Time Course Of Learning Research Articles

Differences in the Time Course of Learning for Hard Compared to Easy Training

Learning is known to facilitate performance in a range of perceptual tasks. Behavioral improvement after training is typically shown after practice with highly similar stimuli that are difficult to discriminate (i.e., hard training), or after exposure to dissimilar stimuli that are highly discriminable (i.e., easy training). However, little is known about the processes that mediate learning after training with difficult compared to easy stimuli. Here we investigate the time course of learning when observers were asked to discriminate similar global form patterns after hard vs. easy training. Hard training required observers to discriminate highly similar global forms, while easy training to judge clearly discriminable patterns. Our results demonstrate differences in learning and transfer performance for hard compared to easy training. Hard training resulted in stronger behavioral improvement than easy training. Further, for hard training, performance improved during single sessions, while for easy training performance improved across but not within sessions. These findings suggest that training with difficult stimuli may result in online learning of specific stimulus features that are similar between the training and test stimuli, while training with easy stimuli involves transfer of learning from highly to less discriminable stimuli that may require longer periods of consolidation.

Changes of Hand Switching Costs during Bimanual Sequential Learning

Many tasks in our daily life demand not only the use of different fingers of one hand in a serial fashion, but also to alternate from one hand to the other. Here, we investigated performance in a bimanual serial reaction time task (SRTT) with particular emphasis on learning-related changes in reaction time (RT) for consecutive button presses for homologous index- and middle fingers. The bimanual SRTT consisted of sequential button presses either with the left or right index- and middle-finger to a series of visual letters displayed on a computer screen. Each letter was assigned a specific button press with one of four fingers. Two outcome measures were investigated: (a) global sequence learning as defined by the time needed to complete a 15-letter SRTT sequence and (b) changes in hand switch costs across learning. We found that bimanual SRTT resulted in a global decrease in RT during the time course of learning that persisted for at least two weeks. Furthermore, RT to a button press showed an increase when the previous button press was associated with another hand as opposed to the same hand. This increase in RT was defined as switch costs. Hand switch costs significantly decreased during the time course of learning, and remained stable over a time of approximately two weeks. This study provides evidence for modulations of switch costs during bimanual sequence learning, a finding that might have important implications for theories of bimanual coordination and learning.

Recent progress in perceptual learning research

Perceptual learning is defined as long-term improvement in perceptual or sensory systems resulting from repeated practice or experience. As the number of perceptual learning studies has increased, controversies and questions have arisen regarding divergent aspects of perceptual learning, including: (1) stages in which perceptual learning occurs, (2) effects of training type, (3) changes in neural processing during the time course of learning, (4) effects of feedback as to correctness of a subject's responses, and (5) double training. Here we review each of these aspects and suggest fruitful directions for future perceptual learning research. WIREs Cogn Sci 2012, 3:293-299. doi: 10.1002/wcs.1175 For further resources related to this article, please visit the WIREs website.

The rapid emergence of stimulus specific perceptual learning.

Is stimulus specific perceptual learning the result of extended practice or does it emerge early in the time course of learning? We examined this issue by manipulating the amount of practice given on a face identification task on Day 1, and altering the familiarity of stimuli on Day 2. We found that a small number of trials was sufficient to produce stimulus specific perceptual learning of faces: on Day 2, response accuracy decreased by the same amount for novel stimuli regardless of whether observers practiced 105 or 840 trials on Day 1. Current models of learning assume early procedural improvements followed by late stimulus specific gains. Our results show that stimulus specific and procedural improvements are distributed throughout the time course of learning.

