Human Reward Research Articles

Social interaction for efficient agent learning from human reward

Learning from rewards generated by a human trainer observing an agent in action has been proven to be a powerful method for teaching autonomous agents to perform challenging tasks, especially for those non-technical users. Since the efficacy of this approach depends critically on the reward the trainer provides, we consider how the interaction between the trainer and the agent should be designed so as to increase the efficiency of the training process. This article investigates the influence of the agent’s socio-competitive feedback on the human trainer’s training behavior and the agent’s learning. The results of our user study with 85 participants suggest that the agent’s passive socio-competitive feedback—showing performance and score of agents trained by trainers in a leaderboard—substantially increases the engagement of the participants in the game task and improves the agents’ performance, even though the participants do not directly play the game but instead train the agent to do so. Moreover, making this feedback active—sending the trainer her agent’s performance relative to others—further induces more participants to train agents longer and improves the agent’s learning. Our further analysis shows that agents trained by trainers affected by both the passive and active social feedback could obtain a higher performance under a score mechanism that could be optimized from the trainer’s perspective and the agent’s additional active social feedback can keep participants to further train agents to learn policies that can obtain a higher performance under such a score mechanism.

Reminders of past choices bias decisions for reward in humans

We provide evidence that decisions are made by consulting memories for individual past experiences, and that this process can be biased in favour of past choices using incidental reminders. First, in a standard rewarded choice task, we show that a model that estimates value at decision-time using individual samples of past outcomes fits choices and decision-related neural activity better than a canonical incremental learning model. In a second experiment, we bias this sampling process by incidentally reminding participants of individual past decisions. The next decision after a reminder shows a strong influence of the action taken and value received on the reminded trial. These results provide new empirical support for a decision architecture that relies on samples of individual past choice episodes rather than incrementally averaged rewards in evaluating options and has suggestive implications for the underlying cognitive and neural mechanisms.

Training an Actor-Critic Reinforcement Learning Controller for Arm Movement Using Human-Generated Rewards.

Functional Electrical Stimulation (FES) employs neuroprostheses to apply electrical current to the nerves and muscles of individuals paralyzed by spinal cord injury to restore voluntary movement. Neuroprosthesis controllers calculate stimulation patterns to produce desired actions. To date, no existing controller is able to efficiently adapt its control strategy to the wide range of possible physiological arm characteristics, reaching movements, and user preferences that vary over time. Reinforcement learning (RL) is a control strategy that can incorporate human reward signals as inputs to allow human users to shape controller behavior. In this paper, ten neurologically intact human participants assigned subjective numerical rewards to train RL controllers, evaluating animations of goal-oriented reaching tasks performed using a planar musculoskeletal human arm simulation. The RL controller learning achieved using human trainers was compared with learning accomplished using human-like rewards generated by an algorithm; metrics included success at reaching the specified target; time required to reach the target; and target overshoot. Both sets of controllers learned efficiently and with minimal differences, significantly outperforming standard controllers. Reward positivity and consistency were found to be unrelated to learning success. These results suggest that human rewards can be used effectively to train RL-based FES controllers.

Nicotine Acutely Enhances Reinforcement from Non-Drug Rewards in Humans.

Preclinical research documents that, aside from the primary and secondary reinforcing effects of nicotine intake itself, nicotine also acutely enhances the reinforcing efficacy of non-drug reinforcers (“rewards”). Study of these effects in humans has largely been overlooked, but very recent findings suggest they may have clinical implications for more fully understanding the persistence of tobacco dependence. This overview first outlines the topic and notes some recent human studies indirectly addressing nicotine effects on related responses (e.g., subjective ratings), explaining why those findings do not directly confirm enhancement of behavioral reinforcement per se due to nicotine. Then, the methodology used in the subsequently presented studies is described, demonstrating how those studies specifically did demonstrate enhancement of reinforced responding for non-drug rewards. The main section focuses on the limited controlled research to date directly assessing nicotine’s acute reinforcement-enhancing effects in humans, particularly as it relates to reinforced behavioral responding for non-drug rewards in non-human animal models. After detailing those few existing human studies, we address potential consequences of these effects for dependence and tobacco cessation efforts and then suggest directions for future research. This research indicates that nicotine per se increases responding in humans that is reinforced by some rewards (auditory stimuli via music, visual stimuli via video), but perhaps not by others (e.g., money). These reinforcement-enhancing effects in smokers are not due to dependence or withdrawal relief and can be restored by a small amount of nicotine (similar to a smoking lapse), including from e-cigarettes, a non-tobacco nicotine product. Future clinical research should examine factors determining which types of rewards are (or are not) enhanced by nicotine, consequences of the loss of these nicotine effects after quitting smoking, potential individual differences in these effects, and the possibility that nicotine via nicotine replacement therapy and non-nicotine quit medications may attenuate loss of these effects upon quitting. Further study with humans of nicotine’s reinforcement-enhancing effects may provide a more complete understanding of smoking persistence and added mechanisms of cessation medication efficacy.

