Learning from feedback: exactingness and incentives.

In a series of five experiments, exactingness, or the extent to which deviations from optimal decisions are punished, is studied within the context of learning a repetitive decision-making task together with the effects of incentives. Results include the findings that (a) performance is an inverted-U shaped function of exactingness, (b) performance is better under incentives when environments are lenient but not when they are exacting, (c) the interaction between exactingness and incentives does not obtain when an incentives function fails to discriminate sharply between good and bad performance, and (d) when the negative effects of exactingness on performance are eliminated, performance increases with exactingness.     

Movement-produced feedback as a mechanism for the temporal anticipation of motor responses

That part of the input hypothesis of motor timing behavior which postulates that movement-produced feedback can serve as a mechanism for the temporal anticipation of motor responses was tested. 32 male college Ss temporally anticipated (no preview) the coincidence of a moving pointer with a stationary one and executed the timing response with his right hand while either a high (HFB) or low (LFB) level of movement-produced feedback was indirectly manipulated in the left arm. The HFB group temporally anticipated with greater accuracy than the LFB group which supported that part of the input hypothesis being tested. Apparently, Ss were relying on proprioceptive feedback about the position of their left arm movement to cue the timing response of their right hand.     

A preliminary report of oral tactile sensation and responses to delayed auditory feedback

Individual speakers vary widely in their reactions to delayed auditory feedback. In this pilot investigation, 21 adults with normal speech and hearing completed standardized readings under simultaneous and 200-msec. delayed auditory feedback. Minimally and maximally affected speakers were identified. Vibrotactile thresholds were obtained from the tongue and thenar eminence of the right hand for a group of five minimally and a group of five maximally affected speakers. Analysis showed that vibrotactile thresholds were lower (more sensitive) at both assessment sites for the group of maximally affected speakers. This finding is discussed in relation to the possible implications of heightened tactile sensitivity in the control of speech under delayed auditory feedback.     

Children do not recalibrate motor‐sensory temporal order after exposure to delayed sensory feedback

Prolonged adaptation to delayed sensory feedback to a simple motor act (such as pressing a key) causes recalibration of sensory‐motor synchronization, so instantaneous feedback appears to precede the motor act that caused it (Stetson, Cui, Montague & Eagleman, 2006). We investigated whether similar recalibration occurs in school‐age children. Although plasticity may be expected to be even greater in children than in adults, we found no evidence of recalibration in children aged 8–11 years. Subjects adapted to delayed feedback for 100 trials, intermittently pressing a key that caused a tone to sound after a 200 ms delay. During the test phase, subjects responded to a visual cue by pressing a key, which triggered a tone to be played at variable intervals before or after the keypress. Subjects judged whether the tone preceded or followed the keypress, yielding psychometric functions estimating the delay when they perceived the tone to be synchronous with the action. The psychometric functions also gave an estimate of the precision of the temporal order judgment. In agreement with previous studies, adaptation caused a shift in perceived synchrony in adults, so the keypress appeared to trail behind the auditory feedback, implying sensory‐motor recalibration. However, school children of 8 to 11 years showed no measureable adaptation of perceived simultaneity, even after adaptation with 500 ms lags. Importantly, precision in the simultaneity task also improved with age, and this developmental trend correlated strongly with the magnitude of recalibration. This suggests that lack of recalibration of sensory‐motor simultaneity after adaptation in school‐age children is related to their poor precision in temporal order judgments. To test this idea we measured recalibration in adult subjects with auditory noise added to the stimuli (which hampered temporal precision). Under these conditions, recalibration was greatly reduced, with the magnitude of recalibration strongly correlating with temporal precision.     

