- Email: [email protected]
Investigating facial affect processing in psychosis: A study using the Comprehensive Affective Testing System
SCHRES-05892; No of Pages 5 Schizophrenia Research xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Schizophrenia Research journal homepage: www.elsevier.com/locate/schres
Investigating facial affect processing in psychosis: A study using the Comprehensive Affective Testing System Susan L. Rossell a,b,c,d,e,⁎, Tamsyn E. Van Rheenen a,b, Nicole R. Joshua c,d, Alison O’Regan c, Andrea Gogos c,d a
Brain and Psychological Sciences Research Centre, Faculty of Health, Arts and Design, Swinburne University, John St, Hawthorn, Victoria 3122, Australia Cognitive Neuropsychiatry Laboratory, Monash Alfred Psychiatry Research Centre (MAPrc), Level 4, 607 St Kilda Rd, Melbourne, Victoria 3004, Australia Mental Health Research Institute of Victoria, Victoria 3053, Australia 1 d The University of Melbourne, Parkville, Victoria 3010, Australia e Psychiatry, St Vincent's Hospital, Melbourne 3065, Australia b c
a r t i c l e
i n f o
Article history: Received 23 September 2013 Received in revised form 10 May 2014 Accepted 20 May 2014 Available online xxxx Keywords: Facial affect processing Social cognition Bipolar disorder Schizophrenia Mania
a b s t r a c t Facial affect processing (FAP) deficits in schizophrenia (SZ) and bipolar disorder (BD) have been widely reported; although effect sizes vary across studies, and there are limited direct comparisons of the two groups. Further, there is debate as to the influence of both psychotic and mood symptoms on FAP. This study aimed to address these limitations by recruiting groups of psychosis patients with either a diagnosis of SZ or BD and comparing them to healthy controls (HC) on a well validated battery of four FAP subtests: affect discrimination, name affect, select affect and match affect. Overall, both groups performed more poorly than controls in terms of accuracy. In SZ, this was largely driven by impairments on three of the four subtests. The BD patients showed impaired performance specifically on the match affect subtest, a task that had a high cognitive load. FAP performance in the psychosis patients was correlated with severity of positive symptoms and mania. This study confirmed that FAP deficits are a consistent finding in SZ that occur independent of task specific methodology; whilst FAP deficits in BD are more subtle. Further work in this group is needed to replicate these results. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Successful social interaction relies heavily on the ability to correctly interpret the emotions of the people with whom we interact. Deficits in the accurate interpretation of facial expressions and/or affective prosody or their integration can result in a significant social handicap, and have substantial implications for global functioning (Hooker and Park, 2002). These deficits have been shown to be present in schizophrenia (SZ: e.g., Addington and Addington, 1998; Rossell and Boundy, 2005; Bozikas et al., 2006a; Shea et al., 2007; Rossell et al., 2013) and more recently, bipolar disorder (BD: e.g., Getz et al., 2003; Bozikas et al., 2006b, 2007; Van Rheenen and Rossell, in press-a, 2013a, 2014). Given that facial information is processed more easily than prosodic stimuli, the examination of facial affect processing (FAP) has garnered much research interest in the respective SZ and BD literatures (Adolphs, 2002; Kohler et al., 2010; Townsend and Altshuler, 2012; Van Rheenen and Rossell, 2013b). A large body of research has
⁎ Corresponding author at: Cognitive Neuropsychiatry Laboratory, Monash Alfred Psychiatry research centre (MAPrc), Level 4, 607 St Kilda Rd, VIC 3004, Australia. E-mail address: [email protected] (S.L. Rossell). 1 Affiliation where research was conducted.
produced consistent findings of impairment in both groups (Kohler et al., 2010; Van Rheenen and Rossell, 2013b); yet progress toward understanding the fundamental nature and origins of these impairments is currently hampered by a number of factors. Firstly, there remains substantial variability in the effect sizes reported within individual empirical studies in these disorders. It is possible that disparities in methodology, such as the use of different stimuli, different target emotions, and differing levels of task difficulty may be responsible for these inconsistent findings. Secondly, although it appears that these cohorts demonstrate similar FAP deficits, there has been little direct comparison using groups well matched on demographic and clinical symptomatology (Van Rheenen and Rossell, 2013b). This is noteworthy given the contribution that such comparative studies can make to our understanding of shared phenomenology and aetiological underpinnings in these disorders. Past comparisons of neurocognitive performance between the disorders suggest quantitative, but not qualitative differences, and it is possible that this pattern extends to FAP in SZ and BD (Schretlen et al., 2007). Thirdly, some studies have included patients with schizoaffective disorder as part of their SZ cohort, whilst others have excluded them (Kohler et al., 2010). Given the mixed symptom profile of such patients, the amalgamation of patients with schizophrenia and schizoaffective disorder might distort the impact on symptoms on FAP, particularly when comparing performance to that of a BD group.
http://dx.doi.org/10.1016/j.schres.2014.05.026 0920-9964/© 2014 Elsevier B.V. All rights reserved.
Please cite this article as: Rossell, S.L., et al., Investigating facial affect processing in psychosis: A study using the Comprehensive Affective Testing System, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.05.026
2
S.L. Rossell et al. / Schizophrenia Research xxx (2014) xxx–xxx
Finally, the impact of clinical symptoms remains a contentious issue. In BD, some researchers have demonstrated state-like FAP impairments during symptomatic but not euthymic phases (Venn et al., 2004; Langenecker et al., 2010), whereas others have demonstrated trait-like impairments during remission (Addington and Addington, 1998). Similarly in SZ, some studies have reported associations between FAP deficits and negative (Mandal et al., 1999), positive (Poole et al., 2000; Hall et al., 2004; Weniger et al., 2004) or both negative and positive symptomatology (Kohler et al., 2000), whilst others have failed to find these relationships (Cramer et al., 1989; Kucharska-Pietura et al., 2005). Clearly, continued investigation of the influence of these symptoms on FAP is necessary. In light of these limitations, the current study aimed to examine FAP in two different, well matched groups of psychosis patients and a group of healthy controls. Specifically, this study sought to determine the relative FAP profiles of SZ and BD patients by comparing them to controls on a battery of FAP assessments from the Comprehensive Affective Testing System (CATS: Froming et al., 2006). The CATS offers a standardised battery employing four subtests with experimental designs commonly used in the literature. It allows for insight to be gained regarding the impact that subtle differences in task methodologies can have on the estimation of FAP. It was predicted that patients with SZ would be impaired on the four subtests of the CATS compared to healthy controls, whilst BD patients would show intermediary performance. Correlational analysis was performed examining for associations between FAP and both psychotic and mood symptoms. Give the lack of clear findings in the literature with this regard, this analysis was considered exploratory.
2. Methods 2.1. Participants The current study included 54 patients with SZ and 43 patients with BD (type I). Patients were recruited via community support groups and community care units and were all out-patients. Diagnosis was ascertained using the Structured Clinical Interview for DSMIV (SCID: First et al., 1996). Current symptomology was acquired using the Positive and Negative Syndrome Scale (PANSS: Kay et al., 1987): global positive, negative and general scores were calculated. Ratings of depression and mania were made using the Beck Depression Inventory (BDI: Beck and Steer, 1987) and the Bech-Rafaelsen Mania Rating Scale (MRS: Bech et al., 1979), respectively. In the BD group, 11 were euthymic (BDI ≤ 10) and 32 depressed (BDI N 10); no patient in any group fit criteria for a current episode of mania. Only patients with no other co-morbid Axis 1 diagnoses (including schizoaffective disorder) were included in the study. Demographic and clinical characteristics are presented in Table 1. Within the SZ group, all 54 were taking antipsychotic medication, 5 were also taking mood stabilisers, and 8 were also taking antidepressants. Of the BD patients, 20 were taking antipsychotic medication, 30 were taking mood stabilisers, 10 were taking antidepressants and 3 participants were medication free.2 One hundred and twelve healthy control participants were recruited via newspaper advertisements. Control participants were excluded if they had any history of psychiatric disorder or a first degree relative with schizoaffective disorder, SZ or BD (SCID: First et al., 1996). Participants from all four groups met the following criteria: a) no history of neurological disorder or head trauma, b) no current
2 For patients taking antipsychotics a chlorpromazine equivalent score was calculated and correlated with FAP: no significant correlations were reported. In addition, patients taking mood stabilisers were compared with those not taking mood stabilisers and no group differences were demonstrated.
