Statistical learning leads to persistent memory: Evidence for one-year consolidation

Consolidation Enhances Sequential Multistep Anticipation but Diminishes Access to Perceptual Features

Many experiences unfold predictably over time. Memory for these temporal regularities enables anticipation of events multiple steps into the future. Because temporally predictable events repeat over days, weeks, and years, we must maintain-and potentially transform-memories of temporal structure to support adaptive behavior. We explored how individuals build durable models of temporal regularities to guide multistep anticipation. Healthy young adults (Experiment 1: N = 99, age range = 18-40 years; Experiment 2: N = 204, age range = 19-40 years) learned sequences of scene images that were predictable at the category level and contained incidental perceptual details. Individuals then anticipated upcoming scene categories multiple steps into the future, immediately and at a delay. Consolidation increased the efficiency of anticipation, particularly for events further in the future, but diminished access to perceptual features. Further, maintaining a link-based model of the sequence after consolidation improved anticipation accuracy. Consolidation may therefore promote efficient and durable models of temporal structure, thus facilitating anticipation of future events.

Identifying Transfer Learning in the Reshaping of Inductive Biases

Transfer learning, the reuse of newly acquired knowledge under novel circumstances, is a critical hallmark of human intelligence that has frequently been pitted against the capacities of artificial learning agents. Yet, the computations relevant to transfer learning have been little investigated in humans. The benefit of efficient inductive biases (meta-level constraints that shape learning, often referred as priors in the Bayesian learning approach), has been both theoretically and experimentally established. Efficiency of inductive biases depends on their capacity to generalize earlier experiences. We argue that successful transfer learning upon task acquisition is ensured by updating inductive biases and transfer of knowledge hinges upon capturing the structure of the task in the inductive bias that can be reused in novel tasks. To explore this, we trained participants on a non-trivial visual stimulus sequence task (Alternating Serial Response Times, ASRT); during the Training phase, participants were exposed to one specific sequence for multiple days, then on the Transfer phase, the sequence changed, while the underlying structure of the task remained the same. Our results show that beyond the acquisition of the stimulus sequence, our participants were also able to update their inductive biases. Acquisition of the new sequence was considerably sped up by earlier exposure but this enhancement was specific to individuals showing signatures of abandoning initial inductive biases. Enhancement of learning was reflected in the development of a new internal model. Additionally, our findings highlight the ability of participants to construct an inventory of internal models and alternate between them based on environmental demands. Further, investigation of the behavior during transfer revealed that it is the subjective internal model of individuals that can predict the transfer across tasks. Our results demonstrate that even imperfect learning in a challenging environment helps learning in a new context by reusing the subjective and partial knowledge about environmental regularities.

The Brain's Sensitivity to Real-world Statistical Regularity Does Not Require Full Attention

Predictive coding accounts of perception state that the brain generates perceptual predictions in the service of processing incoming sensory data. These predictions are hypothesized to be afforded by the brain's ability to internalize useful patterns, that is, statistical regularities, from the environment. We have previously argued that the N300 ERP component serves as an index of the brain's use of representations of (real-world) statistical regularities. However, we do not yet know whether overt attention is necessary in order for this process to engage. We addressed this question by presenting stimuli of either high or low real-world statistical regularity in terms of their representativeness (good/bad exemplars of natural scene categories) to participants who either fully attended the stimuli or were distracted by another task (attended/distracted conditions). Replicating past work, N300 responses were larger to bad than to good scene exemplars, and furthermore, we demonstrate minimal impacts of distraction on N300 effects. Thus, it seems that overtly focused attention is not required to maintain the brain's sensitivity to real-world statistical regularity. Furthermore, in an exploratory analysis, we showed that providing additional, artificial regularities, formed by altering the proportions of good and bad exemplars within blocks, further enhanced the N300 effect in both attended and distracted conditions, shedding light on the relationship between statistical regularities learned in the real world and those learned within the context of an experiment.

Finding Pattern in the Noise: Persistent Implicit Statistical Knowledge Impacts the Processing of Unpredictable Stimuli

Humans can extract statistical regularities of the environment to predict upcoming events. Previous research recognized that implicitly acquired statistical knowledge remained persistent and continued to influence behavior even when the regularities were no longer present in the environment. Here, in an fMRI experiment, we investigated how the persistence of statistical knowledge is represented in the brain. Participants (n = 32) completed a visual, four-choice, RT task consisting of statistical regularities. Two types of blocks constantly alternated with one another throughout the task: predictable statistical regularities in one block type and unpredictable ones in the other. Participants were unaware of the statistical regularities and their changing distribution across the blocks. Yet, they acquired the statistical regularities and showed significant statistical knowledge at the behavioral level not only in the predictable blocks but also in the unpredictable ones, albeit to a smaller extent. Brain activity in a range of cortical and subcortical areas, including early visual cortex, the insula, the right inferior frontal gyrus, and the right globus pallidus/putamen contributed to the acquisition of statistical regularities. The right insula, inferior frontal gyrus, and hippocampus as well as the bilateral angular gyrus seemed to play a role in maintaining this statistical knowledge. The results altogether suggest that statistical knowledge could be exploited in a relevant, predictable context as well as transmitted to and retrieved in an irrelevant context without a predictable structure.

Visual artificial grammar learning across 1 year in 7-year-olds and adults

Acquiring sequential information is of utmost importance, for example, for language acquisition in children. Yet, the long-term storage of statistical learning in children is poorly understood. To address this question, 27 7-year-olds and 28 young adults completed four sessions of visual sequence learning (Year 1). From this sample, 16 7-year-olds and 20 young adults participated in another four equivalent sessions after a 12-month-delay (Year 2). The first three sessions of each year used Stimulus Set 1, and the last session used Stimulus Set 2 to investigate transfer effects. Each session consisted of alternating learning and test phases in a modified artificial grammar learning task. In Year 1, 7-year-olds and adults learned the regularities and showed transfer to Stimulus Set 2. Both groups retained their final performance level over the 1-year period. In Year 2, children and adults continued to improve with Stimulus Set 1 but did not show additional transfer gains. Adults overall outperformed children, but transfer effects were indistinguishable between both groups. The current results suggest that long-term memory traces are formed from repeated sequence learning that can be used to generalize sequence rules to new visual input. However, the current study did not provide evidence for a childhood advantage in learning and remembering sequence rules.

