Linking Affect to Action: Critical Contributions of the Orbitofrontal Cortex

Books

Roberts AC, Robbins TW, Weiskrantz L. 1998. The Prefrontal Cortex: Executive and Cognitive Functions. Oxford University Press, New York.
Amazon

Rolls ET. 2007. Emotion Explained. Oxford University Press, New York.
Amazon

Zald D, Rauch S, eds. 2006. The Orbitofrontal Cortex. Oxford University Press, New York.
Amazon

Journal Articles

Reviews

Baxter MG, Murray EA. 2002. The amygdala and reward. Nat. Rev. Neurosci. 3: 563-573.

Cavada C, Schultz W. 2000. The mysterious orbitofrontal cortex. Cereb. Cortex 10: 205-206. (PDF, 28 KB) FULL TEXT

Chudasama Y, Robbins TW. 2006. Functions of frontostriatal systems in cognition: comparative neuropsychopharmacological studies in rats, monkeys and humans. Biol. Psychol. 73: 19-38.

Dolan RJ. 2007. The human amygdala and orbital prefrontal cortex in behavioural regulation. Philos. Trans. R Soc. Lond. B Biol. Sci. 362: 787-799.

Dolan RJ. 2002. Emotion, cognition, and behavior. Science 298: 1191-1194.

Drevets WC. 2000. Functional anatomical abnormalities in limbic and prefrontal cortical structures in major depression. Prog. Brain Res. 126: 413-431.

Gottfried JA, Zald DH. 2005. On the scent of human olfactory orbitofrontal cortex: meta-analysis and comparison to non-human primates. Brain Res. Brain Res. Rev. 50: 287-304.

Harlow JM. 1848. Passage of an iron rod through the head. Boston Medical and Surgical Journal 39: 389-393. (Republished in J. Neuropsychiatry Clin. Neurosci. 11: 281-283).

Harlow JM. 1868. Recovery after severe injury to the head. Publication of the Massachusetts Medical Society 2: 327-347.

Kringelbach ML, Rolls ET. 2004. The functional neuroanatomy of the human orbitofrontal cortex: evidence from neuroimaging and neuropsychology. Prog. Neurobiol. 72: 341-372.

Kringelbach ML. 2005. The human orbitofrontal cortex: linking reward to hedonic experience. Nat. Rev. Neurosci. 6: 691-702.

Miller EK, Freedman DJ, Wallis JD. 2002. The prefrontal cortex: categories, concepts and cognition. Philos. Trans. R Soc. Lond. B Biol. Sci. 357: 1123-1136. (PDF, 1 MB) FULL TEXT

Price JL. 2005. Free will versus survival: brain systems that underlie intrinsic constraints on behavior. J. Comp. Neurol. 493: 132-139.

Rolls ET. 2006. Brain mechanisms underlying flavour and appetite. Philos. Trans. R Soc. Lond. B Biol. Sci. 361: 1123-1136.

Schoenbaum G, Roesch MR, Stalnaker TA. 2006. Orbitofrontal cortex, decision-making and drug addiction. Trends Neurosci. 29: 116-124.

Schoenbaum G, Roesch M. 2005. Orbitofrontal cortex, associative learning, and expectancies. Neuron 47: 633-636.

Schoenbaum G, Setlow B, Ramus SJ. 2003. A systems approach to orbitofrontal cortex function: recordings in rat orbitofrontal cortex reveal interactions with different learning systems. Behav. Brain Res. 146: 19-29.

Seymour B, Singer T, Dolan RJ. 2007. The neurobiology of punishment. Nat. Rev. Neurosci. 8: 300-311.

Small DM, Prescott J. 2005. Odor/taste integration and the perception of flavor. Exp. Brain Res. 166: 345-357.

Vanderschuren LJ, Everitt BJ. 2005. Behavioral and neural mechanisms of compulsive drug seeking. Eur. J. Pharmacol. 526: 77-88.

Wallis JD. 2007. Orbitofrontal cortex and its contribution to decision-making. Annu. Rev. Neurosci. Apr 6 [Epub ahead of print].

Research Papers

Barbas H, Saha S, Rempel-Clower N, Ghashghaei T. 2003. Serial pathways from primate prefrontal cortex to autonomic areas may influence emotional expression. BMC Neurosci. 4: 25. FULL TEXT

Balleine BW, Dickinson A. 1998. Goal-directed instrumental action: contingency and incentive learning and their cortical substrates. Neuropharmacology 37: 407-419.

Blair RJ, Cipolotti L. 2000. Impaired social response reversal. A case of 'acquired sociopathy'. Brain 123: 1122-1141. FULL TEXT

Braesicke K, Parkinson JA, Reekie Y, et al. 2005. Autonomic arousal in an appetitive context in primates: a behavioural and neural analysis. Eur. J. Neurosci. 21: 1733-1740.

Chudasama Y, Murray E. 2007. Rhesus monkeys with orbital prefrontal cortex lesions can learn to inhibit prepotent responses in the reversed reward contingency task. Cereb. Cortex 17: 1154-1159.

De Martino B, Kumaran D, Seymour B, Dolan RJ. 2006. Frames, biases, and rational decision-making in the human brain. Science 313: 684-687.

Feierstein CE, Quirk MC, Uchida N, et al. 2006. Representation of spatial goals in rat orbitofrontal cortex. Neuron 51: 495-507.

Fellows LK, Farah MJ. 2005. Different underlying impairments in decision-making following ventromedial and dorsolateral frontal lobe damage in humans. Cereb. Cortex 15: 58-63. FULL TEXT

Ferry AT, Ongur D, An X, Price JL. 2000. Prefrontal cortical projections to the striatum in macaque monkeys: evidence for an organization related to prefrontal networks. J. Comp. Neurol. 425: 447-470.

Frey S, Kostopoulos P, Petrides M. 2004. Orbitofrontal contribution to auditory encoding. Neuroimage 22: 1384-1389.

Li W, Luxenberg E, Parrish T, Gottfried JA. 2006. Learning to smell the roses: experience-dependent neural plasticity in human piriform and orbitofrontal cortices. Neuron 52: 1097-1108. FULL TEXT

Hasler G, van der Veen JW, Tumonis T, et al. 2007. Reduced prefrontal glutamate/glutamine and gamma-aminobutyric acid levels in major depression determined using proton magnetic resonance spectroscopy. Arch. Gen. Psychiatry 64: 193-200.

Izquierdo A, Suda RK, Murray EA. 2005. Comparison of the effects of bilateral orbital prefrontal cortex lesions and amygdala lesions on emotional responses in rhesus monkeys. J. Neurosci. 25: 8534-8542. FULL TEXT

Izquierdo A, Murray EA. 2004. Combined unilateral lesions of the amygdala and orbital prefrontal cortex impair affective processing in rhesus monkeys. J. Neurophysiol. 91: 2023-2039. FULL TEXT

Kalisch R, Korenfeld E, Stephan KE, et al. 2006. Context-dependent human extinction memory is mediated by a ventromedial prefrontal and hippocampal network. J. Neurosci. 26: 9503-9511. FULL TEXT

Kim J, Ragozzino M. 2005. The involvement of the orbitofrontal cortex in learning under changing task contingencies. Neurobiol. Learn. Mem. 83: 125-133.

Koenigs M, Young L, Adolphs R, et al. 2007. Damage to the prefrontal cortex increases utilitarian moral judgements. Nature 446: 908-911.

Lough S, Kipps CM, Treise C, et al. 2006. Social reasoning, emotion, and empathy in frontotemporal dementia. Neuropsychologia 44: 950-958.

McAlonan K, Brown VJ. 2003. Orbital prefrontal cortex mediates reversal learning and not attentional set shifting in the rat. Behav. Brain Res. 146: 97-103.

McDannald MA, Saddoris MP, Gallagher M, Holland PC. 2005. Lesions of orbitofrontal cortex impair rats' differential outcome expectancy learning but not conditioned stimulus-potentiated feeding. J. Neurosci. 25: 4626-4632. FULL TEXT

Mesulam MM, Mufson EJ. 1982. Insula of the old world monkey. I. Architectonics in the insulo-orbito-temporal component of the paralimbic brain. J. Comp. Neurol. 212: 1-22.

Morecraft RJ, Geula C, Mesulam MM. 1992. Cytoarchitecture and neural afferents of orbitofrontal cortex in the brain of the monkey. J. Comp. Neurol. 323: 341-358.

Morris JS, Dolan RJ. 2004. Dissociable amygdala and orbitofrontal responses during reversal fear conditioning. Neuroimage 22: 372-380.

