Neural Mechanisms underlying Pupillary
Contagion and subsequent Development
of Trust
An fMRI study investigating social decision-making while including
Eye-tracking, Pupillometry, Skin conductance, Cardiovascular
responses, and Breathing Patterns
University of Amsterdam
Department: Work and Organizational Psychology
Supervisor: Mariska Kret
Co-assessor: Carsten De Dreu
Internship report
Leonie Dühlmeyer – 10437312
1
stNovember 2013
Abstract
Non-verbal communication is vital for human cooperation, yet, the influence of pupillary signals on human cognition and decision-making is largely unknown. Intriguingly, pupil-sizes of two individuals tend to synchronize when they are in direct eye-contact. To invest the influence of “pupillary contagion” on social decision-making, we performed an fMRI study in which five participants made economic trust-decisions, after viewing images of eyes, containing pupils which were either dilating, constricting or static. Using participant’s pupillometry data, we assessed whether their pupil size was modulated by the observed pupil size and in particular, whether there was a mirroring of response, indicating pupillary contagion. First, an index of each participant’s sensitivity to pupillary contagion was determined and used as a regressor to determine brain regions, where activity correlated with this effect. We included skin conductance as a covariate, in order to control for arousal. Second, we performed two localizer tasks to map empathy and threat-related networks individually and look for overlaps with the synchronization patterns. Fourth, we created six whole brain contrasts to investigate the effect of the three conditions of dilating, constricting and static pupils. Concerning the trust-game fMRI results, we observed activation in the precuneous and the cuneous regions during the trust decision after viewing dilating pupils, elucidating an involvement of empathy related areas. Additionally, we surmise that involvement of empathy-related areas contributes during trust decisions, after viewing static pupil stimuli.
Introduction
Living in large groups has contributed to the survival of Homo Sapiens while Homo Neanderthalensis, a species which was stronger and possessed a larger brain, became extinct (Carroll, 2003). Organization of large groups led to numerous advantages, ranging from improved protection against predators, to a greater genetic pool, to division of labor. However, it is highly complex, since members of the same group can be rivals e.g. for food or mating partners (Emery, 2000). Thus, it becomes vital to anticipate an opponent’s intentions upon approach. Nonverbal communication facilitates recognition of a counterpart’s intentions and cooperation at the same time. As hunters, the ancient human depended on non-verbal communication for their basic survival. Evolution of the white sclerae simplified directing and following team members’ gaze (Kobayashi & Kohshima, 1997), as well as reading other’s intentions in their eyes (Emery, 2000) – even unconsciously. The latter seems to be unique for the modern human, since Neanderthal culture lacked the depth of symbolic and progressive thought (McBrearty & Brooks, 2000) and other non-human primates are not able to infer other’s intentions in their eyes. The non-human eyes being so richly informative has presumably led to them capturing more attention than any other area of the face, as well as making them the key aspect in non-verbal communication (Emery, 2000).
We communicate emotions and intentions mostly via the eyes (Baron-Cohen et al., 2001). The eye’s inherent mechanisms such as eye-widening, gaze directing or changing of pupil size, deliver messages significantly faster than articulating messages. Most of these messages are subconsciously sent and perceived, which further accelerates processing (Lamme & Roelfsema, 2000). So far, changes in pupil size are mainly attributed to changes in light and in the physiological system. Darkness and sympathetic activity lead to pupil dilation, brightness and parasympathetic activity to pupil constriction (Lowenstein & Loewenfeld, 1950; Winn et al., 1994). Hess (1965) found that dilated pupils are perceived as attractive, eliciting positive feelings, whereas constricting pupils are related to sadness (Harrison, 2006), eliciting negative feelings (anger, emotional coldness; Hess, 1975).
Recent research has shown that a change of pupil size occurs in a socially interactive context. Observers unconsciously synchronize their pupil size with the pupil size of sad faces (Harrison et al., 2006; Hess, 1975) or neutral faces (Harrison et al., 2007; Kret, Fischer & de Dreu, 2013). The James-Lange theory states that a stimulus evokes a physiological response, which leads to the perception of a distinct emotion (Cannon, 1927). With reference to dilated pupils eliciting positive emotions, both observing and synchronizing seem to be the cause (Kret, Fischer & de Dreu, 2013). The positive feeling is projected on the observed target.
