Stroke of Bad Luck?

Abstract

We hypothesized that specific acute right hemisphere lesions disrupt separate components of valuation and emotional response to winning and losing money and of emotional empathy in observing a partner win or lose money. We measured skin conductance response and ratings of emotions when acute right hemisphere stroke patients or healthy controls won or lost money in roulette, or when they watched a partner win or lose. Percent damage to right anterior insula and frontal operculum negatively correlated with both skin conductance response to winning and losing and difference between rating wins versus losses. Results provide support for our hypotheses.

Introduction

Can a stroke make you a good loser but a poor friend? Right hemisphere stroke survivors often have blunted emotional response and less empathy for their partners’ pain and distress. Extensive research has led to an understanding of emotional empathy as consisting of both emotional contagion (sharing another’s emotion) and affective perspective-taking (making inferences about the emotional state of others) (Decety & Jackson, 2004; Preston & De Waal, 2002). Observing another person experiencing a state activates representations of comparable states in the observer, which activate neural regions that represent interoceptive or emotional states (emotional contagion). The second component of emotional empathy is often considered higher and more cognitive level, because it requires interpretation of the emotional state and attribution to the other person (affective perspective-taking) (Jabbi, Swart, & Keysers, 2007). Affective perspective-taking is one type of cognitive perspective-taking, or “theory of mind” (understanding what another person is thinking or feeling). Emotional contagion develops earlier; e.g. babies cry when their mothers are upset, but only later in development do children attribute their own emotional response of that type to the parent.

Affective perspective-taking requires that a person infers how they would feel in a given situation, and then attribute that feeling to the other person. Inferring how one would feel in a given situation depends on a valuation system – a cerebral network that encodes a value signal and represents rewards of various natures on a common scale (Levy & Glimcher, 2012; Sugrue, Corrado, & Newsome, 2005). Functional imaging studies have indicated that valuation depends on a cerebral network that includes at least the orbitofrontal cortex, medial and lateral prefrontal cortex, and anterior insula (Croxson, Walton, O’Reilly, Behrens, & Rushworth, 2009; O’Doherty, Kringelbach, Rolls, Hornak, & Andrews, 2001; Rushworth, Noonan, Boorman, Walton, & Behrens, 2011; Tom, Fox, Trepel, & Poldrack, 2007); whereas the effort involved in this valuation depends on at least anterior cingulate cortex and ventral striatum (Croxson et al., 2009; Kurniawan, Guitart-Masip, Dayan, & Dolan, 2013).

Functional neuroimaging studies have revealed the regions involved in reward networks are also engaged in various aspects of emotional empathy. Most studies show activation in anterior cingulate and anterior insula in observing others experience pain or negative emotions (e.g. (Gu et al., 2012). In an ALE Meta-analysis of 112 fMRI studies of empathy, significant activation was association with empathy tasks across studies in: bilateral dorsomedial prefrontal cortex, right more than left inferior frontal cortex, bilateral anterior insula, anterior and posterior cingulate, bilateral temporoparietal junction, and right amygdala (Bzdok et al., 2012). In 9 independent fMRI studies of empathy for pain and an ALE meta-analysis of 32 studies of empathy for pain activation was identified in bilateral anterior insula, anterior cingulate, precuneus, and thalamus across studies.(Lamm, Decety, & Singer, 2011). However, these studies did not identify distinct roles of each node in this network in various cognitive processes underlying empathy or empathy for pain.

A resting state fMRI study revealed that dominance of emotional empathy in healthy controls was associated with increased functional connectivity between ventral anterior insula, orbitofrontal cortex, amygdala, and anterior cingulate (Cox et al., 2012). Dominance of cognitive empathy was associated with functional connectivity among between parts of the brainstem, superior temporal sulcus, and ventral anterior insula. Finally, many fMRI studies reveal activation in temporoparietal junction in association with empathy tasks (Saxe & Kanwisher, 2003). But when emotional perspective-taking is compared to third-person visuo-spatial judgments, more activation in temporoparietal junction was associated with third-person visuo-spatial judgments (Schnell, Bluschke, Konradt, & Walter, 2011), indicating that this area may have a general role in assignment of agency or “mentalizing” than in emotional empathy (Decety & Lamm, 2007).

