Information

What does orbital mean in orbital frontal cortex?

What does orbital mean in orbital frontal cortex?



We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

Orbital frontal cortex is where decisions are made.

What does the word orbital there mean?

I looked around in wikipedia and never find it.


Orbitofrontal cortex: the area of the cerebral cortex located at the base of the frontal lobes above the orbits (or eye sockets), involved especially in social and emotional behaviour.


The boney component of the anatomical orbitstrong text (in lay terms the eye socket) is the area bordered by the zygomatic, frontal, maxilla, sphenoid, palatine, lacrimal and ethmoid bones.

The use of the word orbit (red) in orbitofrontal cortex refers to the positioning of this part of the frontal cortex (green) being directly behind the orbit:


Quick search on etymonline for Orbital:

orbital (adj.):

1540s with reference to eye sockets;

1839 with reference to heavenly bodies;

from orbit

(n.) + -al (1).

So my best guess is that it's related to eyes.


The functions of the orbitofrontal cortex

The orbitofrontal cortex contains the secondary taste cortex, in which the reward value of taste is represented. It also contains the secondary and tertiary olfactory cortical areas, in which information about the identity and also about the reward value of odours is represented. The orbitofrontal cortex also receives information about the sight of objects from the temporal lobe cortical visual areas, and neurons in it learn and reverse the visual stimulus to which they respond when the association of the visual stimulus with a primary reinforcing stimulus (such as taste) is reversed. This is an example of stimulus-reinforcement association learning, and is a type of stimulus-stimulus association learning. More generally, the stimulus might be a visual or olfactory stimulus, and the primary (unlearned) positive or negative reinforcer a taste or touch. A somatosensory input is revealed by neurons that respond to the texture of food in the mouth, including a population that responds to the mouth feel of fat. In complementary neuroimaging studies in humans, it is being found that areas of the orbitofrontal cortex are activated by pleasant touch, by painful touch, by taste, by smell, and by more abstract reinforcers such as winning or losing money. Damage to the orbitofrontal cortex can impair the learning and reversal of stimulus-reinforcement associations, and thus the correction of behavioural responses when there are no longer appropriate because previous reinforcement contingencies change. The information which reaches the orbitofrontal cortex for these functions includes information about faces, and damage to the orbitofrontal cortex can impair face (and voice) expression identification. This evidence thus shows that the orbitofrontal cortex is involved in decoding and representing some primary reinforcers such as taste and touch in learning and reversing associations of visual and other stimuli to these primary reinforcers and in controlling and correcting reward-related and punishment-related behavior, and thus in emotion. The approach described here is aimed at providing a fundamental understanding of how the orbitofrontal cortex actually functions, and thus in how it is involved in motivational behavior such as feeding and drinking, in emotional behavior, and in social behavior.


What Is the Orbitofrontal Cortex? (with pictures)

The orbitofrontal cortex is the smallest part of the frontal lobe in the brain. Located within the cranial cavity directly behind the eyes, this region of the prefrontal cortex is involved in a large part of the decision-making process. Part of this brain structure plays a role in the creation of pleasurable or unpleasant sensations evoked by many flavors and smells. During tests that measure brain activity, it is seen to be highly active throughout tasks that involve learning new information.

Different parts of the orbitofrontal cortex control several aspects of learning and behavior. The medial part, or middle of this brain structure, helps the brain process the reward aspect of behavior reinforcement. Lateral or side portions of it helps the brain to process the punishment value of actions. Interactions between the reward and punishment processors in the orbitofrontal cortex are an important factor in a person’s ability to learn from mistakes and change destructive behavior patterns.

Within the orbitofrontal cortex, the emotions and the thinking process combine to influence the daily decisions a person makes. Damage caused by an injury or by the growth of a lesion in it may cause changes in the behavior of a person. Behavior patterns may alter because of the significant impact the cortex has on the valuation of actions, objects, or people. The orbitofrontal cortex is also involved in many of the brain processes that exist in the minds of people with addictions, including the cravings for the unhealthy substance or activity.

Historically, treatment for personality disorders and psychosis involved the surgical excision of portions of the orbitofrontal cortex. The procedure, called a frontal lobotomy, was used when other methods of treatment failed. People that underwent a frontal lobotomy often reported feeling emotionally flat after the operation, however the post-surgical absence of disruptive behavior patterns and a reduction in the symptoms of psychosis often validated the invasive brain surgery.

Another part of the orbitofrontal cortex plays a role in the formation of food preferences. The taste of food can cause a pleasurable response or an unpleasant sensation that originates in the cortex. Even sensory information related to the texture of a food is relayed through it. A decision of whether or not to consume the same food again is an example of the integration of sensory information into the thinking process by this brain structure.

The test used by scientists to measure the activity in the orbitofrontal cortex is called a functional magnetic resonance imaging (fMRI) scan. It measures the changes in blood flow within the brain throughout different types of activities. During a period of increased stimulation, the fMRI images show that blood flow is amplified in the prefrontal cortex.


Materials and methods

Participants

Thirty-four psychiatrically and neurologically healthy participants (21 females mean age, 32.3 years SD, 14.5 years) were subjected to one MRI scanning session after giving their informed consent. The scanning protocol included two resting-state scans of approximately 6.5 min with identical scanning parameters and instructions, separated approximately 10 min in time. Participants were instructed to fixate on a cross at the center of the screen, keep their eyes open, and refrain from intentionally engaging in specific mental tasks or falling asleep during the scan.

FMRI acquisition

Scanning was conducted on a Siemens MAGNETOM Allegra 3T MRI head-only scanner. Head motion was constrained by the use of foam padding. For each participant, 153 T2*-weighted gradient echo planar images (EPI) with 41 slices were acquired (except for 6 participants for whom 203 images were available). EPI can suffer substantial loss of BOLD sensitivity and geometric distortions due to magnetic field inhomogeneity near air tissue interfaces. To minimize MRI signal loss and recover the true spatial signal positions in the OFC, we: (a) used an optimized echo time, (b) tilted the slices (

30° angle), and (c) generated a field map to offline correct susceptibility-related signal displacements. Imaging parameters for the resting-state sequence were as follows: TR, 2500 ms TE, 25 ms flip angle, 90° matrix size, 128 × 96 and FOV, 256 mm distance factor, 20% resulting in a voxel size of 2 × 2 × 3 mm. The gradient echo image used to generate the field map had the same grid and slice orientation as the functional images (TR 704 ms TE 5.11, 7.57 ms flip angle 60°). To enable the localization of functional data, a high-resolution T1-weighted image was acquired with the following parameters: TR 2250 ms TE 2.6 ms flip angle 9° FOV 256 mm slice thickness 1 mm matrix size 256 × 256 number of slices 192 voxel size 1 × 1 × 1 mm.

FMRI preprocessing

Preprocessing of fMRI data was performed using the SPM 5 software (Welcome Trust Center for Neuroimaging, London, UK). The functional data were subjected to the following preprocessing: slice time correction, spatial correction using the field map, realignment, co-registration with the anatomical scan, normalization to the Montreal Neurological Institute (MNI) template (ICBM-152), reslicing to 3 mm isotropic voxels, and smoothing with a 6 mm full width half maximum (FWHM) Gaussian kernel. The T1-weighted images were segmented into grey matter, white matter, and cerebrospinal fluid tissue maps, and these maps were later used in the analyses. Furthermore, we removed non-neuronal contributions from the BOLD signal by regressing the following nuisance variables: the six volume realignment parameters, the average time series in white matter and CSF voxels, the session-specific mean, and the intrinsic autocorrelations. The global brain signal or average grey matter signal was not included as a regressor. Finally, the residual volumes of the multiple regression were Fourier band pass filtered (0.01–0.1 Hz).

