Predictive learning can engage a selective form of cognitive control that biases choice between actions based on information about future outcomes that the learning provides. This influence has been hypothesized to depend on a feedback circuit in the brain through which the basal ganglia modulate activity in the prefrontal cortex; however, direct evidence for this functional circuit has proven elusive. Here, using an animal model of cognitive control, we found that the influence of predictive learning on decision making is mediated by an inhibitory feedback circuit linking the medial ventral pallidum and the mediodorsal thalamus, the activation of which causes disinhibition of the orbitofrontal cortex via reduced activation of inhibitory parvalbumin interneurons during choice. Thus, we found that, for this function, the mediodorsal thalamus serves as a pallidal-cortical relay through which predictive learning controls action selection, which has important implications for understanding cognitive control and its vicissitudes in various psychiatric disorders and addiction.

In real-world decision-making scenarios, negative consequences do not always occur immediately after a choice. This delay between action and outcome drives the underestimation, or \"delay discounting\", of punishment. While the neural substrates underlying sensitivity to immediate punishment have been well-studied, there has been minimal investigation of delayed consequences. Here, we assessed the role of lateral orbitofrontal cortex (LOFC) and basolateral amygdala (BLA), two regions implicated in cost/benefit decision-making, in sensitivity to delayed vs immediate punishment. The delayed punishment decision-making task (DPDT) was used to measure delay discounting of punishment in rodents. During DPDT, rats choose between a small, single pellet reward and a large, three pellet reward accompanied by a mild foot shock. As the task progresses, the shock is preceded by a delay that systematically increases or decreases throughout the session. We observed that rats avoid choices associated with immediate punishment, then shift preference toward these options when punishment is delayed. LOFC inactivation did not influence choice of rewards with immediate punishment, but decreased choice of delayed punishment. We also observed that BLA inactivation reduced choice of delayed punishment for ascending but not descending delays. Inactivation of either brain region produced comparable effects on decision-making in males and females, but there were sex differences observed in omissions and latency to make a choice. In summary, both LOFC and BLA contribute to the delay discounting of punishment and may serve as promising therapeutic targets to improve sensitivity to delayed punishment during decision-making.Significance StatementNegative consequences occurring after a delay are often underestimated, which can lead to maladaptive decision-making. While sensitivity to immediate punishment during reward-seeking has been well-studied, the neural substrates underlying sensitivity to delayed punishment remain unclear. Here, we used the Delayed Punishment Decision-making Task to determine that lateral orbitofrontal cortex and basolateral amygdala both regulate the discounting of delayed punishment, suggesting that these regions may be potential targets to improve decision-making in psychopathology.

Recent findings from this laboratory demonstrate that ethanol reduces the intrinsic excitability of orbitofrontal cortex (OFC) neurons via activation of strychnine-sensitive glycine receptors. Although the mechanism linking ethanol to the release of glycine is currently unknown, astrocytes are a source of neurotransmitters including glycine and activation of dopamine D1-like receptors has been reported to enhance extracellular levels of glycine via a functional reversal of the astrocytic glycine transporter GlyT1. We recently reported that like ethanol, dopamine or a D1/D5 receptor agonist increases a tonic current in lateral OFC (lOFC) neurons. Therefore, in this study, we used whole-cell patch-clamp electrophysiology to examine whether ethanol inhibition of OFC spiking involves the release of glycine from astrocytes and whether this release is dopamine receptor dependent. Ethanol, applied acutely, decreased spiking of lOFC neurons and this effect was blocked by antagonists of GlyT1, the norepinephrine transporter or D1-like but not D2-like receptors. Ethanol enhanced the tonic current of OFC neurons and occluded the effect of dopamine suggesting that ethanol and dopamine may share a common pathway. Altering astrocyte function by suppressing intracellular astrocytic calcium signaling or blocking the astrocyte-specific Kir4.1 potassium channels reduced but did not completely abolish ethanol inhibition of OFC neuron firing. However, when both astrocytic calcium signaling and Kir4.1 channels were inhibited, ethanol had no effect on firing. Ethanol inhibition was also prevented by inhibitors of phospholipase C and conventional isoforms of protein kinase C (cPKC) previously shown to block D1R-induced GlyT1 reversal and PKC inhibition of Kir4.1 channels. Finally, the membrane potential of OFC astrocytes was depolarized by bath application of a Kir4.1 blocker, a D1 agonist or ethanol and ethanol effect was blocked by a D1 antagonist. Together, these findings suggest that acute ethanol inhibits OFC neuron excitability via a D1 receptor-mediated dysregulation of astrocytic glycine transport.

