Transfer of training following discrimination learning with two reinforced responses

Pigeons were given errorless discrimination training between chromatic stimuli with a specific reinforced response associated with each discriminative stimulus. These Ss subsequently acquired a successive go/no-go auditory discrimination (utilizing one response class) faster than Ss given single stimulus training. These results are more compatible with an explanation of transfer of training in successive operant discrimination learning based on general attention than one based on withholding responses.     

Second responses in paired-associate learning

Mathematical models of memory can be evaluated using a continuous paired-associate learning paradigm in which a second response is requested immediately following an error. This procedure provides discriminability among the models on the basis of the relationship between the first- and second-response error probabilities. Theoretical predictions are derived for a variety of models, including finite-state and noncontingent operator models, formulations based on the theory of signal detectability, the multicomponent memory model of Bower (1967b), and the long- and short-term model of Atkinson and Shiffrin (1968). None of these is unequivocally acceptable, and many are decisively rejected by the data of a single-trial experiment. The most satisfactory models are the multicomponent model and the Atkinson and Shiffrin model, both of which assume a greater degree of internal structure than the rejected models. It appears that internal structure of this sort will be a necessary part of any satisfactory model of second responses.     

One trial learning and the suppression of interference

An experiment was made to investigate the conditions under which a subject who has learned a series of items in one order can relearn the same items in a different order in one trial. The hypothesis tested was that one-trial learning is ensured in this situation if interference between the two tasks is reduced to zero. It is shown that complete suppression of such interference, obtained by overtraining on the task which is learned first, is not sufficient to ensure the learning of the second in one trial. Comparison of the present experiment with a similar one reported by Mandler and Heinemann (1956) suggests the hypothesis that the discovery of order in the stimulus situation is necessary, in addition to the suppression of interference, if one trial learning is to occur. The relation of the results to the problem of negative transfer is also discussed.     

Trials per problem and age as factors in learning set formation of children

Discrimination learning set performance was examined in preschool children as a function of age and number of trials per problem. Subjects 3, 4, 5, and 6 years old were trained over three 72-trial sessions. Half of the children at each age were given problems three trials in length, and half received 12-trial problems. As predicted, younger children (3 and 4-year-olds) performed significantly better for 12 than for three trials per problem, whereas no differences in performance were evident for older children. Response pattern analyses revealed that younger children, especially those receiving less exposure within problems, produced a higher proportion of stimulus alternation and position determined response sequences than did older children. These results contradict the commonly held assumption that learning set acquisition is simply a function of the total number of trials presented and indicate that the amount of exposure to individual problems is a factor in problem solution for young human subjects.     

Implicit learning and emotional responses in nine-month-old infants

To study the interplay between motor learning and emotional responses of young infants, we developed a contingent learning paradigm that included two related, difficult, operant tasks. We also coded facial expression to characterise emotional response to learning. In a sample of nine-month-old healthy Chinese infants, 44.7% achieved learning threshold during this challenging arm-conditioning test. Some evidence of learning was observed at the beginning of the second task. The lowest period of negative emotions coincided with the period of maximum movement responses after the initiation of the second task, and movement responses negatively correlated with the frequency of negative emotions. Positive emotions, while generally low throughout the task, increased during peak performance especially for learners. Peak frequency of movement responses was positively correlated with the frequency of positive emotions. Despite the weak evidence of learning this difficult task, our results from the learners would suggest that increasing positive emotions, and perhaps down-regulating negative emotional responses, may be important for improving performance and learning a complex operant task in infancy. Further studies are necessary to determine the role of emotions in learning difficult tasks in infancy.     

Frequency and coding responses in verbal discrimination learning

Ss were given a verbal discrimination task where correct and incorrect stimuli appeared in one, two. or four different pairs in a within-Ss design. For a group tested with the usual anticipation procedure, correct repetitions helped and incorrect repetitions hurt performance. The pattern of results suggested that frequency is a salient cue in verbal discrimination learning but that coding or paired-associate processing may occur at the same time. Data from a second group of Ss for which frequency was a less salient cue gave more direct evidence for coding.     