Ostlund SB, Balleine B. 2005. Lesions of medial prefrontal cortex disrupt the acquisition but not the expression of goal-directed learning. J. Neurosci. 25: 7763-7770. FULL TEXT

Palencia CA, Ragozzino M. 2006. The effect of N-methyl-D-aspartate receptor blockade on acetylcholine efflux in the dorsomedial striatum during response reversal learning. Neuroscience 143: 671-678.

Paton JJ, Belova MA, Morrison SE, Salzman CD. 2006. The primate amygdala represents the positive and negative value of visual stimuli during learning. Nature 439: 865-870.

Padoa-Schioppa C, Assad JA. 2006. Neurons in the orbitofrontal cortex encode economic value. Nature 441: 223-226.

Pickens CL, Saddoris MP, Gallagher M, Holland PC. 2005. Orbitofrontal lesions impair use of cue-outcome associations in a devaluation task. Behav. Neurosci. 119: 317-322. FULL TEXT

Rauch SL, Shin LM, Segal E, et al. 2003. Selectively reduced regional cortical volumes in post-traumatic stress disorder. Neuroreport 14: 913-916.

Roberts AC, Tomic DL, Parkinson CH, et al. 2007. Forebrain connectivity of the prefrontal cortex in the marmoset monkey (Callithrix jacchus): an anterograde and retrograde tract-tracing study. J. Comp. Neurol. 502: 86-112.

Rudebeck PH, Walton ME, Smyth AN, et al. 2006. Separate neural pathways process different decision costs. Nat. Neurosci. 9: 1161-1168.

Rushworth MF, Buckley MJ, Gough PM, et al. 2005. Attentional selection and action selection in the ventral and orbital prefrontal cortex. J. Neurosci. 25: 11628-11636. FULL TEXT

Schoenbaum G, Setlow B, Nugent SL, et al. 2003. Lesions of orbitofrontal cortex and basolateral amygdala complex disrupt acquisition of odor-guided discriminations and reversals. Learn. Mem. 10: 129-140. FULL TEXT

Simmons JM, Richmond BJ. 2007. Dynamic changes in representations of preceding and upcoming reward in monkey orbitofrontal cortex. Cereb. Cortex

Small DM, Gerber JC, Mak YE, Hummel T. 2005. Differential neural responses evoked by orthonasal versus retronasal odorant perception in humans. Neuron 47: 593-605.

Stalnaker TA, Franz TM, Singh T, Schoenbaum G. 2007. Basolateral amygdala lesions abolish orbitofrontal-dependent reversal impairments. Neuron 54: 51-58.

Stalnaker TA, Roesch MR, Franz TM, et al. 2006. Abnormal associative encoding in orbitofrontal neurons in cocaine-experienced rats during decision-making. Eur. J. Neurosci. 24: 2643-2653.

Tremblay L, Schultz W. 1999. Relative reward preference in primate orbitofrontal cortex. Nature 398: 704-708.

Valentin VV, Dickinson A, O'Doherty JP. 2007. Determining the neural substrates of goal-directed learning in the human brain. J. Neurosci. 27: 4019-4026.

Volkow ND, Wang GJ, Ma Y, et al. 2005. Activation of orbital and medial prefrontal cortex by methylphenidate in cocaine-addicted subjects but not in controls: relevance to addiction. J. Neurosci. 25: 3932-3939. FULL TEXT

Winstanley CA, Theobald DE, Cardinal RN, Robbins TW. 2004. Contrasting roles of basolateral amygdala and orbitofrontal cortex in impulsive choice. J. Neurosci. 24: 4718-4722. FULL TEXT

Wallis, JD, Miller EK. Neuronal activity in primate dorsolateral and orbital prefrontal cortex during performance of a reward preference task. Eur. J. Neurosci. 18: 2069-2081.

Sponsorship

This conference and eBriefing were made possible with support from:

The National Institute on Drug Abuse
The National Institute on Aging
The National Institute of Mental Health
The National Institute of Neurological Diseases and Stroke
Autism Speaks
Bowdoin College
Bristol-Myers Squibb
Coulbourn Instruments
Plexon, and

Keynote speaker Ray Dolan (University College London) remarked that, because basic skills can be spared when the OFC is removed, the OFC has been mistaken to be nonfunctional or "silent cortex." He described, with clear horror, how some twentieth-century neurologists perceived damage to this area as irrelevant and psychiatrists prescribed and performed lobotomy (also called leucotomy) of this area as a treatment for a variety of mental illnesses. He contrasted this with more recent research that shows the OFC to be performing some of the most subtle, abstract, and advanced computations of the brain.

The OFC is likely active when we realize that "half-empty" and "half-full" refer to the same volume of water.

One computation that appears to be performed in the circuits of the OFC is the judgment of value. It is known that humans will interpret information about an object or an event in a biased manner depending on how the information is framed. This "framing effect" is clear in the presentation of the same glass as "half-empty" or "half full." Dolan described functional MRI experiments in which his team found that activity in the amygdala was sensitive to the framing bias of information. Interestingly, high levels of activity in the OFC and other frontal areas predicted a reduced susceptibility to the framing bias. That means that our own OFC is likely active when we realize that the descriptions "half-empty" and "half-full" refer to the same volume of water.

Data from animal and human experiments like this one support the idea that the OFC makes or at least participates with other circuits in calculations of value. It also suggests the OFC may be particularly important when these calculations are entangled with emotional or other kinds of information.

Eavesdropping on neurons

Just how are these advanced calculations of value made in the brain? Insertion of metal or glass electrodes into the brain allows us to eavesdrop on neural activity. Given current understanding, it is fairly ridiculous to consider any part of the brain completely inactive. In all parts of the brain neurons are constantly firing "action potentials" that transmit signals via axons to other connected neurons. If a neuron receives enough incoming signals to raise the voltage across its membrane past a threshold, then it will fire an action potential in response, passing along information down a chain of connected neurons. Measurement of these signals, seen as fast spikes in electrical potentials, can inform us about the computation that is being performed by different brain areas.

OFC neurons seem particularly sensitive to events that in the past predicted reward.

It can be easily shown that neurons in auditory cortex fire in response to auditory stimuli, and visual cortex is active when the eyes are presented with visual stimuli. Analysis of the activity in the OFC is more complicated. Cells have been shown to fire in complex patterns that relate to what an animal is doing and what it is expecting based on past experience. OFC neurons seem particularly sensitive to events that in the past predicted delivery of a reward such as food or water.

Early recordings have shown that neurons in the OFC can fire in response to delivery of rewards and also, with experience, to cues that predict reward. Thus, as animals learn to recognize cues that predict reward, elevated firing responses happen increasingly early until they occur with greater strength for cues and events that precede the reward. The activity in OFC appears to code the current value of the expected or impending events. Studies reported at this conference further investigated the sophistication of this reward-related signaling in subjects completing dynamic and complex tasks. Several speakers reported that neural activity in the OFC reflects multiple aspects of expectation and future rewards and even internal states and preferences.

Anticipation of reward

Activity in OFC neurons during a waiting period of three different sessions of an operant task in which monkeys were cued to complete one (top), two (middle), or three trials (bottom) to achieve a reward. The firing rate of OFC neurons increased only in the final episode of each trial when animals had learned to expect a reward (blue lines). In trials where the monkey did not expect a reward (pink lines), but the behavior was identical, firing did not increase.

In recordings from region 13 in the primate OFC and ventral striatum, Janine Simmons (U.S. National Institute of Mental Health) has found neurons that fire in anticipation of reward when learned cues predicted that a trial will be rewarded. It is possible to see distinct differences in the traces from a single neuron after a cue predicts an upcoming rewarded trial as opposed to a non-rewarded trial. When the animal is performing the identical task but the trial will not be rewarded, or may be randomly rewarded, activity in OFC and ventral striatum is low after the cue and before the reward is delivered, signaling no expectancy. However, it may still be elevated after reward is delivered, encoding reward history.

The OFC has a greater proportion of neurons that record reward expectancy, while the ventral striatum responses are largely dominated by reward history. Interestingly, though they are reciprocally connected, upregulation of the OFC and ventral striatal activity does not occur in concert. Reward-related firing in reward-selective OFC neurons occurred earlier in the trial, closer to the cue that triggered the onset of the trial and during the first behavioral requirement (holding a bar), while neurons in the ventral striatum fired later, closer to the time of the "go" cue that prompted an action (letting go of a bar) that was subsequently rewarded.

