Abstract

People tend to assume that outcomes are caused by dispositional factors, for example, a person's constitution or personality, even when the actual cause is due to situational factors, for example, luck or coincidence. This is known as the “correspondence bias.” This tendency can lead normal, intelligent persons to make suboptimal decisions. Here, we used a neuropsychological approach to investigate the neural basis of the correspondence bias, by studying economic decision-making in patients with damage to the ventromedial pFC (vmPFC). Given the role of the vmPFC in social cognition, we predicted that vmPFC is necessary for the normal correspondence bias. In our experiment, consistent with expectations, healthy (n = 46) and brain-damaged (n = 30) comparison participants displayed the correspondence bias during economic decision-making and invested no differently when given dispositional or situational information. By contrast, vmPFC patients (n = 17) displayed a lack of correspondence bias and invested more when given dispositional than situational information. The results support the conclusion that vmPFC is critical for normal social inference and the correspondence bias. The findings help clarify the important (and sometimes disadvantageous) role of social inference in economic decision-making.

INTRODUCTION

Humans tend to make attributions about cause and effect in all manner of situations, a behavior known as “causal attribution.” One way of categorizing causal attributions differentiates between dispositional causes, that is, those attributable directly to the constitution of an object or agent and likely to recur across time and situations, and situational causes, that is, those that are situation-specific, driven by capricious environmental factors and unlikely to recur. In economics, the ability to identify and invest in elements that will cause recurring monetary gain is crucial to successful investment strategies. However, some biases in human cognition, that shape fiscal decision-making, are ill-suited to maximize economic utility. One such potent bias is the tendency to surmise that outcomes are the result of causes that are attributable to an individual, that is, caused by their dispositions, skills, and traits, rather than attributable to environmental causes, that is, situational influences and constraints, luck, and other extraneous factors, and this bias can occur even when outcomes are clearly defined by situational causes. This general phenomenon is known as the correspondence bias (Gawronski, 2004; Gilbert & Malone, 1995).

Cases of gambling addiction exemplify the detrimental effects of misattributing extraneous factors. For example, a big win on a slot machine does not mean one is a “good gambler” or has “figured out” the system; instead one should attribute this windfall to chance. As this example illustrates, the correspondence bias may be a result of making an inferential leap such that situational factors are surmised to indicate dispositional traits.

Along similar lines, the correspondence bias could lead to systematic failure to make optimal decisions when making investments. In fact, poor fiscal decision-making, particularly in the realm of investment, was a major precipitating factor in the severe global economic downturn beginning around 2008. To prevent such economic disasters in the future, it is important to identify economic opportunities that will produce real, sustained gains and avoid those associated with transient, superfluous gains. If the causes of real, sustained economic growth can be identified, then it would make sense to invest in such causes rather than in economic bubbles superfluously associated with gains.

The neural basis of the correspondence bias is not well understood. Several studies have examined the neural correlates of “self-serving attributions,” whereby individuals tend to link positive events to themselves (dispositional causes) and negative events to situational causes (e.g., Miller & Ross, 1975). Among many other regions where activation changes are observed in fMRI and EEG paradigms, regions within the dorsomedial pFC (dmPFC) stand out as particularly important sites associated with causal attribution (Seidel et al., 2010; Krusemark, Keith Campbell, & Clementz, 2008; Mitchell, Cloutier, Banaji, & Macrae, 2006; Harris, Todorov, & Fiske, 2005; Blackwood et al., 2003). Attribution of intention or “theory of mind” has been shown to similarly involve prefrontal cortical regions, including the dmPFC and the region immediately below known as the ventromedial pFC (vmPFC; for a review, see Abu-Akel, 2003). Likewise, moral judgment, a key component of which is attribution of responsibility, has been shown to involve the vmPFC (Koenigs et al., 2007), and so has the judgment of intention, another capacity that is closely tied to the attribution of responsibility (Young et al., 2010). To date, though, the neural basis of the specific attributional phenomenon of the correspondence bias has received little attention in the neuroscience literature. Given the lines of work summarized above and its evolutionary heritage as a center for social evaluation, we speculated that the vmPFC may be particularly important for the correspondence bias.

The medial pFC has been established as an important neural platform for social cognition (Forbes & Grafman, 2010; Moretti, Dragone, & Di Pellegrino, 2009; Van Overwalle, 2009; Lieberman, 2007; Shamay-Tsoory, Tibi-Elhanany, & Aharon-Peretz, 2007; Amodio & Frith, 2006; Saxe, 2006; Adolphs, 2003; Anderson, Bechara, Damasio, Tranel, & Damasio, 1999; Damasio, Tranel, & Damasio, 1990); additionally, the vmPFC sector is associated with representing value and future outcomes (Gläscher et al., 2012; Camille, Griffiths, Vo, Fellows, & Kable, 2011; Lin, Adolphs, & Rangel, 2011; Hare, Camerer, Knoepfle, O'Doherty, & Rangel, 2010; Plassmann, O'Doherty, & Rangel, 2010; Sellitto, Ciaramelli, & di Pellegrino, 2010; Moretti et al., 2009; van den Bos, McClure, Harris, Fiske, & Cohen, 2007; O'Doherty, Kringelbach, Rolls, Hornak, & Andrews, 2001). We have hypothesized that the human vmPFC is likely to be involved in inferring social value due to the role of homologous regions in mammals in social evaluation of conspecifics, and we have labeled this proposition the “inferential brain hypothesis” (Koscik & Tranel, 2012a; Koscik, 2010). The notion that the pFC and its subregions are related in some manner to human and primate evolution is prevalent in the literature (Smaers et al., 2011; Preuss, 2007, Semendeferi, Lu, Schenker, & Damasio, 2002; Semendeferi, Damasio, Frank, & Van Hoesen, 1997), especially given the predominant role of the pFC in social cognition (Adolphs, 2009). The “social brain hypothesis” is a notable example (Dunbar, 1998)—the essence being that an increased need for advanced social cognition due to living in larger social groups is the primary driving force for the evolution of the large human brain. Neural homologues to human vmPFC are critical for social evaluation and attributing value to conspecifics in most mammalian species due to their role in chemical communication via the olfactory and vomeronasal systems. In contrast, humans lack robust chemosignaling between individuals and also have strong visual specialization. On the basis of these evolutionary trends, we predict that the vmPFC remains important for social evaluation as in chemosensory-dependent mammals. Instead of chemosensory, perceptual identification, however, humans have shifted to a more computationally expensive form of inference. We hypothesize that the human vmPFC has been repurposed to fill this role. Humans must gather information about others from multiple sources and sensory modalities including extraneous environmental information, which creates the opportunity to utilize the correspondence bias as an attributional heuristic. For these reasons, we predicted that damage to the vmPFC would result in abnormal causal attribution (lack of a correspondence bias).