Bidirectional gray matter changes after complex motor skill learning

Long-term motor skill learning has been consistently shown to result in functional as well as structural changes in the adult human brain. However, the effect of short learning periods on brain structure is not well understood. In the present study, subjects performed a sequential pinch force task (SPFT) for 20 min on 5 consecutive days. Changes in brain structure were evaluated with anatomical magnetic resonance imaging (MRI) scans acquired on the first and last day of motor skill learning. Behaviorally, the SPFT resulted in sequence-specific learning with the trained (right) hand. Structural gray matter (GM) alterations in left M1, right ventral premotor cortex (PMC) and right dorsolateral prefrontal cortex (DLPFC) correlated with performance improvements in the SPFT. More specifically we found that subjects with strong sequence-specific performance improvements in the SPFT also had larger increases in GM volume in the respective brain areas. On the other hand, subjects with small behavioral gains either showed no change or even a decrease in GM volume during the time course of learning. Furthermore, cerebellar GM volume before motor skill learning predicted (A) individual learning-related changes in the SPFT and (B) the amount of structural changes in left M1, right ventral PMC and DLPFC. In summary, we provide novel evidence that short-term motor skill learning is associated with learning-related structural brain alterations. Additionally, we showed that practicing a motor skill is not exclusively accompanied by increased GM volume. Instead, bidirectional structural alterations explained the variability of the individual learning success.

A reinforcement learning approach to instrumental contingency degradation in rats

Goal-directed action involves a representation of action consequences. Adapting to changes in action-outcome contingency requires the prefrontal region. Indeed, rats with lesions of the medial prefrontal cortex do not adapt their free operant response when food delivery becomes unrelated to lever-pressing. The present study explores the bases of this deficit through a combined behavioural and computational approach. We show that lesioned rats retain some behavioural flexibility and stop pressing if this action prevents food delivery. We attempt to model this phenomenon in a reinforcement learning framework. The model assumes that distinct action values are learned in an incremental manner in distinct states. The model represents states as n-uplets of events, emphasizing sequences rather than the continuous passage of time. Probabilities of lever-pressing and visits to the food magazine observed in the behavioural experiments are first analyzed as a function of these states, to identify sequences of events that influence action choice. Observed action probabilities appear to be essentially function of the last event that occurred, with reward delivery and waiting significantly facilitating magazine visits and lever-pressing respectively. Behavioural sequences of normal and lesioned rats are then fed into the model, action values are updated at each event transition according to the SARSA algorithm, and predicted action probabilities are derived through a softmax policy. The model captures the time course of learning, as well as the differential adaptation of normal and prefrontal lesioned rats to contingency degradation with the same parameters for both groups. The results suggest that simple temporal difference algorithms with low learning rates can largely account for instrumental learning and performance. Prefrontal lesioned rats appear to mainly differ from control rats in their low rates of visits to the magazine after a lever press, and their inability to initially detect weak contingency changes.

Unique characteristics of motor adaptation during walking in young children

Children show precocious ability in the learning of languages; is this the case with motor learning? We used split-belt walking to probe motor adaptation (a form of motor learning) in children. Data from 27 children (ages 8-36 mo) were compared with those from 10 adults. Children walked with the treadmill belts at the same speed (tied belt), followed by walking with the belts moving at different speeds (split belt) for 8-10 min, followed again by tied-belt walking (postsplit). Initial asymmetries in temporal coordination (i.e., double support time) induced by split-belt walking were slowly reduced, with most children showing an aftereffect (i.e., asymmetry in the opposite direction to the initial) in the early postsplit period, indicative of learning. In contrast, asymmetries in spatial coordination (i.e., center of oscillation) persisted during split-belt walking and no aftereffect was seen. Step length, a measure of both spatial and temporal coordination, showed intermediate effects. The time course of learning in double support and step length was slower in children than in adults. Moreover, there was a significant negative correlation between the size of the initial asymmetry during early split-belt walking (called error) and the aftereffect for step length. Hence, children may have more difficulty learning when the errors are large. The findings further suggest that the mechanisms controlling temporal and spatial adaptation are different and mature at different times.