Antipsychotic dose modulates behavioral and neural responses to feedback during reinforcement learning in schizophrenia

Schizophrenia is characterized by an abnormal dopamine system, and dopamine blockade is the primary mechanism of antipsychotic treatment. Consistent with the known role of dopamine in reward processing, prior research has demonstrated that patients with schizophrenia exhibit impairments in reward-based learning. However, it remains unknown how treatment with antipsychotic medication impacts the behavioral and neural signatures of reinforcement learning in schizophrenia. The goal of this study was to examine whether antipsychotic medication modulates behavioral and neural responses to prediction error coding during reinforcement learning. Patients with schizophrenia completed a reinforcement learning task while undergoing functional magnetic resonance imaging. The task consisted of two separate conditions in which participants accumulated monetary gain or avoided monetary loss. Behavioral results indicated that antipsychotic medication dose was associated with altered behavioral approaches to learning, such that patients taking higher doses of medication showed increased sensitivity to negative reinforcement. Higher doses of antipsychotic medication were also associated with higher learning rates (LRs), suggesting that medication enhanced sensitivity to trial-by-trial feedback. Neuroimaging data demonstrated that antipsychotic dose was related to differences in neural signatures of feedback prediction error during the loss condition. Specifically, patients taking higher doses of medication showed attenuated prediction error responses in the striatum and the medial prefrontal cortex. These findings indicate that antipsychotic medication treatment may influence motivational processes in patients with schizophrenia.     

Portable delayed auditory feedback device: A preliminary report

This writing is a preliminary report describing a prototype of a portable delayed auditory feedback device called the DYSTECH 1000. This writing and our observations are presented to stimulate review and feedback by our colleagues of the DYSTECH 1000 as a clinical and generalization tool.     

Speech production mechanisms in aphasia: A delayed auditory feedback study

Data collected from a delayed auditory feedback (DAF) study conducted by F. Boller, P. B. Vrtunski, Y. Kim, and J. L. Mack (1978, Cortex, 14, 212–226) were reanalyzed to determine if speech quality measures which were significantly affected during DAF could be operationalized and to ascertain which type of errors are responsible for the judgement. A second goal of this study was to determine if Conduction aphasics were relatively spared under conditions of DAF, as has been shown in previous studies. Twenty aphasic patients and ten controls were presented with several speech production tasks under two delay conditions and two simultaneous feedback conditions. Results indicated that only the category of vowel length produced a significant number of errors, suggesting that DAF affects articulatory implementation. Conduction aphasics showed the least effect of all the groups. These results support Wernicke's and Geschwind's model, in which Conduction aphasics have a disconnection between the sensory and motor images of words. In contrast, Broca's aphasics were the most affected group under DAF, suggesting that a major component of their disorder is phonetic in nature.     

Nonconscious temporal cognition: Learning rhythms implicitly

Two experiments demonstrate that people can implicitly learn rhythms. Participants responded to a series of fast-paced beeps by pressing a key as soon as possible after each beep. They were not told that the duration (180, 450, or 1,125 msec) between each keypress and the next beep was specified by a repeating sequence. In both experiments, participants responded significantly faster to predictable, sequenced timing than to random timing but did not show more knowledge of the sequence than did control participants on explicit memory measures. This dissociation was obtained even with an explicit memory test in Experiment 2 that maintained the same context and response metric as the implicit task to maximize the transfer of relevant knowledge. Implications for temporal cognition are discussed.     

Speech disruption under delayed auditory feedback in multilingual speakers

The present study investigated the language familiarity hypothesis formulated by Mackay [(1970). How does language familiarity influence stuttering under delayed auditory feedback? Perceptual and Motor Skills, 30, 655-669] that bilinguals speak faster and stutter less under delayed auditory feedback (DAF) when speaking their more familiar language than a less familiar language. Thirty normally fluent native speakers of Dutch (17 males and 13 females, aged between 18;1 and 26;4 years) who were also proficient in French and English read meaningful and nonsense text under DAF in their mother tongue and in the two later acquired languages. The existence of a language familiarity effect was confirmed. The participants required significantly more time and showed significantly more speech disruptions under DAF in the later acquired languages than in the mother tongue, and reading time and number of speech disruptions was significantly higher for the nonsense texts than for the meaningful text for each of the three languages. An additional question addressed was whether or not there were any gender differences in the susceptibility to DAF. Results did not reveal a clear gender difference. EDUCATIONAL OBJECTIVES: The reader will be able to: (1) summarize the importance of language familiarity for the degree of speech disruption experienced by normally fluent multilingual speakers under delayed auditory feedback; and (2) describe gender differences in the susceptibility to delayed auditory feedback.     