Synchronizing pupil size is a form of non-verbal communication that can only occur when partners are in immediate distance from each other –a potentially dangerous situation. Only trusted persons reach this proximity. Proximity and the evocation of positive emotions by dilated pupils suggest an involvement of trust. An accurate judgment of the latter is a vital feature in human interaction - from ancient times until today. In an investment game, Kret et al. (2013) have found that observers invested more money in targets with increasing vs. decreasing pupils, especially when they synchronized with the trustee. Results were independent of the protagonist’s emotional expression and/or group membership. Overall, these studies indicate that perceived pupil dilation is an important cue that can influence social decision-making.
These novel findings point to a poorly understood neurobiological basis of pupil synchronization and its influence on social decision-making and cooperation. Gilzenrat et al. (2011) propose that pupil dilation reflects arousal which is mediated via norepinephrine, elicited in the locus coeruleus. The norepinephrine system is involved intrinsically with the stress response system (Morilak et al., 2005), which could be triggered via positive or negative arousal through pupil dilation. Indeed, previous research shows that the recruited neural networks differ – depending on the context valence - when observing pupil dilation. When observed in happy faces, empathy related areas activate; when observed in an angry faces, motor preparation areas activate. Pupil constriction is mediated via the Edinger-Westphal nuclei, which regulate parasympathetic efferents on the pupil (Harrison, 2006).
However, the findings by Harrison (2006) derive from a small number of subjects and it remains unknown how the neural networks underlying pupillary contagion influence social decision-making. In the current study we investigated how observed pupil size impacts on trust-decisions.
Participants played a trust-game with virtual partners, who were depicted by short video clips of the eye-region. These showed pupils that were dilating, constricting or static. In each round, participants decided how much money they would invest in the partner. The invested money was tripled for the partner, who then made a decision to return all, some, or none of the total money. The trust-decisions were played with real money, the feedback provided after the test session was based on actual decisions by trustees (students of the University of Amsterdam). Regions relevant for empathy or threat can differ greatly between subjects, thus, two localizer tasks were performed. Saxe, Dodell-Feder (2011) and Dufour (2013) have developed a False-belief localizer. This localizer activates empathy-related areas by contrasting activation during False-belief stories - in which figures have inaccurate beliefs about the situation they are in - with activation during False-photograph stories - in which a photograph, map or sign depicts a world state that is incorrect or obsolete. We developed a Threat-localizer after this example. Activation during threatening stories was contrasted with activation during non-threatening stories.
Using each subject’s pupillometry data, we assessed whether the observed pupil size modulates an observer’s own pupil size. In particular, whether there is a mirroring of response, indicating “pupillary contagion”. For the analysis, we first determined an index of each participant’s sensitivity to pupillary contagion and used it as a regressor to determine brain regions where activity correlates. We included skin conductance as a covariate, in order to control for arousal to see the mere effect of synchrony. Second, we mapped these networks on the empathy and threat-related networks that were defined with the two localizer studies. Third, we analyzed whole-brain data for the conditions of dilating pupils (“increase”), constricting pupils (“decrease”) and static pupils (“static”).
Hypotheses
We hypothesized that arousal - modulated via norepinephrine elicited in the locus coerelus – causes pupil-alignment with dilating pupil stimuli. Alignment would activate empathy related areas, as well as threat related areas and increase trust. We expected the Edinger-Westphal nuclei to mediate pupil-alignment with the decreasing pupil stimuli. Alignment would activate sadness-related areas and decrease trust. The static pupil condition served as control condition.
Methods Participants
Five participants (3 females and 2 males; mean age 25.6 years old; range 20-32 years old) took part in the experiment. Participants were right handed and had normal or corrected-to normal-vision. For optimal measuring with the eye-tracker, participants did not wear eye-make up and vision was corrected with contact lenses in 1 participant. For optimal measuring of the cardiac pattern, participants did not wear nail polish. All participants passed the requirements of the Screening
formulier MRI proefpersonen (Attachment 1) of the Spinoza Centre for Neuroimaging and gave
informed consent. Further, they were asked not to drink coffee, smoke or take drugs in the hours before their appointment. The local ethical committee approved the study. Analysis was performed over 5 participants.