Evidence for a competitive relationship between executive functions and statistical learning

The ability of the brain to extract patterns from the environment and predict future events, known as statistical learning, has been proposed to interact in a competitive manner with prefrontal lobe-related networks and their characteristic cognitive or executive functions. However, it remains unclear whether these cognitive functions also possess a competitive relationship with implicit statistical learning across individuals and at the level of latent executive function components. In order to address this currently unknown aspect, we investigated, in two independent experiments (NStudy1 = 186, NStudy2 = 157), the relationship between implicit statistical learning, measured by the Alternating Serial Reaction Time task, and executive functions, measured by multiple neuropsychological tests. In both studies, a modest, but consistent negative correlation between implicit statistical learning and most executive function measures was observed. Factor analysis further revealed that a factor representing verbal fluency and complex working memory seemed to drive these negative correlations. Thus, the antagonistic relationship between implicit statistical learning and executive functions might specifically be mediated by the updating component of executive functions or/and long-term memory access.

The complexity of measuring reliability in learning tasks: An illustration using the Alternating Serial Reaction Time Task

Despite the fact that reliability estimation is crucial for robust inference, it is underutilized in neuroscience and cognitive psychology. Appreciating reliability can help researchers increase statistical power, effect sizes, and reproducibility, decrease the impact of measurement error, and inform methodological choices. However, accurately calculating reliability for many experimental learning tasks is challenging. In this study, we highlight a number of these issues, and estimate multiple metrics of internal consistency and split-half reliability of a widely used learning task on a large sample of 180 subjects. We show how pre-processing choices, task length, and sample size can affect reliability and its estimation. Our results show that the Alternating Serial Reaction Time Task has respectable reliability, especially when learning scores are calculated based on reaction times and two-stage averaging. We also show that a task length of 25 blocks can be sufficient to meet the usual thresholds for minimally acceptable reliability. We further illustrate how relying on a single point estimate of reliability can be misleading, and the calculation of multiple metrics, along with their uncertainties, can lead to a more complete characterization of the psychometric properties of tasks.

Implicit auditory memory in older listeners: From encoding to 6-month retention

Any listening task, from sound recognition to sound-based communication, rests on auditory memory which is known to decline in healthy ageing. However, how this decline maps onto multiple components and stages of auditory memory remains poorly characterised. In an online unsupervised longitudinal study, we tested ageing effects on implicit auditory memory for rapid tone patterns. The test required participants (younger, aged 20-30, and older adults aged 60-70) to quickly respond to rapid regularly repeating patterns emerging from random sequences. Patterns were novel in most trials (REGn), but unbeknownst to the participants, a few distinct patterns reoccurred identically throughout the sessions (REGr). After correcting for processing speed, the response times (RT) to REGn should reflect the information held in echoic and short-term memory before detecting the pattern; long-term memory formation and retention should be reflected by the RT advantage (RTA) to REGr vs REGn which is expected to grow with exposure. Older participants were slower than younger adults in detecting REGn and exhibited a smaller RTA to REGr. Computational simulations using a model of auditory sequence memory indicated that these effects reflect age-related limitations both in early and long-term memory stages. In contrast to ageing-related accelerated forgetting of verbal material, here older adults maintained stable memory traces for REGr patterns up to 6 months after the first exposure. The results demonstrate that ageing is associated with reduced short-term memory and long-term memory formation for tone patterns, but not with forgetting, even over surprisingly long timescales.

Intact predictive processing in autistic adults: evidence from statistical learning

Impairment in predictive processes gained a lot of attention in recent years as an explanation for autistic symptoms. However, empirical evidence does not always underpin this framework. Thus, it is unclear what aspects of predictive processing are affected in autism spectrum disorder. In this study, we tested autistic adults on a task in which participants acquire probability-based regularities (that is, a statistical learning task). Twenty neurotypical and 22 autistic adults learned a probabilistic, temporally distributed regularity for about 40 min. Using frequentist and Bayesian methods, we found that autistic adults performed comparably to neurotypical adults, and the dynamics of learning did not differ between groups either. Thus, our study provides evidence for intact statistical learning in autistic adults. Furthermore, we discuss potential ways this result can extend the scope of the predictive processing framework, noting that atypical processing might not always mean a deficit in performance.

Unconscious knowledge of rewards guides instrumental behaviors via conscious judgments

The demonstration that unconscious learning supports instrumental behaviors (i.e., choosing the stimuli that lead to rewards) is central for the tenet that unconscious cognition sustains human adaptation. Recent studies, using reliable subliminal conditioning paradigms and improved awareness measurements have found evidence against unconscious knowledge sustaining accurate instrumental responses. The present preregistered study proposes a paradigm, in which unconscious processing is stimulated not by subliminally exposing the predictive (conditioned) stimuli, but by employing predictive regularities that are complex and difficult to detect consciously. Participants (N = 211) were exposed to letter strings that, unknown to them, were built from two complex artificial grammars: a "rewarded" or a "non-rewarded" grammar. On each trial, participants memorized a string, and subsequently had to discriminate the memorized string from a distractor. Correct discriminations were rewarded only when the identified string followed the rewarded grammar, but not when it followed the non-rewarded grammar. In a subsequent test phase, participants were presented with new strings from the rewarded and from the unrewarded grammar. Their task was now to directly choose the strings from the rewarded grammar, in order to collect more rewards. A trial-by-trial awareness measure revealed that participants accurately choose novel strings from the rewarded grammar when they had no conscious knowledge of the grammar. The awareness measure also showed that participants were accurate only when the unconsciously learned grammar led to conscious judgments. The present study shows that unconscious knowledge can guide instrumental responses, but only to the extent it supports conscious judgments.