However, it is widely recognized that functional imaging studies reveal correlations between areas of activation and performance of a task, so that they can only reveal areas engaged in a task, rather than areas of the brain that are critical for the task (Fellows et al., 2005; Squire, Wixted, & Clark, 2007). Lesion studies are useful as a complementary approach, to test whether regions activated during a task, such as a measure of cognitive or affective perspective-taking, are indeed necessary for that function.

Impairments in empathy have been studied in a number of neurological conditions, but primarily in neurological disease states with fairly diffuse, bilateral damage or dysfunction, such as frontotemporal dementia (P. J. Eslinger et al., 2005; P. J. Eslinger, Moore, Anderson, & Grossman, 2011; Rankin, Kramer, & Miller, 2005), autism (Baron-Cohen, Jolliffe, Mortimore, & Robertson, 1997; Dziobek et al., 2008), schizophrenia (Hooker, Bruce, Lincoln, Fisher, & Vinogradov, 2011; Lee, Zaki, Harvey, Ochsner, & Green, 2011), and head injury (McDonald & Flanagan, 2004; Neumann et al., 2012). In these diseases it is difficult to attribute the deficits to specific lesions in the brain. There have been a few studies of impaired empathy after stroke and other focal injury, and these have had each reported one or a few patients with impaired empathy due to lesions in: prefrontal cortex (Bramham, Morris, Hornak, Bullock, & Polkey, 2009; Spikman, Timmerman, Milders, Veenstra, & van der Naalt, 2012), inferior frontal gyrus (Samson et al., 2005), amygdala (Hurlemann et al., 2010; Stone, Baron-Cohen, Calder, Keane, & Young, 2003; Hurlemann, et al., 2010), anterior insula (Gu et al., 2012), anterior cingulate (Gu et al., 2012), or temporal pole (Narvid et al., 2009). Some group studies of highly heterogeneous lesions (meningioma, head injury, etc.) and/or heterogeneous time post-onset find similar lesion sites (S. Shamay-Tsoory, Tomer, Berger, & Aharon-Peretz, 2003; S. G. Shamay-Tsoory, Tomer, Berger, Goldsher, & Aharon-Peretz, 2005; S. G. Shamay-Tsoory, Aharon-Peretz, & Perry, 2009; S. Shamay-Tsoory, Tomer, Goldsher, Berger, & Aharon-Peretz, 2004; S. G. Shamay-Tsoory & Aharon-Peretz, 2007). Many of these patient studies have reported double dissociations between abnormal scores on emotional contagion and perspective-taking, relative to healthy controls on self-report measures, such as the Interpersonal Reactivity Index (IRI; (Davis, 1983; P. J. Eslinger, Parkinson, & Shamay, 2002; P. J. Eslinger et al., 2011; P. Eslinger, Satish, & Grattan, 1996; S. G. Shamay-Tsoory et al., 2009).

Most studies have indicated the emotional empathy is more impaired after right hemisphere, compared to left hemisphere, lesions. For example, studies have reported more severe impairment of emotional empathy in patients with right hemisphere than left hemisphere atrophy in frontotemporal dementia (P. J. Eslinger et al., 2011; Rankin et al., 2006) and more common deficits in emotional empathy after right hemisphere than left hemisphere focal lesions (S. Shamay-Tsoory et al., 2003; S. G. Shamay-Tsoory et al., 2005; S. Shamay-Tsoory et al., 2004). One previous study confirmed that acute stroke involving right medial prefrontal cortex, anterior cingulate, anterior insula, temporal pole, amygdala, orbitofrontal cortex, or inferior frontal cortex resulted in impaired affected perspective taking, evaluated by asking participants to making judgments of emotions about others from stories or videos (Leigh et al., 2013). However, it was not possible to determine from that study whether emotional contagion and/or valuation were impaired. Instead, it was noted that participants were impaired in some component of affective perspective-taking (making inferences about the emotions of another). Affective valuation and emotional contagion are often measured by an observer rating the emotional response (Adolphs, 2003) or self-rating scales such as the IRI (Davis, 1983). However, in right hemisphere stroke (RHS) patients, outward expression of emotion and self-rating scales do not provide reliable measures, because RHS often leads to deficits in facial expression, prosody (tone of voice) and interospection (Barrett et al., 2006; Blake, Frymark, & Venedictov, 2013; Ferre, Ska, Lajoie, Bleau, & Joanette, 2011) that would compromise these measures. However, emotional responsiveness and contagion can be objectively and sensitively measured with skin conductance response (SCR) in individuals without autonomic neuropathy (Tranel, 2000).