Our predictions of how the vmPFC contributes to social, causal attribution are consistent with a dual process model of causal attribution (for a review, see Gilbert & Malone, 1995). In this model, all information is initially labeled as dispositional. A secondary, controlled process corrects this initial dispositional attribution for situational constraints. Given that we predict that the vmPFC is involved in assigning values to others, we predict that the vmPFC is necessary for initial, heuristically driven dispositional attribution applied to all incoming information. In contrast, we predict that the vmPFC is not involved in the secondary appraisal process. Thus, damage to the vmPFC will result in a reduction in the bias toward making dispositional attributions, as the first-pass labeling of information as dispositional is damaged. However, the secondary appraisal, making both dispositional and situational attributions, is predicted to remain intact following vmPFC damage, resulting in less biased dispositional and situational attributions.

In our study, participants completed an investment task, in which they could choose to invest in different investors based on either dispositional or situational information. This task allowed us to manipulate these information types in a realistic way, where dispositional information referred directly to the potential investor's attributes whereas situational information did not refer to the potential investor. Simultaneously, this setup allowed us to quantify the use of one type of information versus another, that is, in the amount of money invested. We predicted that typical behavior will reflect the correspondence bias, and participants without damage to the vmPFC would make an heuristically driven inferential leap whereby situational and dispositional information will be considered to be equally diagnostic, and thus, investments would be similar regardless of information type. On the basis of the threads of evidence summarized above regarding the roles of the vmPFC in valuation, attribution, and social cognition more generally, we hypothesized that vmPFC damage would result in a failure to rely on a heuristic response guided by the correspondence bias and would thus differentiate between dispositional information (with which vmPFC patients would invest more) and situational information (with which vmPFC patients would invest less).

METHODS

Participants

We recruited 47 individuals with focal brain injuries from the Patient Registry of the Division of Behavioral Neurology and Cognitive Neuroscience in the Department of Neurology at the University of Iowa. Our target group consisted of participants with damage to the vmPFC group (n = 17; 7 men, 10 women; 2 ACoA aneurysm cases, 4 ischemic stroke cases, 2 resections, 9 benign tumor removals). We also recruited a brain-damaged comparison group whose brain damage did not include vmPFC (BDC group; n = 30; 17 men and 13 women; 24 ischemic stroke cases, 1 resection, 5 benign tumor removals; see Figure 1). Participants whose lesions encompassed mainly either the amygdala or insular cortex were excluded from our sample, given that those structures are known to be important for social cognition and emotion. In addition, we recruited 45 neurologically normal participants (22 men, 23 women) from the Iowa City area as a normal comparison group (NC group; see Table 1). All participants with focal brain damage were recruited from the Patient Registry of the Division of Behavioral Neurology and Cognitive Neuroscience in the Department of Neurology at the University of Iowa. All participants were free of dementia, psychiatric disorder, substance abuse, and significant intellectual impairments. Normal comparison participants were recruited from the Iowa City area through advertisement and were compensated for their participation. All participants provided informed consent before participation in accordance with the Institutional Review Board of the University of Iowa.

Figure 1. 

Lesion overlaps. Lesions of participants in the vmPFC group are shown in warmer colors, where the maximum overlap (15) is within ventromedial prefrontal sector. Lesions in the BDC group cover a much broader swath of cortical territory with a maximum overlap of 4. Purple represents regions of overlap between the vmPFC and BDC groups and covers bilateral dorsal pFC.

Figure 1. 

Lesion overlaps. Lesions of participants in the vmPFC group are shown in warmer colors, where the maximum overlap (15) is within ventromedial prefrontal sector. Lesions in the BDC group cover a much broader swath of cortical territory with a maximum overlap of 4. Purple represents regions of overlap between the vmPFC and BDC groups and covers bilateral dorsal pFC.

Table 1. 

Demographics and Neuropsychology


vmPFC
BDC
NC
n (men, women) 17 (10, 7) 30 (17, 13) 45 (22, 23) 
Handedness (R, L, M) 16, 0, 1 23, 2, 5 41, 4, 0 
Age (years) 60.3 (14.9) 57.5 (13.4) 61.7 (13.3) 
Education (years) 13.7 (1.9)* 14.9 (3.0) 16.2 (2.5) 
Chronicity (years) 12.1 (9.0) 15.0 (12.0) – 
WAIS-R 
 FSIQ 106.6 (11.4) 105.3 (13.4) – 
 PIQ 107.2 (11.3) 105.0 (14.8) – 
 VIQ 105.7 (14.2) 105.7 (16.7) – 
BDI 3.9 (5)* 8.1 (5.8) – 
BAI 4.4 (3.6) 6.0 (4.2) – 
AVLT 13.8 (2.3) 14.0 (1.6) – 
WCST 4.9 (2.1) 5.5 (1.3) – 
Face Disc. 44.7 (3.7) 45.9 (4.2) – 
COWA 39.0 (12.0) 38.5 (12.0) – 
BNT 55.4 (4.2) 54.8 (6.9) – 

vmPFC
BDC
NC
n (men, women) 17 (10, 7) 30 (17, 13) 45 (22, 23) 
Handedness (R, L, M) 16, 0, 1 23, 2, 5 41, 4, 0 
Age (years) 60.3 (14.9) 57.5 (13.4) 61.7 (13.3) 
Education (years) 13.7 (1.9)* 14.9 (3.0) 16.2 (2.5) 
Chronicity (years) 12.1 (9.0) 15.0 (12.0) – 
WAIS-R 
 FSIQ 106.6 (11.4) 105.3 (13.4) – 
 PIQ 107.2 (11.3) 105.0 (14.8) – 
 VIQ 105.7 (14.2) 105.7 (16.7) – 
BDI 3.9 (5)* 8.1 (5.8) – 
BAI 4.4 (3.6) 6.0 (4.2) – 
AVLT 13.8 (2.3) 14.0 (1.6) – 
WCST 4.9 (2.1) 5.5 (1.3) – 
Face Disc. 44.7 (3.7) 45.9 (4.2) – 
COWA 39.0 (12.0) 38.5 (12.0) – 
BNT 55.4 (4.2) 54.8 (6.9) – 

Values are mean (SD). Handedness: R = right-handed, L = left-Handed, M = mixed Handed; WAIS = Wechsler Adult Intelligence Scale-III; FSIQ = Full Scale Intelligence Quotient; PIQ = Performance Intelligence Quotient; VIQ = Verbal Intelligence Quotient; BDI = Beck Depression Inventory-II; BAI = Beck Anxiety Inventory; AVLT = Auditory Verbal Learning Test, 30 min delayed recognition, # correct responses/15; WCST = Wisconsin Card Sorting Task, # of categories completed; Face Disc. = Benton Facial Discrimination Test, raw scores; COWA = Controlled Oral Word Association Test, raw scores; BNT = Boston Naming Test, raw scores (out of 60 maximum). For all neuropsychological measures, references can be found in Tranel (2009).

