Emergent neural dynamics and geometry for generalization in a transitive inference task.

Relational cognition-the ability to infer relationships that generalize to novel combinations of objects-is fundamental to human and animal intelligence. Despite this importance, it remains unclear how relational cognition is implemented in the brain due in part to a lack of hypotheses and predictions at the levels of collective neural activity and behavior. Here we discovered, analyzed, and experimentally tested neural networks (NNs) that perform transitive inference (TI), a classic relational task (if A > B and B > C, then A > C). We found NNs that (i) generalized perfectly, despite lacking overt transitive structure prior to training, (ii) generalized when the task required working memory (WM), a capacity thought to be essential to inference in the brain, (iii) emergently expressed behaviors long observed in living subjects, in addition to a novel order-dependent behavior, and (iv) expressed different task solutions yielding alternative behavioral and neural predictions. Further, in a large-scale experiment, we found that human subjects performing WM-based TI showed behavior inconsistent with a class of NNs that characteristically expressed an intuitive task solution. These findings provide neural insights into a classical relational ability, with wider implications for how the brain realizes relational cognition.

Neural Representations of Food-Related Attributes in the Human Orbitofrontal Cortex during Choice Deliberation in Anorexia Nervosa.

Decisions about what to eat recruit the orbitofrontal cortex (OFC) and involve the evaluation of food-related attributes such as taste and health. These attributes are used differently by healthy individuals and patients with disordered eating behavior, but it is unclear whether these attributes are decodable from activity in the OFC in both groups and whether neural representations of these attributes are differentially related to decisions about food. We used fMRI combined with behavioral tasks to investigate the representation of taste and health attributes in the human OFC and the role of these representations in food choices in healthy women and women with anorexia nervosa (AN). We found that subjective ratings of tastiness and healthiness could be decoded from patterns of activity in the OFC in both groups. However, health-related patterns of activity in the OFC were more related to the magnitude of choice preferences among patients with AN than healthy individuals. These findings suggest that maladaptive decision-making in AN is associated with more consideration of health information represented by the OFC during deliberation about what to eat.SIGNIFICANCE STATEMENT An open question about the OFC is whether it supports the evaluation of food-related attributes during deliberation about what to eat. We found that healthiness and tastiness information was decodable from patterns of neural activity in the OFC in both patients with AN and healthy controls. Critically, neural representations of health were more strongly related to choices in patients with AN, suggesting that maladaptive overconsideration of healthiness during deliberation about what to eat is related to activity in the OFC. More broadly, these results show that activity in the human OFC is associated with the evaluation of relevant attributes during value-based decision-making. These findings may also guide future research into the development of treatments for AN.

The Role of the Cerebellum in Learning to Predict Reward: Evidence from Cerebellar Ataxia.

Recent findings in animals have challenged the traditional view of the cerebellum solely as the site of motor control, suggesting that the cerebellum may also be important for learning to predict reward from trial-and-error feedback. Yet, evidence for the role of the cerebellum in reward learning in humans is lacking. Moreover, open questions remain about which specific aspects of reward learning the cerebellum may contribute to. Here we address this gap through an investigation of multiple forms of reward learning in individuals with cerebellum dysfunction, represented by cerebellar ataxia cases. Nineteen participants with cerebellar ataxia and 57 age- and sex-matched healthy controls completed two separate tasks that required learning about reward contingencies from trial-and-error. To probe the selectivity of reward learning processes, the tasks differed in their underlying structure: while one task measured incremental reward learning ability alone, the other allowed participants to use an alternative learning strategy based on episodic memory alongside incremental reward learning. We found that individuals with cerebellar ataxia were profoundly impaired at reward learning from trial-and-error feedback on both tasks, but retained the ability to learn to predict reward based on episodic memory. These findings provide evidence from humans for a specific and necessary role for the cerebellum in incremental learning of reward associations based on reinforcement. More broadly, the findings suggest that alongside its role in motor learning, the cerebellum likely operates in concert with the basal ganglia to support reinforcement learning from reward.

An objective evaluation of the beholder's response to abstract and figurative art based on construal level theory.

