Human Perceptual Learning Research Articles

Perceptual learning in humans: An active, top-down-guided process.

Deep neural network (DNN) models of human-like vision are typically built by feeding blank slate DNN visual images as training data. However, the literature on human perception and perceptual learning suggests that developing DNNs that truly model human vision requires a shift in approach in which perception is not treated as a largely bottom-up process, but as an active, top-down-guided process.

Visual perceptual learning modulates microsaccade rate and directionality

Microsaccades, incessant “fixational eye movements” (< 1°), are an important window into cognitive functions. Yet, its role in visual perceptual learning (VPL)–improvements in visual discrimination due to practice–remains practically unexplored. Here we investigated whether and how microsaccades change in VPL. Human observers performed a Landolt acuity task for 5 consecutive days and were assigned to the Neutral or Attention group. On each trial, two peripheral Landolt squares were presented briefly along a diagonal. Observers reported the gap side of the target stimulus. Training improved acuity and modified the microsaccade rate; with training, the rate decreased during the fixation period but increased during the response cue. Furthermore, microsaccade direction during the response cue was biased toward the target location, and training enhanced and sped up this bias. Finally, the microsaccade rate during a task-free fixation period correlated with observers’ initial acuity threshold, indicating that the fewer the microsaccades during fixation the better the individual visual acuity. All these results, which were similar for both the Neutral and Attention groups and at both trained and untrained locations, suggest that microsaccades could serve as a physiological marker reflecting functional dynamics in human perceptual learning.

Realize low-power artificial photonic synapse based on (Al,Ga)N nanowire/graphene heterojunction for neuromorphic computing

The fast development of brain-inspired neuromorphic computing systems has stimulated urgent requirements for artificial synapses with low-power consumption. In this work, a photonic synaptic device based on (Al,Ga)N nanowire/graphene heterojunction has been proposed and demonstrated successfully. In the artificial synaptic device, the incident light, the nanowire/graphene heterojunction, and the light-generated carriers play the roles of action potential, pre-synaptic/post-synaptic membrane, and neurotransmitter in a biological synapse, respectively. As a key synaptic function, the paired pulse facilitation index of the photonic synapse can reach 202%, which can be modulated by the interval time between two adjacent light pulses. It is found that the graphene defects, the surface band bending, and the Al vacancies on the surface of (Al,Ga)N nanowires can be the key reasons contributing to the synaptic characteristics of artificial photonic devices. Hence, the dynamic “learning–forgetting” performance of the artificial synaptic device can resemble the “learning–forgetting” behavior of the human brain. Furthermore, the hand-written digits are set up to mimic a typical characteristic of human perceptual learning. After only three training epochs, the simulated network can achieve a high recognition rate of over 90% based on the experimental conductance for long-term potentiation and long-term depression. In supervised learning processes, such few training times are beneficial to reduce energy consumption significantly. Therefore, in the area of neuromorphic computing technology and artificial intelligence systems requiring low-power consumption, this work paves a potential way to develop the optoelectronic synapse based on semiconductor nanowires.

Crossmodal interactions in human learning and memory.

Most studies of memory and perceptual learning in humans have employed unisensory settings to simplify the study paradigm. However, in daily life we are often surrounded by complex and cluttered scenes made up of many objects and sources of sensory stimulation. Our experiences are, therefore, highly multisensory both when passively observing the world and when acting and navigating. We argue that human learning and memory systems are evolved to operate under these multisensory and dynamic conditions. The nervous system exploits the rich array of sensory inputs in this process, is sensitive to the relationship between the sensory inputs, and continuously updates sensory representations, and encodes memory traces based on the relationship between the senses. We review some recent findings that demonstrate a range of human learning and memory phenomena in which the interactions between visual and auditory modalities play an important role, and suggest possible neural mechanisms that can underlie some surprising recent findings. We outline open questions as well as directions of future research to unravel human perceptual learning and memory.