*p < .05.

Participants were predominantly right-handed, and all were of western European descent. There were no differences between groups in age, F(2, 91) = 0.874, p = .421; there was a small but statistically significant difference in years of education, F(2, 91) = 6.411, p = .003. Post hoc tests revealed that the vmPFC group differed significantly from the NC group in education (p = .003); all other comparisons were nonsignificant (see Table 1).

Brain-damaged groups did not differ in lesion chronicity (i.e., time since lesion onset; t(45) = −0.871, p = .388). All neuroanatomical, neuropsychological, and experimental data were collected approximately contemporaneously specifically, during the chronic epoch of recovery (more than 3 months post onset), where neuroanatomical and neuropsychological profiles are stable. Brain-damaged participants had, for the most part, intact psychometric intelligence, memory, executive functions, and verbal abilities (see Table 1). Moreover, there were no significant differences between the vmPFC and BDC groups for any neuropsychological variables (all ps > .327), except for the score on the Beck Depression Inventory (BDI) where the vmPFC group (M = 3.93, SD = 4.95) had lower BDI scores than the BDC group (M = 8.14, SD = 5.76; t(40) = −2.335, p = .025; see Table 1). However, BDI scores of 13 or lower are classified as “minimal depression,” so neither of the groups had average depression in the clinically elevated range.

Procedure

Participants completed an investment decision-making task, in which they freely chose how much to invest with different investors based on a brief description of the investors' prior investing success. These descriptions contained either dispositional information, for example, “Investors made 10% last year due to his hard-working nature,” or situational information, for example, “Investors made 10% last year due to economic growth in China” (see Table 2). On each trial, participants were sequentially presented with two potential investors with whom they could invest some or all of their money with either or both investors however they saw fit. Participants were required to invest a minimum of $10 on each round.

Table 2. 

Descriptions

Description 1
Description 2
Due to increased rates of higher education Because he is a risk-takera 
Due to a rise in commodity prices Because of the expansion of green technologies 
Due to economic growth in China Because he is goal-orienteda 
Because he is highly motivated and ambitiousa Because of the decrease in trade tariffs 
Because he is a conservative investora Due to a recent scientific breakthrough 
Due to his extensive educationa Due to his hard-working naturea 
Due the rise in internet stock trading Because of a rise in the level of tourism 
Because he is very good at solving problemsa Due to his unbiased opinionsa 
Description 1
Description 2
Due to increased rates of higher education Because he is a risk-takera 
Due to a rise in commodity prices Because of the expansion of green technologies 
Due to economic growth in China Because he is goal-orienteda 
Because he is highly motivated and ambitiousa Because of the decrease in trade tariffs 
Because he is a conservative investora Due to a recent scientific breakthrough 
Due to his extensive educationa Due to his hard-working naturea 
Due the rise in internet stock trading Because of a rise in the level of tourism 
Because he is very good at solving problemsa Due to his unbiased opinionsa 

aConnotes dispositional descriptions.

At the beginning of each trial, participants were instructed to complete an unrelated task to earn money to invest. We had two reasons for requiring participants to work for their investment money: (1) it allowed us to introduce a lag between trials (between 5 and 10 min) to assist in separating decisions in time to potential reduce the impact of prior decisions and (2) investing money that was earned rather than given may increase the incentive to invest this money wisely. Participants completed two versions of each of these tasks in the same order and every four trials received money for free, that is, without completing a task, thus resulting in eight trials in total. These unrelated tasks consisted of a hidden covariation detection paradigm (where perceptual properties of stimuli varied systematically with on-screen information and detection of this hidden covariation was probed), an inhibition of return paradigm (an attentional cueing task where participants respond to cued and uncued targets) and an arithmetic task (consisting of simple mathematic problems). The results of these unrelated tasks are as of yet unpublished. Money that participants were to invest was not contingent on performance of these unrelated tasks, rather participants received $100 in facsimile money to invest for each trial.

After earning money on the unrelated tasks, participants read descriptions of two people with whom to potentially invest. Descriptions included a picture of an individual and varied in terms of either containing dispositional or situational information counterbalanced for order of presentation (see Table 2 for descriptions). Pictures consisted of emotionally neutral, all male, European, middle-aged, frontal view, computer-generated faces (using FaceGen Modeller, Singular Inversions, Inc., Toronto, 2009). Dispositional information included a reference to something internal to the individual with an explicit, direct referent to the investor (i.e., his), whereas situational information referred to circumstances that were external to the individual and beyond their control, without an explicit, direct referent to the investor. For example, dispositional information, “John Smith: Investors made 10% last year due to his hard-working nature,” versus situational information, “Bill Johnson: Investors made 10% last year due to economic growth in China.” The percentage of gain was the same for both descriptions on a given trial, but varied between trials in 5% increments from 5% to 20%, in a pseudorandomized order, such that all descriptions were framed to be positive, that is, potential investees were described to have made money for varying reasons. After reading the descriptions, participants allocated their investment however they wanted by clicking on-screen buttons to add or subtract money from each potential investee, and the remainder was placed in a “bank” for them. A new trial then began with another unrelated task to earn money and begin the process again.

Following all of the investment decisions, participants were asked to predict how well each of their investments would do. Finally, participants were given feedback on how well their investments fared (all were associated with the same gain as in the descriptions that were given) and were asked to decide how much they would like to reinvest if given the opportunity. Three trials were catch trials that contained either only dispositional (2 trial) or only situational (1 trial) descriptions, allowing us to examine these two factors in isolation as well. Given that the limited time volunteered by these participants for research and the high demand for them, time constraints only permitted a limited number of trials.

Our main dependent variable from this task is the difference between the amounts invested when given dispositional information and the amounts invested when given situational information. If participants are susceptible to the correspondence bias, then we would expect the difference between the amounts they invested when given dispositional information and the amounts they invested when given situational information should be reduced. In other words, if situational information is considered to be of equal diagnostic value as dispositional information then participants will invest similar amounts regardless of information type. A more “rational” pattern would be to invest more money when given dispositional information, that is, to capitalize on someone's skill, and avoid situational information, that is, to reduce the influence of less predictable loss. We predicted that both BDC and NC groups would display the correspondence bias and show greater deviation from “purely rational” investments, reduced difference between dispositional and situational investments. In contrast, we predicted that vmPFC damage would result in abnormal social evaluation, demonstrated by a lack of the correspondence bias, and would thus follow the “more rational” pattern of investing proportionally more with dispositional information and less with situational information.