The human subthalamic nucleus and globus pallidus internus differentially encode reward during action control.

The subthalamic nucleus (STN) and globus pallidus internus (GPi) have recently been shown to encode reward, but few studies have been performed in humans. We investigated STN and GPi encoding of reward and loss (i.e., valence) in humans with Parkinson's disease. To test the hypothesis that STN and GPi neurons would change their firing rate in response to reward- and loss-related stimuli, we recorded the activity of individual neurons while participants performed a behavioral task. In the task, action choices were associated with potential rewarding, punitive, or neutral outcomes. We found that STN and GPi neurons encode valence-related information during action control, but the proportion of valence-responsive neurons was greater in the STN compared to the GPi. In the STN, reward-related stimuli mobilized a greater proportion of neurons than loss-related stimuli. We also found surprising limbic overlap with the sensorimotor regions in both the STN and GPi, and this overlap was greater than has been previously reported. These findings may help to explain alterations in limbic function that have been observed following deep brain stimulation therapy of the STN and GPi. Hum Brain Mapp 38:1952-1964, 2017. © 2017 Wiley Periodicals, Inc.

Reward and Performance as a Key Element of Motivation

Abstract The crisis and the dramatic unemployment have had an inevitable impact in salary policy and development of organizations. Therefore, considering that every organization is interested for their employees to be more motivated to work, thus, this was for this reason and relying on literature, papers and research to know to what extend the reward influences motivation of the employees. Initially, we shall also understand the notion salary and performance and how important are they in carrying out the work in one organization, and how they have evaluated by the management of the same organization. Afterwards, we shall show the way and the best methods on how to pay employees, how the manager must decide on the level of salaries of the employees, and must decide that salaries are not the same for each employee based on their performance, etc. The time has shown that the performance evaluation provides information that serves the basis for all other important activities of human resources management planning, selection, reward and motivation, continuously leading to the formation of employees. It is worth mentioning the fact that companies in general still has much work to do in connection with programs for compensation of employees, and this relying on the fact that many companies, particularly in the Balkan states where they continue to give salary different bonuses in envelopes, while reducing the transparency of payment and reward in these companies-organizations.

Virtual reality conditioned place preference using monetary reward

Computerized tasks based on conditioned place preference (CPP) methodology offer the opportunity to study learning mechanisms involved in conditioned reward in humans. In this study, we examined acquisition and extinction of a CPP for virtual environments associated with monetary reward ($).Healthy men and women (N=57) completed a computerized CPP task in which they controlled an avatar within a virtual environment. On day 1, subjects completed 6 conditioning trials in which one room was paired with high $ and another with low $. Acquisition of place conditioning was assessed by measuring the time spent in each room during an exploration test of the virtual environments and using self-reported ratings of room liking and preference. Twenty-four hours later, retention and extinction of CPP were assessed during 4 successive exploration tests of the virtual environments.Participants exhibited a place preference for (spent significantly more time in) the virtual room paired with high $ over the one paired with low $ (p=0.015). They also reported that they preferred the high $ room (p<0.001) and liked it significantly more than the low $ room (p<0.001). However, these preferences were short-lived: 24h later subjects did not exhibit a behavioral or subjective preference for the high $ room.These findings show that individuals exhibit transient behavioral and subjective preferences for a virtual environment paired with monetary reward. Variations on this task may be useful to study mechanisms and brain substrates involved in conditioned reward and to examine the influence of drugs upon appetitive conditioning.