Magnitude, probability, and cost of the reward

Jonathan Wallis (University of California, Berkeley) investigated differences in reward coding in the dorsolateral prefrontal cortex (DLPFC) and the OFC of macaques when he varied the magnitude of the reward given (2, 4, or 8 drops of juice) and indicated the upcoming magnitude with a visual cue. He found that activity in the OFC was more closely related to the predicted reward value (drop number) than the direction of eye movement (used to indicate choice of reward). Activity in the DLPFC appeared to reflect a combination of both the direction of eye movement (saccade direction) and predicted reward value. Reward-related activity also appeared earlier in the OFC than the DLPFC suggesting that there is transfer from one area to the other, perhaps following the development of a reward-guided motor plan.

Wallis noted that in the real world, decision-making has to take into account more complex aspects of value than just reward magnitude. These variables are likely to include 1) the probability that the reward will be delivered, 2) the cost of the action that is necessary to obtain the reward, and 3) the eventual benefit of the reward itself.

Wallis designed a second task to dissect these three variables by presenting the macaques with choices from a set of thirty images that predicted gradations of reward probability, reward cost, and size. His group found that approximately 30% of neurons in the lateral PFC, OFC, and medial PFC altered their firing in response to changes in one of these three variables. Neurons responding with properties that represented a compound of these variables were found in greater proportions in the medial PFC (48%), than in the OFC (27%) or lateral PFC (16%).

Neurons show differential firing depending on the reward value associated with the image. Some are sensitive to probability of reward delivery, but not other factors, while others are sensitive to payoff or effort. The modulation of the firing rate of the neuron illustrated here shows that it is most sensitive to changes in probability.

Wallis's experiments suggest that "value" assessment, which incorporates multiple unrelated aspects of reward, may be encoded in the frontal cortex including the OFC. This calculation may be important for efficiently managing decision-making and action selection. Wallis said that complex comparisons between multiple real-life choices cannot be performed quickly by conducting point-by-point analysis of each object. Imagine comparing each of the countless individual pros and cons of the decision to grocery shop, work, or nap based on weather, time, hunger, fatigue, dependents, and relative poverty. Wallis suggested that perhaps these decisions are achieved by creating an extracted value measure for each measure in parallel, summing these, and then selecting the highest rated object, action, or goal as the winner. This process of parallel abstract value assignment, which represents probability, cost, and benefit, may be useful for efficiently making rapid complex comparisons between unrelated objects or actions.

Relative value of rewards

Camillo Padoa-Schioppa (Harvard Medical School) designed a task in which monkeys were asked to make a choice between different quantities of juice. In a typical trial of this task the monkey can choose a small amount of grape juice or a larger amount of green tea. Since monkeys usually like grape juice better, they are more likely to select it, but when the relative amount of tea is much higher and they are very thirsty, then they may choose tea over grape juice.

The activity of OFC neurons may provide a common unit of value for different goods.

Using this economic choice task while recording the firing of neurons, Padoa-Schioppa can follow the behavior of a single neuron while the monkey makes these calculations. Recording from neurons that have been shown to encode value in area 13 of the OFC, he found that some neurons fire for only one of the offered liquids. The neurons' spike rate reflects the relative amount of juice and therefore the relative value of that option. There are also neurons that seem to fire for the taste of one beverage or the other; he calls these taste neurons, as they do not modulate their spiking rate with reference to the amount of liquid to be delivered.

Most interestingly, there is a subset of neurons that fire for both choices. In these cases the spike rate appears to be determined by the relative amount of both liquids offered and the relative preference of the monkey for the two items. The choice behavior of the monkeys and the firing of this subset of neurons reflect each other, as the neurons show a dip in firing at the time when the monkeys reverse their choice from a small quantity of the preferred liquid to the larger quantity of the less preferred liquid. These neurons may also fire for a third proffered choice, and in this case they preserve the transitivity of the three choices.

The activity of neurons may provide a common unit of measure for different goods, possibly underlying the efficient valuation for comparison described by Wallis.

The most frontal and ventral portions of cortex are easy to point to as the general location of the orbitofrontal cortex both in rodents and primates. However, within this general region, parallels between species are harder to draw, and functional subareas are only just beginning to be understood. One mission of the conference was to establish the location of the OFC in rodents and primates and identify homologs between species.

The OFC appears to be divided into sub-networks, whose boundaries are still being established.

After the meeting it was clear that in both rodents and primates the OFC can be found in the stretch between the insula and the frontal pole along the ventral surface of the brain, but it was not clear where the lateral boundaries definitively lie on both the medial wall and lateral extent of each hemisphere. The location of subdivisions of the OFC, particularly the medial OFC still clearly need to be established between species.

Separate medial and orbital networks with caudal–rostral gradients?

Working with tracers and histological markers in the primate brain, Joe Price (Washington University, St. Louis) has fleshed out A. E. Walker's 1940 circuit diagrams of the frontal cortex. He discriminates between two major networks within the OFC: a medial wall network centered around area 32, and a ventral surface "orbital network" centered around area 13. He argues that the "medial network" is largely viscero-motor and the orbital network is dominated by sensory information, especially chemosensory taste and smell.

Orbital and medial metworks in the OMPFC.

Not only do these areas appear to receive different inputs, but they communicate with separate (and adjacent) striato-pallidal networks. The medial network also appears to have greater connections to the hypothalamus and the amygdala. The amygdala also projects to agranular regions (regions without a granular cell layer 4) of the orbital network, but not to core areas of the orbital network, like areas 13m, 13l and also area 11 and 12. Price finds this is also true in rats, where the lateral and medial, but not the central sections of the OFC, receive basolateral amygdala projections.

While the medial and lateral axis is important in Price's work, Michael Petrides (McGill University) suggested that the rostral and caudal portions of the OFC may differ in their functions. Using PET imaging of human subjects, Petrides reports that the processing of visual, tactile, and auditory information can activate the OFC in Price's orbital network. Interestingly, when the stimuli are more emotional or novel, like an image covered with graffiti, a sound of a car crashing, or a gooey tactile stimulus, then the activation of the OFC shifts to adjacent but more caudal regions of the OFC.

This caudal shift for more novel or emotional stimuli can perhaps be explained by the anatomical work of Marcel Mesulam (Northwestern University). Mesulam reported that the more caudal regions of the OFC receive a greater number of inputs from the limbic system, which are likely to be more engaged by these novel or emotional (possibly dangerous) stimuli.

Mesulam also pointed out that in the primate, the caudal OFC regions, including the insula, are agranular, lacking a clear layer 4. The orbital cortex becomes increasingly granular as you move in the rostral direction in a gradient corresponding to the decline in limbic inputs. Mesulam postulated that this gradient may be a living timeline illustrating the evolution of six-layer granular cortex from more primitive allocortical structures.

A brake on the amygdala and other limbic structures?

The OFC not only receives input from the limbic system, but also projects back down to these structures. Mesulam has suggested that the OFC evolved to liberate the brain from limbic and "hypothalamic tyranny." It is possible that some of these descending projections suppress activity in these structures.

Nancy Rempel-Clower (Grinnell College) has documented the connections of the rat OFC with subregions of the amygdala, hypothalamus, and temporal lobe hippocampal system. She concluded that the anterior insula of the rat influences both sensory input and autonomic output. The lateral orbital region (LO) of the rat OFC has more sensory-related influence, and the ventral (VO) and ventral lateral (VLO) regions have more influence on autonomic output. Of particular interest is a projection from the OFC (area LO in rats) to the intercalated nuclei of the amygdala, which serves as an inhibitory brake on the rest of amygdala.] Projections to other regions of the amygdala may both activate neurons in this area or help shut them down. These projections may be particularly important for regulating negative affect or anxiety and are therefore of considerable interest to human psychiatry.

Separate circuits for pleasant and unpleasant tastes? for eating and foraging?

Using fMRI in human subjects, Dana Small (Yale University) has isolated different areas of the OFC that respond when subjects are given pleasant and unpleasant tastes. Pleasant tastes activate more of the right hemisphere OFC, while unpleasant tastes are left lateralized. She suggests that these different signals may be segregated by evolution because pleasant smells or tastes should drive appetitive approach behavior, while unpleasant smells or tastes associated with spoiled food or other dangers should drive avoidance behaviors. The OFC also shows greater activation in situations when subjects are asked to judge the pleasantness of a taste, as opposed to when they are asked merely to perform more passive tasks, such as acknowledging one has tasted something or identifying a taste.

The area of the brain that evolved to process value and reward also processes aspects of taste and smell.

