Signal but not noise changes with perceptual learning

An essential role for the latero-medial secondary visual cortex in the acquisition and retention of visual perceptual learning in mice

Perceptual learning refers to any change in discrimination abilities as a result of practice, a fundamental process that improves the organism's response to the external environment. Visual perceptual learning (vPL) is supposed to rely on functional rearrangements in brain circuity occurring at early stages of sensory processing, with a pivotal role for the primary visual cortex (V1). However, top-down inputs from higher-order visual areas (HVAs) have been suggested to play a key part in vPL, conveying information on attention, expectation and the precise nature of the perceptual task. A direct assessment of the possibility to modulate vPL by manipulating top-down activity in awake subjects is still missing. Here, we used a combination of chemogenetics, behavioral analysis and multichannel electrophysiological assessments to show a critical role in vPL acquisition and retention for neuronal activity in the latero-medial secondary visual cortex (LM), the prime source for top-down feedback projections reentering V1.

Recurrent inhibition refines mental templates to optimize perceptual decisions

Translating sensory inputs to perceptual decisions relies on building internal representations of features critical for solving complex tasks. Yet, we still lack a mechanistic account of how the brain forms these mental templates of task-relevant features to optimize decision-making. Here, we provide evidence for recurrent inhibition: an experience-dependent plasticity mechanism that refines mental templates by enhancing γ-aminobutyric acid (GABA)-mediated (GABAergic) inhibition and recurrent processing in superficial visual cortex layers. We combine ultrahigh-field (7 T) functional magnetic resonance imaging at submillimeter resolution with magnetic resonance spectroscopy to investigate the fine-scale functional and neurochemical plasticity mechanisms for optimized perceptual decisions. We demonstrate that GABAergic inhibition increases following training on a visual (i.e., fine orientation) discrimination task, enhancing the discriminability of orientation representations in superficial visual cortex layers that are known to support recurrent processing. Modeling functional and neurochemical plasticity interactions reveals that recurrent inhibitory processing optimizes brain computations for perpetual decisions and adaptive behavior.

Neural-circuit basis of song preference learning in fruit flies

As observed in human language learning and song learning in birds, the fruit fly Drosophila melanogaster changes its auditory behaviors according to prior sound experiences. This phenomenon, known as song preference learning in flies, requires GABAergic input to pC1 neurons in the brain, with these neurons playing a key role in mating behavior. The neural circuit basis of this GABAergic input, however, is not known. Here, we find that GABAergic neurons expressing the sex-determination gene doublesex are necessary for song preference learning. In the brain, only four doublesex-expressing GABAergic neurons exist per hemibrain, identified as pCd-2 neurons. pCd-2 neurons directly, and in many cases mutually, connect with pC1 neurons, suggesting the existence of reciprocal circuits between them. Moreover, GABAergic and dopaminergic inputs to doublesex-expressing GABAergic neurons are necessary for song preference learning. Together, this study provides a neural circuit model that underlies experience-dependent auditory plasticity at a single-cell resolution.

Reduced categorical learning of faces in dyslexia

The perception of phonological categories in dyslexia is less refined than in typically developing (TD) individuals. Traditionally, this characteristic was considered unique to phonology, yet many studies showed non-phonological perceptual difficulties. Importantly, measuring the dynamics of cortical adaptation, associated with category acquisition, revealed a broadly distributed faster decay of cortical adaptation. Taken together, these observations suggest that the acquisition of perceptual categories in dyslexia may be slower across modalities. To test this, we tested adult individuals with developmental dyslexia (IDDs) and TDs on learning of two unknown faces, yielding face-specific categorization. Initial accuracy was similar in the two groups, yet practice-induced increase in accuracy was significantly larger in TDs. Modeling the learning process (using Drift Diffusion Model) revealed that TDs' steeper learning results from a larger increase in their effective face-specific signal. We propose that IDDs' slower item-specific categorical learning of unknown faces indicates that slower categorical learning in dyslexia is a core, domain-general difficulty.

Improved modeling of human vision by incorporating robustness to blur in convolutional neural networks

Whenever a visual scene is cast onto the retina, much of it will appear degraded due to poor resolution in the periphery; moreover, optical defocus can cause blur in central vision. However, the pervasiveness of blurry or degraded input is typically overlooked in the training of convolutional neural networks (CNNs). We hypothesized that the absence of blurry training inputs may cause CNNs to rely excessively on high spatial frequency information for object recognition, thereby causing systematic deviations from biological vision. We evaluated this hypothesis by comparing standard CNNs with CNNs trained on a combination of clear and blurry images. We show that blur-trained CNNs outperform standard CNNs at predicting neural responses to objects across a variety of viewing conditions. Moreover, blur-trained CNNs acquire increased sensitivity to shape information and greater robustness to multiple forms of visual noise, leading to improved correspondence with human perception. Our results provide multi-faceted neurocomputational evidence that blurry visual experiences may be critical for conferring robustness to biological visual systems.

Contribution of internal noise and calculation efficiency to face discrimination deficits in older adults

Classification images (CIs) measured in a face discrimination task differ significantly between older and younger observers. These age differences are consistent with the hypothesis that older adults sample diagnostic face information less efficiently, or have higher levels of internal noise, compared to younger adults. The current experiments assessed the relative contributions of efficiency and internal noise to age differences in face discrimination using the external noise masking and double-pass response consistency paradigms. Experiment 1 measured discrimination thresholds for faces embedded in several levels of static white noise, and the resulting threshold-vs.-noise curves were used to estimate calculation efficiency and equivalent input noise: older observers had lower efficiency and higher equivalent input noise than younger observers. Experiment 2 presented observers with two identical sequences of faces embedded in static white noise to measure the association between response accuracy and response consistency and estimate the internal:external (i/e) noise ratio for each observer. We found that i/e noise ratios did not differ significantly between groups. These results suggest that age differences in face discrimination are due to differences in calculation efficiency and additive internal noise, but not to age differences in multiplicative internal noise.

Improving iconic memory through contrast detection training with HOA-corrected vision

Iconic memory and short-term memory are not only crucial for perception and cognition, but also of great importance to mental health. Here, we first showed that both types of memory could be improved by improving limiting processes in visual processing through perceptual learning. Normal adults were trained in a contrast detection task for ten days, with their higher-order aberrations (HOA) corrected in real-time. We found that the training improved not only their contrast sensitivity function (CSF), but also their iconic memory and baseline information maintenance for short-term memory, and the relationship between memory and CSF improvements could be well-predicted by an observer model. These results suggest that training the limiting component of a cognitive task with visual perceptual learning could improve visual cognition. They may also provide an empirical foundation for new therapies to treat people with poor sensory memory.