Lesion Analysis

Neuroanatomical analysis was based on MR or CT images obtained in the chronic epoch of recovery. Each brain lesion was reconstructed in three dimensions using Brainvox (Frank, Damasio, & Grabowski, 1997; Damasio & Frank, 1992) and manually warped to a normal template brain using the MAP-3 technique (Damasio, Tranel, Grabowski, Adolphs, & Damasio, 2004). Following manual transfer to the normal template space, the template brain was warped to the MNI152 standard 1 mm T1-weighted atlas (Mazziotta et al., 2001; Collins, Neelin, Peters, & Evans, 1994; Evans, Dai, Collins, Neelin, & Marrett, 1991) to provide a more direct comparison with a large portion of the literature that also uses this standard space. This warping was accomplished using BRAINSDemonWarp (Johnson & Zhao, 2009), which is a high-dimension image registration algorithm that generates displacement vectors for each voxel to define the transform from the moving to fixed image (Thirion, 1998). This transform, from the lesion template to the MNI152 template, was then applied to each of the lesion maps. Lesion maps were then processed with Matlab (r2007b, The Mathworks) to create overlap maps of pertinent participants.

Naturally occurring brain lesions do not respect functional or anatomical boundaries, thus an all-or-none approach to classifying lesions is inappropriate. Likewise, anatomical parcellation schemes give a false impression of distinct, abrupt boundaries between regions; thus, lesions that have their focus in an adjacent, nontarget region may spillover in the periphery of the ROI, but not affect the core of this region to a significant extent. Our solution to this classification problem was twofold. First our recruitment procedures targeted participants with known damage to the vmPFC, our ROI. All but two participants with vmPFC damage had bilateral vmPFC lesions (one patient had unilateral left vmPFC damage and one had unilateral right vmPFC damage). Second, to exclude participants with damage limited to the periphery of our ROI, we set a lower limit on the proportion of damaged voxels for inclusion at 5% of the volume of the vmPFC in either hemisphere. In our sample, of the lesions confined entirely to the vmPFC, the smallest proportion of the vmPFC that is covered is approximately 5.5% (a unilateral lesion confined entirely within left vmPFC); six participants with foci of damage in regions adjacent to the vmPFC have some spillover (with an average 1.1% coverage of bilateral vmPFC, maximum 2.0%). These data were then visualized using MRIcroN (Rorden, 2007, 2008; see Figure 1). Brain damage in the vmPFC group affected significantly larger volumes of neural tissue (M = 4.8, SD = 3.4, expressed as a percentage of all voxels representing neural tissue) compared with the BDC group (M = 2.6, SD = 3.2; U = 131.0, p = .006).

Statistical Analyses

For our primary analysis, we compared vmPFC, BDC, and NC groups on the difference between the amounts of money invested following dispositional descriptions and situational descriptions using a one-way ANOVA procedure. We used Tukey's post hoc test to locate significant between-group differences. To account for between-subject differences in absolute investment behavior (i.e., since some participants tend to invest more than others in general), the amount of money invested with either information type was quantified as a proportion of the total amount invested on that trial.

Given that we observed group differences in Education (i.e., vmPFC had fewer years of education than NCs as reported above), we examined whether or not this affected our results. First, we examined the correlation between Education and our dependent variable. Second, we used an ANCOVA with Education as a covariate to examine the effect of education or our results. Similarly, since we observed group differences in overall lesion size, we examined potential correlation and an ANCOVA to examine whether or not overall lesion size affected our results.

Secondarily, we examined whether or not any group differences were due to differential investments to specific changes to either dispositional or situational investments. For example, if the vmPFC group displayed an increased difference between dispositional and situational investments, this could be due to an increase in the amount dispositional investments alone, a decrease in situational investments alone, or a combination of both. We examined this in two ways: (1) we compared the amounts invested for dispositional and situational information (i.e., the same variables used to calculate the difference score) and (2) we examined the investments made on catch trials where there was no choice between dispositional versus situational information types. The analysis of catch trials was based on absolute investment amount, since calculation of the amount invested by information type as a “proportion” sum, as above, would always result in a value of 1. Since the calculation of values on catch trials does not involve a proportional measure and may partially reflect interindividual differences in investment thresholds, we used an ANCOVA procedure in which the average absolute amount invested was a covariate. To further examine whether or not groups differed in general investment thresholds, we conducted another follow-up analysis comparing absolute investment values including all trial types. This allowed us to determine whether any of the groups were more fiscally conservative or more or less likely to invest their money.

Next, we analyzed whether or not the vmPFC, BDC, and NC groups differed in their predictions of how they felt their investments would yield and their willingness to reinvest money given the outcome of their previous investment. As in our primary analysis, we submitted the difference between dispositional and situational investments to one-way ANOVA, using Tukey's post hoc tests to pinpoint any potential group differences. We also examined whether or not investments differ based on the expected gain as indicated in the description of the investor (i.e., 5%, 10%, 15%, 20%) using one-way ANOVAs at each possible value.

Finally, we performed an exploratory analysis to examine the possibility that there is functional specificity with regards to our results within subregions of the vmPFC. Given that this exploratory analysis relies on a relatively small sample to look for within-group relationships, we limited this analysis to subregions within the pFC where five or more participants in the vmPFC group had any damage to that region. We examined correlations between the extent of damage to a vmPFC subregion and the difference between the amounts invested by information type. Given the exploratory nature of this analysis, we report significant findings at a liberal, α < 0.1, threshold; moreover, we only report vmPFC subregions with a significant relationship.

RESULTS

Our primary analysis of the difference between dispositional and situational investments revealed significant differences between groups, F(2, 91) = 8.461, p < .0005, η2 = 0.160. Post hoc tests revealed that the vmPFC group had a significantly greater difference in investments by information type compared with the BDC group (p < .0005, 95% CI [0.107, 0.405]) and the NC group (p = .006, 95% CI [0.045, 0.326]; see Figure 2). As predicted, participants with vmPFC damage invested more money when given dispositional information than when given situational information. In contrast, comparison groups did not vary investments by information type, a finding that can be taken to indicate the correspondence bias.

Figure 2. 