Using tubes to deliver odors to different regions of the respiratory tract, Small found that the processing of odors was different depending on whether the odor was delivered orthonasally (via the nostrils) or retronasally (via the back of the throat). Small suggested that the perception of odors originating in these different locations may be processed differently, as the orthonasal system may be specialized for foraging for food out in the world, while the retronasal system may be more closely associated with the consumption of food in the mouth and the processing of flavor. Others have reported that the OFC may specially process flavor, such that individual odors or tastes may minimally activate the OFC, but when they are combined, presumably as a fusion of flavor, then superadditive OFC activity is observed.

Small remarked that it is probably not a coincidence that the area of the brain that evolved to process value and reward is the same as that processing higher order aspects of taste and smell. All value can perhaps be traced back to food.

Functional differences in subregions of the frontal cortex

The frontal cortex is often associated with the greatest feats of cognition and mental agility, but the distinct differences between subregions of frontal cortex are still not well understood. Speakers at the conference addressed the differences between the OFC and more dorsal and medial parts of the frontal cortex. A key question here is whether these subregions can be cleanly dissociated from one another.

Conference co-organizer Jay Gottfried (Northwestern University) believes that there is good reason for making a distinction between the OFC and medial prefrontal cortex, pointing out that "unlike the rest of PFC, the OFC is properly paralimbic cortex—anatomically, physiologically, functionally—which means it is intimately connected to emotional processing in all of its guises."

The different anatomical connections seen in comparing the two regions can thus be interpreted as indicators of different roles, suggesting that the OFC interprets and refines the more visceral reflexive behaviors and drives and the mPFC organizes more abstract and less hedonic cognition. In layman's terms, Gottfried explains, "I would say that in comparison to other areas of prefrontal cortex, the OFC is the emotional older sibling."

Behavioral tests can be used to show that different parts of the frontal cortex perform different functions. Michael Ragozzino (University of Illinois at Chicago) has shown that the mPFC (prelimbic region by some nomenclature) and OFC inactivation in rats have different effects during discrimination tasks. As mentioned above, OFC inactivation impairs reversal of cue-reward associations when the rewarded odor is switched (e.g. changing response behavior after a reward that is regularly hidden in a cup of shavings treated with cinnamon is then switched to cups smelling of nutmeg). While, mPFC inactivation (along the medial wall of the brain just adjacent to the OFC) does not disrupt learning this reversal of odor cues. However, mPFC does affect learning if the task is slightly different. If the animals are asked to shift their strategy; that is, to stop using odor cues to and start using place cues (for example, the cup to the left of the cage rather than to the right), then mPFC inactivation will render them impaired. A similar dissociation between the role of OFC in reversal learning vs. "set-shifting" or changes in strategy have been shown previously by others, including Angela Roberts (University of Cambridge) in marmosets and Verity Brown (University of St. Andrews) in rats.

Complete dissociation of the function of medial and orbitofrontal areas may be impossible, however, because the OFC and mPFC likely operate in conjunction with each other. "Linking affect to action," the title of the conference, may also critically require the mPFC. As conference speaker Jonathan Wallis (University of California, Berkeley) pointed out, "The OFC is a region of the frontal lobe that has minimal connections with the motor system." He suggested that while the OFC is involved in decision-making, the medial PFC may be required for motor planning. "In other words, the OFC evaluates the final outcome, while medial PFC evaluates the means to achieve that outcome."

Future anatomical projects

Anatomical projects are still clearly useful and necessary for further investigation of the orbitofrontal cortex. At the meeting, Marcel Mesulam suggested that divisions between anatomical areas should not be marked by single lines but should instead be indicated by gray zones covering the probability of error in the divisional boundary. Gene expression atlases and immunohistological analysis, particularly for modulatory transmitters, may also assist in delineating OFC subareas and in assigning functional correspondence across species. Further questions were raised about the possibility of lateralization and sex differences in the human OFC. It seems likely that many such differences exist, but so far, the answers to these questions are largely unknown.

Panel discussions after each session showcased questions from students and allowed speakers and attendees to ask questions of each other. Here are some of the questions that arose:

Does the Iowa Gambling Task measure impulsivity or impairment in reversal learning?

Can we distinguish between reward and motivation?

What is the difference between choice and learning? What is the difference between emotion and affect?

Do animals need to learn about primary unconditioned stimuli like food and water or are they instinctual?

Is there auditory input to the OFC? Where does it originate?

Where do odors become conscious? What part of olfactory pathways corresponds to retina, thalamus, and primary visual cortex in the visual pathway?

What mechanisms enhance odor discrimination after minutes of exposure?

How do we subdivide the OFC in the medial lateral and anterior posterior axis? Are boundaries between sub-areas sharp or gradual?

How do we draw parallels between different subregions of the OFC in different species?

Is there a difference between active punishment and the absence of reward?

How does the ability to make decisions change with age? Do older people process reward and affect differently?

Is OFC function very different between the two hemispheres in humans? What does each hemisphere specialize in?

Are there sex differences in the OFC?

How can we delineate outcome encoding and outcome updating?

Why do so few cells respond to gustatory inputs in gustatory cortex and OFC?

Is perseveration in reversal tasks due to learned-non-reward? What happens when reversal tasks involve multiple and novel options?

Do physiologists studying neural activity in over-trained animals investigate the same phenomena as researchers following task acquisition as it develops?

What is the role of the OFC in Pavlovian and instrumental conditioning?

How does the OFC differ functionally from the prelimbic, cingulated, insular, and perirhinal cortex? What is the most common direction of flow of information through these circuits?

How does OFC circuitry integrate with motor circuits?

Is obsessive-compulsive disorder an anxiety disorder? How does it differ in the frontal lobe networks implicated, when compared to other anxiety disorders?

How do tasks using abstract cues (like lights and tones) and tasks directly using food reward themselves differ in how they are processed in the brain?

How does neural activity change in the seconds to minutes after a reward has occurred? How are rewards associated retroactively with action and choice?

What is the role of the OFC in drug craving and addictive relapse?

Is a delay a form of punishment? How do they differ? How are they encoded?

Does a damaged OFC lead to diminished ability to consider future consequences (future blindness)?

Why do we see location related activity in the rodent but not primate OFC?

Does OFC activity in the rat that appears to code for location relay information about allocentric locations or egocentric locations?

Are distinct regions of the OFC involved in instrumental and Pavlovian learning? How does this vary by species?

What is the role of instrumental conditioning vs. faulty Pavlovian affective conditioning in addiction?

Is the medial orbitofrontal cortex encoding for the abstract value of the items that are acquired through actions, or of the motor plans associated with getting the items?

What is the role of the amygdala in encoding the value of positive items?

What computational mechanisms are used to compute contingency, or to implement goal-directed learning more generally?

What is the homology of the human mOFC in the rat brain?

What frontal and basal ganglia regions are involved in implementing habit learning in humans?

The most famous patient to suffer orbitofrontal cortex (OFC) damage was Phineas Gage, a railroad worker who survived an 1848 explosion that drove a tamping iron though his frontal lobes. Accounts from the time document a surprising recovery. After the accident, Gage was still functional but, once a polite and quiet person, he became crude and uninhibited, showing poor judgment in social situations. J. M. Harlow, a country doctor who followed Gage's case, remarked that Gage was "impatient of restraint or advice when it conflicts with his desires, at times pertinaciously obstinate, yet capricious and vacillating, devising many plans of future operations, which are no sooner arranged than they are abandoned in turn for others appearing more feasible."

The OFC may be important for inhibition of reflexive urges.

Descriptions of Gage's change in behavior have led to speculation that the OFC is important for inhibition of reflexive urges. Studies in animal models have confirmed this, but they also suggest that more subtle cognitive processes may underlie the changes Harlow observed.

Neuropsychologists working with patients more recently have found that damage to different areas of the frontal lobe lead to subtly different impairment on cognitive tests. Damage to the dorsal frontal lobe leads to impairment in subjects taking the Wisconsin Card Sorting Test, which measures the ability to shift strategies in rule use. People who have sustained more ventral orbitofrontal cortex damage are also impaired on a neuropsychological test called the Iowa Gambling Task (IGT), a test traditionally thought to measure impulsive or risky behavior.

