Chester Clinic

Doi:10.1016/j.schres.2008.05.010

Available online at www.sciencedirect.com Schizophrenia Research 105 (2008) 49 – 60 Static posed and evoked facial expressions of emotions Christian G. Kohler ⁎, Elizabeth A. Martin, Neal Stolar, Fred S. Barrett, Ragini Verma, Colleen Brensinger, Warren Bilker, Raquel E. Gur, Ruben C. Gur Neuropsychiatry Division, Department of Psychiatry, University of Pennsylvania, Philadelphia, PA, United States Received 19 November 2007; received in revised form 28 April 2008; accepted 1 May 2008 Objective: Impaired facial expressions of emotions have been described as characteristic symptoms of schizophrenia. Differencesregarding individual facial muscle changes associated with specific emotions in posed and evoked expressions remain unclear. This studyexamined static facial expressions of emotions for evidence of flattened and inappropriate affect in persons with stable schizophrenia.
Methods: 12 persons with stable schizophrenia and matched healthy controls underwent a standardized procedure for posed andevoked facial expressions of five universal emotions, including happy, sad, anger, fear, and disgust expressions, at three intensitylevels. Subjects completed self-ratings of their emotion experience. Certified raters coded images of facial expressions for presenceof action units (AUs) according to the Facial Action Coding System. Logistic regression analyses were used to examine differencesin the presence of AUs and emotion experience ratings by diagnosis, condition and intensity of expression.
Results: Patient and control groups experienced similar intensities of emotions, however, the difference between posed and evokedemotions was less pronounced in patients. Differences in expression of frequent and infrequent AUs support clinical observationsof flattened and inappropriate affect in schizophrenia. Specific differences involve the Duchenne smile for happy expressions anddecreased furrowed brows in all negative emotion expressions in schizophrenia.
Conclusion: While patterns of facial expressions were similar between groups, general and emotion specific differences support theconcept of impaired facial expressions in schizophrenia. Expression of emotions in schizophrenia could not be explained by impairedexperience. Future directions may include automated measurement, remediation of expressions and early detection of schizophrenia.
2008 Elsevier B.V. All rights reserved.
Keywords: Emotion expression; Schizophrenia; Facial Action Coding System; Affective flattening; Inappropriate affect across species (Abnormal expressionsof emotions have been described as characteristic Facial expressions are shared in humans and animals, symptoms of schizophrenia ( and are central for communication both within and ) and may precede the onset of illness by manyyears (Affective flattening andother negative symptoms are present at onset of illness ⁎ Corresponding author. Neuropsychiatry Section, Psychiatry Department, University of Pennsylvania, 3400 Spruce Street, 10 in males, increase with illness duration Gates Building, Philadelphia, PA 19104, United States.
E-mail address: (C.G. Kohler).
and appear distinct from depression ( 0920-9964/$ - see front matter 2008 Elsevier B.V. All rights reserved.
doi: C.G. Kohler et al. / Schizophrenia Research 105 (2008) 49–60 In contrast to positive symptoms of schizo- limited to global assessment of positive and negative phrenia, negative symptoms may not respond as well to emotion expressions, rather than changes in specific antipsychotics and have been linked to impairment in face regions. Other measurements of emotion expres- psychosocial functioning sions have included recognition rates of expressions Whereas there are widely used and validated instruments that measure and parse aspects of cognitive derived measures without analysis of specific AUs dysfunction and its neurobiology in schizophrenia, clinical assessments of affective flattening and other negative symptoms have been limited to observer based rating scales. The ability to quantify emotional expres- automated methods have included computerized face sion, especially in the face, has been enhanced by work aimed at measuring unique features of universal ) and electromyographic measurements emotions. Six universal emotions are recognized across cultures in facial expressions — happiness, sadness, can measure minute muscle activations, albeit limited to anger, fear, disgust and surprise ( select face regions.
). Based on facial Most studies have supported affective flattening in muscle movement, devel- general, rather than inappropriate affect. Studies that oped the Facial Action Coding System (FACS), which examined specific emotions reported on selective identifies discrete facial muscle movements, called impairment in happy (), sad( Action Units (AUs). FACS has been simplified and adapted for clinical research. Emotion FACS (EMFACS: identifies AUs associated expressions. Laterality differences of emotional expres- with the predicted expression of the particular emotion, sions have not been reported, although acuity of illness and the Facial Expression Coding System (FACES: may be associated with differential impairment in upper versus lower face expressions ( overall dynamic facial changes, according to number of expressions, intensity and duration.
affective flattening is considered characteristic of Examinations of facial expressions beyond clinical schizophrenia, comparisons with psychiatric ( rating scales in schizophrenia have reported on imitative , deliberate or posed control groups have raised questions regarding specificity.