CsPbBr3/graphene nanowall artificial optoelectronic synapses for controllable perceptual learning

The rapid development of neuromorphic computing has stimulated extensive research interest in artificial synapses. Optoelectronic artificial synapses using laser beams as stimulus signals have the advantages of broadband, fast response, and low crosstalk. However, the optoelectronic synapses usually exhibit short memory duration due to the low lifetime of the photo-generated carriers. It greatly limits the mimicking of human perceptual learning, which is a common phenomenon in sensory interactions with the environment and practices of specific sensory tasks. Herein, a heterostructure optoelectronic synapse based on graphene nanowalls and CsPbBr3 quantum dots was fabricated. The graphene/CsPbBr3 heterojunction and the natural middle energy band in graphene nanowalls extend the carrier lifetime. Therefore, a long half-life period of photocurrent decay - 35.59 s has been achieved. Moreover, the long-term optoelectronic response can be controlled by the adjustment of numbers, powers, wavelengths, and frequencies of the laser pulses. Next, an artificial neural network consisting of a 28 × 28 synaptic array was established. It can be used to mimic a typical characteristic of human perceptual learning that the ability of sensory systems is enhanced through a learning experience. The learning behavior of image recognition can be tuned based on the photocurrent response control. The accuracy of image recognition keeps above 80% even under a low-frequency learning process. We also verify that less time is required to regain the lost sensory ability that has been previously learned. This approach paves the way toward high-performance intelligent devices with controllable learning of visual perception.

Microsaccades as a long-term oculomotor correlate in visual perceptual learning.

Human perceptual learning, experience-induced gains in sensory discrimination, typically yields long-term performance improvements. Recent research revealed long-lasting transfer at the untrained location enabled by feature-based attention (FBA), reminiscent of its global effect (Hung & Carrasco, Scientific Reports, 11(1), 13914, (2021)). Visual Perceptual Learning (VPL) is typically studied while observers maintain fixation, but the role of fixational eye movements is unknown. Microsaccades – the largest of fixational eye movements – provide a continuous, online, physiological measure from the oculomotor system that reveals dynamic processing, which is unavailable from behavioral measures alone. We investigated whether and how microsaccades change after training in an orientation discrimination task. For human observers trained with or without FBA, microsaccade rates were significantly reduced during the response window in both trained and untrained locations and orientations. Critically, consistent with long-term training benefits, this microsaccade-rate reduction persisted over a year. Furthermore, microsaccades were biased toward the target location prior to stimulus onset and were more suppressed for incorrect than correct trials after observers’ responses. These findings reveal that fixational eye movements and VPL are tightly coupled and that learning-induced microsaccade changes are long lasting. Thus, microsaccades reflect functional dynamics of the oculomotor system during information encoding, maintenance and readout, and may serve as a reliable long-term physiological correlate in VPL.Supplementary InformationThe online version contains supplementary material available at 10.3758/s13423-022-02151-8.

Neural Substrates of Incidental Associations and Mediated Learning: The Role of Cannabinoid Receptors.

The ability to form associations between different stimuli in the environment to guide adaptive behavior is a central element of learning processes, from perceptual learning in humans to Pavlovian conditioning in animals. Like so, classical conditioning paradigms that test direct associations between low salience sensory stimuli and high salience motivational reinforcers are extremely informative. However, a large part of everyday learning cannot be solely explained by direct conditioning mechanisms – this includes to a great extent associations between individual sensory stimuli, carrying low or null immediate motivational value. This type of associative learning is often described as incidental learning and can be captured in animal models through sensory preconditioning procedures. Here we summarize the evolution of research on incidental and mediated learning, overview the brain systems involved and describe evidence for the role of cannabinoid receptors in such higher-order learning tasks. This evidence favors a number of contemporary hypotheses concerning the participation of the endocannabinoid system in psychosis and psychotic experiences and provides a conceptual framework for understanding how the use of cannabinoid drugs can lead to altered perceptive states.