Does abstract art evoke a different cognitive state than figurative art? To address this question empirically, we bridged art theory and cognitive research and designed an experiment leveraging construal level theory (CLT). CLT is based on experimental data showing that psychologically distant events (i.e., occurring farther away in space or time) are represented more abstractly than closer events. We measured construal level elicited by abstract vs. representational art and asked subjects to assign abstract/representational paintings by the same artist to a situation that was temporally/spatially near or distant. Across three experiments, we found that abstract paintings were assigned to the distant situation significantly more often than representational paintings, indicating that abstract art was evocative of greater psychological distance. Our data demonstrate that different levels of artistic abstraction evoke different levels of mental abstraction and suggest that CLT provides an empirical approach to the analysis of cognitive states evoked by different levels of artistic abstraction.

Reinforcement learning with associative or discriminative generalization across states and actions: fMRI at 3 T and 7 T.

The model-free algorithms of "reinforcement learning" (RL) have gained clout across disciplines, but so too have model-based alternatives. The present study emphasizes other dimensions of this model space in consideration of associative or discriminative generalization across states and actions. This "generalized reinforcement learning" (GRL) model, a frugal extension of RL, parsimoniously retains the single reward-prediction error (RPE), but the scope of learning goes beyond the experienced state and action. Instead, the generalized RPE is efficiently relayed for bidirectional counterfactual updating of value estimates for other representations. Aided by structural information but as an implicit rather than explicit cognitive map, GRL provided the most precise account of human behavior and individual differences in a reversal-learning task with hierarchical structure that encouraged inverse generalization across both states and actions. Reflecting inference that could be true, false (i.e., overgeneralization), or absent (i.e., undergeneralization), state generalization distinguished those who learned well more so than action generalization. With high-resolution high-field fMRI targeting the dopaminergic midbrain, the GRL model's RPE signals (alongside value and decision signals) were localized within not only the striatum but also the substantia nigra and the ventral tegmental area, including specific effects of generalization that also extend to the hippocampus. Factoring in generalization as a multidimensional process in value-based learning, these findings shed light on complexities that, while challenging classic RL, can still be resolved within the bounds of its core computations.

Restrictive eating across a spectrum from healthy to unhealthy: behavioral and neural mechanisms.

BACKGROUND: Restriction of food intake is a central feature of anorexia nervosa (AN) and other eating disorders, yet also occurs in the absence of psychopathology. The neural mechanisms of restrictive eating in health and disease are unclear. METHODS: This study examined behavioral and neural mechanisms associated with restrictive eating among individuals with and without eating disorders. Dietary restriction was examined in four groups of women (n = 110): healthy controls, dieting healthy controls, patients with subthreshold (non-low weight) AN, and patients with AN. A Food Choice Task was administered during fMRI scanning to examine neural activation associated with food choices, and a laboratory meal was conducted. RESULTS: Behavioral findings distinguished between healthy and ill participants. Healthy individuals, both dieting and non-dieting, chose significantly more high-fat foods than patients with AN or subthreshold AN. Among healthy individuals, choice was primarily influenced by tastiness, whereas, among both patient groups, healthiness played a larger role. Dorsal striatal activation associated with choice was most pronounced among individuals with AN and was significantly associated with selecting fewer high-fat choices in the task and lower caloric intake in the meal the following day. CONCLUSIONS: A continuous spectrum of behavior was suggested by the increasing amount of weight loss across groups. Yet, data from this Food Choice Task with fMRI suggest there is a behavioral distinction between illness and health, and that the neural mechanisms underlying food choice in AN are distinct. These behavioral and neural mechanisms of restrictive eating may be useful targets for treatment development.

Memory for individual items is related to nonreinforced preference change.

It is commonly assumed that memories contribute to value-based decisions. Nevertheless, most theories of value-based decision-making do not account for memory influences on choice. Recently, new interest has emerged in the interactions between these two fundamental processes, mainly using reinforcement-based paradigms. Here, we aimed to study the role memory processes play in preference change following the nonreinforced cue-approach training (CAT) paradigm. In CAT, the mere association of cued items with a speeded motor response influences choices. Previous studies with this paradigm showed that a single training session induces a long-lasting effect of enhanced preferences for high-value trained stimuli, that is maintained for several months. We hypothesized that CAT increases memory of trained items, leading to enhanced accessibility of their positive associative memories and in turn to preference changes. In two preregistered experiments, we found evidence that memory is enhanced for trained items and that better memory is correlated with enhanced preferences at the individual item level, both immediately and 1 mo following CAT. Our findings suggest that memory plays a central role in value-based decision-making following CAT, even in the absence of external reinforcements. These findings contribute to new theories relating memory and value-based decision-making and set the groundwork for the implementation of novel nonreinforced behavioral interventions that lead to long-lasting behavioral change.