), spontaneous expressions within Antipsychotics, particularly first-generation, are dyadic interactions associated with extrapyramidal symptoms, but their influence on emotion expression remains unclear. Some expressions associated with emotional film clips studies indicated an adverse effect of medications on facial expression ( or emotional experiences of the examined patients both on and off antipsychotics and found no clear effect Media for capturing facial expressions have included on expressivity.
still photographs ( Previously, we investigated AUs in high intensity evoked expressions of universal emotions expressed by actors and determined AUs, which were essential for accurate recognition and increased recognition, when present in combinations ). The aim and electromyographic recordings of the present study was to extend previous investiga- tions on evidence of impaired affect in schizophrenia Videotaped acquisition offers the advantage of capturing and to examine individual muscle movements in static duration and frequency of emotion expressions. How- facial expressions of emotions in persons with stable ever, analyses of such lengthy data sets have been symptoms. We expected persons with schizophrenia to C.G. Kohler et al. / Schizophrenia Research 105 (2008) 49–60 produce emotion expressions which include fewer AUs for positive symptoms ) (mean total that are frequently present in expressions of controls and score ± SD = 5.8 ± 10.7, range = 0–32) and negative more AU that are infrequently present in controls.
symptoms () (mean total score ± Matched groups of persons with stable schizophrenia SD = 29 ± 18, range = 0–55), including item ratings for and healthy controls underwent a standardized proce- affective flattening (mean score ± SD = 2.3 ± 1.0, dure of eliciting posed and evoked facial expressions of range = 0–4) and inappropriate affect (mean score ± five universal emotions. We applied FACS to examine SD = 0.8 ± 1.1, range = 0–3) at the time of testing.
facial changes based on the presence of AUs within each Symptom assessments were performed by trained raters emotion and condition, i.e. posed and evoked. Stratified meeting inter-rater reliability (icc N .80). All patients by emotion, we examined the sum of frequent and were treated with standard dosages of second genera- infrequent AUs, combinations of frequent AUs, and tion-antipsychotics (n = 12) and augmenting dosages of differences in the presence of individual AUs. We first-generation antipsychotics, specifically long-acting expected different results for posed compared to evoked haloperidol (n = 2), without increase in antipsychotic emotional expressions, as the former are regulated by medication for 3 months. Antipsychotic dosages were cortical systems and are under greater volitional and converted to estimated olanzapine-equivalents (mean cognitive control Specifically, we antici- dosage ± SD = 15.5 mg ± 2.3) and chlorpromazine- pated that impaired emotional expression in schizo- equivalents (mean dosage ± SD = 312.5 mg ± 188).
phrenia is more pronounced for evoked, rather than None of the participants exhibited clinical evidence of posed expressions.
tardive dyskinesia or acute extrapyramidal symptoms.
After complete description of the study to the subjects, written informed consent was obtained.
2.2. Emotion expression There were 12 persons (mean age = 31.50 Our method of acquisition is based on obtaining ± SD = 7.74, range = 21–42) with the DSM-IV diagnosis emotion expressions in actors (Gur et al.2002) and of schizophrenia based on the Diagnostic Interview for controls, and required extensive training of synchro- Genetic Studies ) and 12 healthy nizing the different aspects of image acquisition.
controls, case matched to patients according to gender, Participants were seated in a brightly lit room and ethnicity, age and parental education. There were 7 men instructed to remain within direct view of the digital and 5 women, including 6 Caucasians and 6 African- camera. Research personnel, including the instructor Americans in each group. No participant in this study and photographer, were located beyond the camera in displayed any facial hair, which may have obscured an unlit area. Interaction between research personnel facial movements, such as beard or mustache. Partici- allowed for rapid and repeated acquisition of photo- pants were recruited from the Schizophrenia Center at graphs, if necessary. Expressions were obtained of 5 the University of Pennsylvania, according to the universal emotions that are reliably rated cross following general exclusion criteria: History of sub- culturally and of neutral expressions. We followed stance related disorders (DSM-IV and laboratory data the procedure previously described () including toxicology) within the past 6 months; mental and the order of emotions expressed remained fixed: retardation defined as a full scale IQ b70; not proficient Happiness, Anger, Fear, Disgust and Sadness. A in English; history of any neurologic event or disease priori decision was made to not include surprise, since affecting brain function; medical diseases that may affect surprise can be conceptualized as the abrupt onset of brain function or interfere with participation. In addition, any other universal emotion. The sequence of healthy controls were free of Axis I, Axis II disorders emotions was based on the assumption that happiness and Axis I disorders in first-degree as the only positive emotion is most easily achieved in isolation, i.e. being first, and sadness was Patients were screened and excluded for self-reported positioned last, since its effect in the evoked condition and observed depression, using standard rating scales.
