Conceived and designed the experiments: NC MM CDD DP SFC GR. Performed the experiments: NC CDD PS. Analyzed the data: NC CDD PS. Wrote the paper: NC MM CDD DP SFC GR.
The authors have declared that no competing interests exist.
Previous studies showed that the understanding of others' basic emotional experiences is based on a “resonant” mechanism, i.e., on the reactivation, in the observer's brain, of the cerebral areas associated with those experiences. The present study aimed to investigate whether the same neural mechanism is activated both when experiencing and attending complex, cognitively-generated, emotions. A gambling task and functional-Magnetic-Resonance-Imaging (
From the early stages of cognitive development, humans are able to represent and understand others' mental and emotional states
Although there may be several ways in which others' emotions can be understood, recent studies indicate that one such mechanism is based on the reactivation of the cerebral areas associated with the observer's direct emotional experience
To further advance our understanding of complex emotional processes, the present study investigates whether the understanding of others' negative emotions involves the activation of the same neural mechanism as in the first-person experience. Specifically, we investigated whether a neural resonance system is also engaged in situations involving complex emotions that emerge at the interface with high-level cognitive processing. To this purpose we used
Evidence that regret and disappointment are mediated by neural structures only partially overlapping comes from clinical
In the present work, we extended the studies on regret by investigating whether the same cortical areas involved in the first person experience of regret become active also when the individual is faced with emotional experiences of regret in others. Two
From left to right, schematic depiction of the sequence of events in the conditions IP (“I play”, top) and OP (“Other plays”, bottom). Within each condition there are 5 phases: instruction, evaluation of the wheels, choice of the gamble, outcome evaluation and judgment of the outcome. In the depicted example, the participant chose the loosing wheel. The length in seconds of each sub-event in the two studies is shown, in the inferior-most part of the figure.
As noted above, regret results from a sense of responsibility. Therefore, to address specifically regret, as opposed to disappointment, in two control conditions a computer program randomly chose one of the gambles for the participant or for the other player. In these instances, the computer choices still resulted in real monetary gains or losses for the players but, given the participants' lack of responsibility upon the gamble selection, the game outcome did not result in the feeling of regret
The main difference between the two studies lies in the nature of the participants' task when presented with the outcomes obtained. In the first study, we ensured that the participants' emotional reaction to the results of the gambles was consistent with the actual counterfactual comparison between the obtained and unobtained outcomes (i.e., satisfied or unsatisfied with the outcome). In this way, we could also assess the participants' understanding of the other players' emotional state at outcome evaluation during “Other Plays” condition. More precisely, the participants were asked to indicate, after each trial, whether they were satisfied with their own decision (“I Play” condition) or whether, in their opinion, the other player was satisfied with her/his decision (“Other Plays” condition). Although this response was necessary to unfold the participants' emotional coherence with the actual outcomes in both IP and OP tasks, this type of judgment, by its nature, is likely to prompt an emotional response in the beholder. Since one requirement for a mirror response is its automaticity, to make sure that the observed activations were not affected by the explicit emotional appraisal of the gamble results, in the second study participants were required to give a non-emotional evaluation of the outcomes indicating whether results represented a win or loss.
