اختلال در بازشناسی ابراز هیجانی در اسکیزوفرنی: مطالعه ارتباط خانوادگی کوبایی

Abstract It is well established that schizophrenia is associated with difficulties in recognizing facial emotional expressions, but few studies have reported the presence of this deficit among their unaffected relatives. This study attempts to add new evidence of familial association on an emotional expression processing test. The study evaluated the performance of 93 paranoid schizophrenia patients, 110 first-degree relatives of probands from multiplex schizophrenia families, and 109 nonpsychiatric controls on a facial emotional recognition test using a computer morphing technique to present the dynamic expressions. The task entailed the recognition of a set of facial expressions depicting the six basic emotions presented in 21 successive frames of increasing intensity. The findings indicated that schizophrenia patients were consistently impaired for the recognition of the six basic facial expressions. In contrast, their unaffected relatives showed a selective impairment for the recognition of disgust and fearful expressions. Familial association of selective facial emotional expressions processing deficit may further implicate promising new endophenotypes that can advance the understanding of affective deficits in schizophrenia.

1. Introduction There is growing body of data indicating that an emotional expression processing (EEP) deficit is associated with cognitive domains in schizophrenia (SZ) (Kohler et al., 2000, Edwards et al., 2002, Sachs et al., 2004 and Martin et al., 2005). This deficit is considered a fundamental core disturbance in this illness (Kohler et al., 2004 and Gur et al., 2007a). However, there is mixed evidence regarding whether these deficits are specific to SZ (Blair et al., 2004 and Pinkham et al., 2007). EEP have been reported in both first-episode and remitted SZ patients and do not appear to be attributable to the effects of medication, thus this impairment may be state independent (Gur et al., 2007a). In line with these findings, the study of such EEP impairment in non-affected relatives of SZ patients would support the hypothesis of trait status because relatives are not affected by either the confounding variables associated with chronic illness or the medication effects. First-degree relative's studies suggest that cognitive deficiencies found in the SZ patients' relatives are similar to those found in the patients. These abnormalities are caused by familial predisposition to SZ (Kéri and Janka, 2004, Sitskoorn et al., 2004 and Snitz et al., 2006). This fact led to the familial association concept regarded as the presence of neurocognitive traits in nonaffected relatives of ill probands showing higher deficit rates than the general population (Snitz et al., 2006). Therefore, abnormalities that are overrepresented in unaffected relatives as well as SZ patients are likely to be produced by genes that increase risk of the disorder. This approach has been applied to a number of potential neurobiological SZ markers as a valuable step to explore their utility as endophenotypes (Gottesman and Gould, 2003, Schulze et al., 2003, Schürhoff et al., 2003 and Braff et al., 2007). Additionally, the search for cognitive impairments in relatives from multiplex families has proved to be more profitable due to the genetic vulnerability of these families (Tsuang et al., 2006). The strategy of dividing multiplex families (families having at least two SZ patients) from simplex families (families with only one member with SZ) has been useful in detecting the association of higher genetic loading with more severe cognitive deficits, which are potential markers for genetic susceptibility to SZ, in nonpsychotic relatives of SZ patients (Chang et al., 2009). Few studies have examined the presence of EEP impairment in relatives of people with SZ. Although there are studies that failed to find significant differences on tests of facial affect recognition between relatives and controls (Toomey et al., 1999 and Bolte and Poustka, 2003), positive evidence of affect perception deficits were found in unaffected siblings of individuals with SZ (Van't Wout et al., 2007; Leppänen et al., 2008; Alfimova et al., 2009). Gur et al. (2007a) have reported evidence of significant impairment in the accurate discrimination of emotion intensity among relatives of multiplex SZ families. Furthermore, Bediou et al. (2007) found an impaired performance of facial emotion recognition in first-episode SZ patients (n = 40) and their 30 first-degree