Head motion has been shown to significantly underestimate long-range and overestimating short-range FC connectivity, even after regressing out volume to volume head motion measures (Power et al. 2012 van Dijk et al. 2012). To further reduce this bias, we took the following approach: (1) we identified scans during which the frame-wise displacement exceeded 0.4 mm [13.3% of voxel size i.e., translation in the z direction or rotation in the x direction corresponding to 0.4 mm z-displacement of frontopolar voxels, assuming an x-rotation point 88 mm from the frontal pole Talairach and Tournoux (1988)], (2) we excluded the identified volumes together with the 1-back and 2-forward volumes [to avoid spin history assumptions’ violations caused by movement Power et al. 2012)], and (3) we excluded participants for whom less than 120 volumes [i.e., 5 min van Dijk et al. 2010)] of resting data remained after the correction (mean duration, 6.4 min SD, 0.8 min).

OMPFC intrinsic FC-based parcellation

The parcellation analysis was performed for each participant and each hemisphere separately. For each participant, the voxels selected for parcellation comprised all the voxels that fell both within the person’s normalized grey matter mask (density > 0.5) and a liberal OMPFC ROI mask (left or right hemisphere). Thus, the parcellation mask differed between participants to accommodate anatomical variation and avoid contaminating of the analysis with none-cortical voxels. The liberal ROI mask was constructed from the Automated Anatomical Labeling (AAL) map in MNI space (WFU PickAtlas Maldjian et al. 2003 Rollset al. 2015 Tzourio-Mazoyer et al. 2002). It comprised left-side AAL regions with the following labels: “frontal superior orbital”, “frontal middle orbital”, “frontal inferior orbital”, “frontal medial orbital”, “rectus”, “cingulum anterior”, and “frontal superior medial”. From the region labeled “frontal superior medial”, only a part was included, extending dorsally until the horizontal border defined by the anterior cingulate AAL label (manually drawn using MRIcron Rorden and Brett 2000). The ROI mask was expanded spatially to ensure coverage of grey matter in all participants and to cover also parts of areas boarding the orbital and medial areas of interest. The latter allow empirical delineation of the full extent of areas of interest at the boundary of the mask. The expansion of the mask comprised of a twice repeated 10 mm FWHM Gaussian smoothing followed by high-pass thresholding at 0.2 density.

For the voxels selected for parcellation, a voxel-by-voxel correlation matrix was constructed by computing the Pearson correlation between their cleaned time courses measured during the first resting-state scan. A high-pass absolute weight threshold was applied to the correlation matrix to eliminate the weak, less-significant links that most likely represent spurious connections (Rubinov and Sporns 2010). Because modularity is inversely related to graph density (Goulas et al. 2012) and since our aim was to retrieve the maximal modularity solution for grouping the voxels into modules, we searched for the lowest connection density that still yielded a connected graph. The connection densities investigated were: 0.25, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, and 4.0%.

To partition a thresholded correlation matrix into discrete modules, we employed the Louvain module detection algorithm [Blondel et al. 2008 Brain Connectivity Toolbox (Rubinov and Sporns 2010)], one of the best performing algorithms for fast and efficient detection of modules in extended networks (Lancichinetti and Fortunato 2009). The modularity statistic quantifies how well a network can be subdivided into groups of nodes (voxels) with higher than chance connectivity in between them (Girvan and Newman 2002 Newman 2006). Applied to brain networks, it can be used to delineate neurobiological meaningful functional units (e.g., Goulas et al. 2012 Meunier et al. 2009 Rubinov and Sporns 2010). Modularity is defined as follows:

where e i is the amount of edges (connections) linking nodes (voxels) within module i, d i is the total amount of edges of module i nodes (i.e., degree of module i), and m is the total number of edges in the graph (i.e., network degree). Large Q values indicate the presence of community structure within the graph. To compensate for the stochastic nature of the algorithm, each parcellation analysis of individual data at a particular threshold was repeated 50 times and the solution with the highest Q value was selected as the final solution (Sporns et al. 2007). The parcellation procedure results in the unique classification of every voxel in the OMPFC into one of the modules in the solution.

To test the statistical significance of the observed modularity structure, we compared the obtained Q value with the Q value of null models computed from the individual data sets. Zalesky et al. (2012) have recently drawn attention to the fact that observations of brain networks that use correlation as a measure of connectivity are inherently more clustered than random networks. For this reason, they should be benchmarked against null networks that preserve the transitive structure of correlation networks. We created such null networks by applying the Hirschberger–Qi–Steuer algorithm (for algorithm and details see Zalesky et al. 2012) to the individual correlation matrices. This algorithm generates random null covariance matrices with distributional properties matched to the observed matrices. The resulting null covariance matrices were thresholded to match the density of the original matrices.

Group-wise clustering

Modules obtained from parcellating individual data sets were grouped together to capture their individual transcending commonalities. Because individual parcellations differed in voxel space (within each participant the parcellation took place within his/her unique grey matter map) and had different numbers of modules, the grouping was based on the spatial proximity of the modules’ center of mass (COM), within the Euclidean coordinates of the MNI space. Spatial proximity was defined as the inverse of the Euclidean distance between the COM of modules from two different parcellations. The integration progressed iteratively. At each iteration, first, a cost matrix was computed for matching each parcellation with every other parcellation. The matching cost for a pair of parcellations was the sum of the distances between their assigned modules. Module assignment between pairs of parcellations was based on minimizing the matching cost using the modified Hungarian assignment algorithm (Munkres 1957 Cao 2008). In the second step, the pairs of parcellations with the lowest matching cost were merged by weighted averaging (for details, see Supplementary Information, section “Methods”, point 1), one after the other, and then eliminated from the cost matrix. The merged parcellations entered the next iteration level and the procedure was repeated until a final set of COMs was obtained. Each final COM represented a cluster of modules from individual parcellations merged into this common COM. Note that not all participants will be represented in every final cluster, due to different numbers of modules per participant/parcellation.

Replicability of intrinsic FC parcellation

To investigate replicability of parcellation results, we analyzed for each participants a second, independently resting-state scan acquired in a different run during the same session. The within participant consistency of the obtained modules across the two data sets was compared first to the maximum possible consistency and second to chance consistency. This maximum possible consistency is less than 100% due to the stochastic nature of the modularity detection algorithm used. The maximum replicability was estimated by re-analysing the same data sets twice. To test whether the observed consistency exceeded the chance level, it was compared to the replicability between null models, in which a matching number of modules is positioned purely randomly. These models implement the null hypothesis that the spatial location of the modules is not dependent on information in the connectivity matrix. The creation details of these null models are described in Supplementary Information (“Methods”, point 2).

The measures used to compare parcellation similarity were the Dice similarity coefficient (Crum et al. 2006) and normalized mutual information. Both measures will be around 0.5 for completely mismatching parcellations and 1.0 for perfectly overlapping parcellations. The normalized mutual information has the advantage over the Dice similarity coefficient that it does not require an assignment of modules between the parcellations prior to quantification, which introduces additional room for error. The modified Hungarian assignment algorithm (Munkres 1957 Cao 2008) was used for modules assignment.

Replicability was further investigated by looking at the module boundaries. A boundary in a parcellation was defined in volume space as a center-surround discrepancy, with the center being a single voxel’s module identification number, and the voxel’s surround comprising the averaged value of any of the six voxels touching it on each of its sides. Voxels outside of the parcellation patch were not considered (i.e., surround < six voxels) to prevent voxels on the edge of the parcellation patch from being counted as boundary voxels. For each pair of parcellations, the consistency of boundary voxels was quantified by the number of voxels that were identified as boundary voxel in both parcellations relative to the number of boundary voxels in each parcellation, expressed as a Dice coefficient.

Replicability of boundaries was also assessed at the group level, by counting the incidence of voxels being boundary voxels across participants. This requires summation across participants in voxel space, which can only yield a rough approximation of boundary co-localization. To improve co-localization, consistent boundary images per participant and the similar images derived from the random module models were smoothed with a 6 mm FWHM Gaussian kernel. A voxel-wise t-test of the observed sum of co-localized boundaries against the co-localization of boundaries in the random models was performed and differences evaluated at the alpha level of 0.05, FDR-level correction for multiple comparison (Genovese et al. 2002).