Finally, to shed light on the question of whether the engagement of a resonance mechanism when attending someone else's experience of regret is affected by the individuals' empathic aptitude, we compared brain activations of females and males, under the assumption that females are more empathic than males
The reported activations are based on the contrasts between the conditions where the players (the participant or the actor) made the decision
In line with previous works on the neural correlates of regret processing
The conjunction analysis between IP and OP statistical maps (relative to IF and OF, respectively; p<0.001 uncorrected) revealed significant common parametric activations in the left ventromedial prefrontal cortex (vmPFC), left amygdala and bilaterally in the hippocampus (
Activations linearly and positively related to the objective amount of regret (measured as the difference between the outcomes of the chosen and unchosen gambles) in
H | Anatomical region (BA) | MNI | Z-score | ||
x | y | z | |||
L | vmPFC (11) | −14 | 46 | −14 | 3.62 |
L | Anterior cingulate cortex (24/32) | −8 | 38 | 12 | 3.31 |
R | Anterior cingulate cortex (24/32) | 2 | 38 | 12 | 3.32 |
L | SMA (6) | 0 | −8 | 58 | 3.28 |
L/R | SMA (6) | 4 | −14 | 52 | 3.39 |
R | Middle cingulate cortex (6) | −4 | −6 | 48 | 3.34 |
R | Middle temporal gyrus (21) | 58 | −10 | −18 | 3.40 |
L | Amygdala | −14 | −2 | −18 | 3.26 |
L | Temporal pole (38) | −28 | 4 | −20 | 3.47 |
Amygdala | −24 | 0 | −20 | 3.32 | |
L | Hippocampus | −32 | −34 | −12 | 3.61 |
R | Hippocampus | 32 | −10 | −32 | 3.53 |
Activations linearly and positively related to the objective amount of regret (measured as the difference between the actual outcome and the outcome of the unchosen gamble) in both the IP (
To make sure that these results did not only reflect an emotional response to a negative outcome
Shared effect of the parametric amount of disappointment (measured as the difference between the actual and unobtained outcome of the chosen gamble) across IP (
H | Anatomical region (BA) | MNI | Z-score | ||
x | y | z | |||
L | Postcentral gyrus (2) | −46 | −30 | 42 | 3.20 |
R | Hippocampus | 14 | −28 | −10 | 4.18 |
R | Hippocampus | 26 | −16 | −18 | 3.33 |
L | Parahippocampal gyrus | −18 | −22 | −18 | 4.07 |
L/R | Thalamus/periacqueductal grey matter | 0 | −24 | 10 | 4.13 |
L/R | Cerebellum vermis 3 | −2 | −36 | 2 | 4.66 |
R | Cerebellum (VI) | 30 | −44 | −26 | 3.67 |
L/R | Brainstem | −6 | −20 | −26 | 3.99 |
Activations linearly and positively related to the objective amount of disappointment (measured as the difference between the obtained and unobtained outcomes of the
Like in study 1, here we carried out a conjunction analysis of the parametric effects observed between IP (
H | Anatomical region (BA) | MNI | Z-score | ||
x | y | z | |||
L | vmPFC (11/10) | −10 | 42 | −10 | 4.26 |
L | Lateral OFC/anterior insula (11/38) | −26 | 16 | −20 | 3.98 |
L | Anterior cingulate cortex (24/32) | −14 | 34 | 14 | 3.28 |
R | Middle cingulate cortex (24) | 12 | 4 | 22 | 3.93 |
R | Middle frontal gyrus (46) | 46 | 44 | 18 | 3.80 |
Inferior frontal gyrus (45) | 50 | 40 | 16 | 3.58 | |
R | Amygdala | 28 | −12 | −12 | 3.28 |
Hippocampus | 34 | −20 | −14 | 3.36 | |
R | Dorsal striatum | 12 | 14 | 8 | 3.64 |
Activations linearly and positively related to the objective amount of regret in both the IP (
During a post-scanning session, participants had to complete an Italian translation
Behavioral data from the BEES showed that the mean scores for our participants in study 1 were 34.83 (s.d. = 16.75) for females and 19.33 (s.d. = 18.39) for males. These data were representative of the normal Italian population (female mean = 37, s.d. = 18; male mean = 21, s.d. = 18;
Consistent with these results, direct gender comparisons carried out in the parametric statistical maps of the third-person task (OP
The different linear parametric effect of regret for female
H | Anatomical region (BA) | MNI | Z-score | ||
x | y | z | |||
L | vmPFC (11) | −4 | 40 | −8 | 4.36 |
Anterior cingulate cortex (11/32) | −12 | 42 | 14 | 3.43 | |
L | Supramarginal gyrus (2/40) | −56 | −34 | 36 | 5.18 |
Inferior parietal lobule (2) | −56 | −30 | 42 | 4.08 | |
R | Postcentral gyrus (2/1) | 44 | −40 | 58 | 4.55 |
Inferior parietal lobule (2) | 54 | −40 | 56 | 3.93 | |
L | Hippocampus | −40 | −18 | −16 | 4.75 |
L | Hippocampus | −28 | −14 | −32 | 3.86 |
R | Hippocampus | 40 | −22 | −12 | 4.63 |
R | Hippocampus | 30 | −12 | −26 | 5.02 |