Deficient Goal-Directed Control in a Population Characterized by Extreme Goal Pursuit.

Research in computational psychiatry has sought to understand the basis of compulsive behavior by relating it to basic psychological and neural mechanisms: specifically, goal-directed versus habitual control. These psychological categories have been further identified with formal computational algorithms, model-based and model-free learning, which helps to provide quantitative tools to distinguish them. Computational psychiatry may be particularly useful for examining phenomena in individuals with anorexia nervosa (AN), whose self-starvation appears both excessively goal directed and habitual. However, these laboratory-based studies have not aimed to examine complex behavior, as seen outside the laboratory, in contexts that extend beyond monetary rewards. We therefore assessed (1) whether behavior in AN was characterized by enhanced or diminished model-based behavior, (2) the domain specificity of any abnormalities by comparing learning in a food-specific (i.e., illness-relevant) context as well as in a monetary context, and (3) whether impairments were secondary to starvation by comparing learning before and after initial treatment. Across all conditions, individuals with AN, relative to healthy controls, showed an impairment in model-based, but not model-free, learning, suggesting a general and persistent contribution of habitual over goal-directed control, across domains and time points. Thus, eating behavior in individuals with AN that appears very goal-directed may be under more habitual than goal-directed control, and this is not remediated by achieving weight restoration.

Dopamine is associated with prioritization of reward-associated memories in Parkinson's disease.

Patients with Parkinson's disease have reduced reward sensitivity related to dopaminergic neuron loss, which is associated with impairments in reinforcement learning. Increasingly, however, dopamine-dependent reward signals are recognized to play an important role beyond reinforcement learning. In particular, it has been shown that reward signals mediated by dopamine help guide the prioritization of events for long-term memory consolidation. Meanwhile, studies of memory in patients with Parkinson's disease have focused on overall memory capacity rather than what is versus what isn't remembered, leaving open questions about the effect of dopamine replacement on the prioritization of memories by reward and the time-dependence of this effect. The current study sought to fill this gap by testing the effect of reward and dopamine on memory in patients with Parkinson's disease. We tested the effect of dopamine modulation and reward on two forms of long-term memory: episodic memory for neutral objects and memory for stimulus-value associations. We measured both forms of memory in a single task, adapting a standard task of reinforcement learning with incidental episodic encoding events of trial-unique objects. Objects were presented on each trial at the time of feedback, which was either rewarding or not. Memory for the trial-unique images and for the stimulus-value associations, and the influence of reward on both, was tested immediately after learning and 2 days later. We measured performance in Parkinson's disease patients tested either ON or OFF their dopaminergic medications and in healthy older control subjects. We found that dopamine was associated with a selective enhancement of memory for reward-associated images, but that it did not influence overall memory capacity. Contrary to predictions, this effect did not differ between the immediate and delayed memory tests. We also found that while dopamine had an effect on reward-modulated episodic memory, there was no effect of dopamine on memory for stimulus-value associations. Our results suggest that impaired prioritization of cognitive resource allocation may contribute to the early cognitive deficits of Parkinson's disease.

A comparison of food-based decision-making between restricting and binge-eating/purging subtypes of anorexia nervosa.

OBJECTIVE: By definition, restricting (ANR) and binge-eating/purging (ANBP) subtypes of anorexia nervosa (AN) differ in some manifestations of maladaptive eating behavior. This study aimed to determine whether the groups differ in the choices they make about what to eat, and whether there are differences in valuation related to food choice, using an experimental paradigm. METHOD: Inpatients with ANR (n = 40) and ANBP (n = 46) participated in a Food Choice Task. During the task, participants rated 76 food images for healthiness and tastiness, and choice preferences. Groups were compared in percent selection of high-fat and low-fat foods, value ratings of foods, and engagement of self-control in food choice. RESULTS: There were no differences between AN subtypes in healthiness or tastiness ratings, or in tendency to limit choice of high-fat foods. There was no difference between the groups in measures of self-control in food choice. DISCUSSION: Individuals with ANR and ANBP similarly manifest reduced choices of high-fat foods, with similar tendencies to undervalue the tastiness of high-fat foods. These results suggest that while individuals with ANR and ANBP differ across a range of clinical characteristics, the decision-making process associated with the maladaptive restriction of high-fat foods characteristic of AN is shared by both subtypes.