Diurnal rhythms in psychological reward functioning in healthy young men: ‘Wanting’, liking, and learning

ABSTRACTA range of evidence suggests that human reward functioning is partly driven by the endogenous circadian system, generating 24-hour rhythms in behavioural measures of reward activation. Reward functioning is multifaceted but literature to date is largely limited to measures of self-reported positive mood states. The aim of this study was to advance the field by testing for hypothesised diurnal variation in previously unexplored components of psychological reward: ‘wanting’, liking, and learning using subjective and behavioural measures. Risky decision making (automatic Balloon Analogue Risk Task), affective responsivity to positive images (International Affective Pictures System), uncued self-reported discrete emotions, and learning-contingent reward (Iowa Gambling Task) were measured at 10.00 hours, 14.00 hours, and 19.00 hours in a counterbalanced repeated measures design with 50 healthy male participants (aged 18–30). As hypothesised, risky decision making (unconscious ‘wanting’) and ratings of arousal towards positive images (conscious wanting) exhibited a diurnal waveform with indices highest at 14.00 hours. No diurnal rhythm was observed for liking (pleasure ratings to positive images, discrete uncued positive emotions) or in a learning-contingent reward task. Findings reaffirm that diurnal variation in human reward functioning is most pronounced in the motivational ‘wanting’ components of reward.

Malfunctions in synaptic membrane trafficking in early pathology of Parkinson’s disease: New molecular clues

The midbrain dopaminergic neurons of the substantia nigra and the ventral tegmental area play vital roles in the regulation of voluntary movement, emotion and reward in humans. These neurons are highly metabolic and are under constant oxidative stress. The dopaminergic neurons form extensive synaptic projections to the striatum. When these neurons start dying or when their synaptic connections fail, humans develop Parkinson's disease. This disease is accompanied by the accumulation of toxic α-synuclein-containing protein aggregates in nigrostriatal processes. Synucleins accumulate in a majority of healthy nerve terminals in the central nervous system, but what causes the formation of pathological synuclein aggregates is unclear. Recent studies point out that the interface between membrane trafficking in the nerve terminal and the autophagy–lysosomal pathway is the site for the aggregate assembly. An urgent goal is to find therapeutic targets at early stages of the disease when neurons are still functional.

The hedonic and hungry brain: Nutritional, hormonal, surgical and genetic modulation of human food reward

The hedonic and hungry brain: Nutritional, hormonal, surgical and genetic modulation of human food reward

Non-demented Parkinson's disease patients with apathy show decreased grey matter volume in key executive and reward-related nodes.

Apathy is a common but poorly understood neuropsychiatric disturbance in Parkinson's disease (PD). In a recent study using event-related brain potentials we demonstrated impaired reward processing and compromised mesocortico-limbic pathways in PD patients with clinical symptoms of apathy. Here we aimed to further investigate the involvement of reward circuits in apathetic PD patients by assessing potential differences in brain structure. Using structural magnetic resonance imaging (MRI) and voxel-based morphometry (VBM) we quantified grey matter volume (GMV) in a sample of 18 non-demented and non-depressed PD patients with apathy, and 18 matched non-apathetic patients. Both groups were equivalent in terms of sociodemographic characteristics, disease stage, cognitive performance and L-Dopa equivalent daily dose. Apathetic patients showed significant GMV loss in cortical and subcortical brain structures. Various clusters of cortical GMV decrease were found in the parietal, lateral prefrontal cortex, and orbitofrontal cortex (OFC). The second largest cluster of GMV loss was located in the left nucleus accumbens (NAcc), a subcortical structure that is a key node of the human reward circuit. Isolated apathy in our sample is explained by the combined GMV loss in regions involved in executive functions, and cortical and subcortical structures of the mesolimbic reward pathway. The correlations observed between apathy and cognition suggests apathy as a marker of more widespread brain degeneration even in a sample of non-demented PD patients.