Informational feedback accelerates learning in multi-alternative perceptual judgements of orientation

Experience or training can substantially improve perceptual performance through perceptual learning, and the extent and rate of these improvements may be affected by feedback. In this paper, we first developed a neural network model based on the integrated reweighting theory (Dosher et al., 2013) to account for perceptual learning and performance in n-alternative identification tasks and the dependence of learning on different forms of feedback. We then report an experiment comparing the effectiveness of response feedback (RF) versus accuracy feedback (AF) or no feedback (NF) (full versus partial versus no supervision) in learning a challenging eight-alternative visual orientation identification (8AFC) task. Although learning sometimes occurred in the absence of feedback (NF), RF had a clear advantage above AF or NF in this task. Using hybrid supervision learning rules, a new n-alternative identification integrated reweighting theory (I-IRT) explained both the differences in learning curves given different feedback and the dynamic changes in identification confusion data. This study shows that training with more informational feedback (RF) is more effective, though not necessary, in these challenging n-alternative tasks, a result that has implications for developing training paradigms in realistic tasks.

Improved modeling of human vision by incorporating robustness to blur in convolutional neural networks

Whenever a visual scene is cast onto the retina, much of it will appear degraded due to poor resolution in the periphery; moreover, optical defocus can cause blur in central vision. However, the pervasiveness of blurry or degraded input is typically overlooked in the training of convolutional neural networks (CNNs). We hypothesized that the absence of blurry training inputs may cause CNNs to rely excessively on high spatial frequency information for object recognition, thereby causing systematic deviations from biological vision. We evaluated this hypothesis by comparing standard CNNs with CNNs trained on a combination of clear and blurry images. We show that blur-trained CNNs outperform standard CNNs at predicting neural responses to objects across a variety of viewing conditions. Moreover, blur-trained CNNs acquire increased sensitivity to shape information and greater robustness to multiple forms of visual noise, leading to improved correspondence with human perception. Our results provide novel neurocomputational evidence that blurry visual experiences are very important for conferring robustness to biological visual systems.

Learning spatial frequency identification through reweighted decoding

Perceptual learning, the improvement of perceptual judgments with practice, occurs in many visual tasks. There are, however, relatively fewer studies examining perceptual learning in spatial frequency judgments. In addition, perceptual learning has generally been studied in two-alternative tasks, occasionally in n-alternative tasks, and infrequently in identification. Recently, perceptual learning was found in an orientation identification task (eight-alternatives) and was well accounted for by a new identification integrated reweighting theory (I-IRT) (Liu et al., submitted). Here, we examined perceptual learning in a similar eight-alternative spatial frequency absolute identification task in two different training protocols, finding learning in the majority but not all observers. We fit the I-IRT to the spatial frequency learning data and discuss possible model explanations for variations in learning.

Knowledge generalization and the costs of multitasking

Humans are able to rapidly perform novel tasks, but show pervasive performance costs when attempting to do two things at once. Traditionally, empirical and theoretical investigations into the sources of such multitasking interference have largely focused on multitasking in isolation to other cognitive functions, characterizing the conditions that give rise to performance decrements. Here we instead ask whether multitasking costs are linked to the system's capacity for knowledge generalization, as is required to perform novel tasks. We show how interrogation of the neurophysiological circuitry underlying these two facets of cognition yields further insights for both. Specifically, we demonstrate how a system that rapidly generalizes knowledge may induce multitasking costs owing to sharing of task contingencies between contexts in neural representations encoded in frontoparietal and striatal brain regions. We discuss neurophysiological insights suggesting that prolonged learning segregates such representations by refining the brain's model of task-relevant contingencies, thereby reducing information sharing between contexts and improving multitasking performance while reducing flexibility and generalization. These proposed neural mechanisms explain why the brain shows rapid task understanding, multitasking limitations and practice effects. In short, multitasking limits are the price we pay for behavioural flexibility.

Efficient neural codes naturally emerge through gradient descent learning

Human sensory systems are more sensitive to common features in the environment than uncommon features. For example, small deviations from the more frequently encountered horizontal orientations can be more easily detected than small deviations from the less frequent diagonal ones. Here we find that artificial neural networks trained to recognize objects also have patterns of sensitivity that match the statistics of features in images. To interpret these findings, we show mathematically that learning with gradient descent in neural networks preferentially creates representations that are more sensitive to common features, a hallmark of efficient coding. This effect occurs in systems with otherwise unconstrained coding resources, and additionally when learning towards both supervised and unsupervised objectives. This result demonstrates that efficient codes can naturally emerge from gradient-like learning.

Minimum perturbation theory of deep perceptual learning

Perceptual learning (PL) involves long-lasting improvement in perceptual tasks following extensive training and is accompanied by modified neuronal responses in sensory cortical areas in the brain. Understanding the dynamics of PL and the resultant synaptic changes is important for causally connecting PL to the observed neural plasticity. This is theoretically challenging because learning-related changes are distributed across many stages of the sensory hierarchy. In this paper, we modeled the sensory hierarchy as a deep nonlinear neural network and studied PL of fine discrimination, a common and well-studied paradigm of PL. Using tools from statistical physics, we developed a mean-field theory of the network in the limit of a large number of neurons and large number of examples. Our theory suggests that, in this thermodynamic limit, the input-output function of the network can be exactly mapped to that of a deep linear network, allowing us to characterize the space of solutions for the task. Surprisingly, we found that modifying synaptic weights in the first layer of the hierarchy is both sufficient and necessary for PL. To address the degeneracy of the space of solutions, we postulate that PL dynamics are constrained by a normative minimum perturbation (MP) principle, which favors weight matrices with minimal changes relative to their prelearning values. Interestingly, MP plasticity induces changes to weights and neural representations in all layers of the network, except for the readout weight vector. While weight changes in higher layers are not necessary for learning, they help reduce overall perturbation to the network. In addition, such plasticity can be learned simply through slow learning. We further elucidate the properties of MP changes and compare them against experimental findings. Overall, our statistical mechanics theory of PL provides mechanistic and normative understanding of several important empirical findings of PL.

Current directions in visual perceptual learning

The visual expertise of adult humans is jointly determined by evolution, visual development, and visual perceptual learning. Perceptual learning refers to performance improvements in perceptual tasks after practice or training in the task. It occurs in almost all visual tasks, ranging from simple feature detection to complex scene analysis. In this Review, we focus on key behavioral aspects of visual perceptual learning. We begin by describing visual perceptual learning tasks and manipulations that influence the magnitude of learning, and then discuss specificity of learning. Next, we present theories and computational models of learning and specificity. We then review applications of visual perceptual learning in visual rehabilitation. Finally, we summarize the general principles of visual perceptual learning, discuss the tension between plasticity and stability, and conclude with new research directions.

Auricular Transcutaneous Vagus Nerve Stimulation Diminishes Alpha-Band-Related Inhibitory Gating Processes During Conflict Monitoring in Frontal Cortices

BACKGROUND

Pursuing goals is compromised when being confronted with interfering information. In such situations, conflict monitoring is important. Theoretical considerations on the neurobiology of response selection and control suggest that auricular transcutaneous vagus nerve stimulation (atVNS) should modulate conflict monitoring. However, the neurophysiological-functional neuroanatomical underpinnings are still not understood.