relatives (30) compared to normal subjects (n = 26) using sensitive measures of emotional facial expression recognition using morphed faces. These findings suggest that EEP deficit reflects genetic influences shared among SZ families. However, they highlight the need of further examination of EEP abilities in nonaffected relatives (Gur et al., 2007a). In line with the prediction that the EEP deficit of SZ patients' relatives is between the controls and patients, the authors' goal was to replicate evidence of familial association of EEP impairment in a large sample of multiplex Cuban SZ families using an emotional expression multimorph task. The existence of familial association for the emotion recognition deficit in Cuban SZ families could support this cognitive impairment as an endophenotypic marker useful for genetic analyses of these complex psychiatric disorders in this population

3. Results 3.1. Group demographic comparisons. The three groups differed significantly in terms of educational level (χ2 = 20.96, df = 2, P < 0.001). Multiple comparison analysis indicated that the control group had a higher educational level than the relatives group (P = 0.003) and the group of patients (P < 0.000) who did not differ from each other (P = 0.339). The three groups differed in age (F (2, 307) = 21.58, P < 0.001), being the relatives older than both the patients (P < 0.001) and the controls (P < 0.001) who have similar ages (P = 0.693). The three groups also differed in gender (χ2 = 22.29, df = 2, P < 0.001). The patient group included more males than relatives (P = 0.002) and control groups (P < 0.001). Therefore, in order to show that the significant differences among groups were not attributable to group differences in age, gender, and educational level; MANCOVAs were computed with those variables as covariates. 3.2. Comparison of emotional expression multimorph task performance (EEMT) between groups 3.2.1. Sensitivity for different emotions Fig. 1 shows the average of EEMT sensitivity between SZ patients, their relatives, and controls. As can be seen, the healthy controls showed the lowest threshold levels (higher sensitivity) of correct recognition of emotional stimuli regardless of emotional expressions, where the happy was the easiest emotion to recognize. Performance of emotional expression processing in the studied groups. Fig. 1. Performance of emotional expression processing in the studied groups. Figure options Table 1 shows the EEMT performance by subject group. A 3 ×6 MANCOVA was conducted with group (SZ patients, first-degree relatives, and controls) as the between-factors and mean expression recognition scores as dependent variables (happy, surprised, disgusted, angry, sadness, and fearful). Table 1. Performance on the emotional expression multimorph task by groups, means (± S.D.). SZ patients Relatives Controls P vs C † P vs R † R vs C† Mean number of stage (sensitivity) F (1,283) F (1,283) F (1,283) Happy 9.34(± 3.22) 8.15(± 2.39) 8.07(± 1.60) 7.08 P = 0.008 7.90 P = 0.005 0.005 P = 0.939 Surprised 12.27(± 3.12) 11.01(± 2.60) 10.52(± 2.08) 11.91 P < 0.001 12.24 P < 0.001 0.05 P = 0.809 Angry 13.87(± 2.71) 13.35(± 2.70) 12.29(± 2.14) 5.47 P = 0.019 2.38 P = 0.122 0.95 P = 0.328 Disgusted 14.30(± 2.74) 13.40(± 2.51) 12.31(± 2.03) 16.56 P < 0.001 7.07 P = 0.008 2.98 P = 0.085 Sad 14.74(± 2.52) 13.86(± 2.53) 13.32(± 2.87) 9.19 P = 0.002 9.55 P = 0.002 0.03 P = 0.846 Fearful 14.57(± 2.98) 13.91(± 2.91) 12.34(± 2.31) 17.17 P < 0.001 4.60 P = 0.032 5.34 P = 0.021 Accuracy F (1,275) F (1,275) F (1,275) Happy 0.96 (± 0.13) 0.97(± 0.11) 0.99(± 0.02) 0.76 P = 0.382 0.11 P = 0.730 1.64 P = 0.200 Surprised 0.90 (± 0.19) 0.94(± 0.14) 0.98(± 0.05) 8.43 P = 0.003 3.52 P = 0.061 1.53 P = 0.216 Angry 0.92 (± 0.16) 0.90(± 0.21) 0.95(± 0.12) 0.18 P = 0.671 1.85 P = 0.173 0.82 P = 0.365 Disgusted 0.84 (± 0.21) 0.88 (± 0.17) 0.95(± 0.07) 10.29 P = 0.001 1.48 P = 0.224 4.94 P = 0.026 Sad 0.79 (± 0.20) 0.85 (± 0.18) 0.93 (± 0.11) 11.94 P < 0.001 5.99 P = 0.014 1.58 P = 0.208 Fearful 0.83(± 0.21) 0.87 (± 0.23) 0.98(± 0.05) 17.67 P < 0.001 4.01 P = 0.045 6.27 P = 0.012 MANCOVA analysis with age, sex, and educational level as covariates. †Multivariate test of significance for planned comparisons. P vs C = SZ patients versus controls. P vs R = SZ patients versus unaffected relatives of probands. R vs C = unaffected relatives versus controls. Table options The results of the MANCOVA revealed a main effect of group (Wilks's test = 0.890, F = 2.75, df = 12, P = 0.001), age (Wilks's test = 0.936, F = 2.75, df = 6, P = 0.005), and