Difference in proportional investment by group. Bars represent the difference in the proportion of money invested when given dispositional information minus situational information. Participants with vmPFC damage displayed a greater difference than the other groups.

Figure 2. 

Difference in proportional investment by group. Bars represent the difference in the proportion of money invested when given dispositional information minus situational information. Participants with vmPFC damage displayed a greater difference than the other groups.

A closer examination of these data demonstrates that the difference observed between groups in our main analysis was due to the behavior of the vmPFC group, which had both increased investment for dispositional information, F(2, 91) = 9.055, p < .0005, η2 = 0.169, and decreased investment for situational information, F(2, 91) = 6.473, p = .002, η2 = 0.123 (see Figure 3). However, when we looked at catch trials only, that is, when there was no choice between the different information types, there were no differences between groups for dispositional-only, F(2, 88) = 0.656, p = .522, or situational-only, F(2, 88) = 0.572, p = .567, information (see Figure 4). Group differences were present only under circumstances where a choice between dispositional and situational information was necessary. Thus, we did not observe a systematic avoidance of or attraction to either information type in any of our groups. Moreover, this suggests that participants do not necessarily differ in their understanding of the information types per se; rather, they differ in their use of this information when presented with differing options. It is possible that the correspondence bias does not become activated unless a choice between dispositional and situational information types is required. Future studies explicitly designed to address this possibility should help shed additional light on this issue.

Figure 3. 

Proportional investment by information type. Bars represent the proportion of money invested with dispositional or situational information types. Participants with vmPFC damage invested more with dispositional information and less with situational information compared with the other groups.

Figure 3. 

Proportional investment by information type. Bars represent the proportion of money invested with dispositional or situational information types. Participants with vmPFC damage invested more with dispositional information and less with situational information compared with the other groups.

Figure 4. 

Investments for catch trials. Bars represent the amount invested for dispositional and situational information types on trials where only one of these types of information was given. There are no differences between groups.

Figure 4. 

Investments for catch trials. Bars represent the amount invested for dispositional and situational information types on trials where only one of these types of information was given. There are no differences between groups.

In addition, the absolute amount invested with either information type (as opposed to the proportion of the total amount invested as presented above) did not differ between groups [dispositional: (F(2, 89) = 0.371, p = .691), vmPFC mean = 24.75 SD = 9.93, BDC mean = 25.67 SD = 12.60, NC mean = 23.36 SD = 11.39; situational: (F(2, 89) = 2.129, p = .125), vmPFC mean = 20.50 SD = 8.66, BDC mean = 28.27 SD = 14.21, NC mean = 24.49 SD = 12.71]. Given the relatively large between-subject variability in absolute investment as demonstrated by the relatively large standard deviations, our comparison of dispositional and situational investments as a proportion of the total amount invested helped control for this variability in our main analysis. Importantly, the amount not invested (i.e., banked by the participant) did not differ between groups (vmPFC mean = 77.11 SD = 8.97, BDC mean = 73 SD = 12.84, NC mean = 76 SD = 11.61; F(2, 89) = 0.820, p = .444). This suggests that the groups did not differ in terms of fiscal conservatism and likelihood to invest their money, despite between-participant variability in this regard. In other words, fiscal conservatism was not specific to one group.

Education did not account for performance on the task. Years of education was not correlated with the difference in the amount invested between dispositional and situational information types (r = −0.065, p = .535). Moreover, an ANCOVA analysis revealed no significant effect of Education, F(1, 88) = 0.012, p = .914, and the Group effect remained, F(2, 88) = 8.153, p = .001.

Overall, lesion size was correlated with the difference in the amount invested by information type (r = 0.367, p = .011). However, including lesion size as a covariate in our ANCOVA did not alter the Group (lesion location, vmPFC vs. BDC) effect, F(1, 43) = 8.033, p = .007. There was a weak trend toward lesion size having an effect, F(1, 43) = 3.152, p = .083, but there was no Group × Lesion Size interaction, F(1, 43) = 1.190, p = .281. Overall, lesion size was less important than lesion location.

Additionally, there are no group differences in the amount of money invested depending on expected gain (i.e., the amount of money investors were stated to have earned, max F(2, 91) = 1.726, all ps > .184. We observed no difference between groups in their predictions of how well their investments would do, F(2, 91) = 0.380, p = .685, nor did we observe any differences in reinvestment following feedback, F(2, 91) = 0.267, p = .767.

Finally, our exploratory analysis revealed several subregions within the vmPFC that may be of particular importance (see Figure 5). In the left hemisphere, the subgenual cingulate cortex (r = 0.450, p = .069) and straight gyrus (r = 0.472, p = .056) were related to larger differences in the amount of money invested by information type. No other regions in either hemisphere were related to larger differences (all ps > .1). However, in the right hemisphere, the lateral OFC displayed a negative relationship with the difference in the amount of money invested by information type (r = −0.413, p = .099).

Figure 5. 

Specificity of vmPFC subregions. Each voxel represents the group average difference in the proportion of money invested with dispositional and situational information (dispositional–situational) for participants whose lesion includes that voxel. Warmer colors represent a larger difference. Only voxels where the number of lesion overlaps is greater than 5 are included.

Figure 5. 

Specificity of vmPFC subregions. Each voxel represents the group average difference in the proportion of money invested with dispositional and situational information (dispositional–situational) for participants whose lesion includes that voxel. Warmer colors represent a larger difference. Only voxels where the number of lesion overlaps is greater than 5 are included.

DISCUSSION

The findings support our prediction that damage to the vmPFC would result in a deficit in normal causal attribution and that this would translate into abnormal (advantageous, under the circumstances rigged up in our study) investment decisions. Participants with vmPFC damage were less likely to exhibit the correspondence bias, investing proportionally more money when given dispositional information than when given situational information. In contrast, both brain-damaged and neurologically normal comparison participants exhibited the correspondence bias in their responses by investing similar amounts of money regardless of information type.

In terms of social inference, one way of interpreting the results is that comparison participants made an inference that the situational information reflected some dispositional trait. For example, given the description that gains were “due to economic growth in China,” perhaps comparison participants inferred that this investor was knowledgeable enough to leverage this external factor. Effectively, this would lead to the inference that the situational information reflected a positive, dispositional value. Participants with vmPFC damage did not appear to make this inferential leap and thus invested differentially by information type. This interpretation is consistent with the view that the vmPFC is an important substrate for encoding value from stimuli from various modalities (Rolls, 2000), particularly in relation to social decisions (e.g., Hare et al., 2010).

