The orbitofrontal cortex: reward, emotion and depression

Abstract

The orbitofrontal cortex in primates including humans is the key brain area in emotion, and in the representation of reward value and in non-reward, that is not obtaining an expected reward. Cortical processing before the orbitofrontal cortex is about the identity of stimuli, i.e. 'what' is present, and not about reward value. There is evidence that this holds for taste, visual, somatosensory and olfactory stimuli. The human medial orbitofrontal cortex represents many different types of reward, and the lateral orbitofrontal cortex represents non-reward and punishment. Not obtaining an expected reward can lead to sadness, and feeling depressed. The concept is advanced that an important brain region in depression is the orbitofrontal cortex, with depression related to over-responsiveness and over-connectedness of the non-reward-related lateral orbitofrontal cortex, and to under-responsiveness and under-connectivity of the reward-related medial orbitofrontal cortex. Evidence from large-scale voxel-level studies and supported by an activation study is described that provides support for this hypothesis. Increased functional connectivity of the lateral orbitofrontal cortex with brain areas that include the precuneus, posterior cingulate cortex and angular gyrus is found in patients with depression and is reduced towards the levels in controls when treated with medication. Decreased functional connectivity of the medial orbitofrontal cortex with medial temporal lobe areas involved in memory is found in patients with depression. Some treatments for depression may act by reducing activity or connectivity of the lateral orbitofrontal cortex. New treatments that increase the activity or connectivity of the medial orbitofrontal cortex may be useful for depression. These concepts, and that of increased activity in non-reward attractor networks, have potential for advancing our understanding and treatment of depression. The focus is on the orbitofrontal cortex in primates including humans, because of differences of operation of the orbitofrontal cortex, and indeed of reward systems, in rodents. Finally, the hypothesis is developed that the orbitofrontal cortex has a special role in emotion and decision-making in part because as a cortical area it can implement attractor networks useful in maintaining reward and emotional states online, and in decision-making.

Keywords: decision-making; depression; emotion; orbitofrontal cortex; reward.

Graphical Abstract

Figure 1

Maps of architectonic areas in the orbitofrontal cortex and medial prefrontal cortex of humans. Left, ventral view of the brain: The medial OFC includes areas 13 and 11 (green). The lateral OFC includes area 12 (red). (Area 12 is sometimes termed area 12/47 in humans. This figure shows two architectonic subdivisions of area 12.) Almost all of the human OFC except area 13a is granular. Agranular cortex is shown in dark grey. The part of area 45 shown is the orbital part of the inferior frontal gyrus pars triangularis. Right: the anterior cingulate cortex (medial view) includes the parts shown of areas 32, 25 (subgenual cingulate) and 24. The ventromedial prefrontal cortex includes areas 14 (gyrus rectus) 10 m and 10r. AON—anterior olfactory nucleus; Iai, Ial, Iam, Iapm—subdivisions of the agranular insular cortex [after Öngür et al. (2003)Journal of Comparative Neurology with permission of John Wiley & Sons, Inc., modified from a redrawn version by Passingham and Wise (2012).].

Figure 2

Some of the connections of the taste, olfactory, somatosensory, visual and auditory pathways to the OFC and amygdala in primates. V1, primary visual (striate) cortex; V2 and V4, further cortical visual areas. PFC, prefrontal cortex. The Medial PFC area 10 is part of the VMPFC. Ventro-postero-lateral (VPL) nucleus of the thalamus, which conveys somatosensory information to the primary somatosensory cortex (areas 1, 2 and 3). Ventro-postero-medial nucleus pars parvocellularis (VPMpc) of the thalamus, which conveys taste information to the primary taste cortex. For the purposes of description, the stages can be described as Tier 1, representing what object is present independently of reward value; Tier 2 in which reward value and emotion is represented; and Tier 3 in which decisions between stimuli of different value are taken, and in which value is interfaced to behavioural output systems. A pathway for top-down attentional and cognitive modulation of emotion is shown in purple. Auditory inputs also reach the amygdala (From Rolls, 2019c).