may last longer. Each emotion was obtained in low, All patients lived independently or with family, none medium and high intensity expressions and followed had been hospitalized within 6 months prior and six by a neutral expression to allow the participant to patients – compared to all controls – worked at least achieve neutral emotional state between different 20 h per week or were students. Patients had assessment

C.G. Kohler et al. / Schizophrenia Research 105 (2008) 49–60 Expressions were obtained in two conditions tradi- subjective experience of each emotion at the three tionally used in directing with instructions on avoidance intensities on an 11-point self-rating scale (0 = no of speech, since it would interfere with facial expression emotion, 10 = extreme intensity). Examples of emotion of emotions. During the mechanical approach (English expressions are shown in method) for unfelt or posed expressions participantswere instructed to communicate or signal the target 2.3. Image acquisition emotion through facial expressions. For the posedexpressions, photographs were obtained based on Images utilized in this study were captured with a cuing of emotional intensities by the instructor. Duration Pulnix digital color camera, mounted on a custom-made of acquisition of posed expressions was between 15 and aluminum frame (). Large-screen 25 min. Subsequently, participants identified biographi- floodlights provided soft (diffuse) illumination, impor- cal emotional situations, when each emotion was tant for resulting texture fidelity. Images were trans- experienced in all intensities, and these situations were ferred serially from their buffer in the digital cameras summarized as vignettes with identification of time through a Bitflow multiplexer onto the computer, using points for the three intensities. The genuine approach for a software-controlled hardware interface.
felt or evoked expressions, is based on the acting systemcredited to Constantin Stanislawski that requires the 2.4. Facial Action Coding System (FACS) ratings participant to engage in emotion and sense memoryrelated to a particular event and thus reliving an Captured images of mild, moderate and high emotional experience. Emotional vignettes were intensity expressions were presented via digital video recounted to participants in a narrative manner using projection in pseudo-random order to three certified exact wording derived from vignettes, and participants FACS raters, who were blinded to emotion expressed were instructed to communicate the emotional experi- and participant status. To serve as a baseline compar- ence through facial expression. For the evoked expres- ison, neutral images were presented next to emotional sions, photographs were obtained based on a priori images of the same person. FACS scoring was decisions about which points of the narrative were performed independently by each rater. AUs of Lips associated with the corresponding emotional intensities.
Part and Jaw Drop, which constitute mouth opening, Duration of acquisition of evoked expressions was were collapsed, since they represent differing degrees of between 15 and 25 min. Following each acquisition of the same muscle movement. For purposes of this study, posed and evoked expressions, participants rated their we were interested in the presence, rather than intensity Fig. 1. Examples of facial emotion expressions.
C.G. Kohler et al. / Schizophrenia Research 105 (2008) 49–60 of AUs and intensity ratings were not included in the ranged from .46–.79 for frequent and .32–.36 for analysis. Laterality was collapsed so that if an AU was infrequent AUs.
scored for one side, it was qualified on both. Ratingswere transformed to binomial data, with the presence of an AU recorded with 1, absence with 0. According toagreement of at least 2 raters, an AU was coded as Participants underwent the procedure for posed and present or absent. The number of occurrences for each evoked expressions of emotions without signs of undue AU was calculated for each image, grouped by emotion emotional stress or difficulties. The two groups rated and condition. Based on the distribution of AUs, those similar subjective experience of emotions for each occurring in ≥15% of expressions within a particular emotion and condition (see . As expected, emotion and either condition in the healthy control experiences increased with target intensities (x- group were considered to be frequent AUs (see sq = 26.11, p b .001) and were rated as greater during Conversely, AUs occurring less frequently were con- evoked compared to posed expressions (x-sq = 16.73, sidered to be infrequent AUs for the particular emotion.
df = 1, p b .001). With respect to intensities of experience, Lastly within each emotion, we examined correla- there were interactions for emotion by condition and tions between clinical assessments of affective flattening condition by diagnosis. Differences between experience and inappropriate affect with presence of frequent and of posed and evoked emotions were more pronounced in infrequent AUs.
sad, anger and fearful expressions, and less pronounced inhappy expressions (x-sq = 16.48, df = 4, p = 0.002). Differ- 2.5. Data analysis ences between experience of evoked and posed emotionswere less pronounced in patients (x-sq = 5.11, df = 1, Conditional logistic regression analyses, condition- p = 0.024). Patients had higher intensities of experience ing on each matched patient–control pair, were used to compared to controls for posed emotions (x-sq = 6.34, test for differences in the presence of AUs by diagnosis, df = 1, p = 0.012) but not for evoked emotions (x-sq = 0.17, condition, and intensity of expression. Two-way inter- df = 1, p = 0.68).
actions were included for diagnosis by condition and Within each emotion, results are presented starting diagnosis by intensity. Since diagnosis was of primary with differences in expression of sum total of frequent and interest, the interactions were dropped from the model infrequent AUs, and combination of at least two frequent when not statistically significant, but the main effects for AUs, followed by differences in expression of individual condition and intensity were retained in the model in AUs according to group and group by condition (evoked order to adjust for these effects. Intensity was of versus posed). Within each emotion, group differences in secondary interest to show the internal validity of the AUs broken down by condition are represented in procedure when examining presence of frequent AUs.