Cerebral regions showing significant gender-effects related to the objective amount of attended regret in the OP (
These findings were confirmed in OP condition (
H | Anatomical region (BA) | MNI | Z-score | ||
x | y | z | |||
L/R | ACC/vmPFC (10/11) | −6 | 48 | 0 | 3.57 |
L | Medial OFC (11) | −10 | 54 | −14 | 3.36 |
L | Anterior insula/IFG pars orbitalis (47) | −34 | 28 | −4 | 3.69 |
R | Anterior insula/IFG pars orbitalis (47) | 42 | 28 | −2 | 3.36 |
L | SMA/Middle cingulate cortex (6) | −14 | −22 | 52 | 3.53 |
L | Postcentral gyrus (3b) | −18 | −40 | 58 | 3.42 |
R | Sensorimotor cortex (4a/1) | 40 | −14 | 52 | 3.41 |
L/R | Middle cingulate cortex (23) | −8 | 14 | 28 | 3.96 |
L/R | Superior medial gyrus (9) | −4 | 52 | 36 | 3.34 |
Cerebral regions showing significant gender-effects related to the objective amount of attended regret in the OP (
The aim of the present study was to investigate whether the understanding of complex emotions, like regret, in others involves the reactivation of the cerebral areas associated with the observer's direct emotional experience. Regret is a negative emotion arising from a counterfactual comparison between the outcome of chosen and discarded options, whereby the discarded option would have produced higher benefits to the individual
In this study we controlled for the effect of regret on cerebral activity by means of methodological and statistical measures. Methodologically, we dealt with the players' feeling of responsibility by comparing the conditions in which the participants actively made a deliberate choice (IP, OP) with control conditions in which choices were randomly made by the computer (IF, OF). Statistically, we used a parametric analysis to investigate only those areas whose activity showed a positive relation with increasing levels of regret. Specifically, we modeled the difference between the outcome of the chosen and unchosen gambles, so that also positive outcomes could result in regret if compared to an even more positive unselected outcome. Violation to these assumptions (feeling of the responsibility and counterfactual evaluation) lead to another emotional state, namely disappointment, even when faced with the same amount of loss.
The neural correlates of regret processing have been previously investigated using
What distinguishes the present study from the previous ones is a specific focus to the understanding of the experience of regret when observing someone else experiencing it, i.e. a resonance mirror effect that, to date, has been investigated only with basic-level emotional stimuli. Among the studies addressing mirroring in the emotional system, of particular interest is the
In line with these studies
Our data on the parametric effects common to IP and OP tasks (relative to baseline) in both studies 1 and 2 revealed several activation foci including the ventromedial prefrontal cortex, the dorsal anterior cingulate cortex (ACC) and hippocampus (
Largely on the basis of evidence coming from animal studies, the
These results confirm the view that vmPFC defines the emotional value of the error given by the difference between the obtained outcome and the unselected alternatives that, if chosen, would have produced better results. This error, which emotionally results in the negative feeling of regret, is a necessary drive for behavioral reorganization. Anterior cingulate cortex uses information about the emotional valence of unsuccessful behavior to re-organize future choices accordingly
Core of this study are the common effects observed between the conditions IP and OP (after baseline subtraction), which indicate that vmPFC-ACC and hippocampal activations mediate the processing of regret not only when directly experienced, but also when knowing that someone else is facing a counterfactual negative outcome. More specifically, this finding shows that the understanding of others' regret is mediated by the reactivation of the same core cerebral regions that induce the feeling of regret in the beholder during a first person experience, hence supporting the involvement of a resonance, mirror-like, mechanism in the comprehension of the high-order emotion of regret when experienced by others. Through this mechanism, others' emotional states are mapped onto the same areas that underlie ones' own direct experiences, therefore allowing an automatic understanding of the cognitive/emotional states intrinsic to the complex emotion of regret in others.