Feature-based attention enables robust, long-lasting location transfer in human perceptual learning

Visual perceptual learning (VPL) is typically specific to the trained location and feature. However, the degree of specificity depends upon particular training protocols. Manipulating covert spatial attention during training facilitates learning transfer to other locations. Here we investigated whether feature-based attention (FBA), which enhances the representation of particular features throughout the visual field, facilitates VPL transfer, and how long such an effect would last. To do so, we implemented a novel task in which observers discriminated a stimulus orientation relative to two reference angles presented simultaneously before each block. We found that training with FBA enabled remarkable location transfer, reminiscent of its global effect across the visual field, but preserved orientation specificity in VPL. Critically, both the perceptual improvement and location transfer persisted after 1 year. Our results reveal robust, long-lasting benefits induced by FBA in VPL, and have translational implications for improving generalization of training protocols in visual rehabilitation.

Human perceptual learning is delayed by the N-methyl-D-aspartate receptor partial agonist D-cycloserine.

Background:The optimisation of learning has long been a focus of scientific research, particularly in relation to improving psychological treatment and recovery of brain function. Previously, partial N-methyl-D-aspartate agonists have been shown to augment reward learning, procedural learning and psychological therapy, but many studies also report no impact of these compounds on the same processes.Aims:Here we investigate whether administration of an N-methyl-D-aspartate partial agonist (D-cycloserine) modulates a previously unexplored process – tactile perceptual learning. Further, we use a longitudinal design to investigate whether N-methyl-D-aspartate-related learning effects vary with time, thereby providing a potentially simple explanation for apparent mixed effects in previous research.Methods:Thirty-four volunteers were randomised to receive one dose of 250 mg D-cycloserine or placebo 2 h before tactile sensitivity training. Tactile perception was measured using psychophysical methods before and after training, and 24/48 h later.Results:The placebo group showed immediate within-day tactile perception gains, but no further improvements between-days. In contrast, tactile perception remained at baseline on day one in the D-cycloserine group (no within-day learning), but showed significant overnight gains on day two. Both groups were equivalent in tactile perception by the final testing – indicating N-methyl-D-aspartate effects changed the timing, but not the overall amount of tactile learning.Conclusions:In sum, we provide first evidence for modulation of perceptual learning by administration of a partial N-methyl-D-aspartate agonist. Resolving how the effects of such compounds become apparent over time will assist the optimisation of testing schedules, and may help resolve discrepancies across the learning and cognition domains.

Latent inhibition in young children: A developmental effect?

Previous research by Kaniel & Lubow in 1986 found that young children (aged 4-5 years) exhibited poorer learning (latent inhibition) to preexposed stimuli than older children (aged 7-10 years). The aim of our research was to develop a computer-based, child-friendly study that would replicate and extend the work of Kaniel & Lubow in a way that ruled out other, attention-based explanations of their effect. One hundred and four children and 32 undergraduate students took part in our experiment. This consisted of a preexposure/study phase in which participants were asked to press computer keys in response to clipart pictures of animals and dinosaurs. Each animal or dinosaur picture was preceded by one of 2 "warning signals" that acted as the preexposed stimuli (to which no response was required). In the test phase that followed, the participants had to either press the spacebar or withhold their response to each preexposed stimulus and two novel stimuli. They learned which response was correct by trial and error using the feedback provided. The accuracy and reaction time (RT) of the responses during the test phase were analyzed and indicated that the youngest children showed significantly lower mean accuracy and longer mean response times to the preexposed stimuli than to stimuli they had not been preexposed to. In contrast, the older children showed no significant differences in their responses to preexposed and novel stimuli. These results are consistent with those found by Kaniel & Lubow and as such provide additional evidence for latent inhibition in young children. We discuss the implications for theories of perceptual learning in humans. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

20 Hz Steady-State Response in Somatosensory Cortex During Induction of Tactile Perceptual Learning Through LTP-Like Sensory Stimulation.