Animal research has suggested that damage to the orbitofrontal cortex in animals does not disrupt initial training on many simple discrimination tasks. At first inspection, animals with OFC lesions can seem quite normal: they can learn to discriminate cues, search in a specific area, or press a lever for a reward. For example, several laboratories, including that of Michael Ragozzino (University of Illinois at Chicago) have shown that rats with OFC lesions and intact controls will learn to dig in a cup full of shavings sprinkled with a scent, such as cinnamon, if they have previously found food pellets in these cups. They will also choose to dig in cinnamon-scented cups over nutmeg-scented cups if they have found that there is no reward pellet hidden in nutmeg scented cups. However, when the animals are asked to reverse their choices (when nutmeg-scented cups now contain pellets and cinnamon-scented cups are no longer baited) controls eventually switch their choice behavior. OFC-lesioned rats will continue to choose the cinnamon-scented cup.

This failure to alter choice behavior is called perseveration, or failure to perform reversal learning. OFC damage across species has been shown to cause little deficit in simple task acquisition, but frequently causes a deficit in tasks where there is an element of reversal.

It is unclear whether this failure is due to impulsivity: the lack of inhibition of a learned response (i.e., go to X for reward). It may also reflect an insensitivity to punishment, which would lead to lack of learning about the absence of the reward in the now empty cup. Or it could be due to an inability to overwrite reward values for cues in the brain, so that once cinnamon is a stimulus with a positive valence it cannot be regarded as negative, and vice versa for nutmeg. In this scenario the animal may be sensitive to punishment but simply unable to update learned value abstractions.

The role of the OFC in behavioral inhibition

Although Phineas Gage seemed to behave impulsively after his injury, damage to the OFC does not completely eradicate control over behavior in all situations. The effect of OFC damage on impulse control seems clear for some tests but not all. Emerging studies suggest that the OFC's role in different behaviors and contexts varies in a way we are only beginning to understand.

The OFC's role in different behaviors and contexts varies in a way we are only beginning to understand.

Elisabeth Murray (U.S. National Institute of Mental Health) has shown that monkeys with OFC damage, similar to monkeys with amygdala lesions, will nonchalantly reach for food that is located very near to a plastic snake. Because monkeys have an innate fear of snakes, it is highly unusual for a normal monkey to make such a bold move. The animals are not trained by negative or positive reinforcers in this task (the negative connotation of the plastic snake is immediate and independent of any experience with real snakes), suggesting that without an OFC monkeys do not inhibit their approach response to food with reference to danger signals. At the same time, however, the monkeys with OFC lesions show less defensive behavior (e.g., freezing and eye aversion) in the presence of the snake, so the results cannot be accounted for solely in terms of loss of inhibitory control.

Yogita Chudasama (McGill University) has shown that monkeys with OFC lesions, though impaired in a reversal phase of a task in which they have to correctly select an object predicting reward, were perfectly capable of learning to point to small food rewards (a single peanut) in order to receive a larger food reward (many peanuts). Successful performance on this reverse-reward contingency task by animals with an orbitofrontal lesion provides strong evidence that these animals maintain some mechanisms for inhibitory control.

Other observational studies of animals with OFC lesions do show behaviors that look like impulsivity. Catherine Winstanley (University of British Columbia) has shown that rats with OFC lesions performing a task in which they are required to poke their nose into a port underneath an illuminated light will frequently and indiscriminately poke their nose into any port without waiting for the light cue. Interestingly, this seemingly impulsive misjudgment is also made by rats exposed to cocaine.

Suppression of ongoing autonomic arousal: dependent upon orbitofrontal cortex.

Angela Roberts (University of Cambridge) has investigated the role of the OFC in regulation of affect, particularly suppression of emotional responses. Her studies investigate positive affect and reward anticipation as measured by changes in the autonomic nervous system. Both OFC-lesioned animals and controls show enhanced autonomic arousal in anticipation of a favored reward in comparison with anticipation of a less favored reward (delivery of marshmallows vs. normal chow). However, when the cued reward is not delivered Roberts finds that marmosets with OFC lesions take more than two times longer to extinguish their anticipatory autonomic arousal. It is unclear whether this is due to prolonged disappointment or continued expectation of pleasantness. Either way, this suggests parallels with enhanced duration of emotional responses in people with major depressive disorder, a condition associated with dysfunction in the OFC.

Are we using the right tests to measure impulsivity and reversal learning?

The persistent preference for cards from the high-risk deck in the Iowa Gambling Task (IGT) has been considered evidence of impulsive behavior in humans with orbitofrontal damage. Lesley Fellows (McGill University) questioned if this deficit is due to impulse control issues or a reversal learning error. When Fellows closely analyzed the experience of players in the IGT she noted that initially they receive only rewards from the high-risk deck. This leads both normal and impaired players to initially prefer this high-payoff, high-risk deck.

Impulse control may not actually be the main deficit measured by the Iowa Gambling Task.

When high-loss cards start to appear in selections from this deck, normal players will choose it less, while OFC patients will continue to choose cards from it. It may be that orbitofrontal patients are simply unable to reverse the previously rewarded behavior choices established at the onset of the game. When Fellows shuffled the decks so that some high-loss cards appeared early on in the high-risk deck, then OFC patients behaved similar to controls in the IGT. This suggests that impulse control may not actually be the main deficit measured by the IGT, rather it may measure reversal learning.

Rodent experiment designs which test reversal learning have also been questioned. Working with rodents digging in scented and baited cups as described above, Verity Brown's (University of St. Andrews) data suggests that the simple two-choice structure of many laboratory behavior tasks may be conceptually misleading. Like others, Brown's lab has found that rats with OFC lesions are impaired at altering their strategies for digging in scented cups when the rewarded scent is reversed with the previously unrewarded scent (e.g., from cinnamon to nutmeg as described above). However, she finds when they are asked to learn the reversal task with a cup baited with a new scent (cinnamon to oregano) rather than a previously unbaited scent (nutmeg), the animals are able to learn this new contingency. Brown's data suggest that rats that fail to show reversal learning may have previously learned that "no-reward" was associated with the unbaited cup (nutmeg). Brown calls this "learned non-reward." This distinction is subtle, but it suggests that the source of their error is not about updating a representation of the baited cup, but rather the unbaited cup.

Complex functions lead to complex impairments

In her 2005 paper in Cerebral Cortex, Lesley Fellows remarks, "One of the central challenges of understanding the functions of the human prefrontal cortex is that impairment is most evident when the experimental tasks are complex, but task complexity interferes with our ability to distinguish the different component processes that may be implicated." In sum, the experiments described above suggest that aberrant reversal learning may underlie one of the main tests for impulsive behavior (the IGT) and that the inability to overwrite previously learned stimulus-reward associations may underlie reversal learning. However, measures of impulsivity such as approaching food near a snake, nose poking prior to a "go signal" and aberrant regulation of autonomic responses suggest that the OFC may nonetheless play a role in impulsive behavior and its corollary, response inhibition. The role of the OFC in value encoding and updating reward associations is discussed in the next sections.

The value of a food or liquid reward for an animal is, of course, not fixed. Once an animal is sated or receives food poisoning from a certain food, it is more likely to avoid eating that food. If it is thirsty it will seek out water at the cost of other behaviors such as mating, hunting, or sleeping. How are these ever-changing needs and different behavioral programs managed in the brain?

Keynote speaker Marcel Mesulam (Northwestern University) highlighted the role of the frontal cortex in flexible behavior. He and others have proposed that the evolutionary role of cortex, particularly frontal cortex, is to fine-tune and even override more base, instinctual, or reflexive actions.

Lesion studies have shown that flexibility in learned behavior can be lost when the OFC or other frontal areas are removed. Initially the lesioned animals seem unimpaired and can learn simple discrimination and choice tasks, but they encounter trouble when they are forced to vary what they have learned. Others have shown that neurons in the OFC encode the value of environmental stimuli, associating rewards or outcomes with cues or actions that predict those rewards.

Can the data collected by physiologists be connected with behavioral observation made after the OFC is removed? In an attempt to find how the OFC can encode and enable behavioral flexibility, physiologists have set out to record the firing of individual neurons in the OFC while reward contingencies are reversed. They have found that neurons that encode value in the OFC can also show context-dependent plasticity in their responses to specific stimuli. The timing of these changes is fast enough that it could be the source of the change in the brain that drives the change in behavior as it happens.

The time course of learning exhibited by OFC and amygdala neurons is similar to the time course of behavioral learning.

Using single unit (single neuron) recording, Daniel Salzman (Columbia University) reported that neurons in area 13 of the macaque OFC and the amygdala show sensitivity to visual cues that predict either a juice reward or punishment in the form of a mild airpuff to the face. Responses are even sensitive to a gradient in the amount of reward: some neurons fire at the highest rate for the largest rewards, some fire at medium rates for medium-size rewards, and some fire least for punishments. There are also neurons that fire preferentially for punishments. Intriguingly, he also found that neurons in these areas can reverse their response properties to visual stimuli when visual cues indicating reward and punishment are reversed. Neurons in this case can be seen to 'swap' their preference for cues when a new set of cues starts to be rewarded. The reversal of coding in these neurons may be relevant to reversal of behavior, as the changes in their response properties parallel the change in behavior.