Finally, group-wise replicability was also evaluated from a visual comparison of different group-wise parcellation results. First, the group-integrated parcellation of the left hemisphere was compared to that from the right hemisphere. Therefore, the connectivity-based parcellation and group-wise clustering described above were also applied to the right OMPFC of the same resting-state run. In addition, we tested the replicability within the left hemisphere across runs, by applying the group-wise clustering also to the parcellation results from the left OMPFC of the second resting-state run in the scanning session. We compared the similarity of the group-wise clustering result between each hemisphere and run by plotting their spatial extent to allow visual inspecting of their correspondence with our main results.

OFC signal coverage and inter-run correlation stability

EPI suffers substantial loss of BOLD sensitivity near air tissue interfaces and FC metrics are sensitive to the levels of signal amplitude and signal-to-noise ratio (Golestani and Goodyear 2011).

To quantify the severity of signal dropout in the artifact-susceptible areas, we calculated the relative signal intensity, which is the signal intensity of each voxel (averaged across time) relative to the mean signal intensity of all grey matter voxels (Smits et al. 2007). In addition, we computed the signal-to-noise ratio of the time series (tSNR), defined as the mean intensity of every voxel in the time series divided by its standard deviation across time (Triantafyllou et al. 2011 Golestani and Goodyear 2011).

To assess the impact of signal quality loss on the connectivity metric used, we estimated the stability of the whole-brain FC profiles of each OMPFC voxel across the two separate resting-state runs, using the eta 2 coefficient:

where a i and b i are the correlations of voxel i in the first run and second run, respectively, m i is the mean correlation of both runs at position i, and M is the grand mean of correlations across all locations in both runs. The eta 2 coefficient varies from 0 (no similarity) to 1 (identical) and directly quantifies the difference in the values of the same voxel in the two runs (Cohen et al. 2008). A general linear mixed model analysis was used to test whether the inter-run stability differed significantly between the individual clusters by running. The model used heterogeneous compound symmetry as covariance structure and cluster as a fixed effects factor. The estimation method was restricted maximum-likelihood estimation (maximum iterations = 150). The advantage of general linear mixed models is that they allow the analysis of repeated-measures data in unbalanced designs, as is the case here, since every participant has a module in many but not all group-level clusters of modules. To control the family wise error rate, we used Holm’s sequential rejective Bonferroni correction (Holm 1979 Holland and Copenhaver 1987). All pairwise comparisons among means were adjusted to a corrected alpha of 0.05.

To test whether tSNR had a significant effect on the inter-run correlation stability, we ran the general linear model with the SPSS procedure mixed again this time using tSNR as a covariate. The amount of variance explained by tSNR was calculated as follows:

where r 2 is the proportion of the variance explained, σ is the standard error estimate in the model with tSNR as predictor, and σ 0 is the standard error estimate of the null model.

Similarity of whole-brain FC profiles of the OMPFC subdivisions

For every cluster, a whole-brain FC profile was created by placing a spherical seed of 4 mm radius at the COM of each participant’s module and calculating Pearson’s correlation coefficient between the average seed voxels time course and the time course of every voxels in the rest of the brain. The spherical seed approach was chosen over using the entire module as seed to get a maximal spatial spread of the seeds while avoiding contamination from adjacent modules. This procedure resulted in whole-brain r maps for every module at the individual level. The maps were subsequently r-to-Z transformed using Fisher’s formula. Cluster-wise functional connectivity maps was created by averaging over the r maps of all the modules that had been assigned to the particular cluster.

To examine whether our OMPFC subdivisions could be distinguished into spatially extended networks based in FC profile similarities, we used agglomerative hierarchical cluster analysis which does not involve a priori assumptions about the number of groups present in the data and outputs a “bottom-up” hierarchy of areas in the form of a dendrogram. First, we calculated the (dis)similarity matrix for the whole-brain connectivity profiles of all clusters using correlation as the similarity metric (1 − r) and subsequently created a dendrogram to represent the hierarchies in the data. To select the most appropriate linkage method for the construction of the dendrogram, we ran the analysis using the following linkage methods: centroid, average, single, median, complete, weighted, and Ward. For each of these methods, we computed the cophenetic correlation coefficient of the resulting dendrogram, a measure of how well the original distances in the data are represented. From these, we selected the linkage method that generated a dendrogram which contained clusters proceeding hierarchically (i.e., monotonic) and had the highest cophenetic coefficient.


The Role of the VM in Decision Making

The Gambling Task

The study of the decision-making impairment of patients with VM lesions required an instrument for the detection and measurement of such impairments in the laboratory. The development of a card task known as ‘the gambling task’ ( Bechara et al., 1994) provided this tool. The essential feature of this task is that it mimics real-life situations in the way it factors uncertainty, reward and punishment. The task involves four decks of cards, named A, B, C and D. The goal is to maximize profit on a loan of play money. Subjects are required to make a series of 100 card selections, but are not told ahead of time how many card selections they are going to be allowed to make. Cards can be selected one at a time, from any deck, and subjects are free to switch from any deck to another, at any time and as often as they wish. The decision to select from one deck or another is largely influenced by schedules of reward and punishment. These schedules are pre-programmed and known to the examiner, but not to the subject ( Bechara et al., 1994, 1999a). They are arranged in such a way that every time the subject selects a card from deck A or B, s/he gets $100, and every time deck C or D is selected, the subject gets $50. However, in each of the four decks, subjects encounter unpredictable money loss (punishment). The punishment is set to be higher in the high-paying decks A and B, and lower in the low-paying decks C and D. In decks A and B the subject encounters a total loss of $1250 in every 10 cards. In decks C and D the subject encounters a total loss of $250 in every 10 cards. In the long term, decks A and B are disadvantageous because they cost more, a loss of $250 in every 10 cards. Decks C and D are advantageous because they result in an overall gain in the end, a gain of $250 in every 10 cards.

Insensitivity to Future Consequences following Bilateral Damage of the Prefrontal Cortex

A large sample of normal control subjects (n = 82, balanced in terms of gender, with 8–20 years of education, and between the ages of 20 and 64) has been tested with the original card version of the gambling task described above. Patients with lesions in different sectors of the frontal lobe (n = 45), or with lesions in areas of the lateral temporal cortex or occipital cortex (n = 35) have also been tested. Since the original manual version of the gambling task was described, a new computer version has been devised, and similar numbers of control subjects and patients have been tested with the new computer version. The results from either version of the gambling task are interchangeable. As the task progresses from the first to the 100th trial, normal controls gradually make more selections of cards from the good decks (C and D) and less selections from the bad decks (A and B) (Fig. 2 left). Patients with lesions in the dorsolateral sector of the prefrontal cortex ( Bechara et al., 1998), or in areas outside the prefrontal cortex ( Bechara et al., 1994), perform in a manner similar to that of normal subjects. In sharp contrast, patients with bilateral lesions of the VM do not increase the number of their selection of cards from the good decks (C and D) they persist in selecting more cards from the bad decks (A and B) (Fig. 2 left). The card selection profiles from normal controls show that a typical normal subject initially samples all decks and repeats selections from the bad decks A and B, probably because they pay more. However, eventually the normal subject switches to more and more selections from the good decks C and D, with only occasional returns to decks A and B. On the other hand, a typical VM patient behaves like a normal subject only in the first few selections. The patient begins by sampling all decks and selecting from decks A and B, and then makes several selections from decks C and D, but then soon returns more and more to decks A and B (Fig. 2 right).

In the normal population, performance on the gambling task does not seem to depend on education or gender, although a few preliminary reports suggest that males perform slightly better than females ( LeLand et al., 1998 Reavis et al., 1998). Most intriguing is that, as a group, older adults (above 64 years of age) perform poorly on this task relative to younger adults (i.e. age 26–56) ( Denburg et al., 1999). It should be noted, however, that performance on the gambling task in older adults is dichotomous, i.e. some perform very well and some perform very poorly. This finding raises an important question as to why this happens in some older adults and not others, the answer to which may help explain why some older adults are especially vulnerable to advertising fraud in real life.