We sought to determine whether particular components of the “empathy network” identified by functional imaging studies are critical for emotional contagion and/or valuation of reward or cost (aspects of perspective-taking). We tested the hypothesis that strokes in specific right frontal, temporal, insular, or cingulate cortical regions impair emotional responsiveness and emotional contagion, and/or valuation in a very simple task that reliably generates positive and negative states in healthy controls (winning and losing money). We tested this hypothesis in individuals with acute ischemic right hemisphere stroke, because acute stroke offers the opportunity to test the effects of lesions before reorganization or recovery, in a previously neurologically normal individual. We excluded individuals with impaired level of consciousness, impaired comprehension of the task, or edema on MRI.

Materials and Methods

Participants.

We studied nine patients with acute RHS (within one week of stroke onset) and nine age-matched healthy controls. Inclusion criteria for stroke patients were as follows: right hemisphere ischemic stroke within the past week; ability to provide informed consent; and comprehension of the task and the Likert scale (assessed by answering questions about the task and scale). Participants with any of the following conditions were excluded from the study: stroke restricted to the brainstem or cerebellum; decreased mental status or ongoing sedation; substance use or withdrawal; previous neurological disease affecting the brain; intracerebral hemorrhage on initial scan; and peripheral or autonomic neuropathy by clinical examination or history.

All stroke patients were alert and oriented in all spheres with normal language comprehension. They all had right middle cerebral artery (MCA) distribution strokes, except for one whose stroke involved both MCA and posterior cerebral artery (PCA) distributions. They all had fluent, grammatical speech. None were below normal for their age on forward or backward digit span. All patients had either left hemispatial neglect (on a gap detection task that evaluates for both viewer-centered and object-centered neglect; (Ota, Fujii, Suzuki, Fukatsu, & Yamadori, 2001) or line bisection in various positions with respect to the trunk) or on a test of affective prosody recognition (Ross & Monnot, 2008) (see Tables 1–4 for demographics and scores on neglect and prosody tests). Note that only 3 patients had testing of affective prosody. For all tests, they each used their right, dominant (non-paretic) hand to respond. Likert scales were presented in vertical arrays, so that hemispatial neglect would not influence responses. All responses were also given verbally. Figure 1 shows the lesion overlap of the 9 patients. Controls were right handed without neurological deficits. All participants provided informed consent for the study.

Table 1.

Demographics

Experimental ID	Age	Sex	Education in years	NIHSS Score	Infarct volume mm3
CES5849	23	Female	12	17	286381
DSY2473	58	Male	12	3	9358.6
HBD1527	70	Female	12	2	2669.1
JFN0433	62	Female	12	4	54197.2
PMN3796	55	Female	11	7	21553.8
PSN9712	76	Male	10	4	17091.3
RGR6308	85	Male	12	13	19785.4
TDN5381	38	Female	16	11	45426.5
VET6413	81	Female	16	13	123688

Table 4.

Recognition of Affective Prosody: Given in Percent Error (abnormal scores are in bold)

Experimental ID	Recognition of Emotional Prosody in Neutral Content Sentences	Recognition of Sarcasm	Recognition of Sincere Sentences
HBD1527	62.5	80	15
PMN3796	81.7	--	--
PSN9712	75	25	80

Figure 1.

Overlap of lesions of all 9 right hemisphere stroke patients

Assessment of Emotional Response.

Each player had 5 rounds total; one player finished all five rounds before the other player started. The order of the players (participant first or partner first) was randomized to avoid an order effect.

The roulette wheel had 8 red numbers and 8 black numbers. Only black numbers were used. The player chose numbers as the “winning” numbers, and the remaining numbers became the “losing” numbers. For each round, the player received 10 chips to bet - either to place on winning numbers or to set aside. If the ball landed on one of the winning numbers, the player won the same amount of chips he or she placed on that number. If the ball landed on one of the losing numbers, the player lost all the chips on the table. However, the player could still keep the chips that were set aside. The roulette wheel also included a green 0; if the ball landed on that number, the player got an extra round, thus an additional 10 chips to bet.

After each round, the participant was asked how he or she felt by being asked to point to the corresponding face expressing emotions from happy to sad on a vertical Likert scale. (1=happy to 7=sad, with numeric labels and emoticon faces, arranged vertically to avoid left neglect). When the partner was playing, the patient was asked how the partner would feel after each round, using the same mechanisms. At the end of the game, the total number of chips were counted and rounded to the nearest 10; for every 10 chips, the participant won $1.00. By the end of the game, they could win up to five dollars (but could not lose more than they started with).