I: Effect of delayed auditory feedback on speech and keytapping

The effects of delayed auditory feedback (DAF) on the performance of simple speech and keytapping tasks were compared. In the first experiment, each of the fourteen young adult subjects was asked to repeat the speech sound “b” in groups of three sounds. The subject heard his undelayed speech sounds through earphones. Under the delay condition, the speech sounds were delayed 244 millisec. before being returned to the subject's earphones. In the second experiment, each of the same subjects was asked to tap on a key in groups of three taps. The subjects heard an undelayed click through earphones each time they tapped on the key. Under the delay condition, the clicks were delayed 244 millisec. before being returned to the earphones. Graphic recordings were made of the time and intensity characteristics of tapping under both conditions.

The changes which occurred in speech and keytapping under DAF were qualitatively the same and consisted of increases in intensity and unit-to-unit time of both taps and speech sounds. When performing under DAF, the subjects also exhibited a tendency to repeat more sounds or to tap more times than had been requested. The percentage of change in intensity and time characteristics for speech and keytapping performances which occurred under DAF were not significantly different for the two motor systems. The amount of change in speech performance under DAF did not correlate significantly with the amount of change in keytapping performance.     

Learning how to learn from social feedback: The origins of early vocal development

Infants’ prelinguistic vocalizations reliably organize vocal turn-taking with social partners, creating opportunities for learning to produce the sound patterns of the ambient language. This social feedback loop supporting early vocal learning is well-documented, but its developmental origins have yet to be addressed. When do infants learn that their non-cry vocalizations influence others? To test developmental changes in infant vocal learning, we assessed the vocalizations of 2- and 5-month-old infants in a still-face interaction with an unfamiliar adult. During the still-face, infants who have learned the social efficacy of vocalizing increase their babbling rate. In addition, to assess the expectations for social responsiveness that infants build from their everyday experience, we recorded caregiver responsiveness to their infants’ vocalizations during unstructured play. During the still-face, only 5-month-old infants showed an increase in vocalizing (a vocal extinction burst) indicating that they had learned to expect adult responses to their vocalizations. Caregiver responsiveness predicted the magnitude of the vocal extinction burst for 5-month-olds. Because 5-month-olds show a vocal extinction burst with unfamiliar adults, they must have generalized the social efficacy of their vocalizations beyond their familiar caregiver. Caregiver responsiveness to infant vocalizations during unstructured play was similar for 2- and 5-month-olds. Infants thus learn the social efficacy of their vocalizations between 2 and 5 months of age. During this time, infants build associations between their own non-cry sounds and the reactions of adults, which allows learning of the instrumental value of vocalizing.     

Learning when reward is delayed: a marking hypothesis

Rats were trained on spatial discriminations in which reward was delayed for 1 min. Experiment 1 tested Lett's hypothesis that responses made in the home cage during the delay interval are less likely to interfere with learning than responses made in the maze. Experimental subjects were transferred to their home cages during the delay interval, and control subjects were picked up but then immediately replaced in the maze. Contrary to Lett's hypothesis, both groups learned. Further experiments suggested that handling following a choice response was the crucial variable in producing learning: No learning occurred when handling was delayed (Experiment 2) or omitted (Experiment 3). One possible explanation for the fact that handling facilitated learning is that it served to mark the preceding choice response in memory so that subjects were then more likely to recall it when subsequently reinforced. In accordance with this interpretation, learning was found to be just as strong when the choice response was followed by an intense light or noise as by handling (Experiment 4). The implication of marking for other phenomena such as avoidance, quasi-reinforcement, and the paradoxical effects of punishment is also discussed.     