Choice between delayed food and immediate oxycodone in rats

The choice to seek immediate drug effects instead of more meaningful but delayed rewards is a defining feature of addiction. To develop a rodent model of this behavior, we allowed rats to choose between immediate intravenous delivery of the prescription opioid oxycodone (50μg/kg) and delayed delivery of palatable food pellets. Rats preferred food at delays up to 30s, but they chose oxycodone and food equally at 60-s delay and preferred oxycodone over food at 120-s delay. Comparison of food-drug choice, food-only, and drug-only conditions indicated that food availability decreased drug intake, but drug availability increased food intake. In the food-only condition, food was effective as a reinforcer even when delayed by 120s. Pre-session feeding with chow slowed acquisition of food and drug self-administration, but did not affect choice. To establish procedures for testing potential anti-addiction medications, noncontingent pre-treatment with oxycodone or naltrexone (analogous to substitution and antagonist therapies, respectively) were tested on a baseline in which oxycodone was preferred over delayed food. Naltrexone pre-treatment decreased drug intake and increased food intake. Oxycodone pre-treatment decreased drug intake, but also produced extended periods with no food or drug responding. These findings show that the contingencies that induce preference for drugs over more meaningful but less immediate rewards in humans can be modeled in rodents, and they suggest that the model could be useful for assessing the therapeutic potential of treatments and exploring the underlying behavioral and neural mechanisms involved in addiction.

The µ-opioid system promotes visual attention to faces and eyes.

Paying attention to others' faces and eyes is a cornerstone of human social behavior. The µ-opioid receptor (MOR) system, central to social reward-processing in rodents and primates, has been proposed to mediate the capacity for affiliative reward in humans. We assessed the role of the human MOR system in visual exploration of faces and eyes of conspecifics. Thirty healthy males received a novel, bidirectional battery of psychopharmacological treatment (an MOR agonist, a non-selective opioid antagonist, or placebo, on three separate days). Eye-movements were recorded while participants viewed facial photographs. We predicted that the MOR system would promote visual exploration of faces, and hypothesized that MOR agonism would increase, whereas antagonism decrease overt attention to the information-rich eye region. The expected linear effect of MOR manipulation on visual attention to the stimuli was observed, such that MOR agonism increased while antagonism decreased visual exploration of faces and overt attention to the eyes. The observed effects suggest that the human MOR system promotes overt visual attention to socially significant cues, in line with theories linking reward value to gaze control and target selection. Enhanced attention to others' faces and eyes represents a putative behavioral mechanism through which the human MOR system promotes social interest.

Investigating the Impact of a Genome-Wide Supported Bipolar Risk Variant of MAD1L1 on the Human Reward System.

Recent genome-wide association studies have identified MAD1L1 (mitotic arrest deficient-like 1) as a susceptibility gene for bipolar disorder and schizophrenia. The minor allele of the single-nucleotide polymorphism (SNP) rs11764590 in MAD1L1 was associated with bipolar disorder. Both diseases, bipolar disorder and schizophrenia, are linked to functional alterations in the reward system. We aimed at investigating possible effects of the MAD1L1 rs11764590 risk allele on reward systems functioning in healthy adults. A large homogenous sample of 224 young (aged 18-31 years) participants was genotyped and underwent functional magnetic resonance imaging (fMRI). All participants performed the 'Desire-Reason Dilemma' paradigm investigating the neural correlates that underlie reward processing and active reward dismissal in favor of a long-term goal. We found significant hypoactivations of the ventral tegmental area (VTA), the bilateral striatum and bilateral frontal and parietal cortices in response to conditioned reward stimuli in the risk allele carriers compared with major allele carriers. In the dilemma situation, functional connectivity between prefrontal brain regions and the ventral striatum was significantly diminished in the risk allele carriers. Healthy risk allele carriers showed a significant deficit of their bottom-up response to conditioned reward stimuli in the bilateral VTA and striatum. Furthermore, functional connectivity between the ventral striatum and prefrontal areas exerting top-down control on the mesolimbic reward system was reduced in this group. Similar alterations in reward processing and disturbances of prefrontal control mechanisms on mesolimbic brain circuits have also been reported in bipolar disorder and schizophrenia. Together, these findings suggest the existence of an intermediate phenotype associated with MAD1L1.

Cocaine addiction is associated with abnormal prefrontal function, increased striatal connectivity and sensitivity to monetary incentives, and decreased connectivity outside the human reward circuit.