substance abuse or dependence (previous year), c) English as first language, d) between the ages of 18 and 65 years and e) predicted IQ N 85 as scored by the National Adult Reading Test (NART: Nelson and Willison, 1991). The study was carried out in accordance with the Declaration of Helsinki. The study had ethical approvals from North Western Mental Health, and The University of Melbourne, Melbourne Victoria. Informed consent of all the participants was obtained after the study had been fully explained. 2.2. Tasks Participants were tested with four of the 13 subtests from the CATS battery. The subtests are designed to examine four different aspects of FAP, taking approximately 20 min to complete. None of the subtests have practice items. Subtest order was counterbalanced. For each subtest, overall accuracy and reaction time (RT) to both correct and incorrect answers were recorded. The % correct and RTs for correct answers were used in the analyses for each of the subtests. The four subtests are as follows: (1) Affect Discrimination (AD); where participants are shown two faces at the midline (of the same actor) (N = 22 trials), and they must decide whether the faces show the same or different emotion. The emotions used are happy, sad, angry, fear, surprised, disgusted and neutral. (2) Name Affect (NA): where participants are shown a face and are required to select one of seven labels to describe the emotional expression of the face (N = 16 trials). The emotions used are happy, sad, angry, fear, surprised, disgusted and neutral. (3) Select Affect (SA): where participants are shown five faces (of the same actor), each face expressing a different emotion. A target emotion is named at the top of the screen and they are required to select which of the faces matches the target emotion (N = 20 trials). (4) Match affect (MA): where participants are shown one face at the top of the screen and five faces underneath (different actors), each expressing a different emotion. They are required to decide which of the five faces expresses the same emotion as the face above (N = 20 trials).
3. Statistical analysis Demographic and clinical group differences were assessed via oneway between-groups analysis of variance (ANOVA) or Chi-square tests. Given that the BD group had euthymic and depressed members, a multivariate ANOVA comparing these sub-groups on FAP was conducted in the BD sample. There was no significant main effect of mood, and all subsequent analyses were thus conducted using the BD sample as a whole. Repeated measures ANOVAs with Scheffe post-hoc tests and follow up one-way ANOVAs with simple contrasts were used to examine group related differences in performance on the CATS subtests. Due to group differences in the proportion of males and females across the groups, gender was included as a covariate in the analyses. Similarly, as there were group differences in age and levels of education, a two-fold validity check was performed to examine the effects of these demographic variables; first, by using Pearson's product moment correlations between age, education and accuracy/response time scores on the eight subtests. Even using a stringent alpha of 0.01, the majority of variables were significantly correlated with age and education. Therefore, the repeated measures and follow-up analyses were re-run with age and education as covariates. Relationships between subtest performance and the clinical characteristics from the PANSS, MRS and BDI were investigated using Pearson's product moment correlations in the combined patient sample (using p b .01).
Please cite this article as: Rossell, S.L., et al., Investigating facial affect processing in psychosis: A study using the Comprehensive Affective Testing System, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.05.026
S.L. Rossell et al. / Schizophrenia Research xxx (2014) xxx–xxx
3
Table 1 Demographic and clinical characteristics of the four participant groups (mean (standard deviation) unless otherwise stated). Group
C
SZ
BD
N Age M/F Years in education
112 35.14 (12.40) 37/75 16.82(3.20)
54 42.17 (10.50) 35/19 14.03 (3.54)
43 40.50 (10.64) 16/27 14.67 (3.65)
Group comparisons
Age of onset Years illness Total P Total N Total G MRS BDI Meds
– – – – – – –
23.49 (7.15) 18.83 (10.25) 14.50 (6.05) 11.19 (4.38) 24.26 (6.62) 2.02 (2.60) 15.06 (13.56) 459.05 (358.28)
21.42 (7.00) 19.12 (10.11) 10.79 (3.47) 9.16 (2.94) 24.91 (5.21) 1.86 (2.11) 17.58 (12.49) 173.01 (257.47)
F = 7.52, p b .001; C b SZ, C b BD χ2 = 15.63, p b .001. F = 14.78, p b .001; C N SZ, C N BD F = 2.00, NS F = .02, NS F = 12.80, p b .01; SZ N BD F = 6.74, p ≤ .01; SZ N BD F = .28, NS F = .10, NS F = .84, NS 19.30, p b .001; SZ N BD
C = control, SZ = schizophrenia, BD = bipolar disorder, M/F = male/female, PANSS = Positive and Negative Symptom Scale, Total P = Positive subscale from PANNS, Total N = Negative subscale from PANNS, Total G = Total general from PANNS, MRS = Bech–Rafaelsen Mania Rating Scale, BDI = Beck Depression Inventory, Meds = chlorpromazine equivalents, NS = non-significant.
4. Results 4.1. Demographic analyses As can be seen in Table 1, the SZ patients were significantly less educated and older than controls. Positive symptom ratings for the SZ patients were significantly higher than for BD patients. There were no group differences on negative symptom ratings, mania or depression; and no differences between patient groups on age or years of illness onset. 4.2. Group comparisons Tables 2 and 3 display the mean accuracy and RT data without age and gender correction for the four subtests across the three groups. 4.2.1. Accuracy There was a main effect of task (F(2.80,576.62) = 221.62, p b .001; accuracy decreasing in descending order of AD, SA, NA, MA) and group (F(2,206) = 17.97, p b .001; SZ b C, p b .001; BD b C, p b .05) as well as a significant two way group by task interaction effect (F(5.60,576.62) = 3.85, p ≤ .001); follow-up tests indicated that accuracy was at near ceiling and not significantly different across groups for the AD subtest. In contrast, accuracy was more variable for the other three subtests. In comparison to healthy controls, SZ patients showed a deficit in NA subtest performance; SZ patients also had a reduction in performance on the SA subtest compared with both controls and BD patients, although BD and controls performed equivalently. Both SZ and BD patients showed a reduction in performance on the MA subtest, but did not differ from each other. These group differences remained after controlling for age, education and gender (Table 4). 4.2.2. Reaction times Prior to controlling for age, education and gender, there was a main effect of task (F(2.58,531.76) = 58.32, p b .001; latencies increasing in Table 2 Non age and education corrected accuracy (%) data for the four CATS subtests. C
Affect discrimination Name affect Select affect Match affect
SZ
BD
Group comparisons
M
SD
M
SD
M
SD
95.4 80.4 94.6 83.2
6.5 11 6.4 10.9
92.5 73.1 86.1 72.4
9.3 13.3 13.8 13.5
94.5 76 91.9 76
6.7 11.9 7.8 13.3
ascending order of NA, AD, SA, MA) and group (F (2,206) = 6.67, p b .01; SZ N C, p = .002) as well as a two way task by group interaction effect (F(5.16,531.76), p b .05). However, when covariates were included in the analysis, only the group effect remained significant (F(2,203) = 3.98, p b .02). Thus, no follow-up tests were conducted. 4.3. Correlations Upon examination of the correlations between variables in the study, only two noteworthy sets of associations were found. First, SA and MA response time performance correlated significantly with scores on the PANSS positive scale (SA r = .29, p b .01; MA r = .29, p b .01), such that as positive symptomatology increased RTs increased. Second, response time performance on three of the four CATS subtests correlated with the MRS (AD r = .26, p b .01; SA r = .37, p b .001; MA r = .26, p ≤ .01), such that as mania severity increased RTs also increased (although mania symptoms were low within the cohort). 5. Discussion The current study investigated FAP in two well matched groups of psychosis patients and a group of healthy controls, with a view to comparing their respective FAP profiles on a standardised battery of tasks. In partial agreement with our predictions, when age, education and gender were controlled, SZ patients demonstrated an overall deficit in the speed with which they processed facial emotions. Both the SZ and BD groups also performed less accurately compared to controls overall. In SZ patients this accuracy effect was driven by impaired performance on the SA, NA and MA subtests. In the BD group, it was driven by performance on the MA task only. These results point toward a relatively generalised impairment of FAP in SZ. They also suggest a more specific facial affect matching deficit for those patients with BD. Given the absence of group differences on the AD task, which has an arguably greater affective load relative to the NA, SA and MA tasks that rely more heavily on cognitive processes associated with generating verbal responses or making selections from multiple alternatives, it is possible that the FAP findings observed Table 3 Non age and education corrected RT (seconds) data for the four CATS subtests. C
