is used to generate a prediction vector at each step of learning and these prediction vectors are in turn used to construct semantic representations of words on the fly. The semantic representation of a word is defined as the superposition of prediction vectors that occur prior to the presentation of the word in the sequence. Here we create a formal framework for pTCM and prove several useful results. We explore the effect of manipulating the parameters of the model on learning a sequence of words generated by a bi-gram generating function. In this simple case, we demonstrate that feeding back the constructed semantic representation into the temporal context during learning improves the performance of the model when trained with a finite training sequence from a language with equivalence classes among some words. We also describe , a variant of the model that is identical to pTCM at steady state. has significant computational advantages over pTCM and can improve the quality of its prediction for some training sequences.     

Concurrent learning of temporal and spatial sequences

In a serial reaction time task, stimulus events simultaneously defined spatial and temporal sequences. Responses were based on the spatial dimension. The temporal sequence was incidental to the task, defined by the response-to-stimulus intervals in Experiment 1 and stimulus onset asynchronies in Experiment 2. The two sequences were either of equal length and correlated or of unequal length. In both experiments, spatial learning occurred regardless of sequence length condition. In contrast, temporal learning occurred only in the correlated condition. These results suggest that timing is an integrated part of action representations and that incidental learning for a temporal pattern does not occur independently from the action. Interestingly, sequence learning was enhanced in the correlated condition, reflecting the integration of spatial-temporal information.     

Rapid learning of rapid temporal contexts

In an information-rich visual world and with limited attentional resources, what visual cues allow humans to efficiently navigate their environment? One key environmental characteristic is that stimuli rarely appear in isolation and typically coincide with other specific items that provide cues regarding where and when to guide our attention. Indeed, a predictive spatial context of distractors can enhance the deployment of attention to a target location (Chun &; Jiang, 1998). However, can a predictive, temporal sequence of distractors, which do not enter working memory, cue when to allocate attention? Previous studies addressing this question have employed relatively long (?500 msec/item) stimulus exposure durations. Thus, this temporal cuing may require extensive processing of the distractors. Here, we show that a rapidly presented (～100 msec/item), predictive, temporal context, where stimuli undergo only preliminary analysis, can facilitate the deployment of attention to a specific temporal location.     

The procedural learning of action order is independent of temporal learning

How does learning the timing of actions influence our ability to learn the order of actions? A sequence of responses cued by spatial stimuli was learned in a serial reaction time task where the response-to-stimulus intervals (RSIs) were random, constant, or followed a fixed sequence. In this final sequenced-RSI condition, the response and RSI sequences were consistently matched in phase and could be integrated into a common sequence representation. The main result was that the response sequence was learned to a similar degree in all RSI training conditions, indicating that neither the predictability of RSIs nor the integration of the phase-matched response and timing sequences benefited learning of the response sequence. Nevertheless, temporal learning and integration speeded up performance without strengthening the representation of response order.     

Errorless learning of a conditional temporal discrimination

In the present study we extended errorless learning to a conditional temporal discrimination. Pigeons' responses to a left-red key after a 2-s sample and to a right-green key after a 10-s sample were reinforced. There were two groups: One learned the discrimination through trial and error and the other through an errorless learning procedure. Then, both groups were presented with three types of tests. First, they were exposed to intermediate durations between 2 s and 10 s, and given a choice between both keys (stimulus generalization test). Second, a delay from 1 s to 16 s was included between the offset of the sample and the onset of the choice keys (delay test). Finally, pigeons learned a new discrimination in which the stimuli were switched (reversal test). Results showed that pigeons from the Errorless group made significantly fewer errors than those in the Trial-and-Error group. Both groups performed similarly during the stimulus generalization test and the reversal test, but results of the delay test suggested that, on long stimulus trials, responding in the errorless training group was less disrupted by delays.     

The implicit learning of metrical and nonmetrical temporal patterns

Implicit learning (IL) occurs unintentionally. IL of temporal patterns has received minimal attention, and results are mixed regarding whether IL of temporal patterns occurs in the absence of a concurrent ordinal pattern. Two experiments examined the IL of temporal patterns and the conditions under which IL is exhibited. Experiment 1 examined whether uncertainty of the upcoming stimulus identity obscures learning. Based on probabilistic uncertainty, it was hypothesized that stimulus-detection tasks are more sensitive to temporal learning than multiple-alternative forced-choice tasks because of response uncertainty in the latter. Results demonstrated IL of metrical patterns in the stimulus-detection but not the multiple-alternative task. Experiment 2 investigated whether properties of rhythm (i.e., meter) benefit IL using the stimulus-detection task. The metric binding hypothesis states that metrical frameworks guide attention to periodic points in time. Based on the metric binding hypothesis, it was hypothesized that metrical patterns are learned faster than nonmetrical patterns. Results demonstrated learning of metrical and nonmetrical patterns but metrical patterns were not learned more readily than nonmetrical patterns. However, abstraction of a metrical framework was still evident in the metrical condition. The present study shows IL of auditory temporal patterns in the absence of an ordinal pattern.     