Dynamic Flexibility in Striatal-Cortical Circuits Supports Reinforcement Learning.

Complex learned behaviors must involve the integrated action of distributed brain circuits. Although the contributions of individual regions to learning have been extensively investigated, much less is known about how distributed brain networks orchestrate their activity over the course of learning. To address this gap, we used fMRI combined with tools from dynamic network neuroscience to obtain time-resolved descriptions of network coordination during reinforcement learning in humans. We found that learning to associate visual cues with reward involves dynamic changes in network coupling between the striatum and distributed brain regions, including visual, orbitofrontal, and ventromedial prefrontal cortex (n = 22; 13 females). Moreover, we found that this flexibility in striatal network coupling correlates with participants' learning rate and inverse temperature, two parameters derived from reinforcement learning models. Finally, we found that episodic learning, measured separately in the same participants at the same time, was related to dynamic connectivity in distinct brain networks. These results suggest that dynamic changes in striatal-centered networks provide a mechanism for information integration during reinforcement learning.SIGNIFICANCE STATEMENT Learning from the outcomes of actions, referred to as reinforcement learning, is an essential part of life. The roles of individual brain regions in reinforcement learning have been well characterized in terms of updating values for actions or cues. Missing from this account, however, is an understanding of how different brain areas interact during learning to integrate sensory and value information. Here we characterize flexible striatal-cortical network dynamics that relate to reinforcement learning behavior.

Modulating the Use of Multiple Memory Systems in Value-based Decisions with Contextual Novelty.

With multiple learning and memory systems at its disposal, the human brain can represent the past in many ways, from extracting regularities across similar experiences (incremental learning) to storing rich, idiosyncratic details of individual events (episodic memory). The unique information carried by these neurologically distinct forms of memory can bias our behavior in different directions, raising crucial questions about how these memory systems interact to guide choice and the factors that cause one to dominate. Here, we devised a new approach to estimate how decisions are independently influenced by episodic memories and incremental learning. Furthermore, we identified a biologically motivated factor that biases the use of different memory types-the detection of novelty versus familiarity. Consistent with computational models of cholinergic memory modulation, we find that choices are more influenced by episodic memories following the recognition of an unrelated familiar image but more influenced by incrementally learned values after the detection of a novel image. Together this work provides a new behavioral tool enabling the disambiguation of key memory behaviors thought to be supported by distinct neural systems while also identifying a theoretically important and broadly applicable manipulation to bias the arbitration between these two sources of memories.

Impaired generalization of reward but not loss in obsessive-compulsive disorder.

BACKGROUND: Generalizing from past experiences can be adaptive by allowing those experiences to guide behavior in new situations. Generalizing too much, however, can be maladaptive. For example, individuals with pathological anxiety are believed to overgeneralize emotional responses from past threats, broadening their scope of fears. Whether individuals with pathological anxiety overgeneralize in other situations remains unclear. METHODS: The present study (N = 57) used a monetary sensory preconditioning paradigm with rewards and losses to address this question in individuals with obsessive-compulsive disorder (OCD) and social anxiety disorder (SAD), comparing them to healthy comparison subjects (HC). In all groups, we tested direct learning of associations between cues and reward vs. loss outcomes, as well as generalization of learning to novel choice options. RESULTS: We found no differences between the three groups in the direct learning of stimuli with their outcomes: all subjects demonstrated intact stimulus-response learning by choosing rewarding options and avoiding negative ones. However, OCD subjects were less likely to generalize from rewards than either the SAD or HC groups, and this impairment was not found for losses. Additionally, greater deficits in reward generalization were correlated with severity of threat estimation, as measured by a subscale of the Obsessive Beliefs Questionnaire, both within OCD and across all groups. CONCLUSIONS: These findings suggest that a compromised ability to generalize from rewarding events may impede adaptive behavior in OCD and in those susceptible to high estimation of threat.

Assessment of test-retest reliability of a food choice task among healthy individuals.