In the VM patient population, the decision-making impairment, as measured by the gambling task, is stable over time. When a sample of six VM frontal patients and five normal controls were re-tested after various time intervals (1 month after the first test, 24 h later and for the fourth time, 6 months later), the performance of VM patients did not improve. On the other hand, the performance of normal controls improved significantly over time (Fig. 3 ).

These results demonstrate that the VM patients' performance profile is comparable to their real-life inability to learn from their previous mistakes. This is especially true in personal and social matters, a domain for which in life, as in the gambling task, an exact calculation of the future outcome is not possible and choices must be based on approximations.

Biases Guide Decisions

The results described above prompted the following question: what is the basis for the ‘myopia for the future’ that plagues VM frontal patients? For many years, many theorists of decision making assumed that the feelings triggered when making a decision or a risky choice were not integral to the decision-making process. In this sense, the decision-making theorists assumed that risky decision making was essentially a cognitive activity devoid of an emotional component. These theories suggest that people assess the possible outcomes of their actions through some type of cost–benefit analysis. However, several authors have proposed an alternative theoretical account which highlights the role of the affect experienced during the time of deliberation prior to making decisions ( Schwartz and Clore, 1983 Zajonc, 1984 Damasio et al., 1990).

Evidence in support of this idea comes from studies of normal control subjects and patients with bilateral VM frontal damage during their performance on the gambling task, and the analysis of their psychophysiological activity during task performance ( Bechara et al., 1996). Skin conductance response (SCR) activity has been recorded so far in a large sample of normal subjects (n = 55) and VM patients (n = 15) during the performance of the gambling task. Despite some variations in the methods for collecting the SCR data ( Bechara et al., 1996, 1999a), the general principles remain the same. Every time the subject picks a card, the deck from which that card was picked is recorded, and the magnitude of the SCR in the time window (

5 s) right before the subject picked the card is measured. In addition, the magnitude of the SCR in the time window (

5 s) after the card was picked is also measured. Thus, three types of responses are identified. (i) The reward SCRs, those occurring after turning cards with reward only. (ii) The punishment SCRs, those occurring after turning cards with reward and punishment. (iii) The anticipatory SCRs, those occurring before turning a card from a deck, during the time the subject ponders from which deck to choose ( Bechara et al., 1996).

The results from the psychophysiological experiments conducted so far reveal that normal controls and VM patients generate SCRs as a reaction to reward or punishment. Normal controls, however, as they become experienced with the task, also begin to generate SCRs before the selection of any card. The anticipatory SCRs generated by normal controls: (i) develop over time (i.e. after selecting several cards from each deck, and thus encountering several instances of reward and punishment) and (ii) actually become more pronounced before selecting cards from the disadvantageous decks (A and B). These anticipatory SCRs are absent in the VM patients (Fig. 4 ). This suggests that VM patients have a specific impairment in their ability to generate anticipatory SCRs in response to a possible outcome of their action. Since SCRs are physiological indices of an autonomically controlled change in somatic state, it seems reasonable to conclude that the absence of anticipatory SCRs is an indication that these patients' ability to change somatic states in response to an imagined scenario is severely compromised. In this perspective, the failure to enact a somatic state appropriate to the consequences of a response would be a correlate of their inability to choose advantageously.

Risk Taking versus Impaired Decision Making

None of the bilateral VM patients tested so far have performed advantageously on the gambling task. However, not every normal control subject performs advantageously. Approximately 20% of normal adults who describe themselves as high-risk takers in real life end up selecting more cards from the bad decks relative to the good ones ( Bechara et al., 1999a). When looking at the anticipatory SCRs in these normal individuals, it is often found that the magnitudes of the anticipatory SCRs in relation to the bad decks are slightly lower than those in relation to the good decks ( Bechara et al., 1999a). The opposite is true (i.e. higher anticipatory SCRs with the bad decks relative to the good decks) in normal individuals who play advantageously. The most critical distinction between these normal individuals and the VM patients, however, is that these normal individuals do generate anticipatory SCRs. The VM patients, on the other hand, do not generate anticipatory SCRs at all. These physiological results are very important because they separate individuals with high-risk-taking behavior from individuals with VM frontal lobe dysfunction. When taking a risk, the somatic states signaling the possible negative consequences of the outcome are enacted. However, the individual can override these biases by higher cognitive processes. In the case of VM damage, these biases are never enacted and never enter the decision-making process. This suggests that taking a risk is not the same as having poor judgement and impaired decision making. The issue of risk taking versus decision making was addressed in a previous study that showed that orbitofrontal patients risked significantly less of their accumulated reward than controls, thus suggesting a pattern of conservative behavior ( Rogers et al., 1999). Yet, these same patients made suboptimal choices and spent more time deliberating their choices ( Rogers et al., 1999). Another study suggested that while frontal patients were impulsive and made poor decisions, they did not express a high-risk-taking behavior ( Miller, 1992). This evidence suggests that risk-taking behavior and impaired decision making are not synonymous.

Biases Do Not Need To Be Conscious

Given the important role that biases play in decision making, it is important to determine if these anticipatory responses (biases) develop after the subject knows which decks are good or bad, or if they precede such explicit knowledge. This question was addressed in an experiment in which ten normal subjects and six VM frontal patients were tested on the gambling task, while their SCRs were being recorded as before. However, in this experiment, every time a subject had picked ten cards the game was stopped briefly and the subject was asked to describe whatever s/he knew was going on in the game ( Bechara et al., 1997). The analysis of the subjects' answers suggested that they went through four distinct periods across the task. The first was a pre-punishment period, when subjects sampled the decks, before they encountered any punishment. The second was a pre-hunch period, when subjects began to encounter punishment, but had no clue about what was going on in the game. The third was a hunch period, when subjects began to express a hunch about the decks that were riskier, even if they were not sure about their guess. The fourth was a conceptual period, when subjects knew very well the contingencies in the task, which decks were the good ones, which decks were the bad ones and why this was so (Fig. 5 ). It is interesting that 30% of the control subjects did not reach the fourth, or conceptual, period in this experiment, yet they performed normally on the gambling task.

In normal controls, when the anticipatory SCRs from each of these four periods were examined, it was found that there was no significant anticipatory activity during the pre-punishment period. There was a substantial rise in anticipatory responses during the pre-hunch period, i.e. before any conscious knowledge developed. This anticipatory SCR activity was sustained for the rest of the task. When the type of choice from the different decks was examined for each period, the results revealed that there was a preference for the high-paying decks (A and B) during the pre-punishment period. There was a hint of a shift in the pattern of card selection, away from the bad decks, as early as in the pre-hunch period. This preference for the good decks became more pronounced during the hunch and conceptual periods. Even those 30% of controls who did not reach a full conceptual knowledge of the relative goodness or badness of the decks ended up playing advantageously. The VM frontal patients, on the other hand, never reported a hunch. They also never developed anticipatory SCRs, and continued to choose more cards from decks A and B relative to C and D. However, 50% of VM frontal patients did reach the conceptual period, in which they were able to recognize and identify the bad decks. Even so, they still performed disadvantageously ( Bechara et al., 1997).

These results show that VM frontal patients continue to choose disadvantageously in the gambling task, even after realizing the consequences of their action. This suggests that the anticipatory SCRs represent unconscious biases, probably derived from prior experiences with reward and punishment. These biases help deter the normal subject from pursuing a course of action that is disadvantageous in the future. This biasing effect occurs even before the subject becomes aware of the goodness or badness of the choice s/he is about to make. Even without these biases, the knowledge of what is right and what is wrong may become available, as happened in 50% of the VM patients. However, by itself, such knowledge is not sufficient to ensure an advantageous behavior. Although the frontal patient may be fully aware of what is right and what is wrong, s/he still fails to act accordingly. These patients may ‘say’ the right thing, but ‘do’ the wrong thing.