– SZ b C SZ b C, SZ b BD SZ b C BD b C
Affect discrimination Name affect Select affect Match affect
SZ
BD
M
SD
M
SD
M
SD
4.4 4.6 5.9 6.7
1.4 1.7 1.6 2.3
6.2 5.5 7 7
3.4 2.2 3.9 2.9
5.3 4.9 6.6 7
1.6 1.8 1.6 2.6
Group comparisons SZ N C SZ N C SZ N C –
Please cite this article as: Rossell, S.L., et al., Investigating facial affect processing in psychosis: A study using the Comprehensive Affective Testing System, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.05.026
4
S.L. Rossell et al. / Schizophrenia Research xxx (2014) xxx–xxx
Table 4 Group comparisons re-run with age and education as covariates. Subtest accuracy (%)
FAP × covariate
Corrected group effect
Age Fp
Education Fp
Gender Fp
Group Fp
AD NA SA
1.7 NS .13 NS 0.9 NS
4.84 .03 3.53 NS 1067 .001
5.63 .02 3.32 NS .05 NS
0.9 NS 4.3 .01 9.52 .001
MA
2.36 NS
12.80 .001
4.93 NS
6.87 .001
Significant group contrasts and mean effect size
– C N SZ (p = .004 d = 0.58) C N SZ (p = .001 d = 0.77) BD N SZ (p = .003 d = 0.51) C N SZ (p = .000, d = 0.90) C N BD (p = .03, d = 0.57)
C = healthy control; SZ = schizophrenia; BD = bipolar disorder; AD = affect discrimination; NA = name affect; SA = select affect; MA = match affect; NS = non-significant; age/ education/gender corrected task effect = F(2.78,564.64) = 10.48, p b .001), group effect (F(2,203) = 7.74, p b .001) and group × task interaction effect (F(5.56,564.45) = 2.47, p b .05).
here in both SZ and BD are attributable to a more general cognitive problem as opposed to one specific to emotion. The deficits evident in BD patients in particular, were generated on the subtest that is arguably the most difficult and cognitively demanding of them all; participants had to first recall the target facial expression, generate subsequent labels for the other facial expression choices and then match them under time pressure. As there is an established literature on cognitive difficulties in BD and a general consensus that such cognitive difficulties are simply less intense than those seen in SZ (e.g., Schretlen et al., 2007; Van Rheenen and Rossell, in press-b), it is certainly possible that this specific BD-patient impairment is reflective of an underlying cognitive load or working memory problem that, due to its subtle nature (relative to SZ), is only evident in the context of greater cognitive task demands. Importantly, the notion that the observed FAP deficits observed are consistent with a non-specific cognitive effect which is more expressed in SZ than in BD, fits well with previous unpublished work by our group indicating that neurocognitive dysfunction is a better predictor of FAP difficulties than underlying configural face processing abilities in BD (Joshua, 2010). Other research showing that neurocognition is predictive of social cognition in both SZ and BD (Brekke et al., 2005; Van Rheenen et al., in submission) also gives credence to this hypothesis. The data confirmed that there are trait FAP difficulties in BD during a euthymic phase (e.g., Bozikas et al., 2006b); and that depression does not exacerbate these findings. It was not possible to examine whether a phase of mania influenced performance, as no patient met criteria for mania in the BD (or either of the other psychosis groups). However, the correlational data did indicate that the severity of mania symptoms (albeit mild) was positively correlated with RT in both psychosis groups; such that greater subclinical mania resulted in longer RTs. This suggests that there may be a link between very low level manic symptomatology and performance on the CATS tasks, although further examination of this in an explicitly manic group is needed. The current data also indicated that response latency variable on two of the CATS subtests (SA and MA) were correlated with the positive symptom scale of the PANSS; that is, as positive symptom severity increased RTs also increased. There is other research to have suggested such a relationship between FAP and positive symptoms (Poole et al., 2000; Hall et al., 2004; Weniger et al., 2004). It is noteworthy that these correlations were only significant for the RT variables, and on two of the subtests. Overall, both positive and negative symptomology was within the mild to moderate range in the two clinical groups. It could be that this limited variability influenced the pattern of associations. Testing patients with severe positive and negative symptoms should be incorporated into future studies to continue investigations within the field. There are other limitations to the study beyond those already mentioned. Firstly, the specific subtests used in the current study utilised stimuli representative of seven basic emotions; however, the design of the subtests was limited such that there were too few trials to tease apart any potential deficits relating to specific emotions. Previous research has indicated SZ patients may have specific problems recognising
expressions of fear, sadness and anger (Edwards et al., 2001; Bediou et al., 2005). Others have indicated an overall problem processing all emotions (Joshua, 2010). Unfortunately, this particular assessment tool was not sensitive enough to explore this issue. Secondly, the CATS subtests do not automatically continue between trials as each participant's response is used as the indicator for the next trial to start. Some individuals took a long time to decide upon their answer, which was reflected by the large standard deviations for all the RT data. Given that participants were instructed to try to be as accurate as possible on all subtests, the RT data should also be interpreted with some caution. Finally, as the level of difficulty in the different subtests was not controlled, and as we did not include any neuropsychological tests, the impact of neurocognition on FAP in these cohorts is not known. Future research utilising the CATS FAP subtests to compare these groups would certainly do well to employ cognitive tasks to examine neuropsychological contributions to the observed deficits in BD and SZ patients. In particular, future studies should include measures of working memory and cognitive load into the test battery, and then use the data from these measures as covariates to ascertain whether FAP deficits are still present. In sum, this data has suggested that FAP deficits are present in SZ. They also suggest that severity of positive symptoms and mania exacerbate deficits. FAP impairment in BD, however, is more subtle and requires future research to elaborate on the role of neurocognition on FAP performance in this cohort. Role of funding source This work was funded by the Mental Health Research Institute of Victoria (MHRI), Australia. AG held an Early Career Training Fellowship from the National Health and Medical Research Council (NHMRC), Australia. Neither MHRI nor NHMRC played any further role in the conduct of this research. Contributors Author 1 designed the study, wrote the protocol, ran the analyses and wrote the manuscript. Author 2 managed the literature searches and assisted with the drafting of the manuscript. Authors 3, 4 and 5 were involved in recruitment of participants, and proof read the manuscript. All authors contributed to and have approved the final manuscript. Conflict of interest The authors declare that they have no conflicts of interest. Acknowledgements This research was funded by the Mental Health Research Institute, and the authors wish to thank Professor George Fink for this support.
References Addington, J., Addington, D., 1998. Facial affect recognition and information processing in schizophrenia and bipolar disorder. Schizophr. Res. 32, 171–181. Adolphs, R., 2002. Neural systems for recognizing emotion. Curr. Opin. Neurobiol. 12 (2), 169–177. Bech, P., Bolwig, T.G., Kramp, P., Rafaelsen, O.J., 1979. The Bech–Rafaelsen Mania Scale and the Hamilton Depression Scale. Acta Psychiatr. Scand. 59, 420–430. Beck, A.T., Steer, R.A., 1987. The Beck Depression Inventory Manual. The Psychological Corporation, San Antonio.