Stimuli and behavioral Procedure
Pupil Stimuli
Pupil stimuli originate from Kret et al. (2013), consisting of nine female and nine male photos derived from the validated Amsterdam Dynamic Facial Expression Set (ADFES) (van der Schalk et al., 2011). The authors standardized the photos in Adobe Photoshop (Adobe Systems). First, they were turned to greyscale and cropped to reveal only the eye region. Second, the eye-white, iris and pupil were erased. Third, the average luminance of the photos was calculated and each picture was adjusted to the mean luminance. Fourth, the eyes were filled with new eye-white and irises. To emphasize the convex shape of the eye, the eye-white around the iris was brighter. One iris pair from one photo was cut out and used for all stimuli. In Adobe After Effects, an artificial pupil was created. After static presentation for 1500ms, the pupil either dilated or constricted within the physiological range of
3-8mm for the duration of 1500ms. In the last second of the stimulus presentation, the pupils were static. The change in pupil-size did not start immediately, to create the impression that the change happened in response to the interaction with the participant. A corneal reflection was created in Photoshop and added in Adobe After Effects. For a natural appearance, the reflection was slightly trembling. The pupil dilation or constriction was linear, but the edges were rounded off with an exponential function (the natural formula implemented in Adobe After effects). This way, 54 unique stimuli were created. By using Matlab, Fourier scrambled images were created from the first frame of each video. These images contained the same low-level features including contrast and luminance of the original ones and were presented prior to the stimulus to reduce the light reflex.
Trust-game
Inside the fMRI scanner, the participants first played one run of the trust-game. A scrambled picture appeared for 2500ms + 3000ms jitter, then a fixation cross was presented on top for 800ms after which a video with pupil stimuli appears. One video showed one eye pair (54 videos; 9 male, 9 female) with dilating, constricting or static pupils. After each pupil stimulus, the participant was asked: “How much money do you want to invest”? The participants then had 2000ms to choose 0, 2, 4 or 6 Euros. One run of the trust-game consisted of 54 decisions (18 eye pairs*3 conditions), lasting 9300 to 12300ms. There were three runs separated by the two localizer tasks. Stimuli were viewed on a back-projection screen via a mirror system attached to the MRI head coil.
Fig. 1: Set up of one Trust-game trial
False-belief localizer
The False-belief localizer was developed by Saxe and Dodell-Feder (2011) and Dufour (2013). The stimuli consisted of 20 short stories. 10 of the stories described a situation in which someone held a false belief.
Example False-belief story:
Expecting the game to be postponed because of the rain, the Garcia family took the subway home. The score was tied, 3‐3. During their commute the rain stopped and the game soon ended with a score of 5‐3. Question: The Garcia family arrives home believing the score is 5‐3. True False
The remaining 10 stories were false photograph stories, describing a situation in which there was a false physical representation of the world, such as an out-of-date photograph or advertisement.
Example False-photograph story:
Old maps of the islands near Titan are displayed in the Maritime museum. Erosion has since taken its toll, leaving only the three largest islands.
Question:
Near Titan today there are many islands. True False
Both types of stories require the reader to deal with incorrect or outdated representations of the world, which assimilates them in their meta-representational and logical complexity. Yet, they differ crucially in building a representation of someone else’s mental state. Contrasting those conditions, localizes regions that are recruited particularly for processing mental states (Saxe & Kanwisher, 2003; Dodell-Feder et al., 2011; Dufour et al., 2013).
The False-belief localizer was presented in a mini-block design, counterbalanced starting with either the False-belief or the False-photograph story. The amount of words was matched over the two conditions. One story started with 12 seconds fixation cross, followed by 10 seconds story. After presentation of the story, the participant had 4 seconds to make the decision whether the story was true or false.
Threat-localizer
The Threat-localizer was developed after the example of the False-belief localizer. It consisted of 10 threatening and 10 non-threatening stories, rated by 14 people out of a list of 15 threatening and 16 threatening stories. Stories were rated for threat sensation on a scale from 0 to 10 (0 being non-threatening, 10 being highly threatening), as well as, rated for probability from 0 to 10 (0 being improbable, 10 being highly probable). The 10 most threatening stories had an average threatening value of 8.75 ± 0.73, with a probability of 6.04 ± 0.83. The 10 non-threatening stories had an average of 0.86 ± 0.68, with a probability of 6.94 ± 0.57. The stories were matched in amount of words over the conditions.