A Neurophysiological Basis for Stimulus–Reinforcement Learning and Reversal in the Orbitofrontal Cortex

The neurophysiological evidence and the effects of lesions described suggests that one function implemented by the orbitofrontal cortex is rapid stimulus–reinforcement association learning, and the correction of these associations when reinforcement contingencies in the environment change. To implement this, the orbitofrontal cortex has the necessary representation of primary reinforcers, including taste and somatosensory stimuli. It also receives information about objects, e.g. visual view-invariant information ( Booth and Rolls, 1998), and can associate this at the neuronal level with primary reinforcers such as taste and reverse these associations very rapidly. Another type of stimulus which can be conditioned in this way in the orbitofrontal cortex is olfactory, although here the learning is slower. It is likely that auditory stimuli can be associated with primary reinforcers in the orbitofrontal cortex, though there is less direct evidence of this yet. The orbitofrontal cortex also has neurons which detect non-reward, which are likely to be used in behavioral extinction and reversal. They may do this not only by helping to reset the reinforcement association of neurons in the orbitofrontal cortex, but also by sending a signal to the striatum which could be routed by the striatum to produce appropriate behaviors for non-reward ( Rolls and Johnstone, 1992 Williams et al., 1993 Rolls, 1994b). Indeed, the striatal route may be an important one through which the orbitofrontal cortex influences behavior when the orbitofrontal cortex is decoding reinforcement contingencies and their changes ( Rolls, 1999a). Some of the evidence for this is that neurons with responses that reflect the output of orbitofrontal neurons are found in the ventral part of the head of the caudate nucleus ( Rolls et al., 1983a) and the ventral striatum ( Rolls and Williams, 1987 Schultz et al., 1992 Williams et al., 1993) — parts of the striatum that receive connections from the orbitofrontal cortex — and that lesions of the ventral part of the head of the caudate nucleus impair visual discrimination reversal ( Divac et al., 1967), which is also impaired by orbitofrontal cortex lesions. The relation between orbitofrontal cortex and striatal processing is considered further elsewhere ( Rolls and Johnstone, 1992 Rolls, 1994b, 1999a Rolls and Treves, 1998).

Decoding the reinforcement value of stimuli, which involves for previously neutral (e.g. visual) stimuli learning their association with a primary reinforcer, often rapidly, and which may involve not only rapid learning but also rapid relearning and alteration of responses when reinforcement contingencies change, is then a function proposed for the orbitofrontal cortex. Using this decoding to specify the goals for action would be important in, for example, motivational and emotional behavior. It would be important in, for example, feeding and drinking by enabling primates to learn rapidly about the food reinforcement to be expected from visual stimuli ( Rolls, 1994c, 1999a). This is important, for primates frequently eat more than 100 varieties of food vision by visual–taste association learning can be used to identify when foods are ripe and during the course of a meal, the pleasantness of the sight of a food eaten in the meal decreases in a sensory-specific way ( Rolls et al., 1983b), a function that is probably implemented by the sensoryspecific satiety-related responses of orbitofrontal visual neurons ( Critchley and Rolls, 1996b).

With respect to emotional behavior, decoding and rapidly readjusting the reinforcement value of visual signals is likely to be crucial, for emotions can be described as states elicited by reinforcing signals ( Rolls, 1986a, b, 1990, 1995b, 1999a, 2000b). For example, fear is a state produced by a stimulus or event associated with a punisher such as pain. The ability to perform this learning very rapidly is probably very important in social situations in primates, in which reinforcing stimuli are continually being exchanged, and the reinforcement value of stimuli must be continually updated (relearned), based on the actual reinforcers received and given. Although the functions of the orbitofrontal cortex in implementing the operation of reinforcers such as taste, smell, tactile and visual stimuli including faces are most understood, in humans the rewards processed in the orbitofrontal cortex include quite general learned rewards (i.e. secondary reinforcers) such as working for ‘points', as will be described shortly.

Although the amygdala is concerned with some of the same functions as the orbitofrontal cortex, and receives similar inputs (see Fig. 2 ), there is evidence that it may function less effectively in the very rapid learning and reversal of stimulus reinforcement associations, as indicated by the greater difficulty in obtaining reversal from amygdala neurons ( Sanghera et al., 1979 Rolls, 1992b, 2000a Wilson and Rolls, 2000), and by the greater effect of orbitofrontal lesions in leading to continuing choice of no longer rewarded stimuli ( Jones and Mishkin, 1972). In primates, the necessity for very rapid stimulus–reinforcement re-evaluation and the development of powerful cortical learning systems may result in the orbitofrontal cortex effectively taking over this aspect of amygdala functions ( Rolls, 1992b, 1999a, 2000a).


Is frontal cortex the same as frontal lobe?

The frontal cortex can be defined as the neocortex anterior to the motor somatosensory&ndashcortex border. This is a large region in primates, containing areas involved directly or indirectly in the control of almost every behavior.

Similarly, is the prefrontal cortex part of the frontal lobe? Prefrontal Cortex. The prefrontal cortex is a part of the brain located at the front of the frontal lobe. It is implicated in a variety of complex behaviors, including planning, and greatly contributes to personality development.

In this way, is prefrontal cortex and frontal lobe the same?

The frontal lobe like other lobes has several functional areas. One of the functional area of frontal lobe is prefrontal cortex or prefrontal area. This area is associated with concentration, emotion and other higher brain functions.

What does the frontal lobe do psychology?

The frontal lobe is involved in reasoning, motor control, emotion, and language. It contains the motor cortex, which is involved in planning and coordinating movement the prefrontal cortex, which is responsible for higher-level cognitive functioning and Broca's area, which is essential for language production.


What Do the Terms FTD and FLTD Mean?

One of the challenges shared by patients, families, clinicians, and researchers is confusion about how to classify and label frontotemporal disorders. A diagnosis by one doctor may be called something else by a second, and the same condition or syndrome may be referred to by another name by a pathologist who examines the brain after death.

For many years, scientists and physicians used the term frontotemporal dementia (FTD) to describe this group of illnesses. After further research, FTD is now understood to be just one of several possible variations and is more precisely called behavioral variant frontotemporal dementia, or bvFTD.

This article uses the term frontotemporal disorders to refer to changes in behavior and thinking that are caused by underlying brain diseases collectively called frontotemporal lobar degeneration (FTLD). FTLD is not a single brain disease but rather a family of neurodegenerative diseases, any one of which can cause a frontotemporal disorder. Frontotemporal disorders are diagnosed by physicians and psychologists based on a person’s symptoms and results of brain scans and genetic tests. With the exception of known genetic causes, FTLD can be identified definitively only by brain autopsy after death.


Orbital prefrontal cortex volume predicts social network size: an imaging study of individual differences in humans

The social brain hypothesis, an explanation for the unusually large brains of primates, posits that the size of social group typical of a species is directly related to the volume of its neocortex. To test whether this hypothesis also applies at the within-species level, we applied the Cavalieri method of stereology in conjunction with point counting on magnetic resonance images to determine the volume of prefrontal cortex (PFC) subfields, including dorsal and orbital regions. Path analysis in a sample of 40 healthy adult humans revealed a significant linear relationship between orbital (but not dorsal) PFC volume and the size of subjects' social networks that was mediated by individual intentionality (mentalizing) competences. The results support the social brain hypothesis by indicating a relationship between PFC volume and social network size that applies within species, and, more importantly, indicates that the relationship is mediated by social cognitive skills.

1. Introduction

Primates have larger brains relative to body size than all other vertebrates [1] and it is now widely accepted that the best functional predictor of relative brain size in non-human primates is social group size [2]. This relationship, according to the ‘social brain hypothesis’ [3,4], is founded on the premise that maintaining cohesion and stability through time in the kinds of bonded social groups characteristic of primates is cognitively very demanding [5–8]. However, among primates, differences in brain size do not reflect proportional increases in all brain regions instead, the size of the neocortex accounts for most of the deviation from overall trend lines [9]. Neocortex size rather than total brain size—or the volume of any other brain region—yields the best correlation with social group size across a wide range of primate species [7]. More importantly, there is evidence to suggest that an increasingly better fit is obtained if the more posterior regions of the neocortex (e.g. the visual areas) are excluded from the analysis (neocortex minus primary visual cortex [10] prefrontal cortex only [11]). The underlying assumption is that the neocortex provides the computational power to manage the complex web of social relationships needed to give a social group its cohesion and stability through time.