Geoffrey Schoenbaum (University of Maryland) also reported finding rat OFC and basolateral amygdala neurons that fire preferentially in response to an odor that is rewarded in a Go-NoGo task by sweet and bitter liquids. Recording in freely moving rats, Schoenbaum found that typically these neurons are initially selective for reward delivery and then acquire responsiveness to the rewarded odor with training. Like Salzman's primate OFC and amygdala neurons, these neurons can alter their firing when the rewarded odor is changed, swapping their preference for the new rewarded odor (or the new unrewarded odor). However, the majority of neurons in the rat OFC (but not the amygdala) fail to reverse cue preference, and this is mirrored in the population response.

A go/no-go task for rats with reversal training. (Stalnaker et al. Neuron, 2007.)

The connection between reversal of firing properties in the OFC and reversal of behavioral responding may not be as straightforward as we might initially think. Schoenbaum found that behavioral reversal in a Go-NoGo task takes longer in sessions in which cue-selective OFC neurons reverse their cue preference than in sessions in which cue-selective OFC neurons simply lose selectivity. He hypothesized that reversal learning deficits after OFC lesions are not due to loss of flexible encoding in OFC but rather to the development of inflexible encoding in basolateral amygdala. Consistent with this proposal, when he lesioned the OFC and recorded neuronal firing in the basolateral amygdala, he found that amygdala neurons now also lost their selectivity after reversal training.

It is not just disruption of the OFC, but also disruption of circuits connected with OFC that impairs reversal learning.

Interestingly, OFC lesion-induced reversal deficits are abolished by secondary lesions of these neurons in basolateral amygdala. The seemingly miraculous recovery of reversal learning after an amygdala lesion underlines Schoenbaum's point that it is not just disruption of the OFC, but also disruption of circuits connected with OFC that impairs reversal learning.

In a related experiment, Tom Stalnaker from the Schoenbaum lab showed that cocaine administration and subsequent withdrawal also produces behavioral reversal learning deficits and leads to loss of reversal of cue selectivity in the rat amygdala. Again, this reversal deficit is abolished by basolateral amygdala lesions. This suggests that cocaine has effects similar to an OFC lesion.

Devaluation as a form of updating

Most of the OFC-associated behavioral experiments described so far focused on reversal of cues and behaviors to measure flexibility in learning. Another measure of the ability to update a reward value is to devalue one reward selectively while keeping the value of another constant. This devaluation can be achieved by creating a negative association with a previously positive reward, for example by pairing a food reward with lithium chloride, a chemical that mimics the illness of food poisoning. Another, less invasive method is to feed animals or subjects with one reward item until they are satiated, and then offer them a choice between both the satiated and non-satiated items.

Unlike the simple discrimination and reversal tasks discussed thus far, devaluation paradigms allow one to isolate whether a learned behavior is mediated by the ability of a) a stimulus (cue), or B) an action to evoke a representation of the value of the outcome. This first is called "Pavlovian" learning or conditioning and the second is called "instrumental" learning or conditioning. This distinction is important because classic theories of learning suggest that different processes and circuits may underlie these associations.

In one test of devaluation, Michela Gallagher's lab (Johns Hopkins University) trained rats to approach a food cup in the presence of a light to obtain a reward. After pairing the food with lithium chloride, they found that animals would eat less and show less "approach behavior" to the food cup when the light cue was illuminated. However, animals with OFC lesions reduced their food consumption when it was mixed with lithium chloride, but they still eagerly approached the food cup in the presence of the light. Notably, this deficit was observed when OFC lesions were made at any time prior to the final test in the presence of the light cue, suggesting that the OFC is critical for updating (or devaluing) the light cue in accordance with the new value of the associated food reward.

The rat mPFC is more active in instrumental learning, while the OFC is implicated in Pavlovian learning.

This dissociation between behavior toward the food and the light suggests that unconditioned stimuli (US's), here food, are processed differently, perhaps by different or additional brain circuits, than cues (conditioned stimuli, or CS's). In a two-choice visual discrimination task conducted with primates, Elisabeth Murray (NIMH) found that devaluation of a food reward by overfeeding has a similar outcome. Monkeys with OFC lesions cannot avoid choosing a cue that predicts delivery of a reward upon which they are already sated (and therefore probably do not want). However, if they are given free choice of foods they will preferentially choose the food item upon which they are not sated, and avoid the food with which they are sated.

Gallagher's and Murray's experiments studied ways in which animals use Pavlovian (cue-outcome) associations to mediate their behavior. Bernard Balleine (UCLA) has recently investigated the role of the OFC in an instrumental devaluation paradigm, which forces the animal to use action-outcome (rather than cue-outcome) associations to mediate behavior.

Balleine offered rats two levers: one let them access food pellets and the other, sucrose solution. In the first round of tests they pressed both levers equally. When the rats had been given access to food pellets one hour prior to testing, they showed a preference for pressing the lever that led to sucrose delivery. This suggested that they associated the lever press with the reward that it would normally deliver and adjusted their actions according to their internal satiety.

Rats with OFC lesions were tested on the same task and were found to behave similarly to controls. When they had eaten a large portion of pellets before testing, they also preferred to press the sucrose lever. Interestingly, animals that had a lesion of the prelimbic frontal cortex, mediodorsal striatum, or the mediodorsal thalamus did not show differential responding after they had been sated on food pellets. Thus, in rats, the prelimbic cortex, mediodorsal striatum, and mediodorsal thalamus appear to be more important for devaluation/updating of action-outcome associations (instrumental), while the OFC is more important for updating stimulus-outcome (Pavlovian) associations.

What happens in humans?

John O'Doherty (Caltech) has used functional MRI to detect activation in the OFC in human subjects that have learned to associate different actions with the probability of drinking different liquids, such as chocolate milk, tomato juice, and a flavorless liquid. When subjects are overfed one liquid and then re-perform the task, he observed a change in activity in the medial OFC that indicates that it is sensitive to this change in value of the associated liquid. In contrast to Balleine's study, which showed no effect of OFC lesions on instrumental devaluation, O'Doherty saw differences in activity in the medial orbitofrontal cortex (mOFC) related to satiety, suggesting that the human medial OFC may encode information about instrumental (action-outcome) associations.

The discrepant results of O'Doherty and Balleine suggest that human mOFC may have a broader function than the rodent OFC, in that it may have greater involvement in instrumental learning, as opposed to just Pavlovian learning. Alternatively, it may be that mOFC is more analogous to medial prefrontal cortex (also called prelimbic cortex) in the rat, where Balleine has found that lesions do affect instrumental devaluation.

Learning new habits

Interestingly, once some instrumental behaviors are over-trained they become insensitive to devaluation and are thought to rely on different neural circuitry. This more robust establishment of a behavior pattern is called habit learning, and is thought to rely more heavily on the lateral parts of the striatum. The role of the frontal cortex in the transition of an instrumental behavior into habit is an area of future investigation.

In this age of modern neuroscience, people laugh at the idea that a homunculus (a little man) sits at a control panel inside a human head, peering out from behind the eyes, observing and directing all of a person's decisions. However, funnily enough, the piece of cortex that sits just behind the orbits of the eyes plays an almost analogous role, performing some of the most advanced functions of the brain.

Different parts of the brain are commonly distinguished by their location and shape, but take on particular functions by virtue of their connective wiring. It is this connective wiring that dictates what information courses through their circuits. Some brain areas are specialists, processing only visual, olfactory, or auditory information. Other areas are generalists or "association areas," coordinating signals and relaying disparate types of information.

One important association area is the orbitofrontal cortex (OFC). The OFC is privy to a wealth of information from sensory, emotional, and memory-related brain regions and thus likely serves as an important center for integration and evaluation. Recent research is revealing that the OFC could be a major site, or is at the very least an essential participant in a network of sites, where sensory and memory-related information is evaluated and transformed into predictions of the future used to guide decisions and actions.

On March 11–14, 2007, the New York Academy of Sciences hosted a conference organized by Geoffrey Schoenbaum (University of Maryland School of Medicine), Jay Gottfried (Northwestern University Feinberg School of Medicine), Elisabeth Murray (National Institute of Mental Health), and Seth Ramus (Bowdoin College) that was dedicated to integrating recent research on the orbitofrontal cortex. The overarching goal of the event was to better define the function of the OFC. The proceedings of this conference were also published in a volume of the Annals of the New York Academy of Sciences. (Click for more information.)