What sticks after statistical learning: The persistence of implicit versus explicit memory traces

Statistical learning is a powerful mechanism that extracts even subtle regularities from our information-dense worlds. Recent theories argue that statistical learning can occur through multiple mechanisms-both the conventionally assumed automatic process that precipitates unconscious learning, and an attention-dependent process that brings regularities into conscious awareness. While this view has gained popularity, there are few empirical dissociations of the hypothesized implicit and explicit forms of statistical learning. Here we provide strong evidence for this dissociation in two ways. First, we show in healthy adults (N = 60) that implicit and explicit traces have divergent consolidation trajectories, with implicit knowledge of structure strengthened over a 24-h period, while precise explicit representations tend to decay. Second, we demonstrate that repeated testing strengthens the retention of explicit representations but that implicit statistical learning is uninfluenced by testing. Together these dissociations provide much needed support for the reconceptualization of statistical learning as a multi-component construct.

Lifespan developmental invariance in memory consolidation: evidence from procedural memory

Characterizing ontogenetic changes across the lifespan is a crucial tool in understanding neurocognitive functions. While age-related changes in learning and memory functions have been extensively characterized in the past decades, the lifespan trajectory of memory consolidation, a critical function that supports the stabilization and long-term retention of memories, is still poorly understood. Here we focus on this fundamental cognitive function and probe the consolidation of procedural memories that underlie cognitive, motor, and social skills and automatic behaviors. We used a lifespan approach: 255 participants aged between 7 and 76 years performed a well-established procedural memory task in the same experimental design across the whole sample. This task enabled us to disentangle two critical processes in the procedural domain: statistical learning and general skill learning. The former is the ability to extract and learn predictable patterns of the environment, while the latter captures a general speed-up as learning progresses due to improved visuomotor coordination and other cognitive processes, independent of acquisition of the predictable patterns. To measure the consolidation of statistical and general skill knowledge, the task was administered in two sessions with a 24-h delay between them. Here, we report successful retention of statistical knowledge with no differences across age groups. For general skill knowledge, offline improvement was observed over the delay period, and the degree of this improvement was also comparable across the age groups. Overall, our findings reveal age invariance in these two key aspects of procedural memory consolidation across the human lifespan.

Deterministic and probabilistic regularities underlying risky choices are acquired in a changing decision context

Predictions supporting risky decisions could become unreliable when outcome probabilities temporarily change, making adaptation more challenging. Therefore, this study investigated whether sensitivity to the temporal structure in outcome probabilities can develop and remain persistent in a changing decision environment. In a variant of the Balloon Analogue Risk Task with 90 balloons, outcomes (rewards or balloon bursts) were predictable in the task's first and final 30 balloons and unpredictable in the middle 30 balloons. The temporal regularity underlying the predictable outcomes differed across three experimental conditions. In the deterministic condition, a repeating three-element sequence dictated the maximum number of pumps before a balloon burst. In the probabilistic condition, a single probabilistic regularity ensured that burst probability increased as a function of pumps. In the hybrid condition, a repeating sequence of three different probabilistic regularities increased burst probabilities. In every condition, the regularity was absent in the middle 30 balloons. Participants were not informed about the presence or absence of the regularity. Sensitivity to both the deterministic and hybrid regularities emerged and influenced risk taking. Unpredictable outcomes of the middle phase did not deteriorate this sensitivity. In conclusion, humans can adapt their risky choices in a changing decision environment by exploiting the statistical structure that controls how the environment changes.

A failure of sleep-dependent consolidation of visuoperceptual procedural learning in young adults with ADHD

ADHD has been associated with cortico-striatal dysfunction that may lead to procedural memory abnormalities. Sleep plays a critical role in consolidating procedural memories, and sleep problems are an integral part of the psychopathology of ADHD. This raises the possibility that altered sleep processes characterizing those with ADHD could contribute to their skill-learning impairments. On this basis, the present study tested the hypothesis that young adults with ADHD have altered sleep-dependent procedural memory consolidation. Participants with ADHD and neurotypicals were trained on a visual discrimination task that has been shown to benefit from sleep. Half of the participants were tested after a 12-h break that included nocturnal sleep (sleep condition), whereas the other half were tested after a 12-h daytime break that did not include sleep (wakefulness condition) to assess the specific contribution of sleep to improvement in task performance. Despite having a similar degree of initial learning, participants with ADHD did not improve in the visual discrimination task following a sleep interval compared to neurotypicals, while they were on par with neurotypicals during the wakefulness condition. These findings represent the first demonstration of a failure in sleep-dependent consolidation of procedural learning in young adults with ADHD. Such a failure is likely to disrupt automatic control routines that are normally provided by the non-declarative memory system, thereby increasing the load on attentional resources of individuals with ADHD.