While the social brain hypothesis has been tested in some considerable detail with comparative data across a wide range of primate species [8,11–14], the possibility that it might also apply within species has never been seriously considered. However, at least within humans, there is considerable individual variation in social network size [15,16] and there is evidence to suggest that this variance correlates with competence in the kinds of social cognitive abilities (usually known as ‘theory of mind’ or ‘mentalizing’ [17–20]) that are thought to underpin human sociality [21]. (The terms ‘mentalizing’ and ‘intentionality’ are used interchangeably and refer to the ability to assess the intentions and mental states of others.) In addition, recent magnetic resonance imaging (MRI) studies have shown that various indices of social engagement (including number of Facebook friends) correlate with grey matter volume in the amygdala [22] and the classic theory of mind areas in the temporal and frontal lobes [23,24]. In a previous MRI-based study, we were able to show that individual differences in intentionality competence correlate with differences in the volume of the orbital prefrontal cortex (PFC) [25]. Collectively, these studies suggest that there should be a three-way correlation between social network size, intentionality competence and PFC volume. We test this hypothesis here by applying the Cavalieri method of stereology in combination with point counting on structural T1-weighted MR images from a sample of 40 human subjects.

In this analysis, we focus on the PFC for two reasons. First, the PFC is associated with many cognitive functions that are crucial for social interaction, including social information processing [26,27], planning [28], working memory [29], aspects of language and symbolic behaviour [30–32], and attention [33]. Second, neuroimaging studies have identified areas within the PFC as being central to social cognitive processing, including mentalizing (notably the ventromedial PFC [25,34–38]). This does not exclude the possibility that other regions (especially in the temporal lobe) may also be involved, but we emphasize the PFC simply because it provides a simple yet particularly relevant test case for an analysis that is very time-consuming to do. We used the Cavalieri method in combination with point counting to estimate the volume of four subfields of the PFC (i.e. dorsal and orbital prefrontal regions in each of the two hemispheres). The procedure used to estimate PFC subfields has been repeatedly used [25,39,40] and has been shown to have high test–retest reliability. The logic of the social brain hypothesis is that neural volume determines social cognitive competences, and that it is these in turn that determine social network size. As we have already shown that the first step in this sequence holds [25], we use path analysis on the same sample of subjects to determine whether this relationship extends to social network size as well in a three-step causal model, and then whether the relationship between brain region volume and network size is direct (brain volume and intentionality competences independently influence network size) or indirect (brain volume determines intentionality competence, and these in turn determine network size). The social brain hypothesis implicitly assumes the second.

2. Material and methods

(a) Participants

The original cohort comprised 42 subjects with no history of neurological illness. However, two subjects were subsequently removed due to excessive head movement during MR data acquisition, making an examined cohort of 40 subjects aged between 18 and 47 years (17 males: mean age = 25.47, s.d. = 2.87 23 females: mean age = 25.78, s.d. = 7.83). All participants gave written informed consent. This is the same cohort of subjects as that used by Lewis et al. [24].

(b) Social network size

Prior to scanning, participants were asked to complete a version of the social network questionnaire designed by Stiller & Dunbar [21] to capture the size of the inner layers of an individual's social network (following [15,41]). This involved listing the initials of everyone the participant had personal contact with or communicated with over the previous 7 days, excluding professional contacts (doctor, shopkeepers and teachers) unless the contact could explicitly be considered a genuine social interaction. This has previously been shown to be a reliable way of estimating the size of the inner layers (referred to as clique and sympathy group size) of social networks and has been used in several published studies [15,16,41]. The layers of the social network have been found to scale with each other on a consistent ratio [15,16], such that an index of any one layer in the network is a reasonable estimator for the network as a whole. Intentionality competence scores for the subjects are taken from Powell et al. [25] details outlining the approach used to assess intentionality competence and calculate intentionality score are given therein.

(c) Magnetic resonance image acquisition and analysis

MRI data were acquired using a Siemens Trio 3.0 T whole-body MRI system, with an eight-channel head coil. High-resolution anatomical whole-brain images were obtained using a T1-weighted three-dimensional gradient-echo (i.e. 3D modified driven equilibrium Fourier transform) pulse sequence, with the following parameters acquired in the sagittal plane: T1 = 190, TR = 7.92, TE = 2.48, FOV = 224 × 256, matrix 256 × 256 × 256 mm 3 pixels, flip angle 16°. These structural MR images were used for deriving volume estimates of prefrontal subfields. After MR acquisition, datasets were imported into B rain V oyager software for re-alignment (www.Brainvoyager.com, Brain Innovation, Maastricht, The Netherlands), which required reformatting the image and orienting it to a standardized sagittal plane orthogonal to the bicommissural plane. Details of the reformatting procedure are described elsewhere [39,42].

(d) Parcellation of prefrontal cortex anatomical subfields

The protocol used to demarcate the PFC into anatomically defined subfields is based on the previously established methodology developed by Howard et al. [39] this involves dividing the right and left PFC into dorsolateral (DL), dorsomedial (DM), orbitolateral (OL) and orbitomedial (OM) regions, yielding a total of eight PFC subfields. The bicommisural plane was used to delineate orbital and dorsal subfields. Demarcation of medial from lateral regions used the first axial slice superior to the olfactory sulcus. The genu of the corpus callosum, viewed at sagittal midline, formed the posterior boundary of the DL and DM regions. These provided fixed boundaries, which were marked with editing tools in B rain V oyager software. The posterior boundary of the orbital subfield was visualized by the rater during point counting. For complete details of the demarcation procedure, refer to earlier studies [25,39]. Based on the reviewed evidence on the role of the PFC in social information processing, this study examined both the neocortex volume (grey matter) and white matter volume as a combined volume within the PFC. Whole hemisphere volume was calculated to correct for the effect of relative volume in the model.

(e) Cavalieri method

Following parcellation of PFC anatomical subfields, MR images were imported into E asy M easure software [43,44] for point counting and volume estimation. The technique provides mathematically unbiased volume estimators for which their precision can be computed by applying an error-prediction formula [45–49]. Therefore, volume estimates with comparable levels of precision may be derived for different anatomical structures (see Baddeley & Jensen [50] and references therein). A description of the principles and application of this method to PFC have been reported previously [39]. The technique involves exhaustively ‘sectioning’ (non-invasively) the structure of interest with a series of parallel planes a fixed distance apart, with the first section at a uniform random position within the sectioning interval. A grid of test points for point counting is overlaid with uniform random position and isotropic orientation on every alternate section, beginning randomly on either the first or second section on successive subjects. A grid size of 6 × 6 pixels unit area (36 mm 2 area per point) was used for DL and DM subfields, and a grid size of 4 × 4 pixels unit area (16 mm 2 ) was used for OL and OM regions. Shape coefficients were applied to estimate the corresponding coefficients of error the estimated shape coefficients for DL, DM, OL and OM were 5.65, 5.99, 5.48 and 5.19, respectively. For whole hemisphere volume, a larger grid of test points (i.e. 8 × 8 pixels = 64 mm 2 area per point) on every fifth slice was chosen to accommodate the larger volume of interest and reduce labour intensity during point counting.

The reproducibility and repeatability of the method was assessed by Howard et al. [39] and Cowell et al. [40] across healthy and clinical samples, yielding acceptable limits for total and regional prefrontal volume. For instance, predicted coefficients of error of the volume estimator, based on our data from individual volume estimates, were less than 5 per cent, and regional subfield volume estimates for PFC derived from point-counting methods showed satisfactory 95 per cent limits of agreement between raters (reproducibility between raters) and within rater (repeatability by the same rater). Intraclass correlation coefficients ranged from 0.92 to 0.95 between raters and from 0.93 to 0.99 within rater.

Volume estimates for orbital and dorsal PFC were obtained as the sum of the volumes of the lateral and medial subfields for each hemisphere.