A comprehensive concept of the biological basis of reward, social and emotional behavior, and language requires a deeper understanding of the microstructure and connectivity of the underlying brain regions. Such understanding could provide deeper insights into their role in functional networks, and form the anatomical basis of the functional segregation of this region as shown in recent in vivo imaging studies. Here, we investigated the cytoarchitecture of the lateral orbitofrontal cortex (lateral OFC) in serial histological sections of 10 human postmortem brains by image analysis and a statistically reproducible approach to detect borders between cortical areas. Profiles of the volume fraction of cell bodies were therefore extracted from digitized histological images, describing laminar changes from the layer I/layer II boundary to the white matter. As a result, four new areas, Fo4-7, were identified. Area Fo4 was mainly found in the anterior orbital gyrus (AOG), Fo5 anteriorly in the inferior frontal gyrus (IFG), Fo6 in the lateral orbital gyrus (LOG), and Fo7 in the lateral orbital sulcus. Areas differed in cortical thickness, abundance and size of pyramidal cells in layer III and degree of granularity in layer IV. A hierarchical cluster analysis was used to quantify cytoarchitectonic differences between them. The 3D-reconstructed areas were transformed into the single-subject template of the Montreal Neurological Institute (MNI), where probabilistic maps and a maximum probability map (MPM) were calculated as part of the JuBrain Cytoarchitectonic Atlas. These maps served as reference data to study the functional properties of the areas using the BrainMap database. The type of behavioral tasks that activated them was analyzed to get first insights of co-activation patterns of the lateral OFC and its contribution to cognitive networks. Meta-analytic connectivity modeling (MACM) showed that functional decoding revealed activation in gustatory perception in Fo4; reward and somesthetic perception in Fo5; semantic processing and pain perception in Fo6; and emotional processing and covert reading in Fo7. Together with existing maps of the JuBrain Cytoarchitectonic Atlas, the new maps can now be used as an open-source reference for neuroimaging studies, allowing to further decode brain function.

BACKGROUND: Control of reward-seeking behavior under conditions of punishment is an important function for survival.
OBJECTIVE: We designed a task in which rats could choose to either press a lever and obtain a food pellet accompanied by a footshock or refrain from pressing the lever to avoid footshock, in response to tone presentation. In the task, footshock intensity steadily increased, and the task was terminated when the lever press probability reached < 25% (last intensity). Rats were trained until the last intensity was stable. Subsequently, we investigated the effects of the pharmacological inactivation of the ventromedial prefrontal cortex (vmPFC), lateral orbitofrontal cortex (lOFC), and basolateral amygdala (BLA) on task performance.
RESULTS: Bilateral inactivation of the vmPFC, lOFC, and BLA did not alter lever press responses at the early stage of the task. The number of lever presses increased following vmPFC and BLA inactivation but decreased following lOFC inactivation during the later stage of the task. The last intensity was elevated by vmPFC or BLA inactivation but lowered by lOFC inactivation. Disconnection of the vmPFC-BLA pathway induced behavioral alterations that were similar to vmPFC or BLA inactivation. Inactivation of any regions did not alter footshock sensitivity and anxiety levels.
CONCLUSIONS: Our results demonstrate a strong role of the vmPFC and BLA and their interactions in reward restraint to avoid punishment and a prominent role of the lOFC in reward-seeking under reward/punishment conflict situations.

Alcohol misuse is associated with significant energy deficits. As feeding involves multiple sensory, cognitive, and affective processes, low food intake in problem drinkers likely reflects alterations in both regional and inter-regional responses. To investigate the effects of problem drinking on feeding-related neural activities and connectivities, we examined functional magnetic resonance imaging (fMRI) data in 82 drinkers who viewed palatable food and nonfood images in alternating blocks. Drinking severity was assessed with the Alcohol Use Disorders Identification Test (AUDIT). A whole-brain multiple regression with AUDIT scores as the predictor showed a negative correlation between drinking severity and activation to food vs nonfood cues in the lateral orbitofrontal cortex (lOFC). AUDIT scores were also negatively correlated with the gray matter volume (GMV) of the lOFC and regions that responded preferentially to food stimuli, including the left middle frontal gyrus, bilateral middle insula, and occipital cortices. Connectivity strength between the lOFC and these regions was negatively modulated by drinking severity. In contrast, there was no relationship between AUDIT scores and lOFC connectivity with regions that did not show either selectivity to food images or GMV loss. A mediation analysis further suggested that alcohol misuse may have compromised lOFC\'s structural integrity, which in turn disrupted lOFC interactions with regions that support the processing of visual food cues. Overall, the findings provide evidence for the effects of problem drinking on the brain substrates of feeding, potentially shedding light on the neural mechanisms underlying energy deficits in at-risk drinkers.

The orbitofrontal cortex (OFC) has been proposed to function as a cognitive map of task space: a mental model of the steps involved in a task. This idea has proven popular because it provides a cohesive explanation for a number of disparate findings regarding the OFC\'s role in a broad array of tasks. Concurrently, evidence has begun to reveal the functional heterogeneity of OFC subregions, particularly the medial and lateral OFC. How these subregions uniquely contribute to the OFC\'s role as a cognitive map of task space, however, has not been explored. Here we propose that, in rodents, the lateral OFC represents the agent\'s initial position within that task map (i.e. initial state), determining which actions are available as a consequence of that position, whereas the medial OFC represents the agent\'s future position within the task map (i.e. terminal state), influencing which actions are selected to achieve that position. We argue that these processes are achieved somewhat independently and somewhat interdependently, and are achieved through similar but non-identical circuitry.