METHODS

AtVNS was applied in a randomized crossover study design (n = 45). During atVNS or sham stimulation, conflict monitoring was assessed using a Flanker task. EEG data were recorded and analyzed with focus on theta and alpha band activity. Beamforming was applied to examine functional neuroanatomical correlates of atVNS-induced EEG modulations. Moreover, temporal EEG signal decomposition was applied to examine different coding levels in alpha and theta band activity.

RESULTS

AtVNS compromised conflict monitoring processes when it was applied at the second appointment in the crossover study design. On a neurophysiological level, atVNS exerted specific effects because only alpha-band activity was modulated. Alpha-band activity was lower in middle and superior prefrontal regions during atVNS stimulation and thus lower when there was also a decline in task performance. The same direction of alpha-band modulations was evident in fractions of the alpha-band activity coding stimulus-related processes, stimulus-response translation processes, and motor response-related processes.

CONCLUSIONS

The combination of prior task experience and atVNS compromises conflict monitoring processes. This is likely due to reduction of the alpha-band-associated inhibitory gating process on interfering information in frontal cortices. Future research should pay considerable attention to boundary conditions affecting the direction of atVNS effects.

Dynamic Distortion of Orientation Representation after Learning in the Mouse Primary Visual Cortex

Learning is an essential cognitive mechanism allowing behavioral adaptation through adjustments in neuronal processing. It is associated with changes in the activity of sensory cortical neurons evoked by task-relevant stimuli. However, the exact nature of those modifications and the computational advantages they may confer are still debated. Here, we investigated how learning an orientation discrimination task alters the neuronal representations of the cues orientations in the primary visual cortex (V1) of male and female mice. When comparing the activity evoked by the task stimuli in naive mice and the mice performing the task, we found that the representations of the orientation of the rewarded and nonrewarded cues were more accurate and stable in trained mice. This better cue representation in trained mice was associated with a distortion of the orientation representation space such that stimuli flanking the task-relevant orientations were represented as the task stimuli themselves, suggesting that those stimuli were generalized as the task cues. This distortion was context dependent as it was absent in trained mice passively viewing the task cues and enhanced in the behavioral sessions where mice performed best. Those modifications of the V1 population orientation representation in performing mice were supported by a suppression of the activity of neurons tuned for orientations neighboring the orientations of the task cues. Thus, visual processing in V1 is dynamically adapted to enhance the reliability of the representation of the learned cues and favor generalization in the task-relevant computational space.SIGNIFICANCE STATEMENT Performance improvement in a task often requires facilitating the extraction of the information necessary to its execution. Here, we demonstrate the existence of a suppression mechanism that improves the representation of the orientations of the task stimuli in the V1 of mice performing an orientation discrimination task. We also show that this mechanism distorts the V1 orientation representation space, leading stimuli flanking the task stimuli orientations to be generalized as the task stimuli themselves.

Relative tuning of holistic face processing towards the fovea

Humans quickly detect and gaze at faces in the world, which reflects their importance in cognition and may lead to tuning of face recognition toward the central visual field. Although sometimes reported, foveal selectivity in face processing is debated: brain imaging studies have found evidence for a central field bias specific to faces, but behavioral studies have found little foveal selectivity in face recognition. These conflicting results are difficult to reconcile, but they could arise from stimulus-specific differences. Recent studies, for example, suggest that individual faces vary in the degree to which they require holistic processing. Holistic processing is the perception of faces as a whole rather than as a set of separate features. We hypothesized that the dissociation between behavioral and neuroimaging studies arises because of this stimulus-specific dependence on holistic processing. Specifically, the central bias found in neuroimaging studies may be specific to holistic processing. Here, we tested whether the eccentricity-dependence of face perception is determined by the degree to which faces require holistic processing. We first measured the holistic-ness of individual Mooney faces (two-tone shadow images readily perceived as faces). In a group of independent observers, we then used a gender discrimination task to measured recognition of these Mooney faces as a function of their eccentricity. Face gender was recognized across the visual field, even at substantial eccentricities, replicating prior work. Importantly, however, holistic face gender recognition was relatively tuned-slightly, but reliably stronger in the central visual field. Our results may reconcile the debate on the eccentricity-dependance of face perception and reveal a spatial inhomogeneity specifically in the holistic representations of faces.

The experience of stereoblindness does not improve use of texture for slant perception

Stereopsis is an important depth cue for normal people, but a subset of people suffer from stereoblindness and cannot use binocular disparity as a cue to depth. Does this experience of stereoblindness modulate use of other depth cues? We investigated this question by comparing perception of 3D slant from texture for stereoblind people and stereo-normal people. Subjects performed slant discrimination and slant estimation tasks using both monocular and binocular stimuli. We found that two groups had comparable ability to discriminate slant from texture information and showed similar mappings between texture information and slant perception (biased perception toward frontal surface with texture information indicating low slants). The results suggest that the experience of stereoblindness did not change the use of texture information for slant perception. In addition, we found that stereoblind people benefitted from binocular viewing in the slant estimation task, despite their inability to use binocular disparity information. These findings are generally consistent with the optimal cue combination model of slant perception.

Noise-trained deep neural networks effectively predict human vision and its neural responses to challenging images

Deep neural networks (DNNs) for object classification have been argued to provide the most promising model of the visual system, accompanied by claims that they have attained or even surpassed human-level performance. Here, we evaluated whether DNNs provide a viable model of human vision when tested with challenging noisy images of objects, sometimes presented at the very limits of visibility. We show that popular state-of-the-art DNNs perform in a qualitatively different manner than humans—they are unusually susceptible to spatially uncorrelated white noise and less impaired by spatially correlated noise. We implemented a noise training procedure to determine whether noise-trained DNNs exhibit more robust responses that better match human behavioral and neural performance. We found that noise-trained DNNs provide a better qualitative match to human performance; moreover, they reliably predict human recognition thresholds on an image-by-image basis. Functional neuroimaging revealed that noise-trained DNNs provide a better correspondence to the pattern-specific neural representations found in both early visual areas and high-level object areas. A layer-specific analysis of the DNNs indicated that noise training led to broad-ranging modifications throughout the network, with greater benefits of noise robustness accruing in progressively higher layers. Our findings demonstrate that noise-trained DNNs provide a viable model to account for human behavioral and neural responses to objects in challenging noisy viewing conditions. Further, they suggest that robustness to noise may be acquired through a process of visual learning.

Unlike human observers, deep neural networks fail to recognize objects in severe visual noise. This study develops noise-trained networks and shows that these networks better predict human performance and neural responses in the visual cortex to challenging noisy object images.