The role of feedback in shaping responses to risky road scenarios: Evidence from electrodermal activity

When faced with hazardous driving situations, rapid and effective risk perception and decision-making processes are of crucial importance for avoiding crashes. In these cases, the processes are accompanied and influenced by underlying psychophysiological mechanisms such as electrodermal activity (EDA) modulations. The present work aims to assess the psychophysiological mechanisms underlying participants’ risk perception and decision-making when facing risky road scenarios, as correlated to the feedback role in modulating participants’ behavior. Study 1 (n = 32) explores the behavioral effects of administering a contingent feedback in a decision (decision-making) and an evaluation (risk perception) task in response to a set of risky and not risky images. The behavioral data reveal an effect on the participants’ probability of response, independently from the type of image presented, when the feedback was administered. In the decision task, the effect is accompanied by a change in the amplitude and percentage of the skin conductance responses (SCRs), which are moderated by block of trials. Study 2 (n = 44) better assesses the role of task and block on participants’ physiological activation, as measured by EDA signal. The results show an increase in psychophysiological activation when the feedback is delivered, in the first part of the tasks, both in terms of SCRs amplitude and percentage to the presented road scenarios, followed by a decrease in the second part of the tasks. Moreover, this effect is more evident in the decision task than the evaluation task. These findings suggest that the role exerted by feedback when facing risky traffic images may be described as based on an associative process that, once the correct response has been learned, tends to be reduced as it becomes automatic. Overall, the results of the two studies represent an important step toward the development of training programs aimed at promoting safer behaviors in risky driving contexts.     

Biased reasoning: adaptive responses to health risk feedback

The present study examined reactions toward repeated self-relevant feedback. Participants in a community health screening received feedback about their cholesterol level on two separate occasions. Reactions to the first feedback were examined with regard to feedback valence and expectedness. The findings showed that negative feedback was devalued, but only when it was unexpected. Feedback consistency was incorporated into analyses of the second feedback. Again, results showed that negative feedback was not always devalued-only when it was inconsistent with the first feedback. Furthermore, positive feedback was not unconditionally accepted. When receiving unexpected positive feedback of low consistency, recipients were doubtful about its accuracy. Conversely, expected positive feedback was accepted regardless of its consistency. These results suggest that negative or unexpected positive feedbacks evoke greater sensitivity to feedback consistency, indicating elaborate cognitive processing. Theoretical accounts of these findings are discussed.     

Efficacy of delayed auditory feedback for treating cluttering: Two case studies

Two clutterers were treated with similar preplanned delayed auditory feedback (DAF) procedures, using probe samples in which the DAF was not present to measure treatment efficacy. Whereas both clients met established fluency criteria during treatment, both clutterers had difficulty transferring gains to probe sessions—one more than the other. Differences in results of DAF treatment are discussed from the perspectives of differences in cluttering severity, coexisting disorders, and supplementary clinical techniques.     

Solving the credit assignment problem: explicit and implicit learning of action sequences with probabilistic outcomes

In most problem-solving activities, feedback is received at the end of an action sequence. This creates a credit-assignment problem where the learner must associate the feedback with earlier actions, and the interdependencies of actions require the learner to remember past choices of actions. In two studies, we investigated the nature of explicit and implicit learning processes in the credit-assignment problem using a probabilistic sequential choice task with and without a secondary memory task. We found that when explicit learning was dominant, learning was faster to select the better option in their first choices than in the last choices. When implicit reinforcement learning was dominant, learning was faster to select the better option in their last choices than in their first choices. Consistent with the probability-learning and sequence-learning literature, the results show that credit assignment involves two processes: an explicit memory encoding process that requires memory rehearsal and an implicit reinforcement-learning process that propagates credits backwards to previous choices.