Aberrations in eating patterns constitute a substantial public health burden. Computer-based paradigms that measure responses to images of foods are potentially useful tools for assessing food attitudes and characteristics of eating behavior. In particular, food choice tasks attempt to directly probe aspects of individuals' decisions about what to eat. In the Food Choice Task participants rate the healthiness and tastiness of a variety of food items presented one at a time. Next, participants choose for each food item whether they prefer to eat the item vs. a neutrally rated reference food item. The goal of the current study was to assess the stability and reliability of this Food Choice Task over time and with repeated testing. Secondary analyses were conducted using data from healthy volunteers in two separate studies that administered the task at two time points, separated either by several days or about a month. The overall reliability of the Food Choice Task across multiple administrations was assessed using intra-class correlation coefficients and the reliability of ratings of individual food items was assessed using kappa coefficients. The results indicated that test-retest reliability of the Food Choice Task in healthy volunteers was high at both shorter and longer test-retest intervals. In addition, the reliability of individual food item ratings was good for a majority of items. The proportion of healthy volunteers' high-fat food choices did not change over time in either of the two studies. Thus, the Food Choice Task is suitable for measuring food choices in studies with multiple assessment points. In particular, the task may be well suited to assess restrictive eating, a construct which it has been difficult to assess in experimental settings.

Dopamine selectively remediates 'model-based' reward learning: a computational approach.

Patients with loss of dopamine due to Parkinson's disease are impaired at learning from reward. However, it remains unknown precisely which aspect of learning is impaired. In particular, learning from reward, or reinforcement learning, can be driven by two distinct computational processes. One involves habitual stamping-in of stimulus-response associations, hypothesized to arise computationally from 'model-free' learning. The other, 'model-based' learning, involves learning a model of the world that is believed to support goal-directed behaviour. Much work has pointed to a role for dopamine in model-free learning. But recent work suggests model-based learning may also involve dopamine modulation, raising the possibility that model-based learning may contribute to the learning impairment in Parkinson's disease. To directly test this, we used a two-step reward-learning task which dissociates model-free versus model-based learning. We evaluated learning in patients with Parkinson's disease tested ON versus OFF their dopamine replacement medication and in healthy controls. Surprisingly, we found no effect of disease or medication on model-free learning. Instead, we found that patients tested OFF medication showed a marked impairment in model-based learning, and that this impairment was remediated by dopaminergic medication. Moreover, model-based learning was positively correlated with a separate measure of working memory performance, raising the possibility of common neural substrates. Our results suggest that some learning deficits in Parkinson's disease may be related to an inability to pursue reward based on complete representations of the environment.

Social Cognition as Reinforcement Learning: Feedback Modulates Emotion Inference.

Neuroscientific studies of social cognition typically employ paradigms in which perceivers draw single-shot inferences about the internal states of strangers. Real-world social inference features much different parameters: People often encounter and learn about particular social targets (e.g., friends) over time and receive feedback about whether their inferences are correct or incorrect. Here, we examined this process and, more broadly, the intersection between social cognition and reinforcement learning. Perceivers were scanned using fMRI while repeatedly encountering three social targets who produced conflicting visual and verbal emotional cues. Perceivers guessed how targets felt and received feedback about whether they had guessed correctly. Visual cues reliably predicted one target's emotion, verbal cues predicted a second target's emotion, and neither reliably predicted the third target's emotion. Perceivers successfully used this information to update their judgments over time. Furthermore, trial-by-trial learning signals-estimated using two reinforcement learning models-tracked activity in ventral striatum and ventromedial pFC, structures associated with reinforcement learning, and regions associated with updating social impressions, including TPJ. These data suggest that learning about others' emotions, like other forms of feedback learning, relies on domain-general reinforcement mechanisms as well as domain-specific social information processing.

Dopamine Modulation of Intertemporal Decision-making: Evidence from Parkinson Disease.

Choosing between smaller prompt rewards and larger later rewards is a common choice problem, and studies widely agree that frontostriatal circuits heavily innervated by dopamine are centrally involved. Understanding how dopamine modulates intertemporal choice has important implications for neurobiological models and for understanding the mechanisms underlying maladaptive decision-making. However, the specific role of dopamine in intertemporal decisions is not well understood. Dopamine may play a role in multiple aspects of intertemporal choices--the valuation of choice outcomes and sensitivity to reward delays. To assess the role of dopamine in intertemporal decisions, we tested Parkinson disease patients who suffer from dopamine depletion in the striatum, in either high (on medication, PDON) or low (off medication, PDOFF) dopaminergic states. Compared with both PDOFF and healthy controls, PDON made more farsighted choices and reduced their valuations less as a function of increasing time to reward. Furthermore, reduced discounting in the high dopaminergic state was robust across multiple measures, providing new evidence for dopamine's role in making decisions about the future.

