|Home | About | Journals | Submit | Contact Us | Français|
Conscious awareness of emotion is adaptive and its disruption in schizophrenia can impact social functioning. This study assessed levels of emotional awareness for self and others in social scenarios (LEAS) in 21 individuals with schizophrenia spectrum disorders (SSD) and 20 healthy individuals. Individuals with SSD had lower levels of emotional awareness for others in complex social scenarios, but not simple social scenarios; no difference was found in emotional awareness for self. Higher levels of emotional awareness were associated with better quality of life in patients. Patients also reported higher social anhedonia. Healthy individuals’ higher levels of emotional awareness for self were associated with lower anhedonia, but this relationship was not found in SSD patients. Individuals with schizophrenia have particular difficulty anticipating others’ emotional responses in complex social situations. Further, this deficit is independent of anhedonia in SSD patients, indicating a systemic disruption in the integration of emotional processing.
Studies have described a variety of deficits in emotion processing in individuals with schizophrenia. An extensive review by Tremeau (2006) specifically identified dysfunction in schizophrenia in the domains of emotion expression, emotion experience and emotion recognition. In regard to emotion recognition, individuals with schizophrenia have been shown to demonstrate deficits in recognizing facial expressions, gestures, and prosody (reviewed in Mandal et al, 1998). Notably, most emotion recognition tasks assess the ability to identify emotion via a single channel (i.e.: facial expression, prosody) and thus are unable to capture the complexity of expression through multiple channels, over time, in real-life social situations. It is clearly essential to thoroughly assess abnormalities in each channel, but it is also useful to keep in mind that each channel provides information regarding only some elements involved in the dynamic complexity of social interactions.
Experience of emotions can be assessed in a variety of ways, most commonly intensity of emotion or valence of emotion. However, there are multiple additional aspects of emotional experience and recognition that are often not specifically assessed, such as awareness of emotional state. Damasio (1994) points out that the advantage of conscious awareness of emotion is that it allows emotional information to be integrated with cognitive processes. Conscious awareness of emotion is the conscious metarepresentation of an individual’s emotional state and therefore offers flexibility of emotional response in the moment to help achieve adaptational success (Lane, 2000).
A number of measures have been developed to assess individuals’ ability to recognize their emotional states. The Toronto Alexithymia Scale (TAS-20) is a self-report instrument that measures an individual’s ability to recognize and describe emotional states (Bagby et al, 1994; Taylor et al, 1988). In a study by Cedro et al (2001), 50 individuals diagnosed with schizophrenia demonstrated greater difficulty in two of three factors of the TAS-20 than controls: ‘difficulty identifying feelings’ and ‘difficulty describing feelings’. A second study using the TAS-20 (Serper and Berenbaum, 2008) compared its relationship with specific psychotic symptoms, leading to the conclusion that lower attention to emotion and lower emotion clarity were both associated with more severe hallucinations. These results support the existence of alexithymia in schizophrenia; however, the fact that the TAS-20 scale consists of self-rated agreement statements, and presupposes awareness of the deficit to be reported, raises the concern that emotion recognition deficits that patients do not detect would not be captured by this scale.
When discussing emotional awareness in social scenarios, it is important to consider the current understanding of social cognition in schizophrenia. Frith (1992) hypothesized that many symptoms typical of schizophrenia may be accounted for by a specific cognitive incapacity of schizophrenic patients to accurately attribute mental states to themselves or others (commonly referred to as ‘theory of mind’). A well-developed ‘theory of mind’ allows us to understand ourselves and others not just as sensory objects but also as subjective beings with mental states; we can then anticipate and make use of others’ mental states to guide our own behaviors (Frith and Frith, 1999). Most studies have supported a relationship between ‘theory of mind’ abilities and social functioning in schizophrenia (Roncone et al, 2002; Mazza et al, 2003, but also see Ba et al, 2008).
The concept of ‘affective theory of mind’ involves an understanding and appreciation of the emotional state of the people described in the scenario (measured with ‘irony’ and ‘faux pas’ tasks) rather than a non-emotional, cognitive appreciation of the people’s understanding of the situation (measured with ‘false beliefs’ tasks) (Shamay-Tsoory et al, 2005). Poorer performance on affective ‘theory of mind’ tasks have been found in schizophrenia to be correlated with negative symptoms (Shamay-Tsoory et al, 2007a) and impaired social functioning (Shamay-Tsoory et al, 2007b). Differences in ‘affective’ versus ‘non-affective theory of mind’ tasks have even been postulated to relate to certain behaviors in schizophrenia, such as violence (Abu-Akel and Abushua’leh, 2004). These findings support the idea that certain cognitive functions such as ‘theory of mind’ have an affective dimension that needs to be considered.