According to Schoenbaum, the event was timely because, although research has explored various functions of the OFC, the general function of the region has been underexplained. "The orbitofrontal cortex has been variously described as an olfactory association cortex by those interested in olfactory processing, a prefrontal working memory system by those interested in prefrontal function, and a system for controlling emotions by those interested in limbic function," he says. "Rarely has orbitofrontal cortex been considered as a critical nexus linking these circuits, with an important general function which it contributes within each of these systems."

Flexibility and the ability to reverse a learned association

Jay Gottfried pointed out during his talk that "there are many animals without an OFC that can nevertheless use smells and other sensory stimuli to hunt, forage, find food, avoid predators, and reproduce." This prompts a question, he said: "What advantages does an OFC confer upon mammals endowed with such a region?" The answer could provide insights into the evolution of the human brain.

In his keynote address, Marcel Mesulam (Northwestern University) argued that "a principal function of the OFC in the course of evolution is to liberate the individual organism from hard-wired conservative behaviors based on phylogenetic memories, and to replace them with a repertoire where individual experience and context can promote choice and flexibility." This theory stems from the idea that, because the information flowing into the OFC comes from limbic system, it is relatively reflexive or hard-wired. When this information is reassembled in the evolutionarily newer frontal cortex, it can be re-evaluated and related to a greater repertoire of actions in a more flexible manner.

"A principal function of OFC ... is to liberate the individual organism from hard-wired conservative behaviors."

One of the simplest ways to assess the function of the OFC is to see what happens to a person or animal when this region of the brain is damaged. Lesions of the OFC in rodents and primates do not impair the learning of simple discrimination tasks (such as red means "wait" and green means "go and get a treat"). However, when the lesioned animals are asked to reverse their behavior (red now means "go get a treat" and green means "wait"), they show impairment.

Impairment on reversal learning is now a hallmark of orbitofrontal dysfunction due to accidental damage, drug exposure, or degeneration. This ability or reverse one's behavior when the rules of a task are changed may stem from an inability to update the associations between rewards and learned cues and stimuli (treats and colored cues). When the OFC is intact, the fluid ability to update these associations—either in OFC or possibly in other brain areas that receive input from OFC—may underpin flexible behavior.

Chemosensory and limbic connections inform the coding of value

In the first line of his book The Vendor of Sweets, R. K. Narayan wrote, "Conquer taste, and you will have conquered the self." A neuroscientist might reword this: conquer the OFC and you will have conquered the self. Surprisingly, it may be that the development of moral values, maybe even the development of meaning, piggybacked on the advanced processing of the reward value of taste and smell in the brain.

One often-noted difference between the OFC and other areas of frontal cortex is that it has a greater input from limbic regions, such as the amygdala, the hippocampus, and areas responsible for taste and smell. The amygdala is thought to be particularly important in the processing of emotional events, and the hippocampus is important for memory. Odor and taste are frequently used to pursue and identify food and water, and even mates, enemies, and offspring.

Connections between the orbitofrontal cortex and other major regions in the brain.

The convergence of chemo-sensory, emotion, and memory related inputs within the OFC enables this area to function as a major center for calculating the value of basic rewards and punishments, and relating them to past and future events and behaviors. In animals performing repetitive experimental tasks, physiological activity in the OFC appears to predict or anticipate future events in the task. This kind of activity develops increasingly after experience with cues that reliably predicted events in the past.

The activity of OFC neurons can also reflect the value of anticipated rewards in a context-dependent manner. This means that neural activity in the OFC may change depending on how much of an anticipated food will likely be available or how much the animal desires that food, considering what else is available or what it has already consumed that day. The neurons in the OFC also seem to discriminate between contexts where the same food is available but it will be either easy or hard to obtain.

These neural calculations in the OFC were probably initially involved in choices regarding foraging decisions and navigating basic friend-or-foe-type social discriminations. It is easy to see, though, how advances in tracking of the complex relationships between needs, actions, and events would confer an evolutionary advantage if they enhanced fluid decision-making responsive to internal states and a changing environment.

Conquer the OFC and you will have conquered the self.

The question is, how far does the abstract calculation of value in the OFC go? Is this brain region the resting place or executor of 'values' in the moral sense? As the brain evolved, neurons in the OFC and other parts of the frontal cortex may have initially answered questions like, if I do X after I see Y will I get food or not? It may then have begun to answer questions like, if I do X after Y will it be good or not? The second keynote speaker of the conference, Ray Dolan of University College London pointed out several studies that indicate that the OFC may be necessary to evaluate our sensations and memories, interpret our emotions, and integrate them in rational and even moral decision-making processes. He suggested that though the calculations made by the OFC may be subtle, they likely contribute to what many consider to be our highest and most human thought processes. A recent study in Nature underscores this idea by showing that in ethical thought experiments involving life or death choices, patients with brain damage involving the OFC and surrounding areas show less emotional and more consistently utilitarian (some might say cold or hyperrational) judgments. (See Koenigs et al, 2007.)

Integrating sensation, emotion, and memory with choice is not just important for solving a laboratory task or a philosophical thought experiment. It is something that we do every minute that determines our values in action. Perhaps we don't need to invoke a homunculus to explain this; the anatomy and physiology of the OFC suggests that integration of sensory, limbic, and frontal cortical systems may be enough.

Conference Highlights

Continue reading this eBriefing for a detailed meeting report, a selection of slides and audio, and links to additional resources. Highlights of the conference include the following:

Effects of OFC Damage on Behavior

• OFC damage is associated with impulsive, risky, perseverative, and generally inflexible behavior.
• With such a complex phenotype, it is unclear if there is a single main deficit underlying the development of other deficits, or if an independent spectrum of functions is disrupted by these lesions.
• Animal studies show mixed effects of focal OFC damage. Some studies show reduction of inappropriate response inhibition while others find evidence of intact inhibitory control. Most studies of OFC lesions find perseverative behavior in two-choice discrimination tasks after task rules are reversed. This is called impaired reversal learning.
• OFC-lesioned rats are able to learn to choose novel rewarded stimuli presented during reversal training. This suggests that reversal deficits in OFC-lesioned rats can be explained by inflexibility in "learned non-reward." This means that lesioned animals may not be able to switch responses to two previously encountered stimuli (especially the previously non-rewarded stimuli), as their cue-reward association may not be available for updating.

Encoding Value in the OFC

• Neurons in the OFC encode or calculate aspects of the value of events and associated cues.
• MRI Activity in the human OFC is correlated with reduced susceptibility to a "framing effect." In other words, high OFC activity is correlated with more rational choices in decisions tasks where choice information has been biased by descriptions like "half-empty" or "half-full."
• Neurons in the primate OFC can fire in advance of an expected (reliably cued) reward.
• OFC neurons can be found to code for different properties of the expected reward, including reward magnitude, probability of delivery, and cost to obtain the reward.
• The firing of a single OFC neuron can represent the relative value of two offered rewards and reflect subsequent choice behavior.
• Abstraction and compounding of value parameters in the OFC may enable efficient comparison of choices for complex decision-making.

Updating Associations in the OFC

• The OFC may have evolved to enable flexible behavior.
• Physiologists find that neurons in the rat and primate OFC and amygdala can develop selective responses to rewarded or punished cues.
• When cue-outcome contingencies are reversed, changes in response properties can be induced in the OFC and amygdala.
• Changes in OFC activity may drive changes in connected brain areas like the amygdala.
• Analysis of the timing of physiological changes in the OFC and amygdala suggests they may cause the changes in behavior seen in response to changes in task contingencies.

OFC Anatomy and Function

• The OFC can be subdivided into a viscero-motor "medial network" and a sensory "orbital network" located in more ventral and lateral areas.
• Emotional stimuli are processed in more posterior regions of the OFC when compared to more neutral stimuli of the same type and modality.
• The OFC is reciprocally connected to the amygdala and may serve to modify activity in this and other limbic system structures.
• Different sectors of the OFC respond to pleasant vs. unpleasant tastes and odors detected via the nose vs. the back of the mouth. These circuits may be segregated to serve separate functions: Detection of pleasant and unpleasant odors may be independent as they drive opposite behaviors of approach and avoidance. Foraging for food by smells received via the air may be a separate system than one that monitors foods that are being consumed and therefore smelled via the back of the mouth.