Changes in brain reward and control systems of frontal cortical areas including the orbitofrontal cortex (OFC) are associated with alcohol use disorders (AUD). The OFC is extensively innervated by monoamines, and drugs that target monoamine receptors have been used to treat a number of neuropsychiatric diseases, including AUDs. Recent findings from this laboratory demonstrate that D2, α2-adrenergic and 5HT1A receptors all decrease the intrinsic excitability of lateral OFC (lOFC) neurons in naïve male mice and that this effect is lost in mice exposed to repeated cycles of chronic intermittent ethanol (CIE) vapor. As biological sex differences may influence an individual\'s response to alcohol and contribute to the propensity to engage in addictive behaviors, we examined whether monoamines have similar effects on lOFC neurons in control and CIE exposed female mice. Dopamine, norepinephrine and serotonin all decreased spiking of lOFC neurons in naïve females via activation of Giα-coupled D2, α2-adrenergic and 5HT1A receptors, respectively. Firing was also inhibited by the direct GIRK channel activator ML297, while blocking these channels with barium eliminated the inhibitory actions of monoamines. Following CIE treatment, evoked spiking of lOFC neurons from female mice was significantly enhanced and monoamines and ML297 no longer inhibited firing. Unlike in male mice, the enhanced firing of neurons from CIE exposed female mice was not associated with changes in the after-hyperpolarization and the small-conductance potassium channel blocker apamin had no effect on current-evoked tail currents from either control or CIE exposed female mice. These results suggest that while CIE exposure alters monoamine regulation of OFC neuron firing similarly in males and female mice, there are sex-dependent differences in processes that regulate the intrinsic excitability of these neurons.

Although economic theories suggest that punishment threat is crucial for maintaining social norms, counterexamples are noted in which punishment threat hinders norm compliance. Such discrepancy may arise from the intention behind the threat: unintentionally introduced punishment threat facilitates, whereas intentionally introduced punishment threat hinders, norm compliance. Here, we combined a dictator game and fMRI to investigate how intention modulates the effect of punishment threat on norm compliance and the neural substrates of this modulation. We also investigated whether this modulation can be influenced by brain stimulation. Human participants divided an amount of money between themselves and a partner. The partner (intentionally) or a computer program (unintentionally) decided to retain or waive the right to punish the participant upon selfish distribution. Compared with the unintentional condition, participants allocated more when the partner intentionally waived the power of punishment, but less when the partner retained such power. The right lateral orbitofrontal cortex (rLOFC) showed higher activation when the partner waived compared with when the computer waived or when the partner retained the power. The functional connectivity between the rLOFC and the brain network associated with intention/mentalizing processing was predictive of the allocation difference induced by intention. Moreover, inhibition or activation of the rLOFC by brain stimulation decreased or increased, respectively, the participants\' reliance on the partner\'s intention during monetary allocation. These findings demonstrate that the perceived intention of punishment threat plays a crucial role in norm compliance and that the LOFC is casually involved in the implementation of intention-based cooperative decisions.
Does punishment threat facilitate or hinder norm enforcement? So far, cognitive neuroscience research offers equivocal evidence. By directly manipulating the intention behind punishment threat, we demonstrate that intention modulates the effectiveness of punishment threat. Moreover, we show that inhibition or activation of the right lateral orbitofrontal cortex (rLOFC) decreased or increased the effect of punishment threat in the intentional context, but not in the unintentional context, suggesting the casual involvement of the rLOFC in intention-based cooperative decisions.

In social anxiety disorder (SAD), anxiety reactions are triggered by attentional bias to social threats that automatically appear in social situations. The present study aimed to investigate the neural basis and underlying resting-state pathology of attentional bias toward internal and external social threats as a core element of SAD. Twenty-two patients with SAD and 20 control subjects scanned functional magnetic resonance imaging during resting-state and while performing the visual search task. During the task, participants were exposed to internal threat (hearing participants\' own pulse-sounds) and external threat (crowds in facial matrices). Patients showed activations in the lateral orbitofrontal cortex, rostral anterior cingulate cortex and insula in response to internal threat and activations in the posterior cingulate cortex and middle temporal gyrus in response to external threat. In patients, neural activity related to combined internal and external threats in the posterior cingulate cortex was inversely correlated with the functional connectivity strengths with the default mode network during resting-state. These findings suggest that attentional bias may stem from limbic and paralimbic pathology, and the interactive process of internally- and externally-focused attentional bias in SAD is associated with the self-referential function of resting-state.