The effect of initial performance on motion perception improvements is modulated by training method

Repeated practice of a perceptual task, termed "perceptual learning," can improve visual performance. Previously, the training thresholds were determined in two ways. One is that the stimulus corresponding to a certain level in individually set psychometric functions was selected as the training threshold. The other is that the certain stimulus was selected as the training threshold without consideration of individual differences. However, little is known about how the two training methods modulate perceptual learning. This study aimed to evaluate the effect of initial performance on patterns of motion perceptual learning under two methods-individually set or group averaged-for setting the training threshold. Thirty-six observers were randomly divided into individual and group thresholds. Psychometric functions, with the percentage correct as a function of coherence, were measured using the coherent motion direction identification task. For the individual threshold, each observer was trained at individualized coherence level, targeting 60% correct for each observer's psychometric function. For the group threshold, each observer was trained at one coherence level, targeting 60% correct in the group-averaged psychometric function. The threshold was reduced after training with the method of constant stimulus in both groups, indicating improvements following perceptual learning. Furthermore, observers with a poorer initial performance exhibited greater learning gains independent of the training method. Importantly, the correlation between the initial performance and learning gains was larger in the individual threshold than in the group threshold, suggesting the influence of the initial performance on the learning amount depends on the training method.

Inter-individual variations in internal noise predict the effects of spatial attention

Individuals differ considerably in the degree to which they benefit from attention allocation. Thus far, such individual differences were attributed to post-perceptual factors such as working-memory capacity. This study examined whether a perceptual factor - the level of internal noise - also contributes to this inter-individual variability in attentional effects. To that end, we estimated individual levels of internal noise from behavioral variability in an orientation discrimination task (with tilted gratings) using the double-pass procedure and the perceptual-template model. We also measured the effects of spatial attention in an acuity task: the participants reported the side of a square on which a small aperture appeared. Central arrows were used to engage sustained attention and peripheral cues to engage transient attention. We found reliable correlations between individual levels of internal noise and the effects of both types of attention, albeit of opposite directions: positive correlation with sustained attention and negative correlation with transient attention. These findings demonstrate that internal noise - a fundamental characteristic of visual perception - can predict individual differences in the effects of spatial attention, highlighting the intricate relations between perception and attention.

Perception-Action Integration Is Modulated by the Catecholaminergic System Depending on Learning Experience

BACKGROUND

The process underlying the integration of perception and action is a focal topic in neuroscientific research and cognitive frameworks such as the theory of event coding have been developed to explain the mechanisms of perception-action integration. The neurobiological underpinnings are poorly understood. While it has been suggested that the catecholaminergic system may play a role, there are opposing predictions regarding the effects of catecholamines on perception-action integration.

METHODS

Methylphenidate (MPH) is a compound commonly used to modulate the catecholaminergic system. In a double-blind, randomized crossover study design, we examined the effect of MPH (0.25 mg/kg) on perception-action integration using an established "event file coding" paradigm in a group of n = 45 healthy young adults.

RESULTS

The data reveal that, compared with the placebo, MPH attenuates binding effects based on the established associations between stimuli and responses, provided participants are already familiar with the task. However, without prior task experience, MPH did not modulate performance compared with the placebo.

CONCLUSIONS

Catecholamines and learning experience interactively modulate perception-action integration, especially when perception-action associations have to be reconfigured. The data suggest there is a gain control-based mechanism underlying the interactive effects of learning/task experience and catecholaminergic activity during perception-action integration.

Neural Mechanism of Noise Affecting Face Recognition

Face recognition is one of the most important cognitive functions for humans in social activities. The ability will be negatively affected when the face images deteriorate. However, the neural process of extracting facial information under challenging conditions is still poorly understood. Therefore, it is necessary to further understand the neurophysiological relevance of this effect. We examined patients with multiple subdural electrodes (ECoG) monitored for clinical purposes. During the experimental task, the patients were presented with face and house images with different noise levels and were asked to recognize the faces. We found a striking increase in high gamma band power (HGP; 60-160 Hz) when face images were shown. We localized the face-specific electrodes to the fusiform gyrus (FG) and surrounding cortices. For each subject, the behavioral performance and magnitudes of the HGP for the face-specific sites significantly both fit a sigmoid function and showed similar changes. Additionally, the curve profile of the average HGP magnitude across the face-specific sites was almost equal to the average behavior curve; the former could precisely track the behavioral performance. In general, these results suggest that the HGP in the FG is closely related to the performance of face image recognition.

Differential Circuit Mechanisms of Young and Aged Visual Cortex in the Mammalian Brain

The main goal of this review is to summarize and discuss (1) age-dependent structural reorganization of mammalian visual cortical circuits underlying complex visual behavior functions in primary visual cortex (V1) and multiple extrastriate visual areas, and (2) current evidence supporting the notion of compensatory mechanisms in aged visual circuits as well as the use of rehabilitative therapy for the recovery of neural plasticity in normal and diseased aging visual circuit mechanisms in different species. It is well known that aging significantly modulates both the structural and physiological properties of visual cortical neurons in V1 and other visual cortical areas in various species. Compensatory aged neural mechanisms correlate with the complexity of visual functions; however, they do not always result in major circuit alterations resulting in age-dependent decline in performance of a visual task or neurodegenerative disorders. Computational load and neural processing gradually increase with age, and the complexity of compensatory mechanisms correlates with the intricacy of higher form visual perceptions that are more evident in higher-order visual areas. It is particularly interesting to note that the visual perceptual processing of certain visual behavior functions does not change with age. This review aims to comprehensively discuss the effect of normal aging on neuroanatomical alterations that underlie critical visual functions and more importantly to highlight differences between compensatory mechanisms in aged neural circuits and neural processes related to visual disorders. This type of approach will further enhance our understanding of inter-areal and cortico-cortical connectivity of visual circuits in normal aging and identify major circuit alterations that occur in different visual deficits, thus facilitating the design and evaluation of potential rehabilitation therapies as well as the assessment of the extent of their rejuvenation.

Speech in noise perception improved by training fine auditory discrimination: far and applicable transfer of perceptual learning

A longstanding focus of perceptual learning research is learning specificity, the difficulty for learning to transfer to tasks and situations beyond the training setting. Previous studies have focused on promoting transfer across stimuli, such as from one sound frequency to another. Here we examined whether learning could transfer across tasks, particularly from fine discrimination of sound features to speech perception in noise, one of the most frequently encountered perceptual challenges in real life. Separate groups of normal-hearing listeners were trained on auditory interaural level difference (ILD) discrimination, interaural time difference (ITD) discrimination, and fundamental frequency (F₀) discrimination with non-speech stimuli delivered through headphones. While ITD training led to no improvement, both ILD and F₀ training produced learning as well as transfer to speech-in-noise perception when noise differed from speech in the trained feature. These training benefits did not require similarity of task or stimuli between training and application settings, construing far and wide transfer. Thus, notwithstanding task specificity among basic perceptual skills such as discrimination of different sound features, auditory learning appears readily transferable between these skills and their “upstream” tasks utilizing them, providing an effective approach to improving performance in challenging situations or challenged populations.

Improving motion detection via anodal transcranial direct current stimulation

BACKGROUND

To study motion perception, a stimulus consisting of a field of small, moving dots is often used. Generally, some of the dots coherently move in the same direction (signal) while the rest move randomly (noise). A percept of global coherent motion (CM) results when many different local motion signals are combined. CM computation is a complex process that requires the integrity of the middle-temporal area (MT/V5) and there is evidence that increasing the number of dots presented in the stimulus makes such computation more efficient.