Cocaine addiction has been associated with increased sensitivity of the human reward circuit to drug-related stimuli. However, the capacity of non-drug incentives to engage this network is poorly understood. Here, we characterized the functional sensitivity to monetary incentives and the structural integrity of the human reward circuit in abstinent cocaine-dependent (CD) patients and their matched controls. We assessed the BOLD response to monetary gains and losses in 30 CD patients and 30 healthy controls performing a lottery task in a magnetic resonance imaging scanner. We measured brain gray matter volume (GMV) using voxel-based morphometry and white matter microstructure using voxel-based fractional anisotropy (FA). Functional data showed that, after monetary incentives, CD patients exhibited higher activation in the ventral striatum than controls. Furthermore, we observed an inverted BOLD response pattern in the prefrontal cortex, with activity being highest after unexpected high gains and lowest after losses. Patients showed increased GMV in the caudate and the orbitofrontal cortex, increased white matter FA in the orbito-striatal pathway but decreased FA in antero-posterior association bundles. Abnormal activation in the prefrontal cortex correlated with GMV and FA increases in the orbitofrontal cortex. While functional abnormalities in the ventral striatum were inversely correlated with abstinence duration, structural alterations were not. In conclusion, results suggest abnormal incentive processing in CD patients with high salience for rewards and punishments in subcortical structures but diminished prefrontal control after adverse outcomes. They further suggest that hypertrophy and hyper-connectivity within the reward circuit, to the expense of connectivity outside this network, characterize cocaine addiction.

Two spatiotemporally distinct value systems shape reward-based learning in the human brain.

Avoiding repeated mistakes and learning to reinforce rewarding decisions is critical for human survival and adaptive actions. Yet, the neural underpinnings of the value systems that encode different decision-outcomes remain elusive. Here coupling single-trial electroencephalography with simultaneously acquired functional magnetic resonance imaging, we uncover the spatiotemporal dynamics of two separate but interacting value systems encoding decision-outcomes. Consistent with a role in regulating alertness and switching behaviours, an early system is activated only by negative outcomes and engages arousal-related and motor-preparatory brain structures. Consistent with a role in reward-based learning, a later system differentially suppresses or activates regions of the human reward network in response to negative and positive outcomes, respectively. Following negative outcomes, the early system interacts and downregulates the late system, through a thalamic interaction with the ventral striatum. Critically, the strength of this coupling predicts participants' switching behaviour and avoidance learning, directly implicating the thalamostriatal pathway in reward-based learning.

Using informative behavior to increase engagement while learning from human reward

In this work, we address a relatively unexplored aspect of designing agents that learn from human reward. We investigate how an agent's non-task behavior can affect a human trainer's training and agent learning. We use the TAMER framework, which facilitates the training of agents by human-generated reward signals, i.e., judgements of the quality of the agent's actions, as the foundation for our investigation. Then, starting from the premise that the interaction between the agent and the trainer should be bi-directional, we propose two new training interfaces to increase a human trainer's active involvement in the training process and thereby improve the agent's task performance. One provides information on the agent's uncertainty which is a metric calculated as data coverage, the other on its performance. Our results from a 51-subject user study show that these interfaces can induce the trainers to train longer and give more feedback. The agent's performance, however, increases only in response to the addition of performance-oriented information, not by sharing uncertainty levels. These results suggest that the organizational maxim about human behavior, get what you measure--i.e., sharing metrics with people causes them to focus on optimizing those metrics while de-emphasizing other objectives--also applies to the training of agents. Using principle component analysis, we show how trainers in the two conditions train agents differently. In addition, by simulating the influence of the agent's uncertainty---informative behavior on a human's training behavior, we show that trainers could be distracted by the agent sharing its uncertainty levels about its actions, giving poor feedback for the sake of reducing the agent's uncertainty without improving the agent's performance.

Circadian Rhythms and Psychopathology: From Models of Depression to Rhythms in Clock Gene Expression and Back Again