Do temporal factors affect whether our performance accurately reflects our underlying knowledge? The effects of stimulus presentation rates on the performance versus competence dissociation

Ample evidence shows that the momentary performance can dissociate from the underlying knowledge (competence). Under what circumstances such dissociation occurs, however, remains unclear. Here we tested how temporal factors, and more specifically, the elapsed time between subsequent events affects the dissociation between performance and competence by systematically manipulating the stimulus presentation rates during and after learning. Participants completed a probabilistic sequence learning task with a fast (120 msec) or a slow (850 msec) response-to-stimulus-interval (RSI) during the Learning phase and they were tested with both RSIs 24 h later (Testing phase). We also tested whether they gained explicit knowledge about the sequence or their knowledge remained implicit. Our results revealed higher reaction time learning scores when tested with the fast RSI, irrespective of the RSI during learning, suggesting that faster presentation rates can help better express the acquired knowledge, leading to increased performance measures. For accuracy, participants showed higher learning scores when tested with the same presentation rate as the one that they encountered during learning. The acquired knowledge remained implicit in both groups, suggesting that the observed findings were not confounded by differences in awareness gained in the two groups. Overall, our study highlights that the momentary performance does not always accurately reflect the underlying knowledge, and temporal factors seem to influence this dissociation. Our findings have theoretical, methodological, and translational implications that likely extend beyond learning and memory to other functions and domains as well, including aspects of decision-making, perception, theory of mind, and language.

Tracking human skill learning with a hierarchical Bayesian sequence model

Humans can implicitly learn complex perceptuo-motor skills over the course of large numbers of trials. This likely depends on our becoming better able to take advantage of ever richer and temporally deeper predictive relationships in the environment. Here, we offer a novel characterization of this process, fitting a non-parametric, hierarchical Bayesian sequence model to the reaction times of human participants' responses over ten sessions, each comprising thousands of trials, in a serial reaction time task involving higher-order dependencies. The model, adapted from the domain of language, forgetfully updates trial-by-trial, and seamlessly combines predictive information from shorter and longer windows onto past events, weighing the windows proportionally to their predictive power. As the model implies a posterior over window depths, we were able to determine how, and how many, previous sequence elements influenced individual participants' internal predictions, and how this changed with practice. Already in the first session, the model showed that participants had begun to rely on two previous elements (i.e., trigrams), thereby successfully adapting to the most prominent higher-order structure in the task. The extent to which local statistical fluctuations in trigram frequency influenced participants' responses waned over subsequent sessions, as participants forgot the trigrams less and evidenced skilled performance. By the eighth session, a subset of participants shifted their prior further to consider a context deeper than two previous elements. Finally, participants showed resistance to interference and slow forgetting of the old sequence when it was changed in the final sessions. Model parameters for individual participants covaried appropriately with independent measures of working memory and error characteristics. In sum, the model offers the first principled account of the adaptive complexity and nuanced dynamics of humans' internal sequence representations during long-term implicit skill learning.

Brain-correlates of processing local dependencies within a statistical learning paradigm

Statistical learning refers to the implicit mechanism of extracting regularities in our environment. Numerous studies have investigated the neural basis of statistical learning. However, how the brain responds to violations of auditory regularities based on prior (implicit) learning requires further investigation. Here, we used functional magnetic resonance imaging (fMRI) to investigate the neural correlates of processing events that are irregular based on learned local dependencies. A stream of consecutive sound triplets was presented. Unbeknown to the subjects, triplets were either (a) standard, namely triplets ending with a high probability sound or, (b) statistical deviants, namely triplets ending with a low probability sound. Participants (n = 33) underwent a learning phase outside the scanner followed by an fMRI session. Processing of statistical deviants activated a set of regions encompassing the superior temporal gyrus bilaterally, the right deep frontal operculum including lateral orbitofrontal cortex, and the right premotor cortex. Our results demonstrate that the violation of local dependencies within a statistical learning paradigm does not only engage sensory processes, but is instead reminiscent of the activation pattern during the processing of local syntactic structures in music and language, reflecting the online adaptations required for predictive coding in the context of statistical learning.

The Effects of Cooperative and Competitive Situations on Statistical Learning

Devising cooperative or competitive situations is an important teaching strategy in educational practices. Nevertheless, there is still controversy regarding which situation is better for learning. This study was conducted to explore the effects of cooperative and competitive situations on statistical learning, through the alternating serial reaction time (ASRT) task. Individual cooperative and competitive situations were devised in this study, in which individual situation served as the control condition. Ninety recruited participants were randomly assigned to a cooperative, competitive, or individual group to perform the ASRT task. For general learning, cooperative and competitive situations could indeed make learners respond faster, and there was no significant difference in the RT between the cooperative and competitive groups. Moreover, statistical learning was observed in all three groups. An additional analysis of the early stage of the experiment showed that the learning effect of the competitive group was greater than those of the cooperative and individual groups, in terms of statistical learning. However, the final learning effect was not significantly different among the three groups. Overall, the cooperative and competitive situations had a positive impact on learning and enabled the students to acquire approximately the same learning effect in a shorter time period, compared with the individual situation. Specifically, the competitive situation accelerated the statistical learning process but not the general learning process.

Reduced functional connectivity supports statistical learning of temporally distributed regularities

Statistical learning is a powerful ability that extracts regularities from our environment and makes predictions about future events. Using functional magnetic resonance imaging, we aimed to probe how a wide range of brain areas are intertwined to support statistical learning, characterising its architecture in the whole-brain functional connectivity (FC). Participants performed a statistical learning task of temporally distributed regularities. We used refined behavioural learning scores to associate individuals' learning performances with the FC changed by statistical learning. As a result, the learning performance was mediated by the activation strength in the lateral occipital cortex, angular gyrus, precuneus, anterior cingulate cortex, and superior frontal gyrus. Through a group independent component analysis, activations of the superior frontal network showed the largest correlation with the statistical learning performances. Seed-to-voxel whole-brain and seed-to-ROI FC analyses revealed that the FC between the superior frontal gyrus and the salience, language, and dorsal attention networks were reduced during statistical learning. We suggest that the weakened functional connections between the superior frontal gyrus and brain regions involved in top-down control processes serve a pivotal role in statistical learning, supporting better processing of novel information such as the extraction of new patterns from the environment.