Another important issue is to consider the specific strategy which individuals use when attempting to recognize or infer their own or another individual’s emotional state. Social vignettes are often used as test stimuli, and subjects are typically asked to identify the emotions experienced by characters within the vignettes. However, it is not definitively clear, as of yet, the specific processes that underlie accurate responding on these tasks. Although it often appears to be assumed that individuals attempt to gauge another’s emotional experience by imaging their own in a similar situation (simulation strategy), Siemer and Reisenzen (2007) found that individuals could use theory-based strategies (which focus on recalling or recreating appraisals of the circumstances and individuals in vignettes) to provide accurate responses to such questions. Specifically, Siemer and Reisenzen indicate that, if one can make the assumption that appraisals of situations are proceduralized and unconscious, theory-based strategies using these appraisals would generate faster results than attempts to emotionally simulate the situation described in the vignette. This importantly implies that cognitive capacities, such as ability to efficiently recall appraisals relevant to the current dilemma, rather than emotional skills, such as ability to recreate one’s own or another’s emotional state, would primarily account for performance. Consistent with this perspective, Whittaker et al (2001) found that, although individuals with schizophrenia were impaired in contrast to healthy subjects in accurately naming emotional states based on descriptions of faces and scenarios, this difference appeared to be largely accounted for by IQ differences between groups. In order to at least partially address the important question of whether performance primarily reflects efficient cognitive processes versus emotional simulation processes, in the current study, we review results of analyses both with and without indices of general cognitive capacity included as a covariate.
Emotional domains of social cognition in schizophrenia constitute an area of ongoing interest; it is important to distinguish, however, that social cognition covers a wider spectrum of cognitive characteristics beyond the emotional domains (Eack et al, 2008). ‘Emotional awareness’ focuses directly on knowledge about emotion in oneself or another person. In the current study, we were particularly interested in assessing differences in awareness of emotional state between schizophrenia-spectrum and healthy subjects.
Twenty-one individuals meeting Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994) criteria for schizophrenia spectrum disorders –SSD- (17 schizophrenia; 4 schizoaffective disorder) were recruited at the University of Illinois at Chicago and via ads in the community, special-interest web sites (National Alliance on Mental Illness), and physician referral. All SSD participants were outpatients and completed testing after a minimum of 4 weeks of treatment on a stable medication regimen. Seventeen of the SSD participants were taking second-generation antipsychotic medications, one was taking first-generation antipsychotic medications, two were taking medications other than antipsychotics (anti-depressants, benzodiazepines), and one was unmedicated.
Twenty healthy individuals were recruited via ads in public transportation and public places (supermarkets). For both subject groups, potential participants were excluded if they reported histories of head trauma with loss of consciousness greater than 10 minutes, current substance abuse or substance dependence within the past 6 months, or neurological or systemic illness that might influence cognitive functioning.
Diagnoses for all participants were established with the Structured Clinical Interview for DSM-IV diagnosis (First et al, 1997); a non-patient version of the SCID was used for healthy controls. There were no significant differences between healthy and SSD participants on age, education, parental and subjects’ socioeconomic status, estimated premorbid intelligence (as assessed with the Reading subtest of the Wide Range Achievement Test 3rd Edition, Wilkinson, 1993), race, or sex (see Table 1).
As noted earlier, we were concerned that other scales of emotional awareness may require a level of insight and self-reflection that may be difficult for some of our subjects. We also wanted our measure to take into account the complexity of commonplace social interactions. Based on these concerns, we chose to use the Levels of Emotional Awareness Scale (LEAS) (Lane and Schwartz, 1987), which uses brief stories and qualitative and well as quantitative analyses to assess awareness of one’s own and others’ emotions. The LEAS scale is based on a cognitive-development theory that suggests that emotional awareness is the ability to recognize and describe emotions in oneself and others, and that this ability develops in a process similar to that described by Piaget for cognition in general (Lane and Schwartz, 1987). Emotional awareness in individuals with schizophrenia or SSD has not been tested using this measure. The LEAS has demonstrated decreased emotional awareness in patients with post-traumatic stress disorder (Frewen et al, 2008), eating disorders (Bydlowski et al, 2005) and borderline personality disorder (Levine et al, 1997); there is decreased emotional awareness for others but not self in patients with unipolar depression (Donges et al, 2005) and increased emotional awareness in patients with generalized anxiety disorder (Novick-Kline et al, 2005). Using the LEAS, no difference between controls and individuals with somatoform disorders has been demonstrated (Waller and Scheidt, 2004), but patients with somatoform disorders and psychological factors associated with somatic illness showed lower emotional awareness when compared to other clinical groups (Subic-Wrana et al, 2005). In patients with psoriasis, individuals with low levels of emotional awareness were more reactive to stress but also more responsive to treatment (Consoli et al, 2006); the LEAS has also been shown to capture a change in emotional awareness after treatment in patients with somatoform disorders and psychological factors due to somatic illness (Subic-Wrana et al, 2005).