OBJECTIVE

In this study, we explored whether anodal direct current stimulation (tDCS) over MT/V5 would increase individual performance in a CM task at a low signal-to-noise ratio (SNR, i.e. low percentage of coherent dots) and with a target consisting of a large number of moving dots (high dot numerosity, e.g. >250 dots) with respect to low dot numerosity (<60 dots), indicating that tDCS favour the integration of local motion signal into a single global percept (global motion).

METHOD

Participants were asked to perform a CM detection task (two-interval forced-choice, 2IFC) while they received anodal, cathodal, or sham stimulation on three different days.

RESULTS

Our findings showed no effect of cathodal tDCS with respect to the sham condition. Instead, anodal tDCS improves performance, but mostly when dot numerosity is high (>400 dots) to promote efficient global motion processing.

CONCLUSIONS

The present study suggests that tDCS may be used under appropriate stimulus conditions (low SNR and high dot numerosity) to boost the global motion processing efficiency, and may be useful to empower clinical protocols to treat visual deficits.

A Note on Decomposition of Sources of Variability in Perceptual Decision-making

Information processing underlying human perceptual decision-making is inherently noisy and identifying sources of this noise is important to understand processing. Ratcliff, Voskuilen, and McKoon (2018) examined results from five experiments using a double-pass procedure in which stimuli were repeated typically a hundred trials later. Greater than chance agreement between repeated tests provided evidence for trial-to-trial variability from external sources of noise. They applied the diffusion model to estimate the quality of evidence driving the decision process (drift rate) and the variability (standard deviation) in drift rate across trials. This variability can be decomposed into random (internal) and systematic (external) components by comparing the double-pass accuracy and agreement with the model predictions. In this note, we provide an additional analysis of the double-pass experiments using the linear ballistic accumulator (LBA) model. The LBA model does not have within-trial variability and thus it captures all variability in processing with its across-trial variability parameters. The LBA analysis of the double-pass data provides model-based evidence of external variability in a decision process, which is consistent with Ratcliff et al.'s result. This demonstrates that across-trial variability is required to model perceptual decision-making. The LBA model provides measures of systematic and random variability as the diffusion model did. But due to the lack of within-trial variability, the LBA model estimated the random component as a larger proportion of across-trial total variability than did the diffusion model.

Task rule and choice are reflected by layer-specific processing in rodent auditory cortical microcircuits

The primary auditory cortex (A1) is an essential, integrative node that encodes the behavioral relevance of acoustic stimuli, predictions, and auditory-guided decision-making. However, the realization of this integration with respect to the cortical microcircuitry is not well understood. Here, we characterize layer-specific, spatiotemporal synaptic population activity with chronic, laminar current source density analysis in Mongolian gerbils (Meriones unguiculatus) trained in an auditory decision-making Go/NoGo shuttle-box task. We demonstrate that not only sensory but also task- and choice-related information is represented in the mesoscopic neuronal population code of A1. Based on generalized linear-mixed effect models we found a layer-specific and multiplexed representation of the task rule, action selection, and the animal's behavioral options as accumulating evidence in preparation of correct choices. The findings expand our understanding of how individual layers contribute to the integrative circuit in the sensory cortex in order to code task-relevant information and guide sensory-based decision-making.

Increased Equivalent Input Noise in Glaucomatous Central Vision: Is it Due to Undersampling of Retinal Ganglion Cells?

Purpose

Recent evidence shows that macular damage is common even in early stages of glaucoma. Here we investigated whether contrast sensitivity loss in the central vision of glaucoma patients is due to an increase in equivalent input noise (Neq), a decrease in calculation efficiency, or both. We also examined how retinal undersampling resulting from loss of retinal ganglion cells (RGCs) may affect Neq and calculation efficiency.

Methods

This study included 21 glaucoma patients and 23 age-matched normally sighted individuals. Threshold contrast for orientation discrimination was measured with a sinewave grating embedded in varying levels of external noise. Data were fitted to the linear amplifier model (LAM) to factor contrast sensitivity into Neq and calculation efficiency. We also correlated macular RGC counts estimated from structural (spectral-domain optical coherence tomography) and functional (standard automated perimetry Swedish interactive thresholding algorithm 10-2) data with either Neq or efficiency. Furthermore, using analytical and computer simulation approach, the relative effect of retinal undersampling on Neq and efficiency was evaluated by adding the RGC sampling module into the LAM.

Results

Compared with normal controls, glaucoma patients exhibited a significantly larger Neq without significant difference in efficiency. Neq was significantly correlated with Pelli-Robson contrast sensitivity and macular RGC counts. The results from analytical derivation and model simulation further demonstrated that Neq can be expressed as a function of internal noise and retinal sampling.

Conclusions

Our results showed that equivalent input noise is significantly elevated in glaucomatous vision, thereby impairing foveal contrast sensitivity. Our findings further elucidated how undersampling at the retinal level may increase equivalent input noise.

Perceptual Learning at Higher Trained Cutoff Spatial Frequencies Induces Larger Visual Improvements

It is well known that extensive practice of a perceptual task can improve visual performance, termed perceptual learning. The goal of the present study was to evaluate the dependency of visual improvements on the features of training stimuli (i.e., spatial frequency). Twenty-eight observers were divided into training and control groups. Visual acuity (VA) and contrast sensitivity function (CSF) were measured and compared before and after training. All observers in the training group were trained in a monocular grating detection task near their individual cutoff spatial frequencies. The results showed that perceptual learning induced significant visual improvement, which was dependent on the cutoff spatial frequency, with a greater improvement magnitude and transfer of perceptual learning observed for those trained with higher spatial frequencies. However, VA significantly improved following training but was not related to the cutoff spatial frequency. The results may broaden the understanding of the nature of the learning rule and the neural plasticity of different cortical areas.

Are you from North or South India? A hard face-classification task reveals systematic representational differences between humans and machines

We make a rich variety of judgments on faces, but the underlying features are poorly understood. Here we describe a challenging geographical-origin classification problem that elucidates feature representations in both humans and machine algorithms. In Experiment 1, we collected a diverse set of 1,647 faces from India labeled with their fine-grained geographical origin (North vs. South India), characterized the categorization performance of 129 human subjects on these faces, and compared this with the performance of machine vision algorithms. Our main finding is that while many machine algorithms achieved an overall performance comparable to that of humans (64%), their error patterns across faces were qualitatively different despite training. To elucidate the face parts used by humans for classification, we trained linear classifiers on overcomplete sets of features derived from each face part. This revealed mouth shape to be the most discriminative part compared to eyes, nose, or external contour. In Experiment 2, we confirmed that humans relied the most on mouth shape for classification using an additional experiment in which subjects classified faces with occluded parts. In Experiment 3, we compared human performance for briefly viewed faces and for inverted faces. Interestingly, human performance on inverted faces was predicted better by computational models compared to upright faces, suggesting that humans use relatively more generic features on inverted faces. Taken together, our results show that studying hard classification tasks can lead to useful insights into both machine and human vision.