So far, there was only behavioral evidence to suggest that the mere observation of a negative situation occurring to another individual evokes in the observer the same mental processes as those of the acting individual. These investigations assessed counterfactual reasoning in social contexts by comparing reported mental simulation
Attending another's negative emotion, however, is a complex phenomenon that can elicit different and conflicting reactions in the beholder, as shown by two recent studies that have highlighted some of the several facets related to the understanding of others' emotions. These studies have addressed individuals' emotional responses arising from direct
At a first glance, based on data from both these investigations, one might argue that the neural activations observed in the present study during “Other Plays” condition could relate to the emotion of gloating for the other player's misfortunes, rather than to regret. However, several considerations speak against this interpretation. Firstly, those studies were constructed so to elicit direct social comparisons between individuals by either manipulating participants' specific information or by having individuals playing on same trials. In the present study, the effect of possible social comparisons on the reported results was minimized. In fact, participants played on different trials and, particularly in study 1, the OP trials occurred immediately after the IP ones (direct social comparison) statistically only in 1 out of 32 trials. Additionally, outcomes producing the feelings of regret and relief were counterbalanced, thus further reducing the effect of gloating also when OP trials directly followed IP ones. Moreover, evidence that our results are not spoilt by the effects of gloating is represented by a lack of activation of the ventral striatum in OP task, which Takahashi
A critical factor in the level of an individual's
On the whole, our data suggest that the emotional understanding of regret in others is specifically reflected by the activation of a subset of the regions involved in its direct, first-person, experience. Among these regions, vmPFC appears to be at the core of a counterfactual evaluation of the outcomes, updating the emotional valence of the obtained outcome with respect to that unobtained
Twenty-four healthy right-handed
The participants performed a classical gambling task
In the present investigation, there were two basic experimental conditions (see
As an explicit-baseline, two further conditions were used: in the “
Each trial started with a specific instruction indicating the condition type (1 s), which remained at the bottom of the screen throughout the trial length. All instructions were presented in Italian.
The participants underwent a total of 256 trials (64 for each experimental condition). The complete list of trials was predetermined and identical for all the participants. In each gamble, the 4 possible outcomes resulted from paired combinations of 200, 50, −50 and −200 (arbitrary units), associated with 8 different levels of probability (30-70, 35-65, 40-60, 45-55, 55-45, 60-40, 65-35, 70-30). Thus, the possible combinations of wins and losses gave four potential levels of regret (−100, −150, −250 and −400) and relief (100, 150, 250 and 400). The possible combinations of payoffs and levels of probability were equally balanced across all experimental conditions. In each trial, payoffs and probabilities were associated so that a) one of the gambles was riskier than the other, and b) the difference between the gambles was minimized with regard to the expected-value (i.e., the sum of the probability of the two possible gamble outcomes, each multiplied by the corresponding outcome value). In order to compare the effects of different experienced
The participants underwent a training session and were introduced to the same unknown female actor before the beginning of the study. Moreover, they were informed that both their and the actor's performance in IP/IF and OP/OF tasks, respectively, would have resulted in a financial gain or loss with respect to an initial endowment. Importantly, to constrain a competitive attitude towards the actor's performance, participants were explicitly informed that their potential gains/losses were completely independent of those of the other player. Additionally, when introducing the actor to the participants, the actor's personal profile was purposely kept very low. The participants were informed about their cumulative earnings only outside the scanner, after the functional acquisition.