Measurement of SCR.

SCR was measured through an isolated skin conductance coupler (Model V71-23, Coulbourn Instruments, Allentown, PA) with three electrodes (Compumedisc USA Inc, Charlotte USA) on the ventral distal phalanges of index, middle and ring fingers. The coupler delivered low distortion sine wave excitation voltage of 0.5 volts (mean square root) at 100 Hz across the skin and measured the resulting current flow as SCR. The measured current was further processed by the coupler into an output voltage signal for recording. SCR was alternating current coupled with 5-second time constant before recording. The recording module was Neuro Omega (Alpha Omega, Nazareth, Israel) with sampling rate at 1k Hz. Recorded SCR was downsampled to 50 Hz prior to the subsequent analysis.

Roulette outcome time was identified when the dealer announced the roulette outcome as a win or a loss. Response magnitude was defined as the maximum SCR during the 6-second interval after the roulette outcome time minus the baseline SCR. Baseline SCR was defined as the average of the SCR values during a 2-second interval before the roulette outcome time.

Imaging.

Given the small patient sample (and low power) we did not use a whole brain voxel-based lesion-deficit mapping approach. Rather, because there is a large literature on brain regions involved in the neural circuits supporting empathy, we used a region of interest (ROI) approach to identify which of these areas are critical to valuation and emotional contagion. The 12 ROIs were selected on the basis of functional imaging studies and lesion studies providing evidence for the neural network underlying empathy (Bzdok et al., 2012; Gu et al., 2012; Lamm et al., 2011; Rankin et al., 2005; S. G. Shamay-Tsoory, 2011) or autonomic function (Beissner, Meissner, Bar, & Napadow, 2013): orbitofrontal cortex; medial prefrontal cortex; inferior frontal gyrus pars operularis, pars triangularis, and pars orbicularis; uncinate fasciculus; middle and superior temporal pole; amygdala; and dorsal and subgenual anterior cingulate gyrus. An investigator (KO) masked to behavioral and SCR responses measured percent damage to the 12 ROI on MRI Diffusion Weighted Imaging (DWI) sequences, which are most sensitive to acute infarct. First, the boundaries of the acute stroke lesion(s) were identified, using a threshold of > 30% intensity increase from the unaffected area in the DWI, and then the boundaries were manually modified to avoid false-positive and false-negative areas (Oishi et al., 2009) using RoiEditor (www.MRIstudio.org) to create the lesion map. Then, the DWI b0 sequence was transformed to the JHU-MNI-b0 atlas using affine transformation and large deformation diffeomorphic metric mapping (LDDMM). These matrices were applied to the lesion map for registration. The JHU-MNI Brain Parcellation Map (cmrm.med.jhmi.edu) was then overlaid on the registered lesion map to determine the percent volume of each of the12 ROI infarcted in each case using DiffeoMap (www.MRIstudio.org).

Analysis/Calculation.

Using STATA version 16, we evaluated differences between groups with Wilcoxon Mann-Whitney tests or Kruskal-Wallis tests (for >2 groups) and differences between ratings of emotions for wins versus losses (as well as ratings and SCR for observer versus player) with Wilcoxon signed-rank tests. We evaluated correlations between percent damage to each ROI and (1) SCR to winning or losing, and (2) difference between ratings of emotions for winning versus losing, using Spearman correlations. We used an alpha level of p<0.05.

Results

SCR.

Magnitude of SCR was no different for wins versus losses for either stroke patients (z= −1.09; p= 0.28) or controls (z = −1.06; p = 0.29). This result is consistent with previous studies showing SCR in response to both positive and negative stimuli (Haney & Euse, 1976). SCR was also no different in response to observing another win or lose versus in response to themselves win or lose, for either stroke patients (z=0.42; p=0.68) or healthy controls (z=1.12; p=0.26). This result is consistent with functional imaging studies that show a substantial overlap in brain regions activated in individuals when they watch another experiencing pain or pleasure as when they themselves experience the same pain or pleasure (Lamm et al., 2011). In this respect, the autonomic response seems to parallel the neural response. Because of the lack of differences across these conditions, hereafter SCR is not reported separately for wins versus losses or condition (self versus partner), except in the Figures.