Predicting the Partition of Behavioral Variability in Speed Perception with Naturalistic Stimuli

A core goal of visual neuroscience is to predict human perceptual performance from natural signals. Performance in any natural task can be limited by at least three sources of uncertainty: stimulus variability, internal noise, and suboptimal computations. Determining the relative importance of these factors has been a focus of interest for decades but requires methods for predicting the fundamental limits imposed by stimulus variability on sensory-perceptual precision. Most successes have been limited to simple stimuli and simple tasks. But perception science ultimately aims to understand how vision works with natural stimuli. Successes in this domain have proven elusive. Here, we develop a model of humans based on an image-computable (images in, estimates out) Bayesian ideal observer. Given biological constraints, the ideal optimally uses the statistics relating local intensity patterns in moving images to speed, specifying the fundamental limits imposed by natural stimuli. Next, we propose a theoretical link between two key decision-theoretic quantities that suggests how to experimentally disentangle the impacts of internal noise and deterministic suboptimal computations. In several interlocking discrimination experiments with three male observers, we confirm this link and determine the quantitative impact of each candidate performance-limiting factor. Human performance is near-exclusively limited by natural stimulus variability and internal noise, and humans use near-optimal computations to estimate speed from naturalistic image movies. The findings indicate that the partition of behavioral variability can be predicted from a principled analysis of natural images and scenes. The approach should be extendable to studies of neural variability with natural signals.SIGNIFICANCE STATEMENT Accurate estimation of speed is critical for determining motion in the environment, but humans cannot perform this task without error. Different objects moving at the same speed cast different images on the eyes. This stimulus variability imposes fundamental external limits on the human ability to estimate speed. Predicting these limits has proven difficult. Here, by analyzing natural signals, we predict the quantitative impact of natural stimulus variability on human performance given biological constraints. With integrated experiments, we compare its impact to well-studied performance-limiting factors internal to the visual system. The results suggest that the deterministic computations humans perform are near optimal, and that behavioral responses to natural stimuli can be studied with the rigor and interpretability defining work with simpler stimuli.

Learning Experience Reverses Catecholaminergic Effects on Adaptive Behavior

BACKGROUND

Catecholamines are important for cognitive control and the ability to adapt behavior (e.g., after response errors). A prominent drug that modulates the catecholaminergic system is methylphenidate. On the basis of theoretical consideration, we propose that the effects of methylphenidate on behavioral adaptation depend on prior learning experience.

METHODS

In a double-blind, randomized, placebo-controlled crossover study design, we examined the effect of methylphenidate (0.25 mg/kg) on post error behavioral adaptation processes in a group of n = 43 healthy young adults. Behavioral adaptation processes were examined in a working memory, modulated response selection task. The focus of the analysis was on order effects within the crossover study design to evaluate effects of prior learning/task experience.

RESULTS

The effect of methylphenidate/placebo on post-error behavioral adaptation processes reverses depending on prior task experience. When there was no prior experience with the task, methylphenidate increased post-error slowing and thus intensified behavioral adaptation processes. However, when there was prior task experience, (i.e., when the placebo session was conducted first in the crossover design), methylphenidate even decreased post-error slowing and behavioral adaptation. Effect sizes were large and the power of the observed effects was higher than 95%.

CONCLUSIONS

The data suggest that catecholaminergic effects on cognitive control functions vary as a function of prior learning/task experience. The data establish a close link between learning/task familiarization and catecholaminergic effects for executive functions, which has not yet been studied, to our knowledge, but is of considerable clinical relevance. Theoretical implications are discussed.

Passive perceptual learning modulates motor inhibitory control in superior frontal regions

Response inhibition is of vital importance in the context of controlling inappropriate responses. The role of perceptual processes during inhibitory control has attracted increased interest. Yet, we are far from an understanding of the mechanisms. One candidate mechanism by which perceptual processes may affect response inhibition refers to “gain control” that is closely linked to the signal‐to‐noise ratio of incoming information. A means to modulate the signal‐to‐noise ratio and gain control mechanisms is perceptual learning. In the current study, we examine the impact of perceptual learning (i.e., passive repetitive sensory stimulation) on response inhibition combining EEG signal decomposition with source localization analyses. A tactile GO/NOGO paradigm was conducted to measure action restraint as one subcomponent of response inhibition. We show that passive perceptual learning modulates response inhibition processes. In particular, perceptual learning attenuates the detrimental effect of response automation during inhibitory control. Temporally decomposed EEG data show that stimulus‐related and not response selection processes during conflict monitoring are linked to these effects. The superior and middle frontal gyrus (BA6), as well as the motor cortex (BA4), are associated with the effects of perceptual learning on response inhibition. Reliable neurophysiological effects were not evident on the basis of standard ERPs, which has important methodological implications for perceptual learning research. The results detail how lower level sensory plasticity protocols affect higher‐order cognitive control functions in frontal cortical structures.

Metacognitive Accuracy Improves With the Perceptual Learning of a Low- but Not High-Level Face Property

Experience with visual stimuli can improve their perceptual performance, a phenomenon termed visual perceptual learning (VPL). VPL has been found to improve metacognitive measures, suggesting increased conscious accessibility to the knowledge supporting perceptual decision-making. However, such studies have largely failed to control objective task accuracy, which typically correlates with metacognition. Here, using a staircase method to control this confound, we investigated whether VPL improves the metacognitive accuracy of perceptual decision-making. Across 3 days, subjects were trained to discriminate faces based on their high-level identity or low-level contrast. Holding objective accuracy constant across training days, perceptual thresholds decreased in both tasks, demonstrating VPL in our protocol. However, whilemetacognitive accuracy was not affected by face contrast VPL, it was decreased by face identity VPL. Our findings couldbe parsimoniously explained by a dual-stage signal detection theory-based model involving an initial perceptual decision-making stage and a second confidence judgment stage. Within this model, internal noise reductions for both stages accounts for our face contrast VPL result, while only first stage noise reductions accounts for our face identity VPL result. In summary, we found evidence suggesting that conscious knowledge accessibility was improved by the VPL of face contrast but not face identity.

Catecholaminergic effects on inhibitory control depend on the interplay of prior task experience and working memory demands

BACKGROUND

Catecholamines affect response inhibition, but the effects of methylphenidate on inhibitory control in healthy subjects are heterogenous. Theoretical considerations suggest that working memory demands and learning/familiarization processes are important factors to consider regarding catecholaminergic effects on response inhibition.

AIMS

The purpose of this study was to examine the role of working memory demands and familiarization for methylphenidate effects on response inhibition.

METHODS

Twenty-eight healthy adults received a single dose of methylphenidate (0.5 mg/kg) or placebo in a randomised, double-blind, crossover study design. The subjects were tested using a working memory-modulated response inhibition paradigm that combined a Go/Nogo task with a mental rotation task.