Please cite this article as: Rossell, S.L., et al., Investigating facial affect processing in psychosis: A study using the Comprehensive Affective Testing System, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.05.026
S.L. Rossell et al. / Schizophrenia Research xxx (2014) xxx–xxx Bediou, B., Franck, N., Saoud, M., Baudouin, J.-Y., Tiberghien, G., Daléry, J., d'Amato, T., 2005. Effects of emotion and identity on facial affect processing in schizophrenia. Psychiatry Res. 133 (2–3), 149–157. Bozikas, V.P., Kosmidis, M.H., Anezoulaki, D., Giannakou, M., Andreou, C., Karavatos, A., 2006a. Impaired perception of affective prosody in schizophrenia. J. Neuropsychiatry Clin. Neurosci. 18, 81–85. Bozikas, V.P., Tonia, T., Fokas, K., Karavatos, A., Kosmidis, M.H., 2006b. Impaired emotion processing in remitted patients with bipolar disorder. J. Affect. Disord. 91, 53–56. Bozikas, V.P., Kosmidis, M.H., Tonia, T., Andreou, C., Focas, K., Karavatos, A., 2007. Impaired perception of affective prosody in remitted patients with bipolar disorder. J. Neuropsychiatry Clin. Neurosci. 19 (4), 436–440. Brekke, J., Kay, D.D., Lee, K.S., Green, M.F., 2005. Biosocial pathways to functional outcome in schizophrenia. Schizophr. Res. 80 (2), 213–225. Cramer, P., Weegmann, M., O'Neil, M., 1989. Schizophrenia and the perception of emotions. How accurately do schizophrenics judge the emotional states of others? Br. J. Psychiatry 155, 225–228. Edwards, J., Pattison, P.E., Jackson, H.J., Wales, R.J., 2001. Facial affect and affective prosody recognition in first-episode schizophrenia. Schizophr. Res. 48 (2–3), 235–253. First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B., 1996. Structured Clinical Interview for DSM-IV Axis I Disorders, Clinician Version (SCID-CV). American Psychiatric Press, Washington, D.C., Froming, K., Levy, M., Schaffer, S., Ekman, P., 2006. The Comprehensive Affect Testing System. Psychology Software, Inc., Getz, G.E., Shear, P.K., Strakowski, S.M., 2003. Facial affect recognition deficits in bipolar disorder. J. Int. Neuropsychol. Soc. 9, 623–632. Hall, J., Harris, J.M., Sprengelmeyer, R., Sprengelmeyer, A., Young, A.W., Santos, I.M., Johnstone, E.C., Lawrie, S.M., 2004. Social cognition and face processing in schizophrenia. Br. J. Psychiatry 185, 169–170. Hooker, C., Park, S., 2002. Emotion processing and its relationship to social functioning in schizophrenia patients. Psychiatry Res. 112 (1), 41–50. Joshua, N.R., 2010. Face Processing In Schizophrenia: An Investigation Of Configural Processing And The Relationship With Facial Emotion Processing And Neurocognition. PhD Thesis University of Melbourne. Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The Positive and Negative Syndrome Scale (PANSS) for schizophrenia. Schizophr. Bull. 13, 261–276. Kohler, C.G., Bilker, W., Hagendoorn, M., Gur, R.E., Gur, R.C., 2000. Emotion recognition deficit in schizophrenia: association with symptomatology and cognition. Biol. Psychiatry 48 (2), 127–136. Kohler, C.G., Walker, J.B., Martin, E.A., Healey, K.M., Moberg, P.J., 2010. Facial emotion perception in schizophrenia: a meta-analytic review. Schizophr. Bull. 36 (5), 1009–1019. Kucharska-Pietura, K., David, A.S., Masiak, M., Phillips, M.L., 2005. Perception of facial and vocal affect by people with schizophrenia in early and late stages of illness. Br. J. Psychiatry 187, 523–528.
5
Langenecker, S.A., Saunders, E.F.H., Kade, A.M., Ransom, M.T., McInnis, M.G., 2010. Intermediate: cognitive phenotypes in bipolar disorder. J. Affect. Disord. 122 (3), 285–293. Mandal, M.K., Jain, A., Haque-Nizamie, S., Weiss, U., Schneider, F., 1999. Generality and specificity of emotion-recognition deficit in schizophrenic patients with positive and negative symptoms. Psychiatry Res. 87 (1), 39–46. Nelson, H.E., Willison, J., 1991. National Adult Reading Test manual, 2nd ed. NFER-Nelson, Windsor. Poole, J.H., Tobias, F.C., Vinogradov, S., 2000. The functional relevance of affect recognition errors in schizophrenia. J. Int. Neuropsychol. Soc. 6 (6), 649–658. Rossell, S.L., Boundy, C.L., 2005. Are auditory-verbal hallucinations associated with auditory affective processing deficits? Schizophr. Res. 78 (1), 95–106. Rossell, S.L., Van Rheenen, T.E., Groot, C., Gogos, A., Joshua, N.R., 2013. Investigating affective prosody in psychosis: a study using the Comprehensive Affective Testing System. Psychiatry Res. 210 (3), 896–900. Schretlen, D.J., Cascella, N.G., Meyer, S.M., Kingery, L.R., Testa, S.M., Munro, C.A., Pulver, A.E., Rivkin, P., Rao, V.A., Diaz-Asper, C.M., Dickerson, F.B., Yolken, R.H., Pearlson, G.D., 2007. Neuropsychological functioning in bipolar disorder and schizophrenia. Biol. Psychiatry 62 (2), 179–186. Shea, T.L., Sergejew, A.A., Burnham, D., Jones, C., Rossell, S.L., Copolov, D.L., Egan, G.F., 2007. Emotional prosodic processing in auditory hallucinations. Schizophr. Res. 90 (1–3), 214–220. Townsend, J., Altshuler, L.L., 2012. Emotion processing and regulation in bipolar disorder: a review. Bipolar Disord. 14 (4), 326–339. Van Rheenen, T.E., Rossell, S.L., 2013a. Auditory-prosodic processing in bipolar disorder; from sensory perception to emotion. J. Affect. Disord. 151 (3), 1102–1107. Van Rheenen, T.E., Rossell, S.L., 2013b. Is the non-verbal behavioural emotion-processing profile of bipolar disorder impaired? A critical review. Acta Psychiatr. Scand. 128, 163–178. Van Rheenen, T., Rossell, S., 2014. Multimodal emotion integration in bipolar disorder: an investigation of involuntary cross-modal influences between facial and prosodic channels. J. Int. Neuropsychol. Soc. (in press) (April 11). Van Rheenen, T.E., Rossell, S.L., 2014. Let's face it: facial emotion processing is impaired in bipolar disorder. J. Int. Neuropsychol. Soc. 20 (02), 200–208. Van Rheenen, T.E., Rossell, S.L., 2014. An empirical evaluation of the MATRICS consensus cognitive battery in bipolar disorder. Bipolar Disord. 16 (3), 318–325. http://dx.doi. org/10.1111/bdi.12134. Van Rheenen, T.E., Meyer, D., Rossell, S.L., 2014n. Pathways between neurocognition, social cognition and emotion regulation in bipolar disorder, (in submission). Venn, H.R., Gray, J.M., Montagne, B., Murray, L.K., Michael Burt, D., Frigerio, E., Perrett, D.I., Young, A.H., 2004. Perception of facial expressions of emotion in bipolar disorder. Bipolar Disord. 6 (4), 286–293. Weniger, G., Lange, C., Ruther, E., Irle, E., 2004. Differential impairments of facial affect recognition in schizophrenia subtypes and major depression. Psychiatry Res. 128 (2), 135–146.