Subjective and model-estimated reward prediction: Association with the feedback-related negativity (FRN) and reward prediction error in a reinforcement learning task

In this study, we examined whether the feedback-related negativity (FRN) is associated with both subjective and objective (model-estimated) reward prediction errors (RPE) per trial in a reinforcement learning task in healthy adults ( n = 25). The level of RPE was assessed by 1) subjective ratings per trial and by 2) a computational model of reinforcement learning. As results, model-estimated RPE was highly correlated with subjective RPE ( r = .82), and the grand-averaged ERP waves based on the trials with high and low model-estimated RPE showed the significant difference only in the time period of the FRN component ( p < .05). Regardless of the time course of learning, FRN was associated with both subjective and model-estimated RPEs within subject ( r = .47, p < .001; r = .40, p < .05) and between subjects ( r = .33, p < .05; r = .41, p < .005) only in the Learnable condition where the internal reward prediction varied enough with a behavior–reward contingency.

Effects of Material Emotional Valence on the Time Course of Massive Repetition Priming

Learning through repetition is a fundamental form and also an effective method of language learning critical for achieving proficient and automatic language use. Massive repetition priming as a common research paradigm taps into the dynamic processes involved in repetition learning. Research with this paradigm has so far used only emotionally neutral materials and ignored emotional factors, which seems inappropriate given the well-documented impact of emotion on cognitive processing. The present study used massive repetition priming to investigate whether the emotional valence of learning materials affects implicit language learning. Participants read a list of Chinese words and made speeded perceptual judgments about the spatial configuration of the two characters in a word. Each word was repeated 15 times in the whole learning session. There were three types of words, negative, positive, or neutral in their emotional valence, presented in separate blocks. Although similar levels of asymptotic performance were reached for different valence conditions showing comparable total effects of learning, learning of the positive words was found to be associated with fewer plateaus of shorter durations and to reach saturation earlier, compared with neutral and negative words. The results showed for the first time that the emotional valence of learning materials has significant effects on the time course of learning so that positive materials are learned faster and more efficiently, relative to negative and neutral materials. The study indicates the importance to explicitly consider the role of emotional factors in implicit language learning research.

The use of eye movements in the study of multimedia learning

This commentary focuses on the use of the eye-tracking methodology to study cognitive processes during multimedia learning. First, some general remarks are made about how the method is applied to investigate visual information processing, followed by a reflection on the eye movement measures employed in the studies published in this special issue. It is argued that global eye movement measures indexing attentional and encoding processes during the entire learning period should preferably be complemented with more fine-grained analyses that are either time-locked to important events taking place in an animation or that by other means provide information about the time course of learning. As nicely documented in the present set of studies, it is also of importance to complement the eye-tracking data with offline measures indexing the end product of learning. Such a complementary approach is likely to yield important new insights into the process of multimedia learning.

Learning of a Sequential Motor Skill Comprises Explicit and Implicit Components That Consolidate Differently

The ability to perform accurate sequential movements is essential to normal motor function. Learning a sequential motor behavior is comprised of two basic components: explicit identification of the order in which the sequence elements should be performed and implicit acquisition of spatial accuracy for each element. Here we investigated the time course of learning of these components for a first sequence (SEQA) and their susceptibility to interference from learning a second sequence (SEQB). We assessed explicit learning with a discrete index, the number of correct anticipatory movements, and implicit learning with a continuous variable, spatial error, which decreased during learning without subject awareness. Spatial accuracy to individual sequence elements reached asymptotic levels only when the whole sequence order was known. Interference with recall of the order of SEQA persisted even when SEQB was learned 24 h after SEQA. However, there was resistance to interference by SEQB with increased initial training with SEQA. For implicit learning of spatial accuracy, SEQB interfered at 5 min but not 24 h after SEQA. As in the case of sequence order, prolonged initial training with SEQA induced resistance to interference by SEQB. We conclude that explicit sequence learning is more susceptible to anterograde interference and implicit sequence learning is more susceptible to retrograde interference. However, both become resistant to interference with saturation training. We propose that an essential feature of motor skill learning is the process by which discrete explicit task elements are combined with continuous implicit features of movement to form flawless sequential actions.