RESULTS

Methylphenidate effects were largest in the most challenging mental rotation condition. The direction of effects depended on the extent of the participants' task experience. When performing the task for the first time, methylphenidate impaired response inhibition performance in the most challenging mental rotation condition, as reflected by an increased false alarm rate. In sharp contrast to this, methylphenidate seemed to improve response execution performance in the most challenging condition when performing the task for the second time as reflected by reaction times on Go trials.

CONCLUSION

Effects of catecholamines on inhibitory control processes depend on the interplay of two factors: (a) working memory demands, and (b) learning or familiarization with a task. It seems that the net effect of increases in gain control and decreases in working memory processes determines the methylphenidate effect on response inhibition. Hence, crossover study designs likely underestimate methylphenidate effects on cognitive functions.

Simulation of visual acuity by personalizable neuro-physiological model of the human eye

We present a model of the whole visual train to estimate an individual's visual acuity based on their eye's physical properties. Our simulation takes into account the optics of the eye, neural transmission and noise, as well as the recognition process. Personalized input data are represented by the ocular wavefront aberration and pupil diameter, both either coming from in vivo measurements of a subject or being produced by optical design software using a schematic eye. This flexibility opens the door to a broad range of potential applications, such as objective visual acuity measurements and intraocular lens design. Our algorithm contains only two adjustable neural parameters: additive noise σ, and discrimination range δρ, with their values being experimentally calibrated by fitting the results of simulations to the outcome of real acuity tests performed on healthy young subjects with normal vision (visual acuity: 0…-0.3 logMAR range). It was established that by using fixed values of σ = 0.10 and δρ = 0.0025 for each person examined, the residual of the acuity simulations averaged over the calibration group reached its minimum at 0.045 logMAR.

Assessing the detailed time course of perceptual sensitivity change in perceptual learning

The learning curve in perceptual learning is typically sampled in blocks of trials, which could result in imprecise and possibly biased estimates, especially when learning is rapid. Recently, Zhao, Lesmes, and Lu (2017, 2019) developed a Bayesian adaptive quick Change Detection (qCD) method to accurately, precisely, and efficiently assess the time course of perceptual sensitivity change. In this study, we implemented and tested the qCD method in assessing the learning curve in a four-alternative forced-choice global motion direction identification task in both simulations and a psychophysical experiment. The stimulus intensity in each trial was determined by the qCD, staircase or random stimulus selection (RSS) methods. Simulations showed that the accuracy (bias) and precision (standard deviation or confidence bounds) of the estimated learning curves from the qCD were much better than those obtained by the staircase and RSS method; this is true for both trial-by-trial and post hoc segment-by-segment qCD analyses. In the psychophysical experiment, the average half widths of the 68.2% credible interval of the estimated thresholds from the trial-by-trial and post hoc segment-by-segment qCD analyses were both quite small. Additionally, the overall estimates from the qCD and staircase methods matched extremely well in this task where the behavioral rate of learning is relatively slow. Our results suggest that the qCD method can precisely and accurately assess the trial-by-trial time course of perceptual learning.

Michigan Neural Distinctiveness (MiND) study protocol: investigating the scope, causes, and consequences of age-related neural dedifferentiation

Background

Aging is often associated with behavioral impairments, but some people age more gracefully than others. Why? One factor that may play a role is individual differences in the distinctiveness of neural representations. Previous research has found that neural activation patterns in visual cortex in response to different visual stimuli are often more similar (i.e., less distinctive) in older vs. young participants, a phenomenon referred to as age-related neural dedifferentiation. Furthermore, older people whose neural representations are less distinctive tend to perform worse on a wide range of behavioral tasks. The Michigan Neural Distinctiveness (MiND) project aims to investigate the scope of neural dedifferentiation (e.g., does it also occur in auditory, motor, and somatosensory cortex?), one potential cause (age-related reductions in the inhibitory neurotransmitter gamma-aminobutyric acid (GABA)), and the behavioral consequences of neural dedifferentiation. This protocol paper describes the study rationale and methods being used in complete detail, but not the results (data collection is currently underway).

Methods

The MiND project consists of two studies: the main study and a drug study. In the main study, we are recruiting 60 young and 100 older adults to perform behavioral tasks that measure sensory and cognitive function. They also participate in functional MRI (fMRI), MR spectroscopy, and diffusion weighted imaging sessions, providing data on neural distinctiveness and GABA concentrations. In the drug study, we are recruiting 25 young and 25 older adults to compare neural distinctiveness, measured with fMRI, after participants take a placebo or a benzodiazepine (lorazepam) that should increase GABA activity.

Discussion

By collecting multimodal imaging measures along with extensive behavioral measures from the same subjects, we are linking individual differences in neurochemistry, neural representation, and behavioral performance, rather than focusing solely on group differences between young and old participants. Our findings have the potential to inform new interventions for age-related declines.

Trial registration

This study was retrospectively registered with the ISRCTN registry on March 4, 2019. The registration number is ISRCTN17266136.

Face perception: A brief journey through recent discoveries and current directions

Faces are a rich source of information about the people around us. Identity, state of mind, emotions, intentions, age, gender, ethnic background, attractiveness and a host of other attributes about an individual can be gleaned from a face. When face perception fails, dramatic psycho-social consequences can follow at the individual level, as in the case of prosopagnosic parents who are unable to recognize their children at school pick-up. At the species level, social interaction patterns are shaped by human face perception abilities. The computational feat of recognizing faces and facial attributes, and the challenges overcome by the human brain to achieve this feat, have fascinated generations of vision researchers. In this paper, we present a brief overview of some of the milestones of discovery as well as outline a selected set of current directions and open questions on this topic.

Nostril-specific and structure-based olfactory learning of chiral discrimination in human adults

Practice makes perfect. In human olfaction, such plasticity is generally assumed to occur at the level of cortical synthetic processing that shares information from both nostrils. Here we present findings that challenge this view. In two experiments, we trained human adults unirhinally for the discrimination between odor enantiomers over a course of about 10 to 11 days. Results showed that training-induced perceptual gain was restricted to the trained nostril yet partially generalized to untrained odor enantiomers in a structure- rather than quality- based manner. In other words, learning enhanced the differentiation of chirality (molecular configuration) as opposed to overall odor quality (odor object) per se. These findings argue that, unlike earlier beliefs, one nostril does not readily know what the other learns. Moreover, the initial analytical processing of the structural features of uninarial olfactory input remains plastic in human adults.

Although we may only become consciously aware of our sense of smell when we encounter something pungent, it can greatly influence our quality of life. Smells are processed by our olfactory system, a collection of receptors and nerve cells in the nose and brain.

Odor molecules activate the olfactory system when they bind to receptors in the nostrils. These molecules can have a wide range of chemical and physical properties. However, some odor molecules are mirror images of each other. These variants are known as enantiomers. Some people can naturally smell the difference between enantiomers; others can be taught how to tell them apart. Studying this training process could help us to understand how the olfactory system adapts to new circumstances.