Unpredictable chronic mild stress (UCMS) in rodents induces several changes in behavior and physiology that are similar to major depressive disorder (MDD) in humans. Although the specific brain areas, hormones, and genes that mediate this complex response to UCMS are important points for research and discussion, potential circadian mechanisms have also been considered for some time. Logan et al. (1) describe a compelling correlation between behavioral assays that are relevant to depression and the amplitude of daily locomotion and temperature rhythms in mice. Moreover, they correlate the amplitude of in vitro rhythms of circadian clock gene expression in the nucleus accumbens (NAc) and suprachiasmatic nucleus (SCN), the master circadian clock, with performance on the same tasks. These findings are relevant to two key aspects of MDD. First, MDD is often associated with disrupted circadian rhythms in behavior and physiology, and normalized rhythms often relate to an attenuation of symptoms. Second, MDD has been linked with decreased interest in normally rewarding activities, and the NAc is a key player in the processing of natural and drug-related rewards in rodents and humans (2). Logan et al. (1) describe UCMS-induced deficits on depression-related tasks that are correlated with rhythms of locomotion, core body temperature, and regionspecific clock gene expression. As a result, these findings could be an important step toward describing potential circadian mechanisms underlying MDD and depression-like responses to UCMS. Circadian rhythms in behavior and physiology are generated through the actions of endogenous clocks distributed throughout the brain and body. At the cellular level, these clocks use complex transcriptional-translational feedback loops that depend on the rhythmic expression of a wide array of clock genes (3). However, many researchers have focused on the highly rhythmic expression patterns of period circadian clock 1 or 2 genes (PER1 or PER2) as markers that track the function and phase of the more general circadian loop. Daily rhythms of clock gene expression in the SCN are generally thought to facilitate the clock functions of this brain area. Daily rhythms of PER1 and PER2 expression are also observed in many other brain areas outside of the SCN (4), areas known to be important in learning, memory, motivation, emotion, and reward-related processing. However, the importance of rhythms in clock gene expression per se is still unclear for most of these extra-SCN areas. To remain synchronized with environmental cycles, both SCN and extra-SCN rhythms are influenced by time cues, or zeitgebers. For example, the environmental light-dark cycle is a powerful zeitgeber that adjusts rhythms of period circadian clock gene expression in

An Enterprise Systems Model to Deliver Innovation in the Healthcare Industry Based on Cognitive and Social-Tech Engineering

An optimal enterprise system that fosters innovation is based on people, process and technology. However, current enterprise systems lack a significant involvement of the human aspects. An innovation model can incorporate those three components to provide greater value to stakeholders, drive organizational growth, gain a competitive advantage and achieve greater end-user satisfaction. We investigated how an enterprise systems model for open innovation could be created based on existing innovation socio-technical frameworks and how that model may benefit the UK National Health Service (NHS). We carried out a literature review of enterprise systems models, cognitive and social-tech engineering and innovation management. Two researchers undertook a thematic analysis of 46 academic papers. Themes included business sectors, types of innovation strategy and the use of cognitive and social-tech engineering elements. Results show that little application of innovation models within the healthcare sector has been undertaken to foster innovation. We suggest that existing innovation frameworks such as the Innovation Pentathlon Framework, the NASA's partnering incentive and the Tesco club card scheme may be the basis for the creation of a healthcare innovation model that incorporates substantive human aspects, including human cognition and human reward. Since the cost of missing medical appointments in the NHS is around $1390 million dollars per year, our innovation model may be implemented to improve patient attendance to family doctors and hospital appointments. The model strategy comprises rewarding people with shopping vouchers (Tesco club card points or nectar points) each time they confirm or cancel a medical appointment online or by SMS text message. With this model, the NHS would improve the management of their resources, may improve the quality of service and may have higher user satisfaction. Nevertheless, future challenges may include the adaptation of current NHS business processes and the management of the collaboration with private organizations.

Altered resting-state functional connectivity of the frontal-striatal reward system in social anxiety disorder.

We investigated differences in the intrinsic functional brain organization (functional connectivity) of the human reward system between healthy control participants and patients with social anxiety disorder. Functional connectivity was measured in the resting-state via functional magnetic resonance imaging (fMRI). 53 patients with social anxiety disorder and 33 healthy control participants underwent a 6-minute resting-state fMRI scan. Functional connectivity of the reward system was analyzed by calculating whole-brain temporal correlations with a bilateral nucleus accumbens seed and a ventromedial prefrontal cortex seed. Patients with social anxiety disorder, relative to the control group, had (1) decreased functional connectivity between the nucleus accumbens seed and other regions associated with reward, including ventromedial prefrontal cortex; (2) decreased functional connectivity between the ventromedial prefrontal cortex seed and lateral prefrontal regions, including the anterior and dorsolateral prefrontal cortices; and (3) increased functional connectivity between both the nucleus accumbens seed and the ventromedial prefrontal cortex seed with more posterior brain regions, including anterior cingulate cortex. Social anxiety disorder appears to be associated with widespread differences in the functional connectivity of the reward system, including markedly decreased functional connectivity between reward regions and between reward regions and lateral prefrontal cortices, and markedly increased functional connectivity between reward regions and posterior brain regions.