When cell counts were sparse, exact conditional logistic 3.1. Happy expressions regression was used as implemented in SAS ProcLogistic, using the Exact statement. Linear regression Overall, frequent AUs, which included Cheek Raiser, models were used to test for differences in the emotion Lid Tightener, Lip Corner Pull and Mouth Open were less experience ratings by diagnosis, emotion, and condition.
common in evoked expressions in schizophrenia (OR = .41, Generalized estimating equations (GEE) methodology, p b 0.001), but not in posed expressions (OR= .87, p=.595; using the exchangeable correlation structure, were used test for diagnosis by condition interaction: Wald = 4.1, to adjust the linear regression models for clustering by p = .044). Infrequent AUs were more common in schizo- matched pair. Within the patient group, Spearman phrenia (OR = 2.0, p = .013). Combinations of at least two correlations were conducted between clinical ratings frequent AUs were less common in schizophrenia for affective flattening and inappropriate affect (SANS) (OR = 0.32, p = 0.018). Examining the presence of indivi- and presence of frequent and infrequent AUs within dual AUs, the frequent AUs of Lid Tightener (OR = .37, each emotion. All analyses were performed using SAS p = .018) in both conditions and Cheek Raise limited to version 9.1 (SAS Institute, Cary, NC).
evoked expressions (OR = .15, p = .005) were less commonin SZP. Upper Lip Raiser, an infrequent AU in happy 2.6. Inter-coder reliability expressions, was more common in SZP (OR = 25.8,p b .001). For happy expressions, there was an effect of Cohen's kappa coefficients for agreement on expres- intensity on presence of frequent AUs (Wald = 40.78, sion of individual AUs that differed between groups p b .001) but no interaction with group.

C.G. Kohler et al. / Schizophrenia Research 105 (2008) 49–60

C.G. Kohler et al. / Schizophrenia Research 105 (2008) 49–60 Fig. 3. Subjective experience of posed and evoked emotions.
3.2. Sad expressions Open, and they did not differ in expression of infrequentAUs. Combinations of at least two frequent AUs were Overall, frequent AUs, which included Inner Brow less common in schizophrenia (OR = 0.43, p = 0.027).
Raiser, Brow Lower, Lid Tightener, Lip Corner Depres- Examining the presence of individual AUs, Brow Lower sor, Chin Raiser and Lip Tightener were less common in (OR = .03, p b .001) and Lid Tightener (OR = .39, SZP (OR = .58, p = .003), while infrequent AUs were p = .046) – both frequent AUs – were less common in more common in SZP (OR = 2.46, p b .001). Examining SZP. For fear expressions, there was an effect of the presence of individual AUs, Inner Brow Raise intensity on presence of frequent AUs (Wald = 20.39, (OR = .25, p = .016), Brow Lower (OR = .17, p b .001), p b .001), but no interaction with group.
and Lip Corner Depressor (OR = .03, p b .001) – allfrequent AUs – were less common in SZP. Lid Droop 3.5. Disgust expressions (OR = 4.18, p = .013), an infrequent AU, was morecommon in SZP. For sad expressions, there was an effect Overall, frequent AUs, which included Inner Brow of intensity on presence of frequent AUs (Wald = 10.97, Raiser, Brow Lower, Lid Tightener, Nose Wrinkler, Upper p = .004), but no interaction with group.
Lip Raiser, Lip Stretcher, Lip Tightener and Mouth Open,were less common in SZP (OR = .71, p = .013), while 3.3. Anger expressions infrequent AUs were more common in SZP (OR = 1.77,p = .018). Combinations of at least two frequent AUs Overall, frequent AUs, which included Brow Lower, were less common in schizophrenia (OR = 0.46, p = 0.054).