Different Dynamics of Performance and Brain Activation in the Time Course of Perceptual Learning

Perceptual learning is regarded as a manifestation of experience-dependent plasticity in the sensory systems, yet the underlying neural mechanisms remain unclear. We measured the dynamics of performance on a visual task and brain activation in the human primary visual cortex (V1) across the time course of perceptual learning. Within the first few weeks of training, brain activation in a V1 subregion corresponding to the trained visual field quadrant and task performance both increased. However, while performance levels then saturated and were maintained at a constant level, brain activation in the corresponding areas decreased to the level observed before training. These findings indicate that there are distinct temporal phases in the time course of perceptual learning, related to differential dynamics of BOLD activity in visual cortex.

Neural correlates of short-term perceptual learning in orientation discrimination indexed by event-related potentials

The current work investigated the neural correlates of visual perceptual learning in grating orientation discrimination by recording event-related potentials (ERPs) from human adults. Subjects were trained with a discrimination task of grating orientation in three consecutive training sessions within 2 h. While reaction times (RTs) were shortened gradually across training sessions, the N1 was decreased and the P2 was increased over the parietal and occipital areas. A broadly distributed P3 was increased along with more practices. In addition, the time course of learning reflected in the P2 and P3 amplitudes was in line with the changes of reaction times and exhibited a stable level during later training. The implications of these results to the neural mechanisms subserving perceptual learning were discussed.

Rapid‐rate transcranial magnetic stimulation of animal auditory cortex impairs short‐term but not long‐term memory formation

Bilateral rapid-rate transcranial magnetic stimulation (rTMS) of gerbil auditory cortex with a miniature coil device was used to study short-term and long-term effects on discrimination learning of frequency-modulated tones. We found previously that directional discrimination of frequency modulation (rising vs. falling) relies on auditory cortex processing and that formation of its memory depends on local protein synthesis. Here we show that, during training over 5 days, certain rTMS regimes contingent on training had differential effects on the time course of learning. When rTMS was applied several times per day, i.e. four blocks of 5 min rTMS each followed 5 min later by a 3-min training block and 15-min intervals between these blocks (experiment A), animals reached a high discrimination performance more slowly over 5 days than did controls. When rTMS preceded only the first two of four training blocks (experiment B), or when prolonged rTMS (20 min) preceded only the first block, or when blocks of experiment A had longer intervals (experiments C and D), no significant day-to-day effects were found. However, in experiment A, and to some extent in experiment B, rTMS reduced the within-session discrimination performance. Nevertheless the animals learned, as demonstrated by a higher performance the next day. Thus, our results indicate that rTMS treatments accumulate over a day but not strongly over successive days. We suggest that rTMS of sensory cortex, as used in our study, affects short-term memory but not long-term memory formation.

Learning to discriminate interaural time differences: An exploratory study with amplitude-modulated stimuli