(f) Statistical analysis

Path analysis was performed to test the association between network size, intentional competence, orbital PFC volume and dorsal PFC volume. To do this, two linear regression models were performed. In the first model, network size was the outcome variable, and predictor variables were intentionality score, orbital PFC volume and dorsal PFC volume. In the second model, intentionality score was the outcome variable, and the predictor variables were orbital PFC volume and dorsal PFC volume. Total hemisphere volume was included in the linear regression models, both when group size was the outcome variable and when intentionality was the outcome variable. Hemisphere volume was not significantly associated with either intentionality (coefficient = −0.24, p = 0.2) or clique size (coefficient = −0.068, p = 0.7). These results are therefore not included in the path analysis.

3. Results

Descriptive statistics for the main variables used in the analysis (separated by gender) are summarized in table 1. The mean number of contacts (22.65 ± 11.7 s.d.) is within the range of variation reported in previous studies for the sympathy group layer [21,51]. It is likely that the relatively large number of contacts in the present dataset reflects both the way the social contacts questionnaire was phrased and the fact that the sample consists largely of students (who are likely to have an unusually active social life). As in previous studies [21,51], women have larger networks (mean=25.5 ± 13.5 s.d.) than men (mean = 18.8 ± 7.7 s.d.), though the difference is not statistically significant in the present case, albeit close to the boundary (since equality of variance did not hold, the Welch approximation to degrees of freedom was used: t36 = 1.98, p = 0.06, 95%CI: −0.2, 13.5).

Table 1. Mean (±s.d.) for key variables for the total sample and separated by gender.

We used path analysis to explore the most likely causal relationship between the network size, intentionality and PFC volume (figure 1). The results confirm that orbital PFC volume is the best predictor of network size, with this relationship being mediated via intentionality competence. Dorsal PFC was not significantly related to network size (or intentionality competence). The partial correlation between orbital PFC volume and network size is not significant (p = 0.09), and the slope is negative, indicating that PFC volume does not itself influence network size directly but rather is explicitly mediated via intentionality competence, as predicted by the social brain hypothesis. To check for a possible effect of gender on the results, we fitted a multiple linear regression model with network size as the outcome variable, and gender, intentionality, orbital PFC volume, dorsal PFC volume and interaction terms involving gender, brain region and intentionality as predictor variables: there was no significant effect for gender or either interaction term (p > 0.05), indicating that the main relationship between orbital volume, intentional competence and network size is independent of gender.

Figure 1. Path analysis of the causal relationships between the main prefrontal cortex (PFC) divisions, intentional competences and network size. The numbers on each arrow indicate the slope standardized coefficients obtained from the linear regression models and in parentheses the correlations.

4. Discussion

Orbital PFC volume was significantly correlated with the number of weekly social contacts that an individual had, with this relationship being mediated by mentalizing (i.e. intentionality) competence. The results of this study add important support to the social brain hypothesis by demonstrating that the relationship between brain size and social group size applies not just between species, but even at the level of the individual within species: those individuals with larger orbital PFC have a greater number of weekly (i.e. regular) social contacts (an index of the size of their entire social network). Enlarged frontal lobes, such as are found in humans, are of relatively recent evolutionary origin [8]. More importantly, however, our results confirm an explicit assumption of the social brain hypothesis: that this relationship is mediated by social cognitive abilities. This thus adds an important rider to how we formulate the social brain hypothesis by reminding us that it is not simply a direct causal relationship between brain volume and social group size.

Previous studies have shown an effect of gender on measures of network (or social group) size [14,21,41]. However, the present study did not, although the gender difference in network size is close to the boundary of significance (p = 0.06) when no other factors were taken into account. While our result should not, perhaps, preclude the possibility of a gender effect with a larger sample, the existence of such an effect does not obviate the fact that there is a generic relationship between social network size and PFC volume that applies across both genders (as was also the case in all the cited studies).

The neuropsychological literature suggests a functional distinction between dorsal PFC, which mediates higher-order cognitive functions, and orbital PFC, which is implicated in mood, affective behaviour and social aspects of cognition [52]. This functional division is supported by findings from individuals with dorsolateral PFC lesions, who have been reported to have poorer performance on various subtasks of standardized intelligence tests, as well as deficits in memory, including the Wisconsin Card Sorting Task, and the formation and modification of abstract concepts [53–55]. In contrast, damage to orbital PFC typically leads to impulsive aggressive behaviour [56], as well as impairments of social cognition [57] and risk judgement [58,59]. In healthy adults, the orbitomedial PFC is involved in a range of functions, including affect [60] and olfaction [61]. Our previous study using the same sample of subjects [25] demonstrated a relationship between intentionality capacity (a social cognitive competence) and orbital PFC volume, but no association between short-term memory capacity and orbital PFC volume, suggesting that orbital PFC may be particularly important for high-order social cognition (i.e. orders of intentionality greater than conventional theory of mind).

A previous neuroimaging study reported a relationship between an index of sociality and the volume of the corticobasolateral complex of the amygdala [22], a subcortical unit known to be correlated with sociality in primates [62], but not between sociality and hippocampus volume. This study did not explicitly test for a relationship with neocortex volume, though weak correlations (uncorrected at p = 0.05) were reported from an exploratory analysis of cortical thickness (as opposed to cortical unit volume) in three selected cortical regions of interest (subgenual anterior cingulate cortex, caudal superior frontal gyrus, caudal inferior temporal gyrus), none of which overlap with PFC. Unlike the other two units of the amygdala, the corticobasolateral complex has a direct neural connection to the orbitofrontal cortex, although this may have more to do with inhibition of amygdala emotional responses by the orbitofrontal cortex [63,64]. Similarly, using voxel-based morphometry (VBM), Kanai et al. [23] reported correlations between several indirect indices of online and offline social network size and the volumes of the amygdala and two temporal lobe regions known to be associated with theory of mind competence (temporo-parietal junction and superior temporal sulcus), but did not explicitly test for PFC associations. However, just such a relationship has in fact been demonstrated using VBM [24]. Similar findings relating individual differences in both social group size and social rank to the volumes of subregions of the frontal and temporal lobes have recently been reported for macaques [65], suggesting that these relationships may apply widely across the primates. Our analyses extend these findings by demonstrating that these relationships are in fact explicitly mediated by social cognitive competences (in the case of humans at least, intentionality competence). We argue, and our path analysis convincingly demonstrates, that the causal logic involved is that network size is ultimately determined by social cognitive competences, and that these in turn are underpinned by the amount of neural matter that can be brought to bear on the computational demands involved (reflected here in the volume of key PFC brain regions). Our results thus establish the important point (so far missing in all discussions of the social brain hypothesis) that cognition mediates the relationship between brain volume and social group/network size. As with other imaging studies [22–24], our study is neutral as to the formal causal direction of the relationship between PFC volume and social network size. Since it is known that the speed with which children acquire theory of mind is determined by the size of their sibship [66,67], it is possible that the (developmental) causal arrow runs from network size via intentional competences to PFC volume. Whether (and to what extent) PFC places direct constraints on network size will ultimately require longitudinal studies to determine.

Overall, these studies support both the claim that sociality is cognitively demanding and the claim that the evolution of brain size has been influenced by these demands (see also [6]). Our study specifically adds to this body of literature by demonstrating that social network size (a direct measure of sociality) is related to an increase in relative orbital PFC volume. Understanding the association between increased brain volume and domains of sociality, such as social network size, brings us some way towards understanding the mechanisms that have contributed to the enhanced brain volume of our species.


DISCUSSION

Across species, the OFC plays a crucial role in evaluating relationships between stimuli and predicted outcomes (Murray, O'Doherty, & Schoenbaum, 2007), but its functional organization remains unclear. Here, we investigated a theory of valence selective processing (Kringelbach & Rolls, 2004) by directly recording neural activity and found no consistent selectivity for either valence across the orbital cortex. Many neurons encoded valence, but these were intermingled anatomically. Instead, we found distinctions among orbital areas in coding stimuli and feedback of the same valence and coding the reward bar or trial number within a block. These results suggest that orbital cortex is not organized according to the valence of information being processed but rather by different types of value computations being performed by the different subregions.