Lid Tightener and Lip Tightener, were less common in Examining the presence of individual AUs, Inner Brow SZP (OR = .58, p = .008), while infrequent AUs were Raise (OR = .05, p = .004), Brow Lower (OR = .04, more common in SZP (OR = 1.98, p = .003). Combina- p b .001) and Lip Tightener (OR=.08, p=.031) – all tions of at least two frequent AUs were less common in frequent AUs – were less common in posed expressions in schizophrenia (OR = 0.40, p = 0.017). Examining the SZP. Lower Lip Depressor (OR = 12.0, p = .037), an presence of individual AUs, Brow Lower (OR = .19, infrequent AU, was more common in SZP. Lip Corner p b .001), a frequent AU, was less common in SZP and Depressor (OR = .08, p = .028), an infrequent AU, was Lid Droop (OR = 11.9, p = .010), an infrequent AU, was absent in SZP. For disgust expressions, there was an effect more common in SZP. For angry expressions, there was of intensity on presence of frequent AUs (Wald = 30.37, an effect of intensity on presence of frequent AUs p b .001), but no interaction with group.
(Wald = 16.12, p = b.001), but no interaction withgroup.
3.6. Clinical ratings 3.4. Fear expressions Within the schizophrenia group, ratings for affective flattening did not correlate with presence of frequent Groups did not differ in expression of frequent AUs, AUs for any emotion expressed. Ratings for inappropri- which included Inner and Outer Brow Raiser, Brow ate affect correlated with presence of infrequent AUs for Lower, Upper Lid Raiser, Lid Tightener and Mouth

Fig. 4. Differences in frequencies of action units in emotion expressions.
C.G. Kohler et al. / Schizophrenia Research 105 (2008) 49–60 happy (r = −. 25, p = .036), sad (r = −.44, p b .001) and and misinterpretation as neutral or no emotion, fearful (r = −.24, p = .047) expressions.
inappropriate affect will result in misinterpretation ofthe emotional valence.
Our prediction regarding a more selective impair- ment in evoked expressions was limited to happy Impaired facial expressions of emotions represent expressions. The prediction could not be confirmed in characteristic negative symptoms in schizophrenia and the negative emotions tested, in particular disgust, directly affect interpersonal engagement and social where some frequent AUs were less common in posed functioning. Previous clinical studies have characterized expressions in the patient group.
the enduring and pervasive nature of affective flattening In happy expressions, patients commonly displayed in schizophrenia. Experimental studies have under- upward turned lip corners, as the most common and scored these findings and expanded on this body of characteristic feature of a smile. Other elements, which knowledge by examination of affective flattening in constitute the Duchenne smile and are necessary for a different conditions and in comparison with other smile to appear sincere, such as eye lids tightened and patient groups.
cheeks raised, were less frequent. In all negative In the present study, involving persons with stable emotion expressions, furrowed or lowered brows were schizophrenia and case matched healthy controls, we clearly less common in schizophrenia. The corrugator applied ratings of specific facial changes associated with muscle, as the muscle associated with formation of a posed and evoked expressions of five universal emo- frown or scowl, represents a major constituent in the tions. Overall patterns of facial expressions of emotions expression of negative emotion ().
were similar between patient and control groups in all Different from our finding, reported emotions and, for each emotion, expression at higher increased corrugator muscle activity, as measured by intensities produced increased number of AUs in patient electromyography that can measure small changes and control groups without differential findings. Differ- beyond visual resolution, for both positive and negative ences in expression of AUs between groups support emotional stimuli in schizophrenia.
observations of altered facial expression of emotions in Considering similar subjective emotional experience schizophrenia. Expressions of emotions usually result within the two groups, our findings do not clearly from activation of multiple AUs and for all emotions, support the hypotheses on the unidirectional relation- except sadness, patients displayed fewer expressions ship between emotion experience and expression with combinations of at least two frequent AUs. Most according to either Darwin or James. Perhaps, our experimental studies in schizophrenia have supported findings are more consistent with the concept of affective flattening, rather than inappropriate affect, and neuromotor dysfunction put forth by experimental ratings of affective flattening in dynamic , however, not all muscle movements were less expressions correlated with clinical scales ( frequent and there were specific differences found for each emotion. Although lower than during evoked In our study, experimental measures of affective expressions, levels of emotion experience during posed flattening based on the presence of frequent AUs failed emotion expressions were elevated and in line with a to correlate with clinical ratings affective flattening.
previous study that reported on emotion experience and While this lack of finding was unexpected, it may relate autonomic activity related to coached facial expressions to the research methodology of examining static, rather in actors While subjective than dynamic, expressions and omission of intensity emotional experience cannot be readily quantified, our ratings from data analysis.
findings are in consistent with previous reports Our prediction regarding inappropriate expressions was confirmed and in each emotion infrequent AUs were more commonly expressed in the schizophrenia support the notion that flat affect in schizophrenia does group. In addition, experimental measures of inap- not indicate diminished emotional experience.