educational level (Wilks's test = 0.916, F = 2.07, df = 12, P = 0.017). There was no main effect of sex nor any interaction of group with educational level or sex. It appeared that SZ patients performed more poorly on emotional expression recognition than relatives and control groups for the six emotions. There was a significant main effect for emotion (F = 15.14, df = 5, P < 0.001), but not a significant interaction between group and emotion (F = 1.64, df = 10, P = 0.089) which suggests that the pattern in facial affect recognition (happy is easiest with increasing difficulty for the negative emotions) is similar for the three groups. Pearson's correlation analysis revealed that younger participants presented a lower degree of difficulties when recognizing disgust (r = 0.16, P = 0.004), fear (r = 0.15, P = 0.006), and anger (r = 0.16, P = 0.004). This analysis also showed that people with higher educational level performed better for the recognition of happy (r = − 0.12, P = 0.027; disgust (r = − 0.15, P = 0.005); surprise (r = − 0.17, P = 0.002); fear (r = − 0.19, P = 0.001), and anger (r = − 0.24, P < 0.001). A multivariate test of significance for contrast analysis comparisons was conducted to examine the differences in term of sensitivity to the six basic emotions. These data are presented in Table 1. The results showed that the SZ subjects performed significantly worse in recognizing the basic emotions than the relatives and control groups, except for angry where the relatives and SZ patients groups showed no significant differences (P = 0.122). The planned contrast comparison also indicated that relatives were equally sensitive as controls to the graded intensity of happy, surprised, disgusted, angry, and sad emotional expressions but showed a reduced sensitivity for the recognition of fearful expression (P = 0.021). The Pearson's correlation analysis showed only significant but low familial correlation between SZ patients and their first-degree relatives on fearful expression recognition (r = 0.17, P = 0.015), suggesting a heritability component. 3.2.2. Accuracy for different emotions A 3 × 6 MANCOVA analysis with accuracy EEMT scores as dependent variables showed a significant group effect (Wilks's test = 0.87, F = 3.01, df = 12, P < 0.001), resulting in a distinctive pattern of emotional recognition accuracy in the three defined groups. The results also indicated a significant effect for age (Wilks's test = 0.93, F = 3.34, df = 6, P = 0.003) and educational level (Wilks's test = 0.90, F = 2.23, df = 12, P = 0.009). There was neither a main effect of sex nor any interaction of group with educational level or sex. Analysis by emotion expression showed no main effect of emotion (F = 1.88, df = 5, P = 0.093), and significant emotion expression × age (F = 4.38, df = 5, P < 0.001), emotion expression × group (F = 3.46, df = 10, P < 0.001) and emotion × educational level interactions (F = 2.02, df = 10, P < 0.001). Contrast analysis comparisons indicated that EEMT performance differed between the SZ patients and control groups in the accuracy recognition of surprised, disgusted, sad, and fearful expressions. Relatives' groups were only significantly more likely to make errors for disgust and fearful expressions than the control groups (P = 0.026 and P = 0.021, respectively). Relatives compared to SZ patients showed similar emotion recognition performance in terms of accuracy for disgusted (P = 0.224), angry (P = 0.173), surprised (P = 0.061), and happy (P = 0.730). 3.3. Association of emotion recognition ability with clinical symptoms The Pearson's correlation analysis between the sensitivity to each emotion and PANSS scores for patients showed that all emotions positively correlated with both the negative and the positive PANSS scales (see Table 2). Likewise, in the relatives group, this analysis indicated correlation just for fearful expression emotion recognition with PANSS negative subscale (r = 0.27, P = 0.005). Table 2. Correlation between the performance of emotion recognition and the global scores of negative and positives PANSS subscales. SZ patients Relatives Controls PSS NSS NSS NSS r (P) r (P) r (P) r (P) Mean number of stage (sensitivity) Happy 0.22 (0.031) 0.31(0.003) 0.11 (0.262) − 0.07(0.601) Surprised 0.29 (0.041) 0.31 (0.002) 0.05 (0.723) − 0.13(0.345) Angry 0.26 (0.013) 0.16 (0.073) 0.12 (0.216) − 0.17(0.217) Disgusted 0.21 (0.04) 0.19 (0.063) 0.13 (0.149) − 0.17(0.212) Sad 0.22 (0.013) 0.27 (0.011) 0.17 (0.076) − 0.19(0.176) Fearful 0.26 (0.011) 0.31 (0.002) 0.28 (0.004) − 0.21(0.127)