Interestingly, part of the vmPFC (mesial OFC) is involved in encoding primary gustatory and olfactory rewards (for a review, see Rolls, 2000) and more generally is involved in perceptual chemosensory processing (Gottfried & Zald, 2005). We suggest that brain regions that are necessary for social inference in humans descend from those used for social evaluation in mammals. In mammals, these social evaluations rely heavily on chemosensation. In humans, however, these social evaluations rely very little, if at all, on chemosensation, and instead require inferences drawn from less reliable sources of information, including multiple sensory modalities, over multiple instances, where interference is more easily introduced by factors such as deception, misperception, misapprehension, and loss/lack of attention. This could mitigate the reliability and consistency of pertinent information (Koscik & Tranel, 2012a; Koscik, 2010). Consequently, the vmPFC may be a critical structure involved in social inference, reflecting its conserved role in social evaluation but repurposed for inferential processing. Consistent with this idea, we have demonstrated that damage to the vmPFC leads to a deficit in transitive inference (Koscik & Tranel, 2012b), which, in combination with the current findings, strongly suggests a role for the vmPFC in social inference.

An alternative interpretation of these results is that vmPFC damage resulted in a deficit in utilizing gain information within each trial, instead forcing them to rely on the other forms of information present (i.e., dispositional or situational information). On the other hand, the BDC and NC groups with presumably normal abilities to utilize gain information used this information only, thus investing evenly regardless of situational and dispositional information since expected gain was held constant. However, this account is not parsimonious as expected gain did not seem to have a major affect on participant's investments nor were there group differences in investment based on expected value or predicted outcome.

We also find it quite intriguing that, under the conditions of the experimental protocol used in the current study, patients with vmPFC damage actually made more “rational” investments—investments that were, in fact, economically advantageous. When making real-life investment decisions, it is important to dissociate skill, which may relate to increased likelihood of replicating any gains in the future, and luck, which is unlikely to reproduce gains unless uncontrollable elements of the situation remain constant. The conditions in our experiment were controlled such that all information was positive, that is, all descriptions, either dispositional or situational, involved a gain on the potential investment as did the outcomes of the investments. Under these conditions, comparison participants made less advantageous decisions than did the patients with vmPFC damage. An interesting parallel to this finding was reported by Shiv, Loewenstein, Bechara, Damasio, and Damasio (2005), who found, using a paradigm that was constructed such that long-term outcomes were fiscally positive, that patients with damage to the vmPFC had better fiscal outcomes than did normal, healthy participants (interpreted as due to absence of normal emotions in the vmPFC patients, in the experiment used by Shiv et al.).

If the pattern of investing more when given dispositional information were to be replicated under conditions where money is lost on the investments, the vmPFC participants would likely lose more money than the comparison participants by investing too much with partners who are likely to lose more in the future. This interpretation is consistent with the ” “error management theory” of Haselton and Buss (2000). Essentially the correspondence bias did not evolve in normal individuals as a means to maximize gains; rather introducing this systematic bias reduces the likelihood of falling prey to large, consequential errors and losses at the cost of absorbing small but less consequential losses. We constrained our initial exploration into this topic to positive gains only, to reduce the impact of the potentially confounding influence of loss aversion, for which the vmPFC has been implicated (Tom, Fox, Trepel, & Poldrack, 2007), and a definitive answer to this question awaits extrapolation to investment losses. Additional research is needed to explore the role of causal attributions when information is negative, to examine whether or not this strategy helps normal adults avoid potentially greater losses.

Our data are consistent with the proposal that vmPFC is part of a neural system important for automatic identification of emotional significance from stimuli (Phillips, Drevets, Rauch, & Lane, 2003). Damage to the vmPFC disrupted a portion of the system where automatic, obligatory, dispositional attributions are made for all incoming information. In contrast, the neural systems responsible for making controlled, regulatory appraisals to correct for initial biases were left intact—such systems may include the dmPFC as a critical component (Phillips et al., 2003). It is important to note that additional brain regions are most likely involved in a larger neural system necessary for achieving the complex processing necessary for successful social evaluation, although a full survey is well beyond the scope of our discussion. Given that vmPFC and dmPFC may be critical, complementary nodes in the neural systems necessary for normal attributional processes, it is intriguing that the dmPFC has been implicated in social attribution as well. For example, Mitchell and colleagues (2006) report that dmPFC is activated for both trait diagnostic and nondiagnostic information when explicitly forming impressions; however, when participants were not explicitly forming impressions, the dmPFC was activated to a greater extent for trait diagnostic information. These findings are consistent with the idea that people automatically make dispositional attributions (involving the vmPFC) and only engage in deeper processing to include situational information as an explicit, controlled process (involving the dmPFC; for a review, see Gilbert & Malone, 1995). Assuming that the regulatory process is inadequate to completely correct for initial biases, normal behavior constitutes the correspondence bias. In the case of vmPFC damage, no initial, dispositional attribution is made, leaving only the more accurate controlled process and thus more accurate attributions-and, in something of a paradox, no correspondence bias.

The vmPFC is not a unitary structure, but rather, encompasses a large swath of cortical territory and includes several cytoarchitectonically distinct subregions. Our exploratory analysis suggested that subgenual and medial orbitofrontal regions (including the straight gyrus) may be two intriguing target regions as opposed to lateral OFC, which may be unrelated to attributional processes. These results are preliminary, due to small sample sizes and uneven distribution of lesions across subregions, and further research is needed to address more definitively whether certain subregions of the vmPFC are especially important for various aspects of attributional processes.

In conclusion, we observed that the vmPFC was critical for making normal social attributions in the context of investment decisions. Participants with vmPFC damage actually made more economically “rational” decisions by investing more following trait diagnostic, dispositional information than when given nondiagnostic, situational information. Overall, given that normal healthy persons tend to overuse situational information, our data suggest that people could benefit from being aware of the correspondence bias and carefully reconsider their evaluations of positive information to optimize their fiscal decisions.

Acknowledgments

This work was supported by the National Institute of Neurological Disorders and Stroke (D. T., P50 NS19632), the National Institute on Drug Abuse (D. T., R01 DA022549), and the Natural Sciences and Engineering Research Council of Canada Postgraduate Scholarship (T. R. K., NSERC PGS-D).

Reprint requests should be sent to Timothy R. Koscik, Department of Psychology, University of Toronto, 100 St George Street, Toronto, ON M5S 3G3, or via e-mail: t.koscik@utoronto.ca.