Example Threat story:
Imagine the following situation: At night on the way home, you decide to take a shortcut through the dark park. From the middle of the park, a man with a knife approaches you. You run for your life.
Question:
The situation is threatening True False
The remaining 10 stories were non-threatening.
Example Non-Threat story:
Imagine the following situation: You are watching an animal documentary on TV. The doorbell rings. Your neighbor is at the door and asks whether you have some sugar for her. You go to kitchen to get it for her.
Question:
The situation is threatening True False
The presentation equals the one of the False-belief localizer.
Presentation 16.4 was used to present stimuli and acquire behavioral responses. Participants viewed stimuli on the projector screen over a mirror, which was mounted on the MRI head coil. They responded via a button box held in the right hand.
Procedure
Previous to the appointment, participants received an information brochure, the Beck Depression Index (BDI) (Beck, 1961) and an online link to questionnaires addressing the participant’s empathic abilities and anxiety levels. This questionnaire consists of the Liebowitz Social Anxiety Scale (Liebowitz, 1987) and the Empathic Concern and Perspective Taking - two of four sub
scales of the Interpersonal Reactivity Index (Davis, 1996). Further questions regarding the participants’ suitability can be found in the appendix. Applicants could participate if they were fluent in Dutch, did not rely on glasses to correct to normal vision, had no or removable metal in their body (excluding threads behind the teeth) and described themselves as normally empathic. We met participants a few days before the scanning to give an explanation about the procedure and the tasks. This meeting further took place to secure the participants’ appearance on the scanning day.
By arrival on the scanning appointment, participants signed the informed consent, read an instruction form and answered questions in order to test whether they had understood the rules of the game. After ensuring that they had understood the task, they filled out the State-Trait Anxiety Inventory (Spiegelberger, 1983), received another explanation about the procedure and were asked to use the toilet. Before entering the scanning room, the scanning assistant ensured that there was no more metal on the participant. Then, we attached two electrodes to participants’ left ring and index finger. A pulse oxidation signal (PO) was recorded from the middle finger. Breathing rate was measured with a (the scanner accompanying) band around the participants’ chest from Phillips Achieva. Now, participants entered the scanning room. Lying in the scanner, an electrode to measure heart-rate and breathing was applied to the middle finger of the left hand. First, a sham scan was implemented to ensure the homogeneity of the magnetic field. Second, the T1 anatomical scan followed, during which the participants performed a nine-point calibration of the eye-tracking system. Third, we recorded the functional scans.
T h e participants implemented 5 task units, beginning with the T rust- game. The second task was, counterbalanced, either the False-belief localizer or the Threat-localizer. Next, a second run of the trust-game was played, followed by the remaining localizer and ending with a third run of the Trust-game. Following the scanning session, participants implemented four more tasks in a separate test room. Here, they rated the before shown pupil stimuli for attractiveness and trustworthiness; determined a state of mind a person on a photograph was in (Baron-Cohen et al., 2001); and drew pupils in one happy face and in an angry face (Hess, 1975). Finally, we took a
photograph of the participant to be able to conduct an exploratory analysis to test whether someone’s eye-size influences the extent to which someone synchronizes (Lee, Susskind and Anderson, 2013). After the experiment, we calculated participants’ bonus by choosing one decision of a trial in run 1, which matched their participant number. This decision was compared with previously made decisions by trustees from Kret et al. (2013).
Measurements Eye-tracker
Pupil-size (continuously sampled every 16ms) was down-sampled to 100ms timeslots. We removed outliers (pupil-size between two time-samples changed more than twice the standard deviation from the mean change) and interpolated gaps smaller than 250ms. We smoothed the data with a 10th order low-pass Butterworth filter. A 300ms pre-stimulus baseline was subtracted from each sample during stimulus presentation.
fMRI scanning and analysis
Data were collected on a Philips 3T Achieva scanner. The scanning session started with a T1 scan [T1 turbo field echo, 240*188mm2 field of view (FOV), 220 slices, 1mm slice thickness, 8.2s repetition time (TR), 3.73ms echo time(TE), 8° flip angle (FA), sagittal orientation]. Next, five runs (lasting 12-12.3s, 10s, 12-12.3, 10 and 12-12.3s each) of functional data were collected (2.0s TR, 27.63ms TE, 192*141.24mm2 FOV, 39 slices, 3.3mmslice thickness, 76.1° FA, sagittal orientation) covering the whole brain.