The induction of synaptic plasticity requires the presence of temporally patterned neural activity. Numerous cellular studies in animals and brain slices have demonstrated that long-term potentiation (LTP) enhances synaptic transmission, which can be evoked by high-frequency intermittent stimulation. In humans, plasticity processes underlying perceptual learning can be reliably induced by repetitive, LTP-like sensory stimulation. These protocols lead to improvement of perceptual abilities parallel to widespread remodeling of cortical processing. However, whether maintained rhythmic cortical activation induced by the LTP-like stimulation is also present during human perceptual learning experiments, remains elusive. To address this question, we here applied a 20 Hz intermittent stimulation protocol for 40 min to the index-, middle- and ring-fingers of the right hand, while continuously recording EEG over the hand representation in primary somatosensory cortex in young adult participants. We find that each train of stimulation initiates a transient series of sensory-evoked potentials which accumulate after about 500 ms into a 20 Hz steady-state response persisting over the entire period of the 2-s-train. During the inter-train interval, no consistent evoked activity can be detected. This response behavior is maintained over the whole 40 min of stimulation without any indication of habituation. However, the early stimulation evoked potentials (SEPs) and the event-related desynchronization (ERD) during the steady-state response change over the 40 min of stimulation. In a second experiment, we demonstrate in a separate cohort of participants that the here-applied pneumatic type of stimulation results in improvement of tactile acuity as typically observed for electrically applied 20 Hz intermittent stimulation. Our data demonstrate that repetitive stimulation using a 20 Hz protocol drives rhythmic activation in the hand representation of somatosensory cortex, which is sustained during the entire stimulation period. At the same time, cortical excitability increases as indicated by altered ERD and SEP amplitudes. Our results, together with previous data underlining the dependence of repetitive sensory stimulation effects on NMDA-receptor activation, support the view that repetitive sensory stimulation elicits LTP-like processes in the cortex, thereby facilitating perceptual learning processes.

Компоненты вызванного потенциала в исследовании перцептивного научения

Perceptual learning is defined by increased effectiveness of completing perceptual tasks as a result of experience or training. This review presents the analysis of changes in the components of event-related potentials (ERPs) after visual and auditory perceptual learning in humans. The use of the EEG method, which has a high temporal resolution, makes it possible to trace the spatio-temporal dynamics of changes in the functioning of the brain during learning, which remains hidden in behavioral experimental studies. A review of neurophysiological studies indicates that perceptual learning induces changes across all levels of cortical hierarchy, starting with the early sensory components of ERPs (C1) and ending with the later integrative components (N170, MMN, P2). We also analyzed the short-term and long-term effects of learning. The reviewed neurophysiological data can serve as the basis for the development of new approaches of effective learning, as well as for the objective evaluation of existing methodics by assessing neuronal dynamics at different stages of stimuli processing.

Stimulus comparison: Effects of the pre-exposure schedule and instructions for perceptual learning and attention

This study aimed to test the effects of the preexposure schedule and certain kinds of verbal instructions for pre-exposure in perceptual learning and attention to the distinctive features of the stimuli. For this, in two experiments adult human participants received concurrent or blocked pre-exposure to two stimuli, and were then required to judge them as same or different in a target identification task. During pre-exposure, participants could receive unspecific instructions to look only at the stimuli (Experiments 1 and 2), asking for “same/different” judgments (Experiments 1 and 2), or only for “different” or “same” judgments (Experiment 2), whilst the dwell time on the distinctive and common features of the stimuli was also recorded. Regardless of the instructions, Experiment 1 found a clear benefit of the concurrent schedule relative to the blocked schedule in the ability to differentiate the stimuli, accompanied by a preference to look at the distinctive features of the stimuli. In the concurrent preexposure condition, the instructions asking for same/different judgments boosted stimulus differentiation in general, along with an attentional shift, while a clear effect of instructions was not found for the blocked preexposure conditions. Experiment 2 further found that whilst such effects of the instructions for the concurrent schedule might appear when the participants were asked uniquely whether the stimuli were the same, these were stronger when the instructions asked explicitly whether the stimuli were different. According to these novel findings, the benefit of the concurrent preexposure schedule for human perceptual learning might arise from an attentional shift toward the distinctive elements; the instructions encouraging the search for stimulus differences are not needed for either stimulus differentiation or an attentional shift during concurrent preexposure, even though such instructions might enhance both effects. Several valuable implications of these findings for human and non-human perceptual learning are discussed.