propriate affect in happy, sad and fearful expressions Limitations of our study pertain to subject selec- correlated with clinical ratings in patients. The tion, methodology for acquisition of facial expres- presence of even a single potentially erroneous AU sions and data analysis. The small sample size is in a particular emotion expression should not be mediated by careful matching of patient and control underestimated. Whereas affective flattening may groups, the intensive task design and complex result in lack of recognition of the emotion expressed analysis that accounted for presence and absence of C.G. Kohler et al. / Schizophrenia Research 105 (2008) 49–60 every AU. We chose controls as comparisons, since reliable and sensitive measure of facial expression of actors or coached controls would magnify the emotions. While simple feedback did not enhance difference found amongst groups. Given the limited emotion expression ), more power of our sample size and clinical stability of interactive remediation programs ( patients, the major concern was that the study would ) that utilize information about individual regional fail to detect differences and this did not occur. Due to differences in expression of specific emotions, based on the complexity of data, previous attempts to rate AUs or alternate muscle movement measurement, may dynamic facial expressions using FACS were forced enable remediation of emotion expression.
to limit analysis to global assessment of emotion Lastly, in the past decade identification of prodromal expressions. Therefore to determine differences in states in schizophrenia has received increased attention.
individual muscle movements, we focused on rating Affective flattening may be present well before clinical static facial expressions. Inherent to this methodology onset of schizophrenia (increase is the possibility that photographs may be obtained during the prodromal phase ) and which miss the intended emotion expression, despite represent a marker of vulnerability. In conjunction with extensive training — on synchronizing the different other candidates for endophenotypes, facial expressions aspects of image acquisition. Emotion expressions of emotions in persons with vulnerability to schizo- were obtained in two conditions, posed and evoked.
phrenia may inform us about risk of illness and assist in Advantages for obtaining more spontaneous or future efforts at prevention or postponement of onset of genuine emotion expressions rather than evoked expressions were weighed – against the need tokeep acquisition methods for the two series similar – Role of funding source and the interference of speech on emotion expres- This work was supported by NIMH MH-01839, MH-60722 and sions. The sequence of expressed emotions remained the National Alliance for Research on Schizophrenia and Depression.
fixed and we decided against a randomized sequenceof emotions where happy expressions could be affected by negative emotions and sad experiences Drs. Kohler, R.E. Gur and R.C. Gur designed the study and Dr.
could interfere with other expressions. Lastly, com- Kohler prepared the manuscript. Drs. Stolar and Verma, E. Martin and F.
parisons between groups were based on differences in Barrett were responsible for data acquisition, and C. Brensinger and Dr.
AUs that were frequently or infrequently present in Bilker were responsible for data analysis.
healthy controls, and did not include expressionintensity ratings. Intensity ratings may have produced Conflict of interest interesting findings regarding expression differences There are no potential conflicts of interest to report for Drs. Kohler, amongst groups, but the small sample size precluded Stolar, Bilker, Verma, R.C. Gur and R.E. Gur; C. Brensinger, E. Martin and F. Barrett.
Applying FACS to larger groups with wider range of clinical symptomatology is needed to replicate our findings and investigate the effect of illness acuity and This work was supported by NIMH MH-01839, MH-60722 and symptom clusters on facial expressions. Other future the National Alliance for Research on Schizophrenia and Depression.
directions involving measurement of facial muscle The funding sources had no role in design of study, interpretation of movements may lie in further development of auto- data, manuscript preparation and submission.
mated measurement of facial expressions and therapeu-tic application of measuring facial expressions. Over the Appendix A. Supplementary data past 15 years, several automated programs have beencreated with the aim to better quantitate facial regions Supplementary data associated with this article can be found, in the online version, at morphometry, our group ) developed an automated program that examines facial changes in10 regions that relate to the anatomic areas involved in AUs, as described by FACS. Such computerizedmethods will be able to provide a measure of the subtle Aghevli, M.A., Blanchard, J.J., Horan, W.P., 2003. The expression and changes in facial expression, which FACS is unable to experience of emotion in schizophrenia: a study of social quantify and may replace visual inspection as a more interactions. Psychiatry Res. 119 (3), 261–270.
C.G. Kohler et al. / Schizophrenia Research 105 (2008) 49–60 Andreasen, N.C., 1984a. The Scale for the Assessment of Negative Gur, R.C., Sara, R., Hagendoorn, M., Marom, O., Hughett, P., Macy, Symptoms (SANS). Trans. ed. Vol. The University of Iowa, Iowa City.