Tracking the contribution of inductive bias to individualised internal models

Internal models capture the regularities of the environment and are central to understanding how humans adapt to environmental statistics. In general, the correct internal model is unknown to observers, instead they rely on an approximate model that is continually adapted throughout learning. However, experimenters assume an ideal observer model, which captures stimulus structure but ignores the diverging hypotheses that humans form during learning. We combine non-parametric Bayesian methods and probabilistic programming to infer rich and dynamic individualised internal models from response times. We demonstrate that the approach is capable of characterizing the discrepancy between the internal model maintained by individuals and the ideal observer model and to track the evolution of the contribution of the ideal observer model to the internal model throughout training. In particular, in an implicit visuomotor sequence learning task the identified discrepancy revealed an inductive bias that was consistent across individuals but varied in strength and persistence.

Inhibitory control hinders habit change

Our habits constantly influence the environment, often in negative ways that amplify global environmental and health risks. Hence, change is urgent. To facilitate habit change, inhibiting unwanted behaviors appears to be a natural human reaction. Here, we use a novel experimental design to test how inhibitory control affects two key components of changing (rewiring) habit-like behaviors in healthy humans: the acquisition of new habit-like behavior and the simultaneous unlearning of an old one. We found that, while the new behavior was acquired, the old behavior persisted and coexisted with the new. Critically, inhibition hindered both overcoming the old behavior and establishing the new one. Our findings highlight that suppressing unwanted behaviors is not only ineffective but may even further strengthen them. Meanwhile, actively engaging in a preferred behavior appears indispensable for its successful acquisition. Our design could be used to uncover how new approaches affect the cognitive basis of changing habit-like behaviors.

Surprisingly inflexible: Statistically learned suppression of distractors generalizes across contexts

The present study investigates the flexibility of statistically learned distractor suppression between different contexts. Participants performed the additional singleton task searching for a unique shape, while ignoring a uniquely colored distractor. Crucially, we created two contexts within the experiments, and each context was assigned its own high-probability distractor location, so that the location where the distractor was most likely to appear depended on the context. Experiment 1 signified context through the color of the background. In Experiment 2, we aimed to more strongly differentiate between the contexts using an auditory or visual cue to indicate the upcoming context. In Experiment 3, context determined the appropriate response ensuring that participants engaged the context in order to be able to perform the task. Across all experiments, participants learned to suppress both high-probability locations, even if they were not aware of these spatial regularities. However, these suppression effects occurred independent of context, as the pattern of suppression reflected a de-prioritization of both high-probability locations which did not change with the context. We employed Bayesian analyses to statistically quantify the absence of context-dependent suppression effects. We conclude that statistically learned distractor suppression is robust and generalizes across contexts.

Does an interference task immediately after practice prevent memory consolidation of sequence-specific learning?

[Purpose] Learning of movement procedures (sequence learning) is essential in physical therapy. Studies have shown that sequence-specific learning may be integrated from an early stage. This study examines the effect of an interference task on the retention of sequence-specific learning. [Participants and Methods] Young adults were randomly divided into a control group and an interference task group, and two experiments were performed. In each experiment, the control group practiced task A in both the acquisition phase and the retention phase four to five hours later. The Interference group practiced task A in the acquisition phase followed by task B, which is similar to the interference task, and then performed task A in the retention phase four to five hours later. In Experiment 2, the amount of practice for task A in the practice phase was 25% of that in Experiment 1. [Results] Sequence-specific learning occurred in the early stages of practice. In particular, the performance of Experiment 1 reached the ceiling. The results of the retention test showed no significant interference effect due to similar tasks. [Conclusion] Implicit sequence-specific learning stabilizes performance early and is not affected by interference tasks.

Desirable Difficulties in Language Learning? How Talker Variability Impacts Artificial Grammar Learning

Contending with talker variability has been found to lead to processing costs but also benefits by focusing learners on invariant properties of the signal, indicating that talker variability acts as a desirable difficulty. That is, talker variability may lead to initial costs followed by long-term benefits for retention and generalization. Adult participants learned an artificial grammar affording learning of multiple components in two experiments varying in difficulty. They learned from one, two, or eight talkers and were tested at three time points. The eight-talker condition did not impact learning. The two-talker condition negatively impacted some aspects of learning, but only under more difficult conditions. Generalization of the grammatical dependency was difficult. Thus, we discovered that high and limited talker variability can differentially impact artificial grammar learning. However, talker variability did not act as a desirable difficulty in the current paradigm as the few evidenced costs were not related to long-term benefits.

Cautious or causal? Key implicit sequence learning paradigms should not be overlooked when assessing the role of DLPFC (Commentary on Prutean et al.)

The role of the dorsolateral prefrontal cortex (DLPFC) in implicit sequence/statistical learning has received considerable attention in recent cognitive neuroscience research. Studies have used non-invasive brain stimulation methods to test whether the DLPFC plays a role in the incidental acquisition and expression of implicit sequence learning. In a recent study, Prutean et al. has concluded that stimulating the left or the right DLPFC might not affect the expression of implicit sequence learning measured by the Serial Reaction Time (SRT) task. The authors speculated that the previous results revealing improved implicit sequence learning following DLPFC stimulation might have been found because explicit awareness accumulated with the use of Alternating Serial Reaction Time (ASRT) task. Our response presents solid evidence that the ASRT task measures implicit sequence learning that remains unconscious both at the judgment and structural level. Therefore, contrary to the conclusion of Prutean et al., we argue that the DLPFC could have a crucial effect on implicit sequence learning that may be task-dependent. We suggest that future research should focus on the specific cognitive processes that may be differentially involved in the SRT versus ASRT tasks, and test what the role of the DLPFC is in those specific cognitive processes.