REFERENCES

REFERENCES
Abu-Akel
,
A.
(
2003
).
A neurobiological mapping of theory of mind.
Brain Research Reviews
,
43
,
29
40
.
Adolphs
,
R.
(
2003
).
Cognitive neuroscience of human social behaviour.
Nature Reviews Neuroscience
,
4
,
165
178
.
Adolphs
,
R.
(
2009
).
The social brain: Neural basis of social knowledge.
Annual Review of Psychology
,
60
,
693
.
Amodio
,
D.
, &
Frith
,
C.
(
2006
).
Meeting of minds: The medial frontal cortex and social cognition.
Nature Reviews Neuroscience
,
7
,
268
277
.
Anderson
,
S.
,
Bechara
,
A.
,
Damasio
,
H.
,
Tranel
,
D.
, &
Damasio
,
A.
(
1999
).
Impairment of social and moral behavior related to early damage in human prefrontal cortex.
Nature Neuroscience
,
2
,
1032
1037
.
Blackwood
,
N.
,
Bentall
,
R.
,
ffytche
,
D.
,
Simmons
,
A.
,
Murray
,
R.
, &
Howard
,
R.
(
2003
).
Self-responsibility and the self-serving bias: An fMRI investigation of causal attributions.
Neuroimage
,
20
,
1076
1085
.
Camille
,
N.
,
Griffiths
,
C.
,
Vo
,
K.
,
Fellows
,
L.
, &
Kable
,
J.
(
2011
).
Ventromedial frontal lobe damage disrupts value maximization in humans.
The Journal of Neuroscience
,
31
,
7527
.
Collins
,
D. L.
,
Neelin
,
P.
,
Peters
,
T. M.
, &
Evans
,
A. C.
(
1994
).
Automatic 3D intersubject registration of MR volumetric data in standardized Talairach space.
Journal of Computer Assisted Tomography
,
18
,
192
.
Damasio
,
A.
,
Tranel
,
D.
, &
Damasio
,
H.
(
1990
).
Individuals with sociopathic behavior caused by frontal damage fail to respond autonomically to social-stimuli.
Behavioural Brain Research
,
41
,
81
94
.
Damasio
,
H.
, &
Frank
,
R.
(
1992
).
Three-dimensional in vivo mapping of brain lesions in humans.
Archives of Neurology
,
49
,
137
143
.
Damasio
,
H.
,
Tranel
,
D.
,
Grabowski
,
T.
,
Adolphs
,
R.
, &
Damasio
,
A.
(
2004
).
Neural systems behind word and concept retrieval.
Cognition
,
92
,
179
229
.
Dunbar
,
R.
(
1998
).
The social brain hypothesis.
Evolutionary Anthropology: Issues, News, and Reviews
,
6
,
178
190
.
Evans
,
A.
,
Dai
,
W.
,
Collins
,
L.
,
Neelin
,
P.
, &
Marrett
,
S.
(
1991
).
Warping of a computerized 3-D atlas to match brain image volumes for quantitative neuroanatomical and functional analysis.
Proceedings of the International Society of Optical Engineering (SPIE): Medical Imaging
,
1445
,
236
246
.
Forbes
,
C.
, &
Grafman
,
J.
(
2010
).
The role of the human prefrontal cortex in social cognition and moral judgment.
Annual Review of Neuroscience
,
33
,
299
324
.
Frank
,
R.
,
Damasio
,
H.
, &
Grabowski
,
T.
(
1997
).
Brainvox: An interactive, multimodal visualization and analysis system for neuroanatomical imaging.
Neuroimage
,
5
,
13
30
.
Gawronski
,
B.
(
2004
).
Theory-based bias correction in dispositional inference: The fundamental attribution error is dead, long live the correspondence bias.
European Review of Social Psychology
,
15
,
183
217
.
Gilbert
,
D.
, &
Malone
,
P.
(
1995
).
The correspondence bias.
Psychological Bulletin
,
117
,
21
38
.
Gläscher
,
J.
,
Adolphs
,
R.
,
Damasio
,
H.
,
Bechara
,
A.
,
Rudrauf
,
D.
,
Calamia
,
M.
,
et al
(
2012
).
Lesion mapping of cognitive control and value-based decision-making in the prefrontal cortex.
Proceedings of the National Academy of Sciences, U.S.A.
,
109
,
14681
14686
.
Gottfried
,
J.
, &
Zald
,
D.
(
2005
).
On the scent of human olfactory orbitofrontal cortex: Meta-analysis and comparison to non-human primates.
Brain Research Reviews
,
50
,
287
304
.
Hare
,
T.
,
Camerer
,
C.
,
Knoepfle
,
D.
,
O'Doherty
,
J.
, &
Rangel
,
A.
(
2010
).
Value computations in ventral medial prefrontal cortex during charitable decision making incorporate input from regions involved in social cognition.
The Journal of Neuroscience
,
30
,
583
590
.
Harris
,
L. T.
,
Todorov
,
A.
, &
Fiske
,
S. T.
(
2005
).
Attributions on the brain: Neuro-imaging dispositional inferences, beyond theory of mind.
Neuroimage
,
28
,
763
769
.
Haselton
,
M.
, &
Buss
,
D.
(
2000
).
Error management theory: A new perspective on biases in cross-sex mind reading.
Journal of Personality and Social Psychology
,
78
,
81
91
.
Johnson
,
H.
, &
Zhao
,
Y.
(
2009
).
BRAINSDemonWarp: An application to perform demons registration.
The Insight Journal
.
Jan–June
,
1
16
.
Koenigs
,
M.
,
Young
,
L.
,
Adolphs
,
R.
,
Tranel
,
D.
,
Cushman
,
F.
,
Hauser
,
M.
,
et al
(
2007
).
Damage to the prefrontal cortex increases utilitarian moral judgements.
Nature
,
446
,
908
911
.
Koscik
,
T.
(
2010
).
Social inference and the evolution of the human brain
. In:
Department of Neurology, Division of Behavioral Neurology and Cognitive Neuroscience
(
vol. PhD
,
286
pp).
Iowa
:
University of Iowa
.
Koscik
,
T.
, &
Tranel
,
D.
(
2012a
).
Brain evolution and human neuropsychology: The inferential brain hypothesis.
Journal of the International Neuropsychological Society
,
18
,
394
412
.
Koscik
,
T.
, &
Tranel
,
D.
(
2012b
).
The human ventromedial prefrontal cortex is critical for transitive inference.
Journal of Cognitive Neuroscience
,
24
,
1191
1204
.
Krusemark
,
E. A.
,
Keith Campbell
,
W.
, &
Clementz
,
B. A.
(
2008
).