L., Turner, T., Bajcsy, R., Posner, A., Gur, R.E., 2002. A method Andreasen, N.C., 1984b. The scale for the Assessment of Positive for obtaining 3-dimensional facial expressions and its standardiza- Symptoms (SAPS). Trans. ed. Vol. The University of Iowa, Iowa City.
tion for use in neurocognitive studies. J. Neurosci. Methods 115 Bartlett, M.S., Hager, J.C., Ekman, P., Sejnowski, T.J., 1999.
(2), 137–143.
Measuring facial expressions by computer image analysis.
Ho, B.C., Nopoulos, P., Flaum, M., Arndt, S., Andreasen, N.C., 1998.
Psychophysiology 36 (2), 253–263.
Two-year outcome in first-episode schizophrenia: predictive value Berenbaum, H., 1992. Posed facial expressions of emotion in of symptoms for quality of life. Am. J. Psychiatry 155 (9), schizophrenia and depression. Psychol. Med. 22 (4), 929–937.
Berenbaum, H., Oltmanns, T.F., 1992. Emotional experience and Izard, C.E., 1994. Innate and universal facial expressions: evidence expression in schizophrenia and depression. J. Abnorm. Psychol- from developmental and cross-cultural research. Psychol. Bull. 115 ogy 101 (1), 37–44.
(2), 288–299.
Bleuler, E., 1911. Dementia Praeox oder die Gruppe der Schizo- Kohler, C., Gur, R.C., Swanson, C.L., Petty, R., Gur, R.E., 1998.
phrenien. Trans. ed. Vol.
Depression in schizophrenia: I. Association with neuropsycholo- Cohn, J.F., Zlochower, A.J., Lien, J., Kanade, T., 1999. Automated gical deficits. Biol. Psychiatry 43 (3), 165–172.
face analysis by feature point tracking has high concurrent validity Kohler, C.G., Turner, T., Stolar, N.M., Bilker, W.B., Brensinger, C.M., with manual FACS coding. Psychophysiology 36 (1), 35–43.
Gur, R.E., Gur, R.C., 2004. Differences in facial expressions of Darwin, C., 1872. The expression of emotions in man and animals.
four universal emotions. Psychiatry Res. 128 (3), 235–244.
Trans. ed. Vol. Philosophical Library, New York.
Kring, A.M., Neale, J.M., 1996. Do schizophrenic patients show a Dworkin, R.H., Clark, S.C., Amador, X.F., Gorman, J.M., 1996. Does disjunctive relationship among expressive, experiential, and affective blunting in schizophrenia reflect affective deficit or psychophysiological components of emotion? J. Abnorm. Psychol.
neuromotor dysfunction? Schizophr. Res. 20 (3), 301–306.
105 (2), 249–257.
Earnst, K.S., Kring, A.M., 1999. Emotional responding in deficit and Kring, A.M., Sloan, D.M., 2007. The Facial Expression Coding non-deficit schizophrenia. Psychiatry Res. 88 (3), 191–207.
System (FACES): development, validation, and utility. Psychol.
Earnst, K.S., Kring, A.M., Kadar, M.A., Salem, J.E., Shepard, D.A., Assess. 19 (2), 210–224.
Loosen, P.T., 1996. Facial expression in schizophrenia. Biol.
Kring, A.M., Kerr, S.L., Smith, D.A., Neale, J.M., 1993. Flat affect in Psychiatry 40 (6), 556–558.
schizophrenia does not reflect diminished subjective experience of Edwards, J., McGorry, P.D., Waddell, F.M., Harrigan, S.M., 1999.
emotion. J. Abnorm. Psychol. 102 (4), 507–517.
Enduring negative symptoms in first-episode psychosis: compar- Kring, A.M., Kerr, S.L., Earnst, K.S., 1999. Schizophrenic patients ison of six methods using follow-up data. Schizophr. Res. 40 (2), show facial reactions to emotional facial expressions. Psychophy- siology 36 (2), 186–192.
Eibl-Eibesfeldt, I., 1970. Ethology, the biology of behavior. Trans. ed.
Larsen, J.T., Norris, C.J., Cacioppo, J.T., 2003. Effects of positive and Vol. Holt, Rinehart & Winston, New York.
negative affect on electromyographic activity over zygomaticus Ekman, P., Friesen, W.V., 1975. Unmasking the face. Trans. ed. Vol.
major and corrugator supercilii. Psychophysiology 40 (5), Prentice-Hall, Englewood Cliffs.
Ekman, P., Friesen, W.V., 1978. Manual of the Facial Action Coding Levenson, R.W., Ekman, P., Friesen, W.V., 1990. Voluntary facial System (FACS). Trans. ed. Vol.Consulting Psychologists Press, action generates emotion-specific autonomic nervous system activity. Psychophysiology 27, 363–384.
First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B.W., 1995.