Please cite this article as: Rossell, S.L., et al., Investigating facial affect processing in psychosis: A study using the Comprehensive Affective Testing System, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.05.026
Contents lists available at ScienceDirect
Schizophrenia Research journal homepage: www.elsevier.com/locate/schres
Investigating facial affect processing in psychosis: A study using the Comprehensive Affective Testing System Susan L. Rossell a,b,c,d,e,⁎, Tamsyn E. Van Rheenen a,b, Nicole R. Joshua c,d, Alison O’Regan c, Andrea Gogos c,d a
Brain and Psychological Sciences Research Centre, Faculty of Health, Arts and Design, Swinburne University, John St, Hawthorn, Victoria 3122, Australia Cognitive Neuropsychiatry Laboratory, Monash Alfred Psychiatry Research Centre (MAPrc), Level 4, 607 St Kilda Rd, Melbourne, Victoria 3004, Australia Mental Health Research Institute of Victoria, Victoria 3053, Australia 1 d The University of Melbourne, Parkville, Victoria 3010, Australia e Psychiatry, St Vincent's Hospital, Melbourne 3065, Australia b c
a r t i c l e
i n f o
Article history: Received 23 September 2013 Received in revised form 10 May 2014 Accepted 20 May 2014 Available online xxxx Keywords: Facial affect processing Social cognition Bipolar disorder Schizophrenia Mania
a b s t r a c t Facial affect processing (FAP) deficits in schizophrenia (SZ) and bipolar disorder (BD) have been widely reported; although effect sizes vary across studies, and there are limited direct comparisons of the two groups. Further, there is debate as to the influence of both psychotic and mood symptoms on FAP. This study aimed to address these limitations by recruiting groups of psychosis patients with either a diagnosis of SZ or BD and comparing them to healthy controls (HC) on a well validated battery of four FAP subtests: affect discrimination, name affect, select affect and match affect. Overall, both groups performed more poorly than controls in terms of accuracy. In SZ, this was largely driven by impairments on three of the four subtests. The BD patients showed impaired performance specifically on the match affect subtest, a task that had a high cognitive load. FAP performance in the psychosis patients was correlated with severity of positive symptoms and mania. This study confirmed that FAP deficits are a consistent finding in SZ that occur independent of task specific methodology; whilst FAP deficits in BD are more subtle. Further work in this group is needed to replicate these results. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Successful social interaction relies heavily on the ability to correctly interpret the emotions of the people with whom we interact. Deficits in the accurate interpretation of facial expressions and/or affective prosody or their integration can result in a significant social handicap, and have substantial implications for global functioning (Hooker and Park, 2002). These deficits have been shown to be present in schizophrenia (SZ: e.g., Addington and Addington, 1998; Rossell and Boundy, 2005; Bozikas et al., 2006a; Shea et al., 2007; Rossell et al., 2013) and more recently, bipolar disorder (BD: e.g., Getz et al., 2003; Bozikas et al., 2006b, 2007; Van Rheenen and Rossell, in press-a, 2013a, 2014). Given that facial information is processed more easily than prosodic stimuli, the examination of facial affect processing (FAP) has garnered much research interest in the respective SZ and BD literatures (Adolphs, 2002; Kohler et al., 2010; Townsend and Altshuler, 2012; Van Rheenen and Rossell, 2013b). A large body of research has
⁎ Corresponding author at: Cognitive Neuropsychiatry Laboratory, Monash Alfred Psychiatry research centre (MAPrc), Level 4, 607 St Kilda Rd, VIC 3004, Australia. E-mail address: [email protected] (S.L. Rossell). 1 Affiliation where research was conducted.
produced consistent findings of impairment in both groups (Kohler et al., 2010; Van Rheenen and Rossell, 2013b); yet progress toward understanding the fundamental nature and origins of these impairments is currently hampered by a number of factors. Firstly, there remains substantial variability in the effect sizes reported within individual empirical studies in these disorders. It is possible that disparities in methodology, such as the use of different stimuli, different target emotions, and differing levels of task difficulty may be responsible for these inconsistent findings. Secondly, although it appears that these cohorts demonstrate similar FAP deficits, there has been little direct comparison using groups well matched on demographic and clinical symptomatology (Van Rheenen and Rossell, 2013b). This is noteworthy given the contribution that such comparative studies can make to our understanding of shared phenomenology and aetiological underpinnings in these disorders. Past comparisons of neurocognitive performance between the disorders suggest quantitative, but not qualitative differences, and it is possible that this pattern extends to FAP in SZ and BD (Schretlen et al., 2007). Thirdly, some studies have included patients with schizoaffective disorder as part of their SZ cohort, whilst others have excluded them (Kohler et al., 2010). Given the mixed symptom profile of such patients, the amalgamation of patients with schizophrenia and schizoaffective disorder might distort the impact on symptoms on FAP, particularly when comparing performance to that of a BD group.
http://dx.doi.org/10.1016/j.schres.2014.05.026 0920-9964/© 2014 Elsevier B.V. All rights reserved.
Please cite this article as: Rossell, S.L., et al., Investigating facial affect processing in psychosis: A study using the Comprehensive Affective Testing System, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.05.026
2
S.L. Rossell et al. / Schizophrenia Research xxx (2014) xxx–xxx
Finally, the impact of clinical symptoms remains a contentious issue. In BD, some researchers have demonstrated state-like FAP impairments during symptomatic but not euthymic phases (Venn et al., 2004; Langenecker et al., 2010), whereas others have demonstrated trait-like impairments during remission (Addington and Addington, 1998). Similarly in SZ, some studies have reported associations between FAP deficits and negative (Mandal et al., 1999), positive (Poole et al., 2000; Hall et al., 2004; Weniger et al., 2004) or both negative and positive symptomatology (Kohler et al., 2000), whilst others have failed to find these relationships (Cramer et al., 1989; Kucharska-Pietura et al., 2005). Clearly, continued investigation of the influence of these symptoms on FAP is necessary. In light of these limitations, the current study aimed to examine FAP in two different, well matched groups of psychosis patients and a group of healthy controls. Specifically, this study sought to determine the relative FAP profiles of SZ and BD patients by comparing them to controls on a battery of FAP assessments from the Comprehensive Affective Testing System (CATS: Froming et al., 2006). The CATS offers a standardised battery employing four subtests with experimental designs commonly used in the literature. It allows for insight to be gained regarding the impact that subtle differences in task methodologies can have on the estimation of FAP. It was predicted that patients with SZ would be impaired on the four subtests of the CATS compared to healthy controls, whilst BD patients would show intermediary performance. Correlational analysis was performed examining for associations between FAP and both psychotic and mood symptoms. Give the lack of clear findings in the literature with this regard, this analysis was considered exploratory.
2. Methods 2.1. Participants The current study included 54 patients with SZ and 43 patients with BD (type I). Patients were recruited via community support groups and community care units and were all out-patients. Diagnosis was ascertained using the Structured Clinical Interview for DSMIV (SCID: First et al., 1996). Current symptomology was acquired using the Positive and Negative Syndrome Scale (PANSS: Kay et al., 1987): global positive, negative and general scores were calculated. Ratings of depression and mania were made using the Beck Depression Inventory (BDI: Beck and Steer, 1987) and the Bech-Rafaelsen Mania Rating Scale (MRS: Bech et al., 1979), respectively. In the BD group, 11 were euthymic (BDI ≤ 10) and 32 depressed (BDI N 10); no patient in any group fit criteria for a current episode of mania. Only patients with no other co-morbid Axis 1 diagnoses (including schizoaffective disorder) were included in the study. Demographic and clinical characteristics are presented in Table 1. Within the SZ group, all 54 were taking antipsychotic medication, 5 were also taking mood stabilisers, and 8 were also taking antidepressants. Of the BD patients, 20 were taking antipsychotic medication, 30 were taking mood stabilisers, 10 were taking antidepressants and 3 participants were medication free.2 One hundred and twelve healthy control participants were recruited via newspaper advertisements. Control participants were excluded if they had any history of psychiatric disorder or a first degree relative with schizoaffective disorder, SZ or BD (SCID: First et al., 1996). Participants from all four groups met the following criteria: a) no history of neurological disorder or head trauma, b) no current
2 For patients taking antipsychotics a chlorpromazine equivalent score was calculated and correlated with FAP: no significant correlations were reported. In addition, patients taking mood stabilisers were compared with those not taking mood stabilisers and no group differences were demonstrated.