The study was composed of 8 functional runs. Every run comprised 32 trials (8 for each experimental condition). These were randomly assigned to 8 blocks, each of which contained 4 consecutive trials of the same condition. The order of the functional runs, of the blocks within each run and of the trials within each block were randomized across participants. Null events were also included in every run, to allow estimation of low-level baseline brain activity. In order to desynchronize the timings of event-types with respect to the acquisition of single slices within functional volumes, interstimulus intervals (ISI) between successive trials were presented in different (“jittered”) durations across trials (1350, 1950, and 2550 s, in proportion of 4∶2∶1;
Visual stimuli were viewed via a back-projection screen located in front of the scanner and a mirror placed on the head-coil. The software Presentation 11.0 (Neurobehavioral systems, Albany, CA,
After the scanning, participants were asked to report their personal impressions about the task. Then, they completed an Italian version
Twenty-four healthy right-handed
Three main differences distinguished study 2 from study 1 with regard to the task. Firstly, the emotional component of post-outcome judgment was replaced by a “cold” appraisal of the obtained outcome. Namely, instead of providing a satisfaction-judgment, the participants were required to indicate whether the gamble outcome was a win or a loss. Second, in study 2 participants' response was required in all four conditions (IP, OP, IF, OF) and only on 10% of the trials. Finally, the length of the evaluation phase (gambles presentation) was identical in all four conditions (4.5 s).
Different from study 1, in each gamble the 4 possible outcomes resulted from paired combinations of 200, 50, −50 and −200 (arbitrary units), associated with only 3 different levels of probability (25-75, 50-50, 75-25). However, the possible combinations of wins and losses still gave four potential levels of regret (−100, −150, −250 and −400) and relief (100, 150, 250 and 400).
All participants underwent a training session, and were introduced to an unknown actor. In study 2, half of them (50% females and 50% males) were presented to a female actor and the other half to a male actor.
Anatomical T1-weighted and functional T2*-weighted MR images were acquired with a 3 Tesla Philips Achieva scanner (Philips Medical Systems, Best, NL), using an 8-channels Sense head coil (sense reduction factor = 2). Functional images were acquired using a T2*-weighted gradient-echo, echo-planar (EPI) pulse sequence (38 interleaved coronal slices covering the whole brain, TR = 2200 ms, TE = 30 ms, flip-angle = 85 degrees, FOV = 240 mm×240 mm, inter-slice gap = 0.5 mm, slice thickness = 4 mm, in-plane resolution 2.5 mm×2.5 mm). Each scanning sequence comprised 215 sequential volumes. Immediately after the functional scanning a high-resolution T1-weighted anatomical scan (150 slices, TR = 600 ms, TE = 20 ms, slice thickness = 1 mm, in-plane resolution 1 mm×1 mm) was acquired for each participants.
Image pre-processing and statistical analysis were performed using SPM5 (Wellcome Department of Cognitive Neurology,
Statistical maps were generated using a random-effect model, implemented in a 2-levels procedure
At the first level, two sets of analyses were performed. Firstly, outcome trials were partitioned according to the 4 conditions (IP, IF, OP, OF) which were separately modeled as mini-epoch lasting 3 s. For each of the 4 conditions, one additional regressor modeled a linear parametric modulation of the outcome-related activity by the degree of objective amount of
At the second (group) level, these two types of contrast-image were used to perform separate parametric (i.e., dependent on the degree of either regret or disappointment) analyses. Furthermore, since we aimed at investigating also potential gender effects on “mirror-like” cerebral activity, the 1st-level contrast images for “IP
In order to assess common effects across IP and OP tasks, we carried out a conjunction analysis on the IP (
The location of the activation foci in terms of Brodmann Areas (BAs) was determined in the stereotaxic space of Talairach and Tournoux
Behavioral results in study 1
(0.03 MB DOC)
Cerebral activations in IP condition in study 1
(0.08 MB DOC)
Cerebral activations in OP condition in study 1
(0.06 MB DOC)
Cerebral activations resulting from the direct comparisons IP versus OP conditions in study 1
(0.05 MB DOC)
Cerebral activations in the IP (minus baseline) and OP (minus baseline) conditions in study 1
(0.92 MB TIF)
We wish to thank Prof. Vittorio Girotto and Dr. Raffaele D'Isa for their help in the preparation of the study.