Access to Procedural Memories After One Year: Evidence for Robust Memory Consolidation in Tourette Syndrome

Tourette syndrome is a childhood-onset neurodevelopmental disorder characterized by motor and vocal tics. On the neural level, tics are thought to be related to the disturbances of the cortico-basal ganglia-thalamo-cortical loops, which also play an important role in procedural learning. Several studies have investigated the acquisition of procedural information and the access to established procedural information in TS. Based on these, the notion of procedural hyperfunctioning, i.e., enhanced procedural learning, has been proposed. However, one neglected area is the retention of acquired procedural information, especially following a long-term offline period. Here, we investigated the 5-hour and 1-year consolidation of two aspects of procedural memory, namely serial-order and probability-based information. Nineteen children with TS between the ages of 10 and 15 as well as 19 typically developing gender- and age-matched controls were tested on a visuomotor four-choice reaction time task that enables the simultaneous assessment of the two aspects. They were retested on the same task 5 hours and 1 year later without any practice in the offline periods. Both groups successfully acquired and retained the probability-based information both when tested 5 hours and then 1 year later, with comparable performance between the TS and control groups. Children with TS did not acquire the serial-order information during the learning phase; hence, retention could not be reliably tested. Our study showed evidence for short-term and long-term retention of one aspect of procedural memory, namely probability-based information in TS, whereas learning of serial-order information might be impaired in this disorder.

Lower-level associations in Gilles de la Tourette syndrome: Convergence between hyperbinding of stimulus and response features and procedural hyperfunctioning theories

Gilles de la Tourette syndrome (GTS) can be characterized by enhanced cognitive functions related to creating, modifying and maintaining connections between stimuli and responses (S-R links). Specifically, two areas, procedural sequence learning and, as a novel finding, also event file binding, show converging evidence of hyperfunctioning in GTS. In this review, we describe how these two enhanced functions can be considered as cognitive mechanisms behind habitual behaviour, such as tics in GTS. Moreover, the presence of both procedural sequence learning and event file binding hyperfunctioning in the same disorder can be treated as evidence for their functional connections, even beyond GTS. Importantly though, we argue that hyperfunctioning of event file binding and procedural learning are not interchangeable: they have different time scales, different sensitivities to potential impairment in action sequencing and distinguishable contributions to the cognitive profile of GTS. An integrated theoretical account of hyperbinding and hyperlearning in GTS allows to formulate predictions for the emergence, activation and long-term persistence of tics in GTS.

Statistical learning occurs during practice while high-order rule learning during rest period

Knowing when the brain learns is crucial for both the comprehension of memory formation and consolidation and for developing new training and neurorehabilitation strategies in healthy and patient populations. Recently, a rapid form of offline learning developing during short rest periods has been shown to account for most of procedural learning, leading to the hypothesis that the brain mainly learns during rest between practice periods. Nonetheless, procedural learning has several subcomponents not disentangled in previous studies investigating learning dynamics, such as acquiring the statistical regularities of the task, or else the high-order rules that regulate its organization. Here we analyzed 506 behavioral sessions of implicit visuomotor deterministic and probabilistic sequence learning tasks, allowing the distinction between general skill learning, statistical learning, and high-order rule learning. Our results show that the temporal dynamics of apparently simultaneous learning processes differ. While high-order rule learning is acquired offline, statistical learning is evidenced online. These findings open new avenues on the short-scale temporal dynamics of learning and memory consolidation and reveal a fundamental distinction between statistical and high-order rule learning, the former benefiting from online evidence accumulation and the latter requiring short rest periods for rapid consolidation.

Statistical and sequence learning lead to persistent memory in children after a one-year offline period

Extraction of environmental patterns underlies human learning throughout the lifespan and plays a crucial role not only in cognitive but also perceptual, motor, and social skills. At least two types of regularities contribute to acquiring skills: (1) statistical, probability-based regularities, and (2) serial order-based regularities. Memory performance of probability-based and/or serial order-based regularities over short periods (from minutes to weeks) has been widely investigated across the lifespan. However, long-term (months or year-long) memory performance of such knowledge has received relatively less attention and has not been assessed in children yet. Here, we aimed to test the long-term memory performance of probability-based and serial order-based regularities over a 1-year offline period in neurotypical children between the age of 9 and 15. Participants performed a visuomotor four-choice reaction time task designed to measure the acquisition of probability-based and serial order-based regularities simultaneously. Short-term consolidation effects were controlled by retesting their performance after a 5-h delay. They were then retested on the same task 1 year later without any practice between the sessions. Participants successfully acquired both probability-based and serial order-based regularities and retained both types of knowledge over the 1-year period. The successful retention was independent of age. Our study demonstrates that the representation of probability-based and serial order-based regularities remains stable over a long period of time. These findings offer indirect evidence for the developmental invariance model of skill consolidation.

Implicit anticipation of probabilistic regularities: Larger CNV emerges for unpredictable events

Anticipation of upcoming events plays a crucial role in automatic behaviors. It is, however, still unclear whether the event-related brain potential (ERP) markers of anticipation could track the implicit acquisition of probabilistic regularities that can be considered as building blocks of automatic behaviors. Therefore, in a four-choice reaction time (RT) task performed by young adults (N = 36), the contingent negative variation (CNV) as an ERP marker of anticipation was measured from the onset of a cue stimulus until the presentation of a target stimulus. Due to the probability structure of the task, target stimuli were either predictable or unpredictable, but this was unknown to participants. The cue did not contain predictive information on the upcoming target. Results showed that the CNV amplitude during response preparation was larger before the unpredictable than before the predictable target stimuli. In addition, although RTs increased, the P3 amplitude decreased for the unpredictable as compared with the predictable target stimuli, possibly due to the stronger response preparation that preceded stimulus presentation. These results suggest that enhanced attentional resources are allocated to the implicit anticipation and processing of unpredictable events. This might originate from the formation of internal models on the probabilistic regularities of the stimulus stream, which primarily facilitates the processing of predictable events. Overall, we provide ERP evidence that supports the role of implicit anticipation and predictive processes in the acquisition of probabilistic regularities.