The temporal structure of goal-directed and habitual operant behavior

Operant behavior can reflect the influence of goal-directed and habitual processes. These can be distinguished by changes to response rate following devaluation of the reinforcing outcome. Whether a response is goal directed or habitual depends on whether devaluation affects response rate. Response rate can be decomposed into frequencies of bouts and pauses by analyzing the distribution of interresponse times. This study sought to characterize goal-directed and habitual behaviors in terms of bout-initiation rate, within-bout response rate, bout length, and bout duration. Data were taken from three published studies that compared sensitivity to devaluation following brief and extended training with variable-interval schedules. Analyses focused on goal-directed and habitual responding, a comparison of a habitual response to a similarly trained response that had been converted back to goal-directed status after a surprising event, and a demonstration of contextual control of habit and goal direction in the same subjects. Across experiments and despite responses being clearly distinguished as goal directed and habitual by total response rate, analyses of bout-initiation rate, within-bout rate, bout length, and bout duration did not reveal a pattern that distinguished goal-directed from habitual responding.     

Human sensitivity to temporal proximity: The role of spatial and temporal speed gradients

Estimates of temporal proximity (sometimes called time-to-collision) from random-dot flow patterns are shown to be based upon retinal speed, rather than upon changes in dot density. Neither the spatial nor the temporal gradient of motion is essential to the task, but estimates can be made from either alone. Performance is unaffected by the addition of rotational motion, suggesting that observers axe capable of extracting the radial component of motion, which contains all the relevant information, from complex stimuli.     

Scaling behavior in the temporal context model

The Temporal Context Model (TCM) postulates a distributed representation of temporal context that provides the cue for episodic recall tasks. TCM, coupled with the Luce Choice Rule for determining probability of recall, a conjunction referred to as TCMFR, is able to explain the existence of the long-term recency effect, as well as predicting the persistence of associative effects even with the inclusion of a delay between items. Here, quantitative predictions of TCMFR such as the magnitude of the delay interval is increased in continuous-distractor free recall are developed. The magnitude of the recency effect is operationally defined as the ratio of the probability of first recall (PFR) of the last list item to the PFR of the next-to-last item. Properties of associative effects are operationalized by using analogous measures derived from conditional response probability (CRP) curves. TCMFR predicts a decrease in recency with increasing delay. The rate of this decay and the qualitative pattern of change with increasing delay depend on the rate of contextual drift. For a range of values of the rate of contextual drift, TCMFR also predicts a transient increase in the recency effect as the length of the delay increases from zero. The model predicts that contiguity effects in free recall should follow a similar pattern, but that associative asymmetry, ubiquitously observed in free recall, should decay monotonically with increases in the delay interval.     

The intrinsic link between motor behavior and temporal cognition

The debate about the cognitive mechanisms behind human temporal processing has raged for decades without a clear resolution. The theory presented here describes a different perspective to the traditional accounts on the issue, namely, that motor behaviors or sequences of motor behaviors provide a means of reproducing time intervals. Evidence behind this perspective includes tapping strategies (exemplified by musicians), counting strategies, and neuropsychological results showing activation of motor areas during temporal cognitive tasks. I propose that motor behaviors aid human timing by offering a set of processes that consistently take a set amount of time to accomplish. Motor behaviors also allow segmentation of larger intervals into smaller intervals that are easier to estimate. I conclude with a discussion of implications of this perspective on temporal cognition.     

Species differences in temporal control of behavior

Temporal control of rats' and pigeons' responding was analyzed and compared in detail on fixed-interval and fixed-time schedules with parameters of 30, 60, and 120 seconds. On fixed-time schedules, rats' responding decreased greatly or ceased, whereas pigeons continued to respond, especially on low schedule values. The running rate of responses (calculated by excluding the postreinforcement pause) was related to the duration of the preceding postreinforcement pause for rats but not for pigeons. Changes in response rate in successive segments of the interval were best described by normal curves. The relationship between midpoints of the normal curves and schedule value was a power function, with an exponent of less than one for pigeons but greater than one for rats. These differences could be explained in terms of a basic difference between the key-peck and lever-press responses, the two being differently affected by the response-eliciting properties of food.     

Human simulations of vocabulary learning

The work reported here experimentally investigates a striking generalization about vocabulary acquisition: Noun learning is superior to verb learning in the earliest moments of child language development. The dominant explanation of this phenomenon in the literature invokes differing conceptual requirements for items in these lexical categories: Verbs are cognitively more complex than nouns and so their acquisition must await certain mental developments in the infant. In the present work, we investigate an alternative hypothesis; namely, that it is the information requirements of verb learning, not the conceptual requirements, that crucially determine the acquisition order. Efficient verb learning requires access to structural features of the exposure language and thus cannot take place until a scaffolding of noun knowledge enables the acquisition of clause-level syntax. More generally, we experimentally investigate the hypothesis that vocabulary acquisition takes place via an incremental constraint-satisfaction procedure that bootstraps itself into successively more sophisticated linguistic representations which, in turn, enable new kinds of vocabulary learning. If the experimental subjects were young children, it would be difficult to distinguish between this information-centered hypothesis and the conceptual change hypothesis. Therefore the experimental "learners" are adults. The items to be "acquired" in the experiments were the 24 most frequent nouns and 24 most frequent verbs from a sample of maternal speech to 18-24-month-old infants. The various experiments ask about the kinds of information that will support identification of these words as they occur in mother-to-child discourse. Both the proportion correctly identified and the type of word that is identifiable changes significantly as a function of information type. We discuss these results as consistent with the incremental construction of a highly lexicalized grammar by cognitively and pragmatically sophisticated human infants, but inconsistent with a procedure in which lexical acquisition is independent of and antecedent to syntax acquisition.