Figure 3

Connectivity shown on surface maps of the brain of the different parcels or subdivisions of the human orbitofrontal cortex (OFC). The parcels were based on the functional connectivity of every OFC voxel with each of the 94 automated anatomical labelling atlas 2 brain regions. Six divisions of the OFC are shown, with the approximate correspondence of each division with the cytoarchitectonic areas defined by Öngür et al. (2003) as shown in Fig. 1 as follows: 1—the gyrus rectus (much of it area 14); 2—medial OFC (area 13 m); 3—posterior OFC (area 13 l); 4—anterior OFC (area 11 l); 5—lateral OFC, posterior (area 12 m); 6—lateral OFC, anterior (area 12r). Surface maps showing the cortical connectivity of each parcel are shown. The functional connectivities have been thresholded at 0.3, and were obtained in resting-state fMRI with 654 participants. Quantitative evidence on the connectivity with different brain regions of each parcel is provided by Du et al. (2020b) and Hsu et al. (2020) (after Du et al., 2020a).

Figure 4

Evidence that the human lateral OFC is activated by non-reward. Activation of the lateral OFC in a visual discrimination reversal task on reversal trials, when a face was selected but the expected reward was not obtained, indicating that the subject should select the other face in future to obtain the reward. (A) A ventral view of the human brain with indication of the location of the two coronal slices (A, C) and the transverse slice (d). The activations with the red circle in the lateral OFC (peaks at [42 42 −8] and [−46 30 −8]) show the activation on reversal trials compared to the non-reversal trials. For comparison, the activations with the blue circle show the fusiform face area produced just by face expressions, not by reversal, which are also indicated in the coronal slice in C. (B) A coronal slice showing the activation in the right OFC on reversal trials. Activation is also shown in the supracallosal anterior cingulate region (Cingulate, green circle) that is also known to be activated by many punishing, unpleasant, stimuli (Grabenhorst and Rolls, 2011) (from Kringelbach and Rolls, 2003). (B) Activations in the human lateral OFC are related to a signal to change behaviour in the stop-signal task. In the task, a left or right arrow on a screen indicates which button to touch. However, on some trials, an up-arrow then appears, and the participant must change the behaviour and stop the response. There is a larger response on trials on which the participant successfully changes the behaviour and stops the response, as shown by the contrast stop–success—stop–failure, in the ventrolateral prefrontal cortex in a region including the lateral OFC, with peak at [−42 50 −2] indicated by the cross-hairs, measured in 1709 participants. There were corresponding effects in the right lateral OFC [42 52 −4]. Some activation in the dorsolateral prefrontal cortex in an area implicated in attention is also shown (after Deng et al., 2017). (C) Non-reward error-related neurons maintain their firing after non-reward is obtained. Responses of an OFC neuron that responded only when the macaque licked to a visual stimulus during reversal, expecting to obtain fruit juice reward, but actually obtained the taste of aversive saline because it was the first trial of reversal (trials 3, 6 and 13). Each vertical line represents an action potential; each L indicates a lick response in the Go-NoGo visual discrimination task. The visual stimulus was shown at time 0 for 1 s. The neuron did not respond on most reward (R) or saline (S) trials, but did respond on the trials marked S x, which were the first or second trials after a reversal of the visual discrimination on which the monkey licked to obtain reward, but actually obtained saline because the task had been reversed. The two times at which the reward contingencies were reversed are indicated. After responding to non-reward, when the expected reward was not obtained, the neuron fired for many seconds, and was sometimes still firing at the start of the next trial. It is notable that after an expected reward was not obtained due to a reversal contingency being applied, on the very next trial the macaque selected the previously non-rewarded stimulus. This shows that rapid reversal can be performed by a non-associative process, and must be rule-based. (After Thorpe et al., 1983). (D) BOLD signal in the macaque lateral orbitofrontal related to win-stay/lose-shift performance, that is, to reward reversal performance (after Chau et al., 2015).