FEAT (FMRI Expert Analysis Tool) version 6.0, part of FSL [Oxford Centre for Functional MRI of the Brain (FMRIB) Software Library (www.fmrib.ox.ac.uk/fsl)] was used to analyze the fMRI data. We aligned functional data to the structural image of the subject, and transformed each subject’s data to the standard space of the MNI (Montreal Neurological Institute) using FLIRT. Next, the functional data were spatially smoothed using a 5 mm (full-width-at half-maximum) Gaussian kernel and high-pass
filtered in the temporal domain (high pass frequency cutoff point: 100s). Finally, the functional data were pre-whitened using FSL (Woolrich et al., 2001).
The conditions were separately modeled by convolution with a double-gamma hemodynamic response function in a general linear model. For the trust game, we modeled the conditions “increase”, “decrease” and “static”. In the False-belief localizer, the conditions were “ToM” (Theory of Mind) and “Neutral”. With the Threat-localizer containing “Threat” and “Neutral”. Runs were pooled per subject using a fixed-effects model. Subsequently, a fixed-effects group analysis was conducted using the FMRIB FLAME stage 1, in which relevant lower-level contrasts were combined.
We looked at highest whole-brain activation during each of the three conditions by contrasting activation in voxels during “increase” being stronger active than during “static” (Incr>Sta); “increase” stronger than “decrease” (Incr>Decr); and “decrease” stronger than “static” (Decr>Sta). FEAT calculates the inverse effect for each contrast. All cortical regions with a height threshold of z= 2.3 and a cluster probability of p= 0.05 were reported. This corrects for whole-brain multiple comparisons by using Gaussian random field theory (GRFT) (Worsely, 2001). Activation maps are overlaid over the MNI-2mm brain and regions will be determined using the Harvard-Oxford Cortical Structural Atlas that accompanies FSL.
Results
fMRI
Activation during Increase Condition
The increasing pupil condition showed no regions with significantly stronger activation in the brain as compared to the static Pupil condition (Incr>Sta). This equally counts for the inverse effect (Sta>Incr). However, making the trust-decision after viewing the increasing pupils, led to a significant activation in the parietal and occipital lobe, as compared to making the trust-decision after viewing the decreasing pupils. The main activation lied in the Precuneous Cortex, followed by Cuneal Cortex.
Figure 1: Statistical maps of the whole brain analysis of the contrast Increase>Decrease (p=0.005, GRFT) rendered on the MNI-2mm brain, horizontal view. The Precuneous Cortex and the Cuneous Cortex are activated.
Table 1. Voxelwhise activation during the increase condition contrasted with the decrease decision
Hemisphere Anatomical region MNI coordinates X Y Z
Z value Size in Voxels
Left Precuneous Cortex 14 -78 38 3.57 1
Left Cuneal Cortex 18 -68 20 3.41 1 ↓
Left Cuneal Cortex 4 -72 28 3.09 1 ↓
Left Precuneous Cortex 14 -70 30 3.06 1 ↓
Left Precuneous Cortex 18 -64 28 2.97 1 ↓
Left Precuneous Cortex 6 -64 22 2.97 1 ↓
Discussion
We hypothesized that pupil-alignment of the subject with the dilating pupil-stimuli would activate empathy related areas (previously localized with the False-belief localizer), as well as threat related areas (previously localized with the Threat-localizer). Our comparative study reveals differences between the neural basis of perceiving increasing or decreasing pupil stimuli in the context of an economic trust-game. The major finding is the activation of the precuneous and the cuneous during the increase condition, contrasted with the decrease condition. We have, however, not found significant activation specific for the other contrasts.
Our study pioneers the investigation of pupillary signals in the context of trust. The left precuneous and left cuneous cortex were stronger active for dilating than for constricting pupils, which somewhat supports the hypothesis that perceiving increasing pupils activates empathy related areas. In an fMRI study, Farrow et al. (2001) proved the precuneous’ engagement in empathy by showing that the left precuneous is active during empathic and forgivability judgments. Ochser et al. (2004) confirmed this with whole-brain analysis when directly comparing one’s own and another individuals’ emotional state. Both self and other judgments activated the precuneous. The cuneous is part of the visual cortex and involved in visual processing (Kosslyn et al., 1999). However, Kosslyn et al. (1999) have found a role of the cuneous in visual mental imagery; Sabbagh et al., found some evidence that individual differences (in estimated current density) in the cuneous predict preschool children’s performance in representational theory of mind tasks.