Feng and Zhou trained volunteers to distinguish between odor enantiomers – but only using one nostril. After training, the volunteers were better able to tell the difference between different enantiomers – even for certain scents they had not been trained to discriminate – when sniffing through the trained nostril. However, they got no better at distinguishing between enantiomers when they sniffed them using the other nostril.

Overall, the results reported by Feng and Zhou confirm that human adults can learn how to process the structural features of odorants. However, in contrast to earlier beliefs, they also suggest that one nostril does not readily know what the other learns. This new understanding of how the olfactory system adapts could ultimately help us to develop therapies to restore a lost sense of smell.

The empirical characteristics of human pattern vision defy theoretically-driven expectations

Contrast is the most fundamental property of images. Consequently, any comprehensive model of biological vision must incorporate this attribute and provide a veritable description of its impact on visual perception. Current theoretical and computational models predict that vision should modify its characteristics at low contrast: for example, it should become broader (more lowpass) to protect from noise, as often demonstrated by individual neurons. We find that the opposite is true for human discrimination of elementary image elements: vision becomes sharper, not broader, as contrast approaches threshold levels. Furthermore, it suffers from increased internal variability at low contrast and it transitions from a surprisingly linear regime at high contrast to a markedly nonlinear processing mode in the low-contrast range. These characteristics are hard-wired in that they happen on a single trial without memory or expectation. Overall, the empirical results urge caution when attempting to interpret human vision from the standpoint of optimality and related theoretical constructs. Direct measurements of this phenomenon indicate that the actual constraints derive from intrinsic architectural features, such as the co-existence of complex-cell-like and simple-cell-like components. Small circuits built around these elements can indeed account for the empirical results, but do not appear to operate in a manner that conforms to optimality even approximately. More generally, our results provide a compelling demonstration of how far we still are from securing an adequate computational account of the most basic operations carried out by human vision.

Image reconstruction by domain-transform manifold learning

Image reconstruction is essential for imaging applications across the physical and life sciences, including optical and radar systems, magnetic resonance imaging, X-ray computed tomography, positron emission tomography, ultrasound imaging and radio astronomy. During image acquisition, the sensor encodes an intermediate representation of an object in the sensor domain, which is subsequently reconstructed into an image by an inversion of the encoding function. Image reconstruction is challenging because analytic knowledge of the exact inverse transform may not exist a priori, especially in the presence of sensor non-idealities and noise. Thus, the standard reconstruction approach involves approximating the inverse function with multiple ad hoc stages in a signal processing chain, the composition of which depends on the details of each acquisition strategy, and often requires expert parameter tuning to optimize reconstruction performance. Here we present a unified framework for image reconstruction-automated transform by manifold approximation (AUTOMAP)-which recasts image reconstruction as a data-driven supervised learning task that allows a mapping between the sensor and the image domain to emerge from an appropriate corpus of training data. We implement AUTOMAP with a deep neural network and exhibit its flexibility in learning reconstruction transforms for various magnetic resonance imaging acquisition strategies, using the same network architecture and hyperparameters. We further demonstrate that manifold learning during training results in sparse representations of domain transforms along low-dimensional data manifolds, and observe superior immunity to noise and a reduction in reconstruction artefacts compared with conventional handcrafted reconstruction methods. In addition to improving the reconstruction performance of existing acquisition methodologies, we anticipate that AUTOMAP and other learned reconstruction approaches will accelerate the development of new acquisition strategies across imaging modalities.

Perceptual learning and recognition confusion reveal the underlying relationships among the six basic emotions

The six basic emotions (disgust, anger, fear, happiness, sadness, and surprise) have long been considered discrete categories that serve as the primary units of the emotion system. Yet recent evidence indicated underlying connections among them. Here we tested the underlying relationships among the six basic emotions using a perceptual learning procedure. This technique has the potential of causally changing participants' emotion detection ability. We found that training on detecting a facial expression improved the performance not only on the trained expression but also on other expressions. Such a transfer effect was consistently demonstrated between disgust and anger detection as well as between fear and surprise detection in two experiments (Experiment 1A, n = 70; Experiment 1B, n = 42). Notably, training on any of the six emotions could improve happiness detection, while sadness detection could only be improved by training on sadness itself, suggesting the uniqueness of happiness and sadness. In an emotion recognition test using a large sample of Chinese participants (n = 1748), the confusion between disgust and anger as well as between fear and surprise was further confirmed. Taken together, our study demonstrates that the "basic" emotions share some common psychological components, which might be the more basic units of the emotion system.

Tracing the Trajectory of Sensory Plasticity across Different Stages of Speech Learning in Adulthood

Although challenging, adults can learn non-native phonetic contrasts with extensive training [1, 2], indicative of perceptual learning beyond an early sensitivity period [3, 4]. Training can alter low-level sensory encoding of newly acquired speech sound patterns [5]; however, the time-course, behavioral relevance, and long-term retention of such sensory plasticity is unclear. Some theories argue that sensory plasticity underlying signal enhancement is immediate and critical to perceptual learning [6, 7]. Others, like the reverse hierarchy theory (RHT), posit a slower time-course for sensory plasticity [8]. RHT proposes that higher-level categorical representations guide immediate, novice learning, while lower-level sensory changes do not emerge until expert stages of learning [9]. We trained 20 English-speaking adults to categorize a non-native phonetic contrast (Mandarin lexical tones) using a criterion-dependent sound-to-category training paradigm. Sensory and perceptual indices were assayed across operationally defined learning phases (novice, experienced, over-trained, and 8-week retention) by measuring the frequency-following response, a neurophonic potential that reflects fidelity of sensory encoding, and the perceptual identification of a tone continuum. Our results demonstrate that while robust changes in sensory encoding and perceptual identification of Mandarin tones emerged with training and were retained, such changes followed different timescales. Sensory changes were evidenced and related to behavioral performance only when participants were over-trained. In contrast, changes in perceptual identification reflecting improvement in categorical percept emerged relatively earlier. Individual differences in perceptual identification, and not sensory encoding, related to faster learning. Our findings support the RHT-sensory plasticity accompanies, rather than drives, expert levels of non-native speech learning.

Towards a whole brain model of Perceptual Learning

A hallmark of modern Perceptual Learning (PL) is the extent to which learning is specific to the trained stimuli. Such specificity to orientation, spatial location and even eye of training has been used as psychophysical evidence of the neural basis of learning. This argument that specificity of PL implies regionalization of brain plasticity implicitly assumes that examination of a singular locus of PL is an appropriate approach to understand learning. However, recent research shows that learning effects once thought to be specific depend on subtleties of the training paradigm and that within even a simple training procedure there are multiple aspects of the task and stimuli that are learned simultaneously. Here, we suggest that learning on any task involves a broad network of brain regions undergoing changes in representations, read-out weights, decision rules, attention and feedback processes as well as oculomotor changes. However, importantly, the distribution of learning across the neural system depends upon the details of the training procedure and the characterstics of the individual being trained. We propose that to advance our understanding of PL, the field must move towards understanding how distributed brain processes jointly contribute to behavioral learning effects.