Translating OFC Research

• Breakdown of frontal-striatal and limbic circuits may underlie multiple types of mental illness. Dysfunction in one area may lead to dysfunction in others.
• OFC activity is found to be abnormal in studies of psychopathy, depression, anxiety, obsessive compulsive disorder, and addiction.
• Cocaine exposure alters the OFC at the physiological and molecular level. These changes may account for behavioral changes in addicts.
• The OFC may be particularly vulnerable to age-related decline. OFC dysfunction may leave older adults vulnerable to fraudulent schemes.

Because the OFC is implicated in emotional behavior and decision-making, it is of natural interest to the disciplines of psychiatry, psychology, the social sciences, and the newly emerging field of neuroeconomics.

The reciprocal connections from the OFC to known emotional centers of the brain, for example, suggest that the OFC may be an important site for modulation of emotional reactions. It could therefore be a possible site of dysfunction or a target for amelioration of emotional disorders. In the case of drug addiction, which involves disruption of motivation, emotion, memories, and somatosensory and chemosensory systems, understanding the OFC should also inform addiction treatment and prevention. Finally, understanding how complex streams of information are processed by the brain should also inform our understating of rational and irrational human behavior in society and economic markets. The decline of the OFC in older adults may even explain why this population is particularly sensitive to fraudulent schemes and advertising.

OFC and mental illness

James Blair (U.S. National Institute of Mental Health) discussed the role of the OFC in psychopathy, a developmental disorder characterized by callous and unemotional traits. Diagnosed psychopaths or children showing conduct disorder and callous/unemotional traits have been found to show reversal impairments similar to those with amygdala and frontal lobe OFC damage. Functional MRI studies of these subjects confirm that activity in the amygdala and medial portions of the OFC is aberrant during reversal learning. Blair also showed that coordination of activity between the amygdala and mOFC may be impaired in psychopathy. He suggested that dysfunction in this circuit could disrupt understanding of fear and punishment and thus prevent the establishment of a moral sense of "badness."

The OFC is implicated in OCD, anxiety disorders, psychopathy, and major depressive illness.

The OFC is also implicated in obsessive-compulsive disorder (OCD), anxiety disorders, and major depressive illness. Wayne Drevets (NIMH) catalogued multiple abnormalities found in the OFC in depressed patients. Glucose metabolism and cerebral blood flow is increased in the OFC in a subject in a depressed phase, and reduced during a remitted phase. Gray matter volume, glial cell counts, and serotonin receptors are all found to be reduced in the OFC in depressed subjects. Hemodynamic resposes are also altered compared to normals in reward/loss processing and after negative feedback during a reversal task. Interestingly, patients with major depressive disorder may be hypersensitive to negative feedback, showing enhanced switching behavior after a negative outcome. Similar to James Blair, Drevets proposes that major depressive illness is due to dysregulation of emotional expression due to breakdown of frontal-striatal-limbic circuits.

Scott Rauch (Massachusetts General Hospital) discussed anxiety disorders and the possible role of the OFC in extinction of fear. In rodents extinction of a learned fear response is dependent on active processes involving the infralimbic cortex. In humans the infralimbic cortex may correspond to the ventromedial prefrontal cortex or medial OFC. Regulation of fear expression by top-down modulation of the amygdala may be one of the many ways that the OFC enables flexibility in behavior above and beyond that coded by the limbic circuits. Patients suffering from post-traumatic stress disorder do indeed show decreased function and volume of the medial OFC and other frontal areas. The function of the OFC can also be linked to the severity of symptoms in phobias and OCD.

Different subregions of the OFC may be involved in reward and punishment. (Kringelbach and Rolls, 2004; Kringelbach, 2005)

Rauch presented evidence from a collection of papers showing that human medial OFC function correlates with reward and positive affective states while activity in the lateral OFC occurs in relation to punishment and negative affective states. He proposed that diminished medial activity could diminish the capacity for fear extinction and contribute to anxiety disorders. Antero/lateral OFC may mediate more negative emotions, and activity here could lead to obsessions. He concludes that therapies should seek to enhance medial OFC activity and function, while reducing activity of the lateral OFC.

Addiction and the OFC

Disrupted assessment of value, impulsivity, and inflexible control of behavior are classic hallmarks of addictive behavior. Therefore, it may be no surprise that OFC circuits are involved in the development of addiction and relapse after withdrawal. In accord with the previous days' discussion of the role of the OFC in coding value, Nora Volkow (NIDA) proposed that one of the key problems in addiction may be that the abused drug acquires an inappropriately large value at the expense of other more normal or healthy reinforcers. Barry Everitt (University of Cambridge) showed that rats chronically exposed to drug use are less sensitive to adverse outcomes and do not suppress their drug seeking activity when punished. He also confirmed the role of the OFC in this behavior by showing impairment in the modulation of drug-seeking behavior by conditioned reinforcers in animals with OFC lesions.

Viral transfection of deltaFosB in the OFC can mimic the effects of cocaine use on impulsive behavior.

Catherine Winstanley has shown an upregulation of impulsive behaviors in rats after initial cocaine use or after withdrawal of chronic cocaine administration. This is correlated with upregulation of a transcription factor, deltaFosB, in the OFC. Amazingly, viral transfection of this transcription factor in the OFC can mimic the effects of cocaine use on these impulsive behavior tests.

As mentioned previously, using cocaine exposed rats, Tom Stalnaker (University of Maryland) has shown abnormal reversal learning and abnormal reversal of cue selectivity in neurons recorded in the amygdala after onset of reversal training. This reversal of cue selectivity in the amygdala was also shown to be OFC dependent.

Ageing and the OFC

Using data collected as part of the Baltimore Longitudinal Study of Ageing, Susan Resnick (Institute of Psychiatry, King's College London) reported that the OFC may be especially vulnerable to age-related declines in size and function compared to other areas of the brain. It may also be one of the first areas to show Alzheimer's disease-related plaque deposition, along with the temporal lobe.

Plaque and tangle (NFT) development can be seen in the OFC in some of the earliest stages of Alzheimer's Disease (AD). (Adapted from Braak & Braak, 1997)

Interestingly, while performing a task that involved performing two opposite choices with a delay in between ("delayed non-match-to-sample," a test of memory for events and choices), older adults showed an unexpected pattern of fMRI brain activation when compared to younger adults. Older adults had lower OFC activation and showed greater activity in more posterior association areas of the brain. Resnick suggested that perhaps this change in activity was the result of compensation for age-related OFC dysfunction.

This alteration in circuit use may have important implications for cognition ability in older adults. Natalie Denburg (University of Iowa) reported that, in a separate Iowa study, decreased frontal lobe function was found in about 35%–45% of normal older adults. She showed evidence that this may contribute to poor decision making in this population, perhaps leaving them especially vulnerable to false advertising, fraudulent sweepstakes, or ploys for banking information.

What did we learn from the meeting?

"There is now agreement that the OFC is critically important for representing the expected value of different outcomes."

A wealth of information was presented at the conference and it was absorbed by an equally diverse group of psychologists, neuroscientists, psychiatrists, and economists. Prior to the conference, co-organizer Jay Gottfried remarked, "I look forward to learning whether all of the seemingly disparate functions of the OFC can be reformulated under a single grand theory of function." After the meeting, Peter Rudebeck, a postdoctoral researcher at NIH and meeting attendee, felt "there is now agreement that the OFC is critically important for representing the expected value of different outcomes it is still unclear if this is true for all classes of stimuli or situations." He added, "It is clear that this is true for stimuli related to choice behavior; however, it is unclear if social stimuli or situations are represented in a similar way." Rudebeck thinks that understanding the valuation of social stimuli "will be critically important for understanding disorders such as depression and schizophrenia."

When asked what her impression of the conference, Blair Simpson, a psychiatrist at Columbia University wrote:

"It was one of the more stimulating conferences that I have been to in years. However, I left wanting more, specifically, I wanted a:

• better understanding of neurocircuitry and neurochemistry of OFC in human, nonhuman primates, rats, and mice (yes mice, since these are the gene knockout animal)
• better understanding of how the varied tasks used in animal and human studies tap into these different brain circuits—there seems to be great variability among the tasks. I think greater standardization would help. It makes it very hard to compare findings across studies and across species.
• better connection between human neuroimaging literature and primate neuroanatomy—what is actually meant by VM OFC, medial OFC, lateral OFC relative to Walker's areas?"

Simpson also said she "left very excited, feeling that there is great promise to make important translations between basic neuroscience and clinical approaches in psychiatry over the next decade." She added, "This is what I hoped would happen in psychiatry during my career."