SCR to winning or losing negatively correlated only with percent damage to right anterior insula (rho= −0.56; p=0.015), right medial prefrontal cortex (rho=−0.54; p=0.022), and right frontal operculum (rho= −0.48; p=0.045). SCR magnitude was lower in response to wins and losses in patients with right anterior insula, compared to those without these lesions (z = −6.30; p< 0.00001; Figure 2). SCR magnitude was also lower in response to wins and losses in patients with right medial prefrontal lesions, compared to those without these lesions (z = −5.36; p< 0.00001). Similarly, SCR magnitude was lower in response to wins and losses in patients with right frontal operculum lesions, compared to those without these lesions (z = −2.71; p = 0.0067).

Figure 2.

Median SCR to winning or losing for participants in each group.

Panel A. Median SCR magnitude of response to winning or losing for participants with right anterior insula lesions versus participants without anterior insula lesions

Panel B. Median SCR magnitude of response to winning or losing for participants with right frontal operculum lesions versus participants without right frontal operculum lesions

Panel C. Median SCR magnitude of response to winning or losing for participants with right medial prefrontal lesions versus participants without medial prefrontal lesions

Error bars show the range from 10^th percentile to 90^th percentile. Data are shown separately for trials in which the participant played themselves (self) and when they observed a partner play (partner).

Emotional Ratings.

Patients with lesions in right anterior insula or frontal operculum (only) also showed a diminished difference between their ratings of how happy or sad they or their partner was when they won or lost, compared to healthy controls. Controls and patients without anterior insula or frontal operculum lesions rated partners and themselves as significantly happier when they won versus lost (z= −3.06; p=0.002) on a 7-point scale; but patients with anterior insular and frontal opercular lesions showed no difference in ratings for wins versus losses (z= −1.57; p=0.12). The difference between ratings for wins compared to losses was lower patients with right anterior insular and frontal opercular lesions, compared to those without these lesions (Figure 3); the difference across groups was significant (Χ² (1) = 7.65; df1; p = 0.0057) by Kruskal-Wallis. Moreover, the difference between ratings for wins versus losses negatively correlated with percent damage to right anterior insula (rho=−0.616; p=0.006) and frontal operculum (rho=−.724; p=0.0007) but not total infarct volume (rho=−0.46; ns) or percent damage to other ROI.

Figure 3.

Median difference between wins versus losses for participants with anterior insula lesions, right opercular lesions, and controls or right hemisphere lesions not including the anterior insula or right operculum.

Darker gray represents trials in which the participants rated their own emotions when they won and lost; lighter gray represents trials in which they rated their partner’s emotion when their partner won or lost.

Discussion

Several results from this study converge in support of the hypothesis that acute lesions of right anterior insula and frontal operculum disrupt emotional contagion (autonomic response both to winning and losing money oneself and to observing another win or lose) as well as one aspect of affective perspective-taking (ability to rate another’s emotional response to winning or losing). Not only were lesions associated with inability to rate another’s emotional response, but also inability to rate their own emotional response. This impaired valuation may underlie both deficits in emotional contagion and affective perspective-taking (components of emotional empathy) in some individuals with right insular and opercular lesions. However, it is unclear from our results whether failure to recognize or code an event as reward or cost leads diminished autonomic response to the event, or whether diminished autonomic response contributes to impaired recognition of the reward/cost. We identified a single dissociation: patients with right medial prefrontal lesions showed diminished SCR, but not diminished rating of emotional response to winning and losing. No patients were impaired in rating emotions but had intact SCR, in this study. In our very simple task of “roulette”, rating of emotion showed a narrower range of performance than the SCR, and rating may have been relatively insensitive to deficits compared to SCR.

Functional imaging studies of healthy controls also show that anterior insula and frontal operculum are engaged in emotional empathy (Jabbi et al., 2007; Jabbi & Keysers, 2008). One study showed that anterior insula-frontal operculum activation in response to observing others engaged in both negative and positive gustatory behavior (drinking disgusting, neutral, or pleasant drinks) correlated with self-rating scores on an empathy scale (Jabbi et al., 2007). The authors proposed that anterior insula-frontal operculum activation represented transformation of observed states into emotional states, based on converging evidence from functional imaging studies demonstrating its role in representing empathy for pain (Singer et al., 2004) and perceiving one’s own visceral state, and its anatomical connections needed for such transformations.(Mesulam & Mufson, 1982) A previous study showed that right anterior insula lesions are associated with impairments of perspective-taking (Leigh et al., 2013), but emotional contagion was not studied. Our study provides the first evidence that unilateral (right) lesions in anterior insula or frontal operculum can cause impairments in emotional responsiveness and emotional contagion as well as valuation or ratings of emotions in oneself and others.