substance abuse or dependence (previous year), c) English as first language, d) between the ages of 18 and 65 years and e) predicted IQ N 85 as scored by the National Adult Reading Test (NART: Nelson and Willison, 1991). The study was carried out in accordance with the Declaration of Helsinki. The study had ethical approvals from North Western Mental Health, and The University of Melbourne, Melbourne Victoria. Informed consent of all the participants was obtained after the study had been fully explained. 2.2. Tasks Participants were tested with four of the 13 subtests from the CATS battery. The subtests are designed to examine four different aspects of FAP, taking approximately 20 min to complete. None of the subtests have practice items. Subtest order was counterbalanced. For each subtest, overall accuracy and reaction time (RT) to both correct and incorrect answers were recorded. The % correct and RTs for correct answers were used in the analyses for each of the subtests. The four subtests are as follows: (1) Affect Discrimination (AD); where participants are shown two faces at the midline (of the same actor) (N = 22 trials), and they must decide whether the faces show the same or different emotion. The emotions used are happy, sad, angry, fear, surprised, disgusted and neutral. (2) Name Affect (NA): where participants are shown a face and are required to select one of seven labels to describe the emotional expression of the face (N = 16 trials). The emotions used are happy, sad, angry, fear, surprised, disgusted and neutral. (3) Select Affect (SA): where participants are shown five faces (of the same actor), each face expressing a different emotion. A target emotion is named at the top of the screen and they are required to select which of the faces matches the target emotion (N = 20 trials). (4) Match affect (MA): where participants are shown one face at the top of the screen and five faces underneath (different actors), each expressing a different emotion. They are required to decide which of the five faces expresses the same emotion as the face above (N = 20 trials).
3. Statistical analysis Demographic and clinical group differences were assessed via oneway between-groups analysis of variance (ANOVA) or Chi-square tests. Given that the BD group had euthymic and depressed members, a multivariate ANOVA comparing these sub-groups on FAP was conducted in the BD sample. There was no significant main effect of mood, and all subsequent analyses were thus conducted using the BD sample as a whole. Repeated measures ANOVAs with Scheffe post-hoc tests and follow up one-way ANOVAs with simple contrasts were used to examine group related differences in performance on the CATS subtests. Due to group differences in the proportion of males and females across the groups, gender was included as a covariate in the analyses. Similarly, as there were group differences in age and levels of education, a two-fold validity check was performed to examine the effects of these demographic variables; first, by using Pearson's product moment correlations between age, education and accuracy/response time scores on the eight subtests. Even using a stringent alpha of 0.01, the majority of variables were significantly correlated with age and education. Therefore, the repeated measures and follow-up analyses were re-run with age and education as covariates. Relationships between subtest performance and the clinical characteristics from the PANSS, MRS and BDI were investigated using Pearson's product moment correlations in the combined patient sample (using p b .01).
Please cite this article as: Rossell, S.L., et al., Investigating facial affect processing in psychosis: A study using the Comprehensive Affective Testing System, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.05.026
S.L. Rossell et al. / Schizophrenia Research xxx (2014) xxx–xxx
3
Table 1 Demographic and clinical characteristics of the four participant groups (mean (standard deviation) unless otherwise stated). Group
C
SZ
BD
N Age M/F Years in education
112 35.14 (12.40) 37/75 16.82(3.20)
54 42.17 (10.50) 35/19 14.03 (3.54)
43 40.50 (10.64) 16/27 14.67 (3.65)
Group comparisons
Age of onset Years illness Total P Total N Total G MRS BDI Meds
– – – – – – –
23.49 (7.15) 18.83 (10.25) 14.50 (6.05) 11.19 (4.38) 24.26 (6.62) 2.02 (2.60) 15.06 (13.56) 459.05 (358.28)
21.42 (7.00) 19.12 (10.11) 10.79 (3.47) 9.16 (2.94) 24.91 (5.21) 1.86 (2.11) 17.58 (12.49) 173.01 (257.47)
F = 7.52, p b .001; C b SZ, C b BD χ2 = 15.63, p b .001. F = 14.78, p b .001; C N SZ, C N BD F = 2.00, NS F = .02, NS F = 12.80, p b .01; SZ N BD F = 6.74, p ≤ .01; SZ N BD F = .28, NS F = .10, NS F = .84, NS 19.30, p b .001; SZ N BD
C = control, SZ = schizophrenia, BD = bipolar disorder, M/F = male/female, PANSS = Positive and Negative Symptom Scale, Total P = Positive subscale from PANNS, Total N = Negative subscale from PANNS, Total G = Total general from PANNS, MRS = Bech–Rafaelsen Mania Rating Scale, BDI = Beck Depression Inventory, Meds = chlorpromazine equivalents, NS = non-significant.
4. Results 4.1. Demographic analyses As can be seen in Table 1, the SZ patients were significantly less educated and older than controls. Positive symptom ratings for the SZ patients were significantly higher than for BD patients. There were no group differences on negative symptom ratings, mania or depression; and no differences between patient groups on age or years of illness onset. 4.2. Group comparisons Tables 2 and 3 display the mean accuracy and RT data without age and gender correction for the four subtests across the three groups. 4.2.1. Accuracy There was a main effect of task (F(2.80,576.62) = 221.62, p b .001; accuracy decreasing in descending order of AD, SA, NA, MA) and group (F(2,206) = 17.97, p b .001; SZ b C, p b .001; BD b C, p b .05) as well as a significant two way group by task interaction effect (F(5.60,576.62) = 3.85, p ≤ .001); follow-up tests indicated that accuracy was at near ceiling and not significantly different across groups for the AD subtest. In contrast, accuracy was more variable for the other three subtests. In comparison to healthy controls, SZ patients showed a deficit in NA subtest performance; SZ patients also had a reduction in performance on the SA subtest compared with both controls and BD patients, although BD and controls performed equivalently. Both SZ and BD patients showed a reduction in performance on the MA subtest, but did not differ from each other. These group differences remained after controlling for age, education and gender (Table 4). 4.2.2. Reaction times Prior to controlling for age, education and gender, there was a main effect of task (F(2.58,531.76) = 58.32, p b .001; latencies increasing in Table 2 Non age and education corrected accuracy (%) data for the four CATS subtests. C
Affect discrimination Name affect Select affect Match affect
SZ
BD
Group comparisons
M
SD
M
SD
M
SD
95.4 80.4 94.6 83.2
6.5 11 6.4 10.9
92.5 73.1 86.1 72.4
9.3 13.3 13.8 13.5
94.5 76 91.9 76
6.7 11.9 7.8 13.3
ascending order of NA, AD, SA, MA) and group (F (2,206) = 6.67, p b .01; SZ N C, p = .002) as well as a two way task by group interaction effect (F(5.16,531.76), p b .05). However, when covariates were included in the analysis, only the group effect remained significant (F(2,203) = 3.98, p b .02). Thus, no follow-up tests were conducted. 4.3. Correlations Upon examination of the correlations between variables in the study, only two noteworthy sets of associations were found. First, SA and MA response time performance correlated significantly with scores on the PANSS positive scale (SA r = .29, p b .01; MA r = .29, p b .01), such that as positive symptomatology increased RTs increased. Second, response time performance on three of the four CATS subtests correlated with the MRS (AD r = .26, p b .01; SA r = .37, p b .001; MA r = .26, p ≤ .01), such that as mania severity increased RTs also increased (although mania symptoms were low within the cohort). 5. Discussion The current study investigated FAP in two well matched groups of psychosis patients and a group of healthy controls, with a view to comparing their respective FAP profiles on a standardised battery of tasks. In partial agreement with our predictions, when age, education and gender were controlled, SZ patients demonstrated an overall deficit in the speed with which they processed facial emotions. Both the SZ and BD groups also performed less accurately compared to controls overall. In SZ patients this accuracy effect was driven by impaired performance on the SA, NA and MA subtests. In the BD group, it was driven by performance on the MA task only. These results point toward a relatively generalised impairment of FAP in SZ. They also suggest a more specific facial affect matching deficit for those patients with BD. Given the absence of group differences on the AD task, which has an arguably greater affective load relative to the NA, SA and MA tasks that rely more heavily on cognitive processes associated with generating verbal responses or making selections from multiple alternatives, it is possible that the FAP findings observed Table 3 Non age and education corrected RT (seconds) data for the four CATS subtests. C