Figure 5

The lateral OFC is activated by not winning, and the medial OFC by winning, in the monetary incentive delay task. The lateral OFC region in which activations increased towards no reward (No Win) in the monetary incentive delay task are shown in red in 1140 participants at age 19 and in 1877 overlapping participants at age 14. The conditions were Large Win (10 points) to Small Win (2 points) to No Win (0 points) (at 19; sweets were used at 14). The medial OFC region in which activations increased with increasing reward from No Win to Small Win to High Win) is shown in green. The parameter estimates are shown from the activations for the participants (mean ± sem) with the lateral orbitofrontal in red and medial OFC in green. The interaction term showing the sensitivity of the medial OFC to reward and the lateral OFC to non-reward was significant at P = 10⁻⁵⁰ at age 19 and P < 10⁻⁷² at age 14. In a subgroup with depressive symptoms as shown by the Adolescent Depression Rating Scale, it was further found that there was a greater activation to the No Win condition in the lateral OFC; and the medial OFC was less sensitive to the differences in reward value (modified from Xie et al., 2020).

Figure 6

Some of the emotions associated with different reinforcement contingencies. Intensity increases away from the centre of the diagram, on a continuous scale. The classification scheme created by the different reinforcement contingencies consists with respect to the action of (1) the delivery of a reward (S+), (2) the delivery of a punisher (S−), (3) the omission of a reward (S−) (extinction) or the termination of a reward (S+!) (time out) and (4) the omission of a punisher (S−) (avoidance) or the termination of a punisher (S−!) (escape). It is noted that the vertical axis describes emotions associated with the delivery of a reward (up) or punisher (down). The horizontal axis describes emotions associated with the non-delivery of an expected reward (left) or the non-delivery of an expected punisher (right). For the contingency of non-reward (horizontal axis, left), different emotions can arise depending on whether an active action is possible to respond to the non-reward, or whether no action is possible, which is labelled as the passive condition. In the passive condition, non-reward may produce depression. The diagram summarizes emotions that might result for one reinforcer as a result of different contingencies. Every separate reinforcer has the potential to operate according to contingencies such as these. This diagram does not imply a dimensional theory of emotion, but shows the types of emotional state that might be produced by a specific reinforcer. Each different reinforcer will produce different emotional states, but the contingencies will operate as shown to produce different specific emotional states for each different reinforcer.

Figure 7

Resting-state functional connectivity in depression. The medial and lateral OFC networks that show different functional connectivity in patients with depression. A decrease in functional connectivity is shown by blue arrows, and an increase by red arrows. MedTL—medial temporal lobe from the parahippocampal gyrus to the temporal pole; MidTG21R—middle temporal gyrus area 21 right; OFC13—medial OFC area 13; OFC47/12R—lateral OFC area 47/12 right. The lateral OFC cluster in OFC47/12 is visible on the ventral view of the brain anterior and lateral to the OFC13 clusters (from Cheng et al., 2016).

Figure 8

Functional connectivity (FC) differences of the medial and lateral OFC in major depressive disorder. Higher functional connectivity in depression is shown by red connecting lines, and includes higher functional connectivity of the non-reward/punishment-related lateral OFC with the precuneus, posterior cingulate cortex (PCC), pregenual anterior cingulate cortex (ACC), angular gyrus, and inferior frontal gyrus. Lower functional connectivity in depression is shown by blue connecting lines, and includes lower functional connectivity of the medial OFC with the parahippocampal gyrus memory system (PHG), amygdala, temporal cortex and supracallosal anterior cingulate cortex (ACC). The part of the medial OFC in which voxels were found with lower functional connectivity in depression is indicated in green. The areas apart from the medial OFC shown are as defined in the automated anatomical labelling atlas 2 (Rolls et al., 2015a), although the investigations that form the basis for the summary were at the voxel level.