Concurrent intramodal learning enhances multisensory responses of symmetric crossmodal learning in robotic audio-visual tracking

Tracking an audio-visual target involves integrating spatial cues about target position from both modalities. Such sensory cue integration is a developmental process in the brain involving learning, with neuroplasticity as its underlying mechanism. We present a Hebbian learning-based adaptive neural circuit for multi-modal cue integration. The circuit temporally correlates stimulus cues within each modality via intramodal learning as well as symmetrically across modalities via crossmodal learning to independently update modality-specific neural weights on a sample-by-sample basis. It is realised as a robotic agent that must orient towards a moving audio-visual target. It continuously learns the best possible weights required for a weighted combination of auditory and visual spatial target directional cues that is directly mapped to robot wheel velocities to elicit an orientation response. Visual directional cues are noise-free and continuous but arising from a relatively narrow receptive field while auditory directional cues are noisy and intermittent but arising from a relatively wider receptive field. Comparative trials in simulation demonstrate that concurrent intramodal learning improves both the overall accuracy and precision of the orientation responses of symmetric crossmodal learning. We also demonstrate that symmetric crossmodal learning improves multisensory responses as compared to asymmetric crossmodal learning. The neural circuit also exhibits multisensory effects such as sub-additivity, additivity and super-additivity.     

Stimulus generalization in two associative learning processes

Recent studies involving nonlinear discrimination problems suggest that stimuli in human associative learning are represented configurally with narrow generalization, such that presentation of stimuli that are even slightly dissimilar to stored configurations weakly activate these configurations. The authors note that another well-known set of findings in human associative learning, cue-interaction phenomena, suggest relatively broad generalization. Three experiments show that current models of human associative learning, which try to model both nonlinear discrimination and cue interaction as the result of 1 process, fail because they cannot simultaneously account for narrow and broad generalization. Results suggest that human associative learning involves (a) an exemplar-based process with configural stimulus representation and narrow generalization and (b) an adaptive learning process characterized by broad generalization and cue interaction.     

Concurrent schedules: How responses per reinforcer affects estimates of sensitivity to reinforcement

Conventionally, when choice is measured under concurrent schedules, all responses are included. However, the class of all responses consists of 2 sub-classes which are discriminable by their properties: Reinforced responses always equal obtained reinforcers, while unreinforced responses are free to vary. As a result, the inclusion of reinforced responses in choice measures results in sensitivity values in generalized matching that are biased toward larger values, and this bias becomes extreme in some combinations of overall response and reinforcer rates. Different ways of varying concurrent-schedule values also affect estimated sensitivity and the linearity between choice based on all responses and reinforcer ratios. To avoid spurious results and comparisons, and as a matter of good practice, generalized-matching fits and measures should be done using only unreinforced responses.     

The impact of motor responses on serial-pattern learning

Subjects are able to learn even very complex serial patterns in serial-reaction-time tasks. The investigation of the learning processes behind this phenomenon has yielded contradictory results. Some studies have come to the conclusion the subjects had learned the sequence of stimuli. Other studies have assumed that the sequence of responses had been learned or a combination of both stimuli and responses. The present experiments stress the impact of motor responses on serial-pattern learning. The subjects had to respond to serial targets that were presented within a matrix of distractors. The position of each target could be predicted from the identity and position of the previous target. If the subjects were to learn this pattern, they would be able to speed up the search for the target and give faster responses. The results indicated that the relation between the target identity and the position of the next target was acquired much better by those subjects who had to respond to each target with a special motor response. If the same response was required for the relevant targets, knowledge of the rule was somewhat fragmentary. To explain these results, mechanisms of motor learning and motor planning are discussed. It is assumed that learning of the rules occurs if the position changes appear to be effects of different motor responses.