Specifically, the LEAS consists of twenty evocative interpersonal situations, each described in two to four sentences, and each involving at least two persons. One scenario was presented per page, followed by two questions: “How would you feel?” and “How would the other person feel?” at the top of each page. Subjects wrote their response in the remainder of each page. Examples of items in the scale include “You are walking through the desert with a guide. You ran out of water hours ago. The nearest well is two miles away according to the guide’s map. How would you feel? How would the guide feel?” and “Someone who has been critical of you in the past pays you a compliment. How would you feel? How would the other person feel?” The responses were open-ended and participants were told their responses could be as short or as long as necessary to express how they would feel. Each item was given one score for the answer to how the participant would feel (Self) and one score for the answer to how the other person in the scenario would feel (Other). Responses were rated and quantified following specific structural criteria explained in the scoring manual provided by the author of the scale (Lane and Schwartz, 1987). Scores ranged between 0 and 4, with a higher score indicating more ability to describe emotional experiences. Level 0 indicated no emotional response was given in the answer or that the statement was reflective of an act of cognition (i.e.: “I think this is fair”). Level 1 indicated that the description included bodily or physical sensations, such as “pain” or “nausea”. Level 2 referred to a response describing an action tendency, such as “I’d feel like punching a wall” or undifferentiated emotions, such as “good” or “bad”. Level 3 indicated the use of well-differentiated, specific emotions (i.e.: “happy”). Level 4 indicated a response that blends different types of specific emotions, such as “sad and grateful”. To ensure reliability in scoring, the LEAS was independently scored by two raters and in the rare event of a disagreement in the scoring, a mutual agreement was determined as the final score for the item.
We split both the “self” and “other” scales into subscales reflecting the level of complexity of the scenarios described; designation of complexity level was based on ratings of an independent sample of healthy individuals. Complex items tended to present situations involving more people than simple items or invoke situations from the past or potential future that might have an effect in the present whereas simple items were focused on the present or involved fewer people. Along with scores for LEAS Self and LEAS Other full scales each participant also received separate Self and Other scores for the “simple” and “complex” items. Administration of LEAS was not timed.
Each participant completed the revised Chapman scales assessing physical and social anhedonia (Chapman et al, 1976 and Eckblad et al, 1982). The physical anhedonia scale assesses the degree to which individuals report that they have found physical sensations to be pleasurable; the revised social anhedonia scale assesses the degree to which individuals report that they have found social engagement to be pleasurable.
Finally, clinical symptoms were assessed using the Positive and Negative Symptom Scale (PANSS; Kay et al., 1987) and the Hamilton Depression Ratings Scale (HDRS; Hamilton, 1960). Adaptive functioning was assessed using the Heinrichs-Carpenter Quality of Life Scale (HCQL; Heinrichs et al., 1984), which assesses quality of life in four domains: social activity (Interpersonal Scale), productive activity either in the workplace or at home (Instrumental Scale), use of objects and participation in common activities (Common Scale), and intrapsychic functioning as indicated by motivation, anhedonia, curiosity, and empathy (Intrapsychic Scale). Information for the HCQL scoring was obtained from the participants.
The patient and control groups did not differ in age, sex distribution, socioeconomic status or premorbid intellectual functioning. Premorbid intellectual functioning is assessed using the Wide Range Achievement Test (WRAT) which assesses specific aspects of verbal functioning that have been demonstrated to be influenced by education and intelligence levels, but which are not significantly influenced by psychiatric illness. In this sample, this estimated premorbid IQ fell within a normal range for both groups. (Table 1)
In all analyses, premorbid IQ (as assessed by the WRAT) was included as a covariate because it was found to be significantly correlated with some emotion measures; particularly WRAT scores were significantly correlated with social anhedonia (r=−.49 p<.05) and physical anhedonia (r=−.68, p<.01) in the patient sample, and with control subject scores on the simple LEAS Other items only (r=.46, p<.05). Sex was also included as a factor in all analyses, as females have shown to display higher levels of emotional awareness in a previous study using non-clinical samples (Barrett et al, 2000), but not in the current study. Self and Other scores on the LEAS were compared across groups with a univariate analysis of variance. The LEAS Other complex subscale had a statistically significantly lower score (F(1,35) = 4.41, p<.05; see Table 2) in patients compared to controls; there were no significant differences between groups in the remaining LEAS full scales or subscales. The number of words used for describing the emotional responses did not differ between the groups when considering the full scale (F (1,35) = .49, ns) or when dividing the scale in simple (F (1,35) = .27, ns) and complex (F (1,35) = .76, ns) items. Within each group, ratings for simple and complex items were not significantly different from each other despite presenting more socially complex scenarios. The LEAS score distribution was not different between groups (for self, other or all items) when age, sex, WRAT, physical and social anhedonia ratings were used as covariates.
A univariate analysis of variance was also used to compare levels of physical and social anhedonia across the two groups. Age was shown to correlate with both physical and social anhedonia levels and thus was included as a covariate in the anhedonia analyses. There was no significant difference between groups in levels of physical anhedonia (F(1