Attributions, deception, and event related potentials: An investigation of the self serving bias.
Psychophysiology
,
45
,
511
515
.
Lieberman
,
M.
(
2007
).
Social cognitive neuroscience: A review of core processes.
Annual Review of Psychology
,
58
,
259
289
.
Lin
,
A.
,
Adolphs
,
R.
, &
Rangel
,
A.
(
2011
).
Social and monetary reward learning engage overlapping neural substrates.
Social Cognitive and Affective Neuroscience
,
7
,
274
281
.
Mazziotta
,
J.
,
Toga
,
A.
,
Evans
,
A.
,
Fox
,
P.
,
Lancaster
,
J.
,
Zilles
,
K.
,
et al
(
2001
).
A probabilistic atlas and reference system for the human brain: International Consortium for Brain Mapping (ICBM).
Philosophical Transactions of the Royal Society of London, Series B, Biological Sciences
,
356
,
1293
.
Miller
,
D. T.
, &
Ross
,
M.
(
1975
).
Self-serving biases in the attribution of causality: Fact or fiction?
Psychological Bulletin
,
82
,
213
.
Mitchell
,
J. P.
,
Cloutier
,
J.
,
Banaji
,
M. R.
, &
Macrae
,
C. N.
(
2006
).
Medial prefrontal dissociations during processing of trait diagnostic and nondiagnostic person information.
Social Cognitive and Affective Neuroscience
,
1
,
49
55
.
Moretti
,
L.
,
Dragone
,
D.
, &
Di Pellegrino
,
G.
(
2009
).
Reward and social valuation deficits following ventromedial prefrontal damage.
Journal of Cognitive Neuroscience
,
21
,
128
140
.
O'Doherty
,
J.
,
Kringelbach
,
M.
,
Rolls
,
E.
,
Hornak
,
J.
, &
Andrews
,
C.
(
2001
).
Abstract reward and punishment representations in the human orbitofrontal cortex.
Nature Neuroscience
,
4
,
95
102
.
Phillips
,
M. L.
,
Drevets
,
W. C.
,
Rauch
,
S. L.
, &
Lane
,
R.
(
2003
).
Neurobiology of emotion perception I: The neural basis of normal emotion perception.
Biological Psychiatry
,
54
,
504
514
.
Plassmann
,
H.
,
O'Doherty
,
J.
, &
Rangel
,
A.
(
2010
).
Appetitive and aversive goal values are encoded in the medial orbitofrontal cortex at the time of decision making.
The Journal of Neuroscience
,
30
,
10799
10808
.
Preuss
,
T.
(
2007
).
Primate brain evolution in phylogenetic context.
In J. Kaas (Ed.)
,
Evolution of nervous systems: A comprehensive review
(
Vol. 4
).
New York
:
Elsevier
, pp.
1
34
.
Rolls
,
E.
(
2000
).
The orbitofrontal cortex and reward.
Cerebral Cortex
,
10
,
284
294
.
Rorden
,
C.
(
2007
).
MRIcroN
.
Rorden
,
C.
(
2008
).
MRIcroGL
[Software]. Available from www.mccauslandcenter.sc.edu/mricrogl/
.
Saxe
,
R.
(
2006
).
Uniquely human social cognition.
Current Opinion in Neurobiology
,
16
,
235
239
.
Seidel
,
E. M.
,
Eickhoff
,
S. B.
,
Kellermann
,
T.
,
Schneider
,
F.
,
Gur
,
R. C.
,
Habel
,
U.
,
et al
(
2010
).
Who is to blame? Neural correlates of causal attribution in social situations.
Social Neuroscience
,
5
,
335
350
.
Sellitto
,
M.
,
Ciaramelli
,
E.
, &
di Pellegrino
,
G.
(
2010
).
Myopic discounting of future rewards after medial orbitofrontal damage in humans.
The Journal of Neuroscience
,
30
,
16429
16436
.
Semendeferi
,
K.
,
Damasio
,
H.
,
Frank
,
R.
, &
Van Hoesen
,
G.
(
1997
).
The evolution of the frontal lobes: A volumetric analysis based on three-dimensional reconstructions of the magnetic resonance scans of human and ape brains.
Journal of Human Evolution
,
32
,
375
388
.
Semendeferi
,
K.
,
Lu
,
A.
,
Schenker
,
N.
, &
Damasio
,
H.
(
2002
).
Humans and great apes share a large frontal cortex.
Nature Neuroscience
,
5
,
272
276
.
Shamay-Tsoory
,
S.
,
Tibi-Elhanany
,
Y.
, &
Aharon-Peretz
,
J.
(
2007
).
The green-eyed monster and malicious joy: The neuroanatomical bases of envy and gloating (schadenfreude).
Brain
,
130
,
1663
1678
.
Shiv
,
B.
,
Loewenstein
,
G.
,
Bechara
,
A.
,
Damasio
,
H.
, &
Damasio
,
A. R.
(
2005
).
Investment behavior and the negative side of emotion.
Psychological Science
,
16
,
435
439
.
Smaers
,
J.
,
Steele
,
J.
,
Case
,
C.
,
Cowper
,
A.
,
Amunts
,
K.
, &
Zilles
,
K.
(
2011
).
Primate prefrontal cortex evolution: Human brains are the extreme of a lateralized ape trend.
Brain, Behavior and Evolution
,
77
,
67
78
.
Thirion
,
J. P.
(
1998
).
Image matching as a diffusion process: An analogy with Maxwell's demons.
Medical Image Analysis
,
2
,
243
260
.
Tom
,
S.
,
Fox
,
C.
,
Trepel
,
C.
, &
Poldrack
,
R.
(
2007
).
The neural basis of loss aversion in decision-making under risk.
Science
,
315
,
515
.
Tranel
,
D.
(
2009
).
The Iowa-Benton school of neuropsychological assessment.
In I. Grant & K. M. Adams (Eds.)
,
Neuropsychological Assessment of Neuropsychiatric and Neuromedical Disorders
(
66
83
).
New York
:
Oxford University Press
.
van den Bos
,
W.
,
McClure
,
S.
,
Harris
,
L.
,
Fiske
,
S.
, &
Cohen
,
J.
(
2007
).
Dissociating affective evaluation and social cognitive processes in the ventral medial prefrontal cortex.
Cognitive, Affective, & Behavioral Neuroscience
,
7
,
337
346
.
Van Overwalle
,
F.
(
2009
).
Social cognition and the brain: A meta-analysis.
Human Brain Mapping
,
30
,
829
858
.
Young
,
L.
,
Bechara
,
A.
,
Tranel
,
D.
,
Damasio
,
H.
,
Hauser
,
M.
, &
Damasio
,
A.
(
2010
).
Damage to ventromedial prefrontal cortex impairs judgment of harmful intent.
Neuron
,
65
,
845
851
.