Visual Perceptual Learning and Models.

Visual perceptual learning through practice or training can significantly improve performance on visual tasks. Originally seen as a manifestation of plasticity in the primary visual cortex, perceptual learning is more readily understood as improvements in the function of brain networks that integrate processes, including sensory representations, decision, attention, and reward, and balance plasticity with system stability. This review considers the primary phenomena of perceptual learning, theories of perceptual learning, and perceptual learning's effect on signal and noise in visual processing and decision. Models, especially computational models, play a key role in behavioral and physiological investigations of the mechanisms of perceptual learning and for understanding, predicting, and optimizing human perceptual processes, learning, and performance. Performance improvements resulting from reweighting or readout of sensory inputs to decision provide a strong theoretical framework for interpreting perceptual learning and transfer that may prove useful in optimizing learning in real-world applications.

The stress hormone cortisol blocks perceptual learning in humans

Cortisol, the primary glucocorticoid (GC) in humans, influences neuronal excitability and plasticity by acting on mineralocorticoid and glucocorticoid receptors. Cellular studies demonstrated that elevated GC levels affect neuronal plasticity, for example through a reduction of hippocampal long-term potentiation (LTP). At the behavioural level, after treatment with GCs, numerous studies have reported impaired hippocampal function, such as impaired memory retrieval. In contrast, relatively little is known about the impact of GCs on cortical plasticity and perceptual learning in adult humans. Therefore, in this study, we explored the impact of elevated GC levels on human perceptual learning. To this aim, we used a training-independent learning approach, where lasting changes in human perception can be induced by applying passive repetitive sensory stimulation (rss), the timing of which was determined from cellular LTP studies. In our placebo-controlled double-blind study, we used tactile LTP-like stimulation to induce improvements in tactile acuity (spatial two-point discrimination). Our results show that a single administration of hydrocortisone (30mg) completely blocked rss-induced changes in two-point discrimination. In contrast, the placebo group showed the expected rss-induced increase in two-point discrimination of over 14%. Our data demonstrate that high GC levels inhibit rss-induced perceptual learning. We suggest that the suppression of LTP, as previously reported in cellular studies, may explain the perceptual learning impairments observed here.

The role of instructions in perceptual learning using complex visual stimuli.

Although modeled on procedures used with nonhuman animals, some recent studies of perceptual learning in humans, using complex visual stimuli, differ in that they usually instruct participants to look for differences between the to-be-discriminated stimuli. This could encourage the use of mechanisms not available to animal subjects. To investigate the role of instructions, in 2 experiments, participants were given preexposure to checkerboards that were similar except for the presence of a small distinctive feature on each. For participants instructed to look for differences, performance on a same-different test was enhanced by preexposure in which the critical stimuli were presented on alternate trials-the usual perceptual learning effect. No such effect was found in 2 other preexposure conditions: when participants were told only to look at the stimuli and not explicitly told to look for differences; and when participants were instructed on an alternative task requiring attention to the stimuli. These results indicate a role for a learning process reinforced by success in finding stimulus differences; they challenge previous interpretations of results from studies using complex visual stimuli in the study of perceptual learning. (PsycINFO Database Record