– SZ b C SZ b C, SZ b BD SZ b C BD b C
Affect discrimination Name affect Select affect Match affect
SZ
BD
M
SD
M
SD
M
SD
4.4 4.6 5.9 6.7
1.4 1.7 1.6 2.3
6.2 5.5 7 7
3.4 2.2 3.9 2.9
5.3 4.9 6.6 7
1.6 1.8 1.6 2.6
Group comparisons SZ N C SZ N C SZ N C –
Please cite this article as: Rossell, S.L., et al., Investigating facial affect processing in psychosis: A study using the Comprehensive Affective Testing System, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.05.026
4
S.L. Rossell et al. / Schizophrenia Research xxx (2014) xxx–xxx
Table 4 Group comparisons re-run with age and education as covariates. Subtest accuracy (%)
FAP × covariate
Corrected group effect
Age Fp
Education Fp
Gender Fp
Group Fp
AD NA SA
1.7 NS .13 NS 0.9 NS
4.84 .03 3.53 NS 1067 .001
5.63 .02 3.32 NS .05 NS
0.9 NS 4.3 .01 9.52 .001
MA
2.36 NS
12.80 .001
4.93 NS
6.87 .001
Significant group contrasts and mean effect size
– C N SZ (p = .004 d = 0.58) C N SZ (p = .001 d = 0.77) BD N SZ (p = .003 d = 0.51) C N SZ (p = .000, d = 0.90) C N BD (p = .03, d = 0.57)
C = healthy control; SZ = schizophrenia; BD = bipolar disorder; AD = affect discrimination; NA = name affect; SA = select affect; MA = match affect; NS = non-significant; age/ education/gender corrected task effect = F(2.78,564.64) = 10.48, p b .001), group effect (F(2,203) = 7.74, p b .001) and group × task interaction effect (F(5.56,564.45) = 2.47, p b .05).
here in both SZ and BD are attributable to a more general cognitive problem as opposed to one specific to emotion. The deficits evident in BD patients in particular, were generated on the subtest that is arguably the most difficult and cognitively demanding of them all; participants had to first recall the target facial expression, generate subsequent labels for the other facial expression choices and then match them under time pressure. As there is an established literature on cognitive difficulties in BD and a general consensus that such cognitive difficulties are simply less intense than those seen in SZ (e.g., Schretlen et al., 2007; Van Rheenen and Rossell, in press-b), it is certainly possible that this specific BD-patient impairment is reflective of an underlying cognitive load or working memory problem that, due to its subtle nature (relative to SZ), is only evident in the context of greater cognitive task demands. Importantly, the notion that the observed FAP deficits observed are consistent with a non-specific cognitive effect which is more expressed in SZ than in BD, fits well with previous unpublished work by our group indicating that neurocognitive dysfunction is a better predictor of FAP difficulties than underlying configural face processing abilities in BD (Joshua, 2010). Other research showing that neurocognition is predictive of social cognition in both SZ and BD (Brekke et al., 2005; Van Rheenen et al., in submission) also gives credence to this hypothesis. The data confirmed that there are trait FAP difficulties in BD during a euthymic phase (e.g., Bozikas et al., 2006b); and that depression does not exacerbate these findings. It was not possible to examine whether a phase of mania influenced performance, as no patient met criteria for mania in the BD (or either of the other psychosis groups). However, the correlational data did indicate that the severity of mania symptoms (albeit mild) was positively correlated with RT in both psychosis groups; such that greater subclinical mania resulted in longer RTs. This suggests that there may be a link between very low level manic symptomatology and performance on the CATS tasks, although further examination of this in an explicitly manic group is needed. The current data also indicated that response latency variable on two of the CATS subtests (SA and MA) were correlated with the positive symptom scale of the PANSS; that is, as positive symptom severity increased RTs also increased. There is other research to have suggested such a relationship between FAP and positive symptoms (Poole et al., 2000; Hall et al., 2004; Weniger et al., 2004). It is noteworthy that these correlations were only significant for the RT variables, and on two of the subtests. Overall, both positive and negative symptomology was within the mild to moderate range in the two clinical groups. It could be that this limited variability influenced the pattern of associations. Testing patients with severe positive and negative symptoms should be incorporated into future studies to continue investigations within the field. There are other limitations to the study beyond those already mentioned. Firstly, the specific subtests used in the current study utilised stimuli representative of seven basic emotions; however, the design of the subtests was limited such that there were too few trials to tease apart any potential deficits relating to specific emotions. Previous research has indicated SZ patients may have specific problems recognising
expressions of fear, sadness and anger (Edwards et al., 2001; Bediou et al., 2005). Others have indicated an overall problem processing all emotions (Joshua, 2010). Unfortunately, this particular assessment tool was not sensitive enough to explore this issue. Secondly, the CATS subtests do not automatically continue between trials as each participant's response is used as the indicator for the next trial to start. Some individuals took a long time to decide upon their answer, which was reflected by the large standard deviations for all the RT data. Given that participants were instructed to try to be as accurate as possible on all subtests, the RT data should also be interpreted with some caution. Finally, as the level of difficulty in the different subtests was not controlled, and as we did not include any neuropsychological tests, the impact of neurocognition on FAP in these cohorts is not known. Future research utilising the CATS FAP subtests to compare these groups would certainly do well to employ cognitive tasks to examine neuropsychological contributions to the observed deficits in BD and SZ patients. In particular, future studies should include measures of working memory and cognitive load into the test battery, and then use the data from these measures as covariates to ascertain whether FAP deficits are still present. In sum, this data has suggested that FAP deficits are present in SZ. They also suggest that severity of positive symptoms and mania exacerbate deficits. FAP impairment in BD, however, is more subtle and requires future research to elaborate on the role of neurocognition on FAP performance in this cohort. Role of funding source This work was funded by the Mental Health Research Institute of Victoria (MHRI), Australia. AG held an Early Career Training Fellowship from the National Health and Medical Research Council (NHMRC), Australia. Neither MHRI nor NHMRC played any further role in the conduct of this research. Contributors Author 1 designed the study, wrote the protocol, ran the analyses and wrote the manuscript. Author 2 managed the literature searches and assisted with the drafting of the manuscript. Authors 3, 4 and 5 were involved in recruitment of participants, and proof read the manuscript. All authors contributed to and have approved the final manuscript. Conflict of interest The authors declare that they have no conflicts of interest. Acknowledgements This research was funded by the Mental Health Research Institute, and the authors wish to thank Professor George Fink for this support.
References Addington, J., Addington, D., 1998. Facial affect recognition and information processing in schizophrenia and bipolar disorder. Schizophr. Res. 32, 171–181. Adolphs, R., 2002. Neural systems for recognizing emotion. Curr. Opin. Neurobiol. 12 (2), 169–177. Bech, P., Bolwig, T.G., Kramp, P., Rafaelsen, O.J., 1979. The Bech–Rafaelsen Mania Scale and the Hamilton Depression Scale. Acta Psychiatr. Scand. 59, 420–430. Beck, A.T., Steer, R.A., 1987. The Beck Depression Inventory Manual. The Psychological Corporation, San Antonio.