Episodic memory encoding interferes with reward learning and decreases striatal prediction errors.

Learning is essential for adaptive decision making. The striatum and its dopaminergic inputs are known to support incremental reward-based learning, while the hippocampus is known to support encoding of single events (episodic memory). Although traditionally studied separately, in even simple experiences, these two types of learning are likely to co-occur and may interact. Here we sought to understand the nature of this interaction by examining how incremental reward learning is related to concurrent episodic memory encoding. During the experiment, human participants made choices between two options (colored squares), each associated with a drifting probability of reward, with the goal of earning as much money as possible. Incidental, trial-unique object pictures, unrelated to the choice, were overlaid on each option. The next day, participants were given a surprise memory test for these pictures. We found that better episodic memory was related to a decreased influence of recent reward experience on choice, both within and across participants. fMRI analyses further revealed that during learning the canonical striatal reward prediction error signal was significantly weaker when episodic memory was stronger. This decrease in reward prediction error signals in the striatum was associated with enhanced functional connectivity between the hippocampus and striatum at the time of choice. Our results suggest a mechanism by which memory encoding may compete for striatal processing and provide insight into how interactions between different forms of learning guide reward-based decision making.

Transfer of learning relates to intrinsic connectivity between hippocampus, ventromedial prefrontal cortex, and large-scale networks.

An important aspect of adaptive learning is the ability to flexibly use past experiences to guide new decisions. When facing a new decision, some people automatically leverage previously learned associations, while others do not. This variability in transfer of learning across individuals has been demonstrated repeatedly and has important implications for understanding adaptive behavior, yet the source of these individual differences remains poorly understood. In particular, it is unknown why such variability in transfer emerges even among homogeneous groups of young healthy participants who do not vary on other learning-related measures. Here we hypothesized that individual differences in the transfer of learning could be related to relatively stable differences in intrinsic brain connectivity, which could constrain how individuals learn. To test this, we obtained a behavioral measure of memory-based transfer outside of the scanner and on a separate day acquired resting-state functional MRI images in 42 participants. We then analyzed connectivity across independent component analysis-derived brain networks during rest, and tested whether intrinsic connectivity in learning-related networks was associated with transfer. We found that individual differences in transfer were related to intrinsic connectivity between the hippocampus and the ventromedial prefrontal cortex, and between these regions and large-scale functional brain networks. Together, the findings demonstrate a novel role for intrinsic brain dynamics in flexible learning-guided behavior, both within a set of functionally specific regions known to be important for learning, as well as between these regions and the default and frontoparietal networks, which are thought to serve more general cognitive functions.

Dynamic shifts in brain network activation during supracapacity working memory task performance.

Despite significant advances in understanding how brain networks support working memory (WM) and cognitive control, relatively little is known about how these networks respond when cognitive capabilities are overtaxed. We used a fine-grained manipulation of memory load within a single trial to exceed WM capacity during functional magnetic resonance imaging to investigate how these networks respond to support task performance when WM capacity is exceeded. Analyzing correct trials only, we observed a nonmonotonic (inverted-U) response to WM load throughout the classic WM network (including bilateral dorsolateral prefrontal cortex, posterior parietal cortex, and presupplementary motor areas) that peaked later in individuals with greater WM capacity. We also observed a relative increase in activity in medial anterior prefrontal cortex, posterior cingulate/precuneus, and lateral temporal and parietal regions at the highest WM loads, and a set of predominantly subcortical and prefrontal regions whose activation was greatest at the lowest WM loads. At the individual subject level, the inverted-U pattern was associated with poorer performance while expression of the early and late activating patterns was predictive of better performance. In addition, greater activation in bilateral fusiform gyrus and right occipital lobe at the highest WM loads predicted better performance. These results demonstrate dynamic and behaviorally relevant changes in the level of activation of multiple brain networks in response to increasing WM load that are not well accounted for by present models of how the brain subserves the cognitive ability to hold and manipulate information on-line.