A neural network model for top-down regulation of sensory plasticity

Auditory cortex displays remarkable plasticity, but how such plasticity relates to human perceptual learning remains unclear. Here, we present a neural network model that attempts to link the two. The model consists of an encoding layer, a decoding/decision layer, and an interneuron that receives both top-down regulation and bottom-up stimulation. The interneuron, once activated, implements a neural learning mechanism such as reorganization of the tonotopic map or lateral inhibition. During active training, top-down regulation sensitizes the interneuron and initiates the learning process. Depending on the relative influence of top-down regulation and bottom-up stimulation, learning ranges from partially to completely stimulus specific, consistent with empirical observations. For stimulus-specific learning, we have recently shown that passive exposure to an untrained frequency after, but not before active training induces transfer to that frequency. The model displays similar behaviors regardless of the specific neural learning mechanism (remapping or inhibition) implemented. Overall, the simulation results indicate that auditory cortical plasticity is subject to top-down regulation and that the plasticity, once unlocked, can afford perceptual learning with bottom-up stimulation only.

Instruction-driven processing in human perceptual learning.

We conducted a series of experiments to determine the relative contribution of unsupervised versus controlled mechanisms to the intermixed-blocked effect. In Experiment 1, participants received pre-exposure instructions prompting a search for differences between stimuli, in keeping with past studies, and the intermixed-blocked effect was observed. In the remaining experiments, participants did not receive the aforementioned instructions, but instead were instructed either to simply observe the stimuli (Experiment 2) or in relation to a masking task (Experiment 3). None of the latter experiments produced an intermixed-blocked effect, suggesting that the effect found in Experiment 1 was driven by the instructions to search for differences, consistent with a controlled processing account of the effect. Moreover, we tested a prediction assuming the operation of a search strategy against one assuming the operation of a short-term habituation mechanism and found evidence more consistent with the search strategy hypothesis. We formulate a new account of the intermixed-blocked effect in humans based on an instruction-driven search and discuss how the account could explain many findings in the human literature.

Switching off perceptual learning: Anodal transcranial direct current stimulation (tDCS) at Fp3 eliminates perceptual learning in humans.

Perceptual learning can be acquired as a result of experience with stimuli that would otherwise be difficult to tell apart, and is often explained in terms of the modulation of feature salience by an error signal based on how well that feature can be predicted by the others that make up the stimulus. In this article we show that anodal transcranial Direct Current Stimulation (tDCS) at Fp3 directly influences this modulation process so as to eliminate and possibly reverse perceptual learning. In 2 experiments, anodal stimulation disrupted perceptual learning (indexed by an inversion effect) compared with sham (Experiment 1) or cathodal (Experiment 2) stimulation. Our findings can be interpreted as showing that anodal tDCS severely reduced or even abolished the modulation of salience based on error, greatly increasing generalization between stimuli. This result supports accounts of perceptual learning based on variations in salience as a consequence of pre-exposure, and opens up the possibility of controlling this phenomenon. (PsycINFO Database Record

The effect of additional exposure to the unique features in a perceptual learning task can be attributed to a location bias.

It has been suggested that human perceptual learning could be explained in terms of a better memory encoding of the unique features during intermixed exposure. However, it is possible that a location bias could play a relevant role in explaining previous results of perceptual learning studies using complex visual stimuli. If this were the case, the only relevant feature would be the location, rather than the content, of the unique features. To further explore this possibility, we attempted to replicate the results of Lavis, Kadib, Mitchell, and Hall (2011, Experiment 2), which showed that additional exposure to the unique elements resulted in better discrimination than simple intermixed exposure. We manipulated the location of the unique elements during the additional exposure. In one experiment, they were located in the same position as that when presented together with the common element. In another experiment, the unique elements were located in the center of the screen, regardless of where they were located together with the common element. Our results showed that additional exposure only improved discrimination when the unique elements were presented in the same position as when they were presented together with the common element. The results reported here do not provide support for the explanation of the effects of additional exposure of the unique elements in terms of a better memory encoding and instead suggest an explanation in terms of location bias.