Positive and Negative Valence Processing

Positive and negative outcomes can be operationally identified as those that an animal will work to obtain or avoid (Seymour, Singer, & Dolan, 2007). Here, subjects learned arbitrary stimulus-response mappings to obtain or avoid losing secondary reinforcers. To learn a correct response to a positive picture, subjects must have been motivated only to obtain secondary reinforcement, because there was never a potential for loss on these trials. Likewise, on negative picture trials, subjects were motivated to avoid a loss, because these pictures were never associated with gains. Although it is true that completing a negative picture trial advanced the subject within a block, the same result could be obtained without learning the stimulus-response mapping and executing an arbitrary response, because no penalties other than loss of a secondary reinforcer were imposed. Therefore, loss of a secondary reinforcer was an aversive outcome that they learned to avoid. Similar results have been obtained using paradigms studying competitive games. Gain and loss of secondary reinforcers had approximately equal and opposite effects on monkeys' choices in a mixed-strategy game, such that gains were rewarding and losses were aversive (Seo & Lee, 2009).

Despite this, we found no evidence that mOFC preferentially responded to rewards or that IC responded to punishment. This discrepancy between our results and the theory of valence selectivity in orbital cortex could have a number of causes. The theory is based on imaging studies in humans (primarily fMRI), whereas our data consist of single neuron activity recorded in monkeys. It is impossible to rule out species' differences in function or anatomy. However, current evidence suggests remarkable homology between human and monkey OFC (Jbabdi, Lehman, Haber, & Behrens, 2013 Wallis, 2012 Mackey & Petrides, 2010). In contrast, the relationship between the fMRI BOLD signal and spiking activity is less straightforward (Sirotin & Das, 2009 Goense & Logothetis, 2008 Logothetis, Pauls, Augath, Trinath, & Oeltermann, 2001). In addition, recent neuroimaging studies have cast doubt on a valence-based organization (Elliott et al., 2010). For example, mOFC BOLD correlates with both appetitive and aversive food values (Plassmann, O'Doherty, & Rangel, 2010) or monetary wins and losses (Tom, Fox, Trepel, & Poldrack, 2007). Finally, one neurophysiology study did find separate groups of neurons responding to appetitive and aversive stimuli, but they were in close proximity to one another within the most posterior regions of mOFC and ventromedial PFC (Monosov & Hikosaka, 2012). We did not find similar distinctions however, our recordings were anterior to these sites. Overall, our results confirm emerging data suggesting that orbital subregions are not organized according to the valence of information they process.

Nonetheless, there is reason to believe that different neural circuits should be involved in evaluating appetitive and aversive stimuli. In particular, these stimuli have opposite effects on behavior. Rewarding stimuli promote approach responses punishing stimuli encourage avoidance (Huys et al., 2011). This is illustrated by classic studies of rats, who easily learn to press a lever for reward but have difficulty learning to lever press to avoid a footshock (Bolles, 1970). In contrast, they readily learn to run or jump to a safe location to escape the same shock. Thus, learning based on negative stimuli is hampered when the necessary response is approach (lever press) and facilitated when the necessary response is avoidance (run away, hide, withhold responding), suggesting that dissociable circuits in the brain are prepared to learn certain responses to oppositely valenced stimuli (also see Hershberger, 1986). It stands to reason that this valence separation could occur in higher brain areas. Whereas this was not the case in orbital cortex, neurons in a region of subgenual anterior cingulate cortex appear to respond selectively to negative values (Amemori & Graybiel, 2012). Such neurons may play a role in enabling different behavioral “modules” (Amemori, Gibb, & Graybiel, 2011), such as avoidance responses, through connections with the striatum (Eblen & Graybiel, 1995). In contrast, orbital areas project to the ventral striatum (Haber, Kunishio, Mizobuchi, & Lynd-Balta, 1995 Selemon & Goldman-Rakic, 1985), important in value-based learning and decision making. In this circuit, hard-wired valence-specific responses may be undesirable. Whereas valence-specific motor responses can safeguard an organism from approaching potential harm, decision making is more nuanced and often involves accepting an undesirable cost to obtain a desired benefit.

Functional Differences between Orbitofrontal Areas

Although our results did not reveal valence-specific processing in orbital regions, we did observe interesting area differences. Few mOFC neurons encoded any variables assessed, suggesting that it was relatively less engaged by the task. The precise functions of mOFC are unclear, but recent data point to a role in comparing option values (Noonan et al., 2010), or coding internal motivational values, particularly in the absence of external prompts (Bouret & Richmond, 2010). In either case, mOFC may not have been engaged by the present task, because subjects were not presented with stimulus choices and all responses were cued.

Neurons in both OFC and IC were active during the task and showed interesting distinctions. First, OFC tracked progress through trial blocks but did not encode the size of the reward bar only IC showed appreciable encoding of bar size. Second, single OFC neurons encoded pictures and feedback with the same valence, and at the time of feedback, they lost selectivity when an unexpected outcome occurred. In the present task, whether an outcome is unexpected is confounded with whether it is the less preferred outcome, and further studies are needed to resolve this issue. However, a previous study found that rat OFC neurons show little or no response to unexpected rewards as well as unexpected omission of reward (Takahashi et al., 2009), supporting the view that it is the degree to which an outcome is expected or not that accounts for our results.

We interpret the overall pattern of encoding in OFC as follows. Although OFC is associated with reward processing, reward-related responses can be heavily dependent on the task in which the subject is engaged (Luk & Wallis, 2013). Such observations support the idea that OFC uses knowledge about the task structure and environment to make outcome predictions (Jones et al., 2012 Takahashi et al., 2011). From this view, OFC neurons with consistent valence encoding may play a role in predicting feedback based on sample pictures. Selectivity at feedback time represents the realization of these predictions, and when they are not met, selective coding diminishes. Finally, OFC tracking of trial number may reflect knowledge of the task structure or predictions about when primary reward will be obtained.

However, this raises the question as to why OFC neurons showed only weak encoding of the reward bar, because it predicts the amount of juice to be delivered. One explanation compatible with this account is that the bar is interpreted as an outcome rather than as an outcome-predictive cue. Supporting this, our behavioral analysis showed that the effect of the bar size on subjects' behavior was independent of how close they were to exchanging it for juice. If the bar were treated as a prediction of juice, one would expect that it should be temporally discounted like other reward-predicting stimuli. A related interpretation is that, although the bar reinforced subjects' behavior, its value (how much juice it predicted) was independent of the task at hand. Supporting this, the value of the bar was tracked preferentially by IC neurons, and it was precisely in this population that we observed a significant loss of value-related encoding during juice delivery. We believe that this happened because the amount of juice was fully predicted by the reward bar, and it was the bar, not the delivery of juice, that reinforced specific behaviors in the task.

In contrast to OFC, IC neurons strongly encoded the reward bar and showed no evidence of consistent coding between sample and feedback epochs. These observations suggest that IC neurons do not perform the same functions as OFC. IC receives highly processed sensory information from temporal and parietal cortex (Petrides & Pandya, 2002 Carmichael & Price, 1995 Cavada & Goldman-Rakic, 1989), and lesions impair the use of visual information to guide motor responses and behavioral strategies (Baxter, Gaffan, Kyriazis, & Mitchell, 2009 Bussey, Wise, & Murray, 2001, 2002). IC may play a role in attention processes (Kennerley & Wallis, 2009) or in determining the behavioral significance of stimuli (Rushworth et al., 2005). As such, IC neurons may assign meaning to stimuli such as the reward bar.

Conclusion

In contrast to a commonly held notion that medial and lateral orbital areas process positive and negative stimuli, at the single neuron level, we found no evidence that orbital processing is organized with respect to valence. Instead, we report differences between distinct orbital regions in how they represent aspects of the task that support the view that different areas use value information in markedly different ways.