Divided attention does not affect the acquisition and consolidation of transitional probabilities

Statistical learning facilitates the efficient processing and prediction of environmental events and contributes to the acquisition of automatic behaviors. Whereas a minimal level of attention seems to be required for learning to occur, it is still unclear how acquisition and consolidation of statistical knowledge are affected when attention is divided during learning. To test the effect of divided attention on statistical learning and consolidation, ninety-six healthy young adults performed the Alternating Serial Reaction Time task in which they incidentally acquired second-order transitional probabilities. Half of the participants completed the task with a concurrent secondary intentional sequence learning task that was applied to the same stimulus stream. The other half of the participants performed the task without any attention manipulation. Performance was retested after a 12-h post-learning offline period. Half of each group slept during the delay, while the other half had normal daily activity, enabling us to test the effect of delay activity (sleep vs. wake) on the consolidation of statistical knowledge. Divided attention had no effect on statistical learning: The acquisition of second-order transitional probabilities was comparable with and without the secondary task. Consolidation was neither affected by divided attention: Statistical knowledge was similarly retained over the 12-h delay, irrespective of the delay activity. Our findings can contribute to a better understanding of the role of attentional processes in and the robustness of visuomotor statistical learning and consolidation.

Retrieval of a well-established skill is resistant to distraction: Evidence from an implicit probabilistic sequence learning task

The characteristics of acquiring new sequence information under dual-task situations have been extensively studied. A concurrent task has often been found to affect performance. In real life, however, we mostly perform a secondary task when the primary task is already well acquired. The effect of a secondary task on the ability to retrieve well-established sequence representations remains elusive. The present study investigates whether accessing well-acquired probabilistic sequence knowledge is affected by a concurrent task. Participants acquired non-adjacent regularities in an implicit probabilistic sequence learning task. After a 24-hour offline period, participants were tested on the same probabilistic sequence learning task under dual-task or single-task conditions. Here, we show that although the secondary task significantly prolonged the overall reaction times in the primary (sequence learning) task, access to the previously learned probabilistic representations remained intact. Our results highlight the importance of studying the dual-task effect not only in the learning phase but also during memory access to reveal the robustness of the acquired skill.

Speed or Accuracy Instructions During Skill Learning do not Affect the Acquired Knowledge

A crucial question in skill learning research is how instruction affects the performance or the underlying representations. Little is known about the effects of instructions on one critical aspect of skill learning, namely, picking-up statistical regularities. More specifically, the present study tests how prelearning speed or accuracy instructions affect the acquisition of non-adjacent second-order dependencies. We trained 2 groups of participants on an implicit probabilistic sequence learning task: one group focused on being fast and the other on being accurate. As expected, we detected a strong instruction effect: accuracy instruction resulted in a nearly errorless performance, and speed instruction caused short reaction times (RTs). Despite the differences in the average RTs and accuracy scores, we found a similar level of statistical learning performance in the training phase. After the training phase, we tested the 2 groups under the same instruction (focusing on both speed and accuracy), and they showed comparable performance, suggesting a similar level of underlying statistical representations. Our findings support that skill learning can result in robust representations, and they highlight that this form of knowledge may appear with almost errorless performance. Moreover, multiple sessions with different instructions enabled the separation of competence from performance.

Predictability Changes What We Remember in Familiar Temporal Contexts

The human brain constantly anticipates the future based on memories of the past. Encountering a familiar situation reactivates memory of previous encounters, which can trigger a prediction of what comes next to facilitate responsiveness. However, a prediction error can lead to pruning of the offending memory, a process that weakens its representation in the brain and leads to forgetting. Our goal in this study was to evaluate whether memories are spared from such pruning in situations that allow for accurate predictions at the categorical level, despite prediction errors at the item level. Participants viewed a sequence of objects, some of which reappeared multiple times ("cues"), followed always by novel items. Half of the cues were followed by new items from different (unpredictable) categories, while others were followed by new items from a single (predictable) category. Pattern classification of fMRI data was used to identify category-specific predictions after each cue. Pruning was observed only in unpredictable contexts, while encoding of new items was less robust in predictable contexts. These findings demonstrate that how associative memories are updated is influenced by the reliability of abstract-level predictions in familiar contexts.

Tracking the implicit acquisition of nonadjacent transitional probabilities by ERPs

The implicit acquisition of complex probabilistic regularities has been found to be crucial in numerous automatized cognitive abilities, including language processing and associative learning. However, it has not been completely elucidated how the implicit extraction of second-order nonadjacent transitional probabilities is reflected by neurophysiological processes. Therefore, this study investigated the sensitivity of event-related brain potentials (ERPs) to these probabilistic regularities embedded in a sequence of visual stimuli without providing explicit information on the structure of the stimulus stream. Healthy young adults (N = 32) performed a four-choice RT task that included a sequential regularity between nonadjacent trials yielding a complex transitional probability structure. ERPs were measured relative to both stimulus and response onset. RTs indicated the rapid acquisition of the sequential regularity and the transitional probabilities. The acquisition process was also tracked by the stimulus-locked and response-locked P3 component: The P3 peak was larger for the sequence than for the random stimuli, while the late P3 was larger for less probable than for more probable short-range relations among the random stimuli. According to the RT and P3 effects, sensitivity to the sequential regularity is assumed to be supported by the initial sensitivity to the transitional probabilities. These results suggest that stimulus-response contingencies on the probabilistic regularities of the ongoing stimulus context are implicitly mapped and constantly revised. Overall, this study (1) highlights the role of predictive processes during implicit memory formation, and (2) delineates a potential to gain further insight into the dynamics of implicit acquisition processes.