In a study investigating whether observed pupil size modulates perception of other’s emotional expression, Harrison et al. (2006) found activation of the precuneous cortex only during viewing of photographs with decreasing pupils in a happy face, increasing pupils in an angry face and decreasing pupils in a neutral face. This does not contradict our findings, because these combinations of expression and pupillary state do not occur in real life. Interpreting the state of mind in those photographs is unfamiliar, leading one to wonder about the person’s intentions and activating
Theory of Mind related brain regions.
The small participant number likely caused no significant activation in threat-related areas during the increasing pupil condition, as well as sadness-related areas during the decreasing pupil condition. Another reason for not finding sadness-related areas could be the smaller adaptive value of synchronizing with sadness compared with arousal. Yet, the relevance of regarding trust-decreasing cues could still lead to a significant behavioral effect. Another aspect is that Harrison (2006) used an unnaturally small pupil-size which could have led to the found, sadness-related activation patterns. In this case, no relation of decreasing pupil-size and the sensation of sadness is given. Yet, unnatural body language could create distrust in the observers.
We have already collected data of 28 more participants and will analyze it in January 2014. An only significantly greater activation in the precuneous and cuneous during the incr>decr, but not during the incr>sta contrast shows that the static condition likely activates precuneous and cuneous as well - although to a smaller extent. Hence, exclusively looking into another person’s eyes in the context of a trust-decision leads to engaging empathy related areas.
Conclusions
Our study found activation in the precuneous and cuneous, prior to making a trust-decision after viewing dilating pupil stimuli. Empathy related areas being significantly engaged in only the dilating pupil condition shows that pupillary signals are a vital part of non-verbal communication - especially social decision-making and cooperation.
References
Baron‐Cohen, S., Wheelwright, S., Hill, J., Raste, Y., & Plumb, I. (2001). The “Reading the mind in the eyes” test revised version: A study with normal adults, and adults with asperger syndrome or high‐ functioning autism. Journal of Child Psychology and Psychiatry, 42(2), 241-251.
Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression. Archives of general psychiatry, 4(6), 561.
Bradley, M. M., Miccoli, L., Escrig, M. A., & Lang, P. J. (2008). The pupil as a measure of emotional arousal and autonomic activation. Psychophysiology,45(4), 602-607.
Buck, R. (1980). Nonverbal behavior and the theory of emotion: The facial feedback hypothesis. Journal of Personality and Social Psychology, 38(5), 811.
Cannon, W. B. (1927). The James-Lange theory of emotions: A critical examination and an alternative theory. The American Journal of Psychology,39(1/4), 106-124.
Carroll, S. B. (2003). Genetics and the making of Homo sapiens. Nature,422(6934), 849-857.
Davis, M. (1996). Interpersonal reactivity index. Empathy: A Social Psychological Approach, 55-6.
Dodell-Feder, D., Koster-Hale, J., Bedny, M., & Saxe, R. (2011). fMRI item analysis in a theory of mind task. NeuroImage, 55(2), 705-712.
Dufour, N., Redcay, E., Young, L., Mavros, P. L., Moran, J. M., Triantafyllou, C., ... & Saxe, R. (2013). Similar Brain Activation during False Belief Tasks in a Large Sample of Adults with and without
Autism. PLOS ONE, 8(9), e75468.
Emery, N. J. (2000). The eyes have it: the neuroethology, function and evolution of social gaze. Neuroscience & Biobehavioral Reviews, 24(6), 581-604.
Farrow, T. F., Zheng, Y., Wilkinson, I. D., Spence, S. A., Deakin, J. W., Tarrier, N., … & Woodruff, P. W. (2001). Investigating the functional anatomy of empathy and forgiveness. Neuroreport, 12(11), 2433-2438.