The advent of bilateral cochlear implants (CIs) has increased interest in learning on binaural tasks, and studies in normal-hearing listeners provide important background information. However, few studies have considered learning with discrimination of interaural time difference (ITD). Here, learning with ITD was explored using stimuli that are more relevant to bilateral CIs than used previously. Inexperienced listeners were trained with envelope-based ITD using high-frequency amplitude-modulated tones with or without an interaural carrier frequency difference (IFD), the former to simulate asymmetrical bilateral CI insertions. All were tested with and without IFD before and after training. In most listeners, ITD thresholds improved substantially with training, not necessarily reaching asymptote after 3000 trials. In these, the magnitude and time-course of learning was larger than anticipated from a previous study with low-frequency ITD. Learning generalized across IFD and the effect of IFD on ITD thresholds at post-test was smaller than reported previously. These results have implications for studies of bilateral CIs, such as the need to provide extensive training to avoid over-estimating any apparent ‘impairment’.SumarioEl advenimiento de los implantes cocleares bilaterales (CI) ha aumentado el interés en aprender sobre tareas binauriculares, y los estudios en sujetos oyentes normales han aportado importante información pertinente. Sin embargo, pocos estudios han considerado el estudio del aprendizaje con base en la discriminación de diferencias interauriculares de tiempo (ITD). Aquí, el aprendizaje con ITD fue explorado utilizando estímulos más relevantes para los CI bilaterales que los usados previamente. Se entrenó a oyentes sin experiencia con una ITD con base en envolventes que comprendían tonos de alta frecuencia de amplitud modulada, con o sin diferencia interauricular de la frecuencia portadora (IFD), esta última para simular inserciones bilaterales asimétricas de CI. Todos fueron evaluados con y sin IFD antes y después del entrenamiento, sin alcanzar necesariamente la asíntota después de 3000 intentos. En ellos, la magnitud y el tiempo del aprendizaje fue mayor de lo anticipado en un estudio con una ITD de baja frecuencia. El aprendizaje en general en todas las IFD y el efecto de la IFD sobre los umbrales de ITD después de la prueba fue similar a los reportes previos. Estos resultados tienen implicaciones en los estudios de CI bilaterales, tales como la necesidad de proveer amplio entrenamiento para evitar la sobre-estimación de cualquier ‘impedimento’ aparente.

Who's on first? What's on second? The time course of learning in corticostriatal systems

The prefrontal cortex and basal ganglia are known to be crucial for learning arbitrary sensorimotor associations (e.g. knowing to stop at red traffic lights). However, little is known about the timing of learning-related activity in these brain systems. Conventional wisdom suggests that the prefrontal cortex should drive learning-related changes in the basal ganglia. However, it is possible that the basal ganglia are instead responsible for the development of learning-related activity in prefrontal cortex. Indeed, recent work using methods for recording in the prefrontal cortex and basal ganglia simultaneously suggests that learning-related activity emerges first in the basal ganglia. Here, these studies are reviewed and integrated with the known anatomy of corticostriatal connections. Testable hypotheses regarding corticostriatal interactions during learning are proposed.

The time course of learning during a vowel discrimination task by hearing-impaired and masked normal-hearing listeners

Vowel discrimination was compared between a group of young, well-trained listeners with mild-to-moderate sensorineural hearing impairment (YHI), and a matched group of normal hearing, noise-masked listeners (YNH). Unexpectedly, discrimination of F1 and F2 in the YHI listeners was equal to or better than that observed in YNH listeners in three conditions of similar audibility [Davis et al., J. Acoust. Soc. Am. 109, 2501 (2001)]. However, in the same time interval, the YHI subjects completed an average of 55% more blocks of testing than the YNH group. New analyses were undertaken to examine the time course of learning during the vowel discrimination task, to determine whether performance was affected by number of trials. Learning curves for a set of vowels in the F1 and F2 regions showed no significant differences between the YHI and YNH listeners. Thus while the YHI subjects completed more trials overall, they achieved a level of discrimination similar to that of their normal-hearing peers within the same number of blocks. Implications of discrimination performance in relation to hearing status and listening strategies will be discussed. [Work supported by NIHDCD-02229.]

Parallel cortico-basal ganglia mechanisms for acquisition and execution of visuomotor sequences - a computational approach.