Lesions in right medial prefrontal cortex were associated with impairments in autonomic response to winning or losing (emotional responsiveness and emotional contagion), but they were not associated with impairments of valuation in this study. We assessed valuation by asking the participant to rate how happy their partner was when they won or lost (and how happy they themselves were when they themselves won or lost). One limitation of our study was that rating of the partner’s emotion might have been based on either: (1) assessment of the partner’s emotional facial expression and tone of voice, or (2) judgment of whether the partner should be happy or sad because they won or lost (based on how they themselves would feel in the same situation). Participants with right anterior insula or frontal opercular lesions failed to rate the emotions in the same manner as controls using either strategy. The participants with right medial prefrontal lesions did rate emotions similarly to controls, but it is unclear which strategy they used. While valuation is one component of affective perspective-taking (i.e. it is critical to understand how one would feel in order to make inferences about how another would feel in the same situation), affective perspective-taking also requires attribution and mental flexibility. Previous studies have reported that patients with acute right medial prefrontal lesions are impaired in affective perspective-taking, when the task requires making a judgment of emotions of others from stories read to them or from videos (Leigh et al., 2013). Chronic right medial prefrontal lesions have also been associated with impaired affective perspective-taking (S. G. Shamay-Tsoory & Aharon-Peretz, 2007).

A major limitation of this study is that the limited number of participants and variety of lesion sites did not allow us to evaluate the role of all regions likely to be critical components of the neural networks underlying reward and emotional empathy. For example, other studies have shown that SCR correlates with activation not only in anterior insula, but also in amygdalae, orbitofrontal gyrus, and cingulate cortex during emotional states, such as fear (Mériau et al., 2009; Piche, Arsenault, & Rainville, 2010). Only three of our patients had amygdala damage. In our study, patients with damage involving right amygdala, anterior cingulate cortex, temporal pole, and orbitofrontal gyrus also had reduced SCR compared to healthy controls, but they did not have reduced SCR compared to participants without such lesions. Furthermore, the percent damage to these structures did not correlate with the SCR or with the difference in emotional rating for wins versus losses (a measure of perspective-taking). These areas may also be critical to one or both components of empathy (Chakrabarti, Bullmore, & Baron-Cohen, 2006; P. Eslinger et al., 1996; Gu et al., 2012; Olson, Plotzker, & Ezzyat, 2007; Rankin et al., 2005; S. G. Shamay-Tsoory, 2011), but we may have had inadequate power to demonstrate their role. We did show that acute lesions of right anterior insula or frontal operculum had a detrimental effect on SCR and also on rating of emotions of others, relative to healthy controls. But larger studies are needed to identify all of the areas critical for these components of valuation and emotional empathy.

Another limitation of our study is that we did not include patients with similar size acute lesions that include left anterior insula, medial prefrontal cortex, and/or frontal operculum. We have attempted to test such patients, but their testing was complicated by language deficits that interfered with their understanding of the game and/or rating procedure. Patients with isolated lesions to these structures should have adequate comprehension for the task; and future studies will evaluate whether acute lesions to left anterior insula, medial prefrontal cortex, and/or frontal operculum cause similar impairments.

It is possible that the deficits in emotional contagion and understanding of positive and negative emotions of self and others become mitigated over time; other neural regions may be able to assume the role of damaged right anterior insula and frontal operculum during recovery. However, persistently impaired emotional empathy (with inability to experience or express emotions or to use emotional processing to relate to others) has been reported after a right middle cerebral artery stroke that included anterior insula and operculum, as well as anterior temporal cortex (P. J. Eslinger & Geder, 2000). In our study, one patient who was retested one month later showed an increase in SCR to wins and losses (from mean 4.1 to 8.6 above baseline; median 3.5 to 7.7 above baseline), but not to the level of normal controls. She continued to show no difference (<1 on a Likert scale) between ratings of wins versus losses (her own or partner’s). She had a second stroke prior to her planned testing at 6 months post-stroke. Future studies will evaluate which areas are critical to recovery of distinct components of empathy.