Gilzenrat, M. S., Nieuwenhuis, S., Jepma, M., & Cohen, J. D. (2010). Pupil diameter tracks changes in control state predicted by the adaptive gain theory of locus coeruleus function. Cognitive, Affective,
& Behavioral Neuroscience,10(2), 252-269.
Harrison, N. A., Singer, T., Rotshtein, P., Dolan, R. J., & Critchley, H. D. (2006). Pupillary contagion: central mechanisms engaged in sadness processing. Social Cognitive and Affective Neuroscience, 1(1), 5-17.
Harrison, N. A., Wilson, C. E., & Critchley, H. D. (2007). Processing of observed pupil size modulates perception of sadness and predicts empathy.Emotion, 7(4), 724.
Hess, E. H. (1965). Attitude and pupil size. Scientific American, 212, 46-54.
Hess, E. H. (1975). The role of pupil size in communication. Scientific American, 233(5), 110.
Kosslyn, S. M., Pascual-Leone, A., Felician, O., Camposano, S., Keenan, J. P., Ganis, G., ... & Alpert, N. M. (1999). The role of area 17 in visual imagery: convergent evidence from PET and
rTMS. Science, 284(5411), 167-170.
Kobayashi, H., & Kohshima, S. (1997). Unique morphology of the human eye. Nature, 387, 767-768.
Kret, M.E., Fischer, A. & de Dreu, C.K.W. (in preparation, 2013).: The eye of the beholder: pupil synchronization induces trust
Lamme, V. A., & Roelfsema, P. R. (2000). The distinct modes of vision offered by feedforward and recurrent processing. Trends in neurosciences, 23(11), 571-579.
Lee, D. H., Susskind, J. M., & Anderson, A. K. (2013). Social Transmission of the Sensory Benefits of Eye Widening in Fear Expressions. Psychological science, 24(6), 957-965.
Liebowitz MR. Social Phobia. Mod Probl Pharmacopsychiatry 1987;22:141-173
Lowenstein, O., & Loewenfeld, I. E. (1950). Role of sympathetic and parasympathetic systems in reflex dilatation of the pupil: Pupillographic studies. Archives of Neurology and Psychiatry, 64(3), 313.
McBrearty, S., & Brooks, A. S. (2000). The revolution that wasn't: a new interpretation of the origin of modern human behavior. Journal of human evolution, 39(5), 453-563.
Morilak, D. A., Barrera, G., Echevarria, D. J., Garcia, A. S., Hernandez, A., Ma, S., & Petre, C. O. (2005). Role of brain norepinephrine in the behavioral response to stress. Progress in
Neuro-Psychopharmacology and Biological Psychiatry, 29(8), 1214-1224.
(2004). Reflecting upon feelings: an fMRI study of neural systems supporting the attribution of emotion to self and other. Journal of cognitive neuroscience, 16(10), 1746-1772.
Saxe, R., & Kanwisher, N. (2003). People thinking about thinking people: The role of the temporo-parietal junction in "theory of mind". NeuroImage, 19, 1835-1842.
van der Schalk, J., Hawk, S. T., Fischer, A. H., & Doosje, B. (2011). Moving faces, looking places: validation of the Amsterdam Dynamic Facial Expression Set (ADFES). Emotion, 11(4), 907.
Sterpenich, V., D’Argembeau, A., Desseilles, M., Balteau, E., Albouy, G., Vandewalle, G., ... & Maquet, P. (2006). The locus ceruleus is involved in the successful retrieval of emotional memories in humans. The Journal of neuroscience, 26(28), 7416-7423.
Sabbagh, M. A., Bowman, L. C., Evraire, L. E., & Ito, J. (2009). Neurodevelopmental correlates of theory of mind in preschool children. Child Development, 80(4), 1147-1162.
Spielberger, C. D., Gorsuch, R. L., Lushene, R. E., & Vagg, P. R. (1983). State-Trait Anxiety Inventory (STAI). BiB 2010, 180.
Winn, B., Whitaker, D., Elliott, D. B., & Phillips, N. J. (1994). Factors affecting light-adapted pupil size in normal human subjects. Investigative Ophthalmology & Visual Science, 35(3), 1132-1137.
Woolrich, M. W., Ripley, B. D., Brady, M., & Smith, S. M. (2001). Temporal autocorrelation in univariate linear modeling of FMRI data. NeuroImage, 14(6), 1370-1386.