Experimental studies have suggested that many brain areas, including the basal ganglia (BG), contribute to procedural learning. Focusing on the basal ganglia-thalamocortical (BG-TC) system, we propose a computational model to explain how different brain areas work together in procedural learning. The BG-TC system is composed of multiple separate loop circuits. According to our model, two separate BG-TC loops learn a visuomotor sequence concurrently but using different coordinates, one visual, and the other motor. The visual loop includes the dorsolateral prefrontal (DLPF) cortex and the anterior part of the BG, while the motor loop includes the supplementary motor area (SMA) and the posterior BG. The concurrent learning in these loops is based on reinforcement signals carried by dopaminergic (DA) neurons that project divergently to the anterior ("visual") and posterior ("motor") parts of the striatum. It is expected, however, that the visual loop learns a sequence faster than the motor loop due to their different coordinates. The difference in learning speed may lead to inconsistent outputs from the visual and motor loops, and this problem is solved by a mechanism called a "coordinator," which adjusts the contribution of the visual and motor loops to a final motor output. The coordinator is assumed to be in the presupplementary motor area (pre-SMA). We hypothesize that the visual and motor loops, with the help of the coordinator, achieve both the quick acquisition of novel sequences and the robust execution of well-learned sequences. A computational model based on the hypothesis is examined in a series of computer simulations, referring to the results of the 2 x 5 task experiments that have been used on both monkeys and humans. We found that the dual mechanism with the coordinator was superior to the single (visual or motor) mechanism. The model replicated the following essential features of the experimental results: (1) the time course of learning, (2) the effect of opposite hand use, (3) the effect of sequence reversal, and (4) the effects of localized brain inactivations. Our model may account for a common feature of procedural learning: A spatial sequence of discrete actions (subserved by the visual loop) is gradually replaced by a robust motor skill (subserved by the motor loop).

ERP correlates of associative learning

We examined changes of event-related potentials (ERPs) while participants learned stimulus-to-stimulus relations in an S1-S2 task. The design allowed for separating processes of associative learning from nonspecific effects. Participants had to respond to S2 by a left or right key-press dependent on S2 identity (letter W or M). Preparation for S2 could be improved by using the associative information given by S1. The S1 was an arrow pointing to the left or right. In combination with its color, arrow direction was informative about location and identity of S2, but participants were not informed about the relevance of color. Arrows in two of the colors were fully predictive for the S2 whereas the third color gave no valid information. This third stimulus controlled for habituation and procedural learning. Six blocks with 200 trials each and all three S1 colors in random order were presented. Behavioral and ERP differences in each block between “learning” and control trials were used to identify processes of associative learning. Several effects of associative learning were identified indicating the involvement of specific stages of information processing: a continuous increase of P3 amplitude evoked by S1 was accompanied by a decrease of P3 evoked by S2. These changes reflected the modifications of stimulus weights for response selection and the strengthened association between the two stimulus complexes in the time course of learning. The related motor preparation benefited from learning too, expressed in a decrease of CNV amplitude and an increase of LRP amplitude. Finally a decrease of N1 amplitude evoked by S2 indicated the reduced need to allocate spatial attention to the S2 location according to the learned meaning of S1.

ERP correlates of associative learning

We examined changes of event-related potentials (ERPs) while participants learned stimulus-to-stimulus relations in an S1-S2 task. The design allowed for separating processes of associative learning from nonspecific effects. Participants had to respond to S2 by a left or right key-press dependent on S2 identity (letter W or M). Preparation for S2 could be improved by using the associative information given by S1. The S1 was an arrow pointing to the left or right. In combination with its color, arrow direction was informative about location and identity of S2, but participants were not informed about the relevance of color. Arrows in two of the colors were fully predictive for the S2 whereas the third color gave no valid information. This third stimulus controlled for habituation and procedural learning. Six blocks with 200 trials each and all three S1 colors in random order were presented. Behavioral and ERP differences in each block between "learning" and control trials were used to identify processes of associative learning. Several effects of associative learning were identified indicating the involvement of specific stages of information processing: a continuous increase of P3 amplitude evoked by S1 was accompanied by a decrease of P3 evoked by S2. These changes reflected the modifications of stimulus weights for response selection and the strengthened association between the two stimulus complexes in the time course of learning. The related motor preparation benefited from learning too, expressed in a decrease of CNV amplitude and an increase of LRP amplitude. Finally a decrease of N1 amplitude evoked by S2 indicated the reduced need to allocate spatial attention to the S2 location according to the learned meaning of S1.