Our results show that acute right anterior insula or frontal operculum stroke can cause impaired autonomic response and impaired perception of the emotional state of oneself and others, perhaps due to impaired transformation of experienced and observed states into emotional states (Jabbi et al., 2007). These deficits are manifest as a blunted emotional responsiveness (being a “good loser” or at least unperturbed by losing) and a loss of emotional empathy (being a poor friend).

Table 2.

Gray and white matter structures (parcels on the JHU-MNI Brain Parcellation Map; cmrm.med.jhmi.edu) in the right hemisphere where at least 10% of the parcel was lesioned

Experimental ID
CES5849: superior frontal gyrus (SFG; posterior segment & prefrontal cortex); middle frontal gyrus (MFG; posterior segment & dorsal prefrontal cortex);inferior frontal gyrus (IFG; pars opercularis; pars orbitralis; pars triangularis);lateral orbitofrontal gyrus (LOFG); precentral gyrus (PreCG); superior parietal gyrus (SPG); supramarginal gyrus (SMG); angular gyrus (AG); pre-cuneus (PC); superior temporal gyrus (STG; posterior & pole); middle temporal gyrus (MTG; posterior); inferior occipital gyrus (IOG); dorsal anterior cingulate cortex (dACC); anterior insula; posterior insula;amygdala; superior corona radiata (supCR); body of corpus callosum (bCC); posterior limb of internal capsule (postIC); external capsule (EC); cingulum; superior fronto-occipital fasciculus (SFOF); superior longitudinal fasciculus (SLF);uncinate fasciculus;
DSY2473: SFG (posterior segment & pole); MFG (posterior segment), posterior corona radiata (postCR)
HBD1527: dACC, anterior & posterior insula; amygdala; SCR; pIC; bCC; EC, cingulum; SFOF; SLF; uncinate fasciculus
JFN0433: ITG; parahippocampal gyrus (PHG); fusiform gyrus (FG); middle occipital gyrus (MOG); inferior occipital gyrus (IOG); cuneus; lingual gyrus; posterior cingulate gyrus (PCC); hippocampus; cingulum; fornix; posterior thalamic radiation (PTR); sagittal stratum (SS)
PMN3796: post-central gyrus (postCG); preCG; SMG; STG (posterior); anterior insula; EC; SFOF (<10% of IFG opercularis)
PSN9712: supCR; postCR; PTR; SLF (<10% of IFG orbitalis and anterior insula)
RGR6308: postCG; SMG; anterior insula; posterior insula; supCR; postCR; EC; SLF (<10% of IFG opercularis);
TDN5381: gyrus rectus; rACC. dACC; subgenual & subcallosal ACC; anterior insula; posterior insula; nucleus innominata of Mynert; nucleus accumbens; genu of corpus callosum; bCC; cingulum; IFO; SLF; uncinate fasciculus; olfactory radiation (<10% of IFG opercularis)
VET6413: MFG (posterior segment); IFG (opercularis, orbitalis, & triangularis); preCG; postCG; SMG; STG (posterior & pole); MTG (posterior); IOG; anterior insula; posterior insula; caudate; putamen; antCR; postCR&superior corona radiata; tapatum; anterior, posterior & retrolenticular IC; EC; IFO; PTR; SFO; SLF; uncinate fasciculus

Table 3.

Performance on Neglect Tests: Given in Percent Error/Percent Deviation to Right (for Line Bisection)

Experimental ID	Left –Right Gap Detection (Object Centered Neglect)	Left – Right Circle Detection (Viewer Centered Neglect)	Line Bisection right of trunk	Line Bisection center of trunk	Line Bisection left of trunk
CES5849	12.5 – 0	80 -13.3	--	--	--
DSY2473	0 – 0	0 -0	8.9	8.2	0.74
HBD1527	0 – 0	60 - 0	15.6	14.8	5.9
JFN0433	0 – 0	100 –60	49.6	79.3	100
PMN3796	0 – 0	0 - 0	9.6	12.6	8.15
PSN9717	55.6 – 3.3	17.4 -23.3	7.4	0.7	51.2
RGR6308	0 – 0	0 -0	−2.2	8.3	17.2
TDN5381	0 – 0	86.7-16.6	1.2	14.8	33.0
VET6413	0 – 0	100 -43	34.8	43.7	25.2

(abnormal scores are in bold)