Please cite this article as: Rossell, S.L., et al., Investigating facial affect processing in psychosis: A study using the Comprehensive Affective Testing System, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.05.026
S.L. Rossell et al. / Schizophrenia Research xxx (2014) xxx–xxx Bediou, B., Franck, N., Saoud, M., Baudouin, J.-Y., Tiberghien, G., Daléry, J., d'Amato, T., 2005. Effects of emotion and identity on facial affect processing in schizophrenia. Psychiatry Res. 133 (2–3), 149–157. Bozikas, V.P., Kosmidis, M.H., Anezoulaki, D., Giannakou, M., Andreou, C., Karavatos, A., 2006a. Impaired perception of affective prosody in schizophrenia. J. Neuropsychiatry Clin. Neurosci. 18, 81–85. Bozikas, V.P., Tonia, T., Fokas, K., Karavatos, A., Kosmidis, M.H., 2006b. Impaired emotion processing in remitted patients with bipolar disorder. J. Affect. Disord. 91, 53–56. Bozikas, V.P., Kosmidis, M.H., Tonia, T., Andreou, C., Focas, K., Karavatos, A., 2007. Impaired perception of affective prosody in remitted patients with bipolar disorder. J. Neuropsychiatry Clin. Neurosci. 19 (4), 436–440. Brekke, J., Kay, D.D., Lee, K.S., Green, M.F., 2005. Biosocial pathways to functional outcome in schizophrenia. Schizophr. Res. 80 (2), 213–225. Cramer, P., Weegmann, M., O'Neil, M., 1989. Schizophrenia and the perception of emotions. How accurately do schizophrenics judge the emotional states of others? Br. J. Psychiatry 155, 225–228. Edwards, J., Pattison, P.E., Jackson, H.J., Wales, R.J., 2001. Facial affect and affective prosody recognition in first-episode schizophrenia. Schizophr. Res. 48 (2–3), 235–253. First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B., 1996. Structured Clinical Interview for DSM-IV Axis I Disorders, Clinician Version (SCID-CV). American Psychiatric Press, Washington, D.C., Froming, K., Levy, M., Schaffer, S., Ekman, P., 2006. The Comprehensive Affect Testing System. Psychology Software, Inc., Getz, G.E., Shear, P.K., Strakowski, S.M., 2003. Facial affect recognition deficits in bipolar disorder. J. Int. Neuropsychol. Soc. 9, 623–632. Hall, J., Harris, J.M., Sprengelmeyer, R., Sprengelmeyer, A., Young, A.W., Santos, I.M., Johnstone, E.C., Lawrie, S.M., 2004. Social cognition and face processing in schizophrenia. Br. J. Psychiatry 185, 169–170. Hooker, C., Park, S., 2002. Emotion processing and its relationship to social functioning in schizophrenia patients. Psychiatry Res. 112 (1), 41–50. Joshua, N.R., 2010. Face Processing In Schizophrenia: An Investigation Of Configural Processing And The Relationship With Facial Emotion Processing And Neurocognition. PhD Thesis University of Melbourne. Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The Positive and Negative Syndrome Scale (PANSS) for schizophrenia. Schizophr. Bull. 13, 261–276. Kohler, C.G., Bilker, W., Hagendoorn, M., Gur, R.E., Gur, R.C., 2000. Emotion recognition deficit in schizophrenia: association with symptomatology and cognition. Biol. Psychiatry 48 (2), 127–136. Kohler, C.G., Walker, J.B., Martin, E.A., Healey, K.M., Moberg, P.J., 2010. Facial emotion perception in schizophrenia: a meta-analytic review. Schizophr. Bull. 36 (5), 1009–1019. Kucharska-Pietura, K., David, A.S., Masiak, M., Phillips, M.L., 2005. Perception of facial and vocal affect by people with schizophrenia in early and late stages of illness. Br. J. Psychiatry 187, 523–528.
5
Langenecker, S.A., Saunders, E.F.H., Kade, A.M., Ransom, M.T., McInnis, M.G., 2010. Intermediate: cognitive phenotypes in bipolar disorder. J. Affect. Disord. 122 (3), 285–293. Mandal, M.K., Jain, A., Haque-Nizamie, S., Weiss, U., Schneider, F., 1999. Generality and specificity of emotion-recognition deficit in schizophrenic patients with positive and negative symptoms. Psychiatry Res. 87 (1), 39–46. Nelson, H.E., Willison, J., 1991. National Adult Reading Test manual, 2nd ed. NFER-Nelson, Windsor. Poole, J.H., Tobias, F.C., Vinogradov, S., 2000. The functional relevance of affect recognition errors in schizophrenia. J. Int. Neuropsychol. Soc. 6 (6), 649–658. Rossell, S.L., Boundy, C.L., 2005. Are auditory-verbal hallucinations associated with auditory affective processing deficits? Schizophr. Res. 78 (1), 95–106. Rossell, S.L., Van Rheenen, T.E., Groot, C., Gogos, A., Joshua, N.R., 2013. Investigating affective prosody in psychosis: a study using the Comprehensive Affective Testing System. Psychiatry Res. 210 (3), 896–900. Schretlen, D.J., Cascella, N.G., Meyer, S.M., Kingery, L.R., Testa, S.M., Munro, C.A., Pulver, A.E., Rivkin, P., Rao, V.A., Diaz-Asper, C.M., Dickerson, F.B., Yolken, R.H., Pearlson, G.D., 2007. Neuropsychological functioning in bipolar disorder and schizophrenia. Biol. Psychiatry 62 (2), 179–186. Shea, T.L., Sergejew, A.A., Burnham, D., Jones, C., Rossell, S.L., Copolov, D.L., Egan, G.F., 2007. Emotional prosodic processing in auditory hallucinations. Schizophr. Res. 90 (1–3), 214–220. Townsend, J., Altshuler, L.L., 2012. Emotion processing and regulation in bipolar disorder: a review. Bipolar Disord. 14 (4), 326–339. Van Rheenen, T.E., Rossell, S.L., 2013a. Auditory-prosodic processing in bipolar disorder; from sensory perception to emotion. J. Affect. Disord. 151 (3), 1102–1107. Van Rheenen, T.E., Rossell, S.L., 2013b. Is the non-verbal behavioural emotion-processing profile of bipolar disorder impaired? A critical review. Acta Psychiatr. Scand. 128, 163–178. Van Rheenen, T., Rossell, S., 2014. Multimodal emotion integration in bipolar disorder: an investigation of involuntary cross-modal influences between facial and prosodic channels. J. Int. Neuropsychol. Soc. (in press) (April 11). Van Rheenen, T.E., Rossell, S.L., 2014. Let's face it: facial emotion processing is impaired in bipolar disorder. J. Int. Neuropsychol. Soc. 20 (02), 200–208. Van Rheenen, T.E., Rossell, S.L., 2014. An empirical evaluation of the MATRICS consensus cognitive battery in bipolar disorder. Bipolar Disord. 16 (3), 318–325. http://dx.doi. org/10.1111/bdi.12134. Van Rheenen, T.E., Meyer, D., Rossell, S.L., 2014n. Pathways between neurocognition, social cognition and emotion regulation in bipolar disorder, (in submission). Venn, H.R., Gray, J.M., Montagne, B., Murray, L.K., Michael Burt, D., Frigerio, E., Perrett, D.I., Young, A.H., 2004. Perception of facial expressions of emotion in bipolar disorder. Bipolar Disord. 6 (4), 286–293. Weniger, G., Lange, C., Ruther, E., Irle, E., 2004. Differential impairments of facial affect recognition in schizophrenia subtypes and major depression. Psychiatry Res. 128 (2), 135–146.
Please cite this article as: Rossell, S.L., et al., Investigating facial affect processing in psychosis: A study using the Comprehensive Affective Testing System, Schizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.05.026