Malla, A.K., Takhar, J.J., Norman, R.M., Manchanda, R., Cortese, L., Structured Clinical Interview for DSM-IV Axis I Disorders, Non- Haricharan, R., Verdi, M., Ahmed, R., 2002. Negative symptoms Patient Edition (SCID-NP). Trans. ed. Vol. New York State in first episode non-affective psychosis. Acta Psychiatr. Scand. 105 Psychiatric Institute/Biometrics Research Department, New York.
(6), 431–439.
Friesen, W.V., 1986. Recent developments in FACS-EMFACS. Facial Mattes, R.M., Schneider, F., Heimann, H., Birbaumer, N., 1995.
Measurement Newsletter 1, 1–2.
Reduced emotional response of schizophrenic patients in remission Frommann, N., Streit, M., Wolwer, W., 2003. Remediation of facial during social interaction. Schizophr. Res. 17 (3), 249–255.
affect recognition impairments in patients with schizophrenia: a Nurnberger, J.I., Blehar, M.C., Kaufmann, C.A., et al., 1994.
new training program. Psychiatry Res. 117 (3), 281–284.
Diagnostic interview for genetic studies: rationale, unique features, Gaebel, W., Wolwer, W., 2004. Facial expressivity in the course of and training. Arch. Gen. Psychiatry 51, 949.
schizophrenia and depression. Eur. Arch. Psychiatry Clin.
Putnam, K.M., Kring, A.M., 2007. Accuracy and intensity of posed Neurosci. 254 (5), 335–342.
emotional expressions in unmedicated schizophrenia patients: Gelber, E.I., Kohler, C.G., Bilker, W.B., Gur, R.C., Brensinger, C., vocal and facial channels. Psychiatry Res. 151 (1–2), 67–76.
Siegel, S.J., Gur, R.E., 2004. Symptom and demographic profiles Rinn, W.E., 1991. Neuropsychology of facial expression. In: Feldman, in first-episode schizophrenia. Schizophr. Res. 67 (2–3), R.S., Rime, B. (Eds.), Fundamentals of Nonverbal Behavior.
Cambridge University Press, Cambridge, UK, pp. 31–72.
Gottheil, E., Paredes, A., Exline, R.V., Winkelmayer, R., 1970.
Schneider, F., Heimann, H., Himer, W., Huss, D., Mattes, R., Adam, Communication of affect in schizophrenia. Arch. Gen. Psychiatry B., 1990. Computer-based analysis of facial action in schizo- 22 (5), 439–444.
phrenic and depressed patients. Eur. arch. psychiatry clin. neurosci.
Gottheil, E., Thornton, C.C., Exline, R.V., 1976. Appropriate and 240, 67–76.
background affect in facial displays of emotion. Comparison of Schneider, F., Ellgring, H., Friedrich, J., Fus, I., Beyer, T., Heimann, normal and schizophrenic males. Arch. Gen. Psychiatry 33 (5), H., Himer, W., 1992. The effects of neuroleptics on facial action in schizophrenic patients. Pharmacopsychiatry 25 (5), 233–239.
C.G. Kohler et al. / Schizophrenia Research 105 (2008) 49–60 Schwartz, B.L., Mastropaolo, J., Rosse, R.B., Mathis, G., Deutsch, S.I., expressiveness in patients with schizophrenia compared to 2006. Imitation of facial expressions in schizophrenia. Psychiatry depressed patients and nonpatient comparison subjects. Am. J.
Res. 145 (2–3), 87–94.
Psychiatry 162 (1), 92–101.
Shtasel, D.L., Gur, R.E., Gallacher, F., Heimberg, C., Gur, R.C., 1992.
Verma, R., Davatzikos, C., Loughead, J., Indersmitten, T., Hu, R., Kohler, Gender differences in the clinical expression of schizophrenia.
C., Gur, R.E., Gur, R.C., 2005. Quantification of facial expressions Schizophr. Res. 7 (3), 225–231.
using high-dimensional shape transformations. J. Neurosci. Methods Steimer-Krause, E., Krause, R., Wagner, G., 1990. Interaction 141 (1), 61–73.
regulations used by schizophrenic and psychosomatic patients: Walker, E.F., Grimes, K.E., Davis, D.M., Smith, A.J., 1993. Childhood studies on facial behavior in dyadic interactions. Psychiatry 53 (3), precursors of schizophrenia: facial expressions of emotion. Am. J.
Psychiatry 150 (11), 1654–1660.
Tremeau, F., Malaspina, D., Duval, F., Correa, H., Hager-Budny, M., Coin-Bariou, L., Macher, J.P., Gorman, J.M., 2005. Facial

Source: http://frederickbarrett.kimusubi.net/sites/default/files/frederickbarrett.kimusubi.net/Kohler2008SczRes2.pdf