Input Complexity Affects Long-Term Retention of Statistically Learned Regularities in an Artificial Language Learning Task

Statistical learning (SL) involving sensitivity to distributional regularities in the environment has been suggested to be an important factor in many aspects of cognition, including language. However, the degree to which statistically-learned information is retained over time is not well understood. To establish whether or not learners are able to preserve such regularities over time, we examined performance on an artificial second language learning task both immediately after training and also at a follow-up session 2 weeks later. Participants were exposed to an artificial language (Brocanto2), half of them receiving simplified training items in which only 20% of sequences contained complex structures, whereas the other half were exposed to a training set in which 80% of the items were composed of complex sequences. Overall, participants showed signs of learning at the first session and retention at the second, but the degree of learning was affected by the nature of the training they received. Participants exposed to the simplified input outperformed those in the more complex training condition. A GLMM was used to model the relationship between stimulus properties and participants' endorsement strategies across both sessions. The results indicate that participants in the complex training condition relied more on an item's chunk strength than those in the simple training condition. Taken together, this set of findings shows that statistically learned regularities are retained over the course of 2 weeks. The results also demonstrate that training on input featuring simple items leads to improved learning and retention of grammatical regularities.

Is there more room to improve? The lifespan trajectory of procedural learning and its relationship to the between- and within-group differences in average response times

Characterizing the developmental trajectories of cognitive functions such as learning, memory and decision making across the lifespan faces fundamental challenges. Cognitive functions typically encompass several processes that can be differentially affected by age. Methodological issues also arise when comparisons are made across age groups that differ in basic performance measures, such as in average response times (RTs). Here we focus on procedural learning–a fundamental cognitive function that underlies the acquisition of cognitive, social, and motor skills–and demonstrate how disentangling subprocesses of learning and controlling for differences in average RTs can reveal different developmental trajectories across the human lifespan. Two hundred-seventy participants aged between 7 and 85 years performed a probabilistic sequence learning task that enabled us to separately measure two processes of procedural learning, namely general skill learning and statistical learning. Using raw RT measures, in between-group comparisons, we found a U-shaped trajectory with children and older adults exhibiting greater general skill learning compared to adolescents and younger adults. However, when we controlled for differences in average RTs (either by using ratio scores or focusing on a subsample of participants with similar average speed), only children (but not older adults) demonstrated superior general skill learning consistently across analyses. Testing the relationship between average RTs and general skill learning within age groups shed light on further age-related differences, suggesting that general skill learning measures are more affected by average speed in some age groups. Consistent with previous studies of learning probabilistic regularities, statistical learning showed a gradual decline across the lifespan, and learning performance seemed to be independent of average speed, regardless of the age group. Overall, our results suggest that children are superior learners in various aspects of procedural learning, including both general skill and statistical learning. Our study also highlights the importance to test, and control for, the effect of average speed on other RT measures of cognitive functions, which can fundamentally affect the interpretation of group differences in developmental, aging and clinical psychology and neuroscience studies.

Deconstructing Procedural Memory: Different Learning Trajectories and Consolidation of Sequence and Statistical Learning

Procedural learning is a fundamental cognitive function that facilitates efficient processing of and automatic responses to complex environmental stimuli. Here, we examined training-dependent and off-line changes of two sub-processes of procedural learning: namely, sequence learning and statistical learning. Whereas sequence learning requires the acquisition of order-based relationships between the elements of a sequence, statistical learning is based on the acquisition of probabilistic associations between elements. Seventy-eight healthy young adults (58 females and 20 males) completed the modified version of the Alternating Serial Reaction Time task that was designed to measure Sequence and Statistical Learning simultaneously. After training, participants were randomly assigned to one of three conditions: active wakefulness, quiet rest, or daytime sleep. We examined off-line changes in Sequence and Statistical Learning as well as further improvements after extended practice. Performance in Sequence Learning increased during training, while Statistical Learning plateaued relatively rapidly. After the off-line period, both the acquired sequence and statistical knowledge was preserved, irrespective of the vigilance state (awake, quiet rest or sleep). Sequence Learning further improved during extended practice, while Statistical Learning did not. Moreover, within the sleep group, cortical oscillations and sleep spindle parameters showed differential associations with Sequence and Statistical Learning. Our findings can contribute to a deeper understanding of the dynamic changes of multiple parallel learning and consolidation processes that occur during procedural memory formation.

How can the depressed mind extract and remember predictive relationships of the environment? Evidence from implicit probabilistic sequence learning

A growing body of evidence suggests that emotion and cognition are fundamentally intertwined; impairments in explicit, more effortful and attention-dependent cognitive functions have widely been observed in negative mood. Here we aimed to test how negative mood affects implicit cognition that is less susceptible to motivational and attentional factors associated with negative mood. Therefore, we examined implicit learning and retention of predictive relationships in patients with major depressive episode (MDE). Additionally, we directly compared subgroups of patients with major depressive disorder (MDD) vs. bipolar disorder (BD) in order to gain a deeper understanding of how implicit cognition is affected by these conditions. Implicit probabilistic sequence learning was measured by the Alternating Serial Reaction Time Task. The acquired knowledge was retested after a 24-hour delay period. Consistent with the frontostriatal deficits frequently reported in depression, we found weaker learning in patients with MDE, with a more pronounced deficit in patients with MDD compared to BD. After the 24-hour delay, MDE patients (both subgroups) showed forgetting, while the controls retained the previously acquired knowledge. These results cannot be explained by alterations in motivation, attention and reward processing but suggest more profound impairments of implicit learning and retention of predictive relationships among neutral stimuli in depression. To the best of our knowledge, this is the first study investigating retention of implicitly acquired sequential knowledge and reporting deficits in this domain in MDE. Our findings not only contribute to a better understanding of the complex interplay between affect and cognition but can also help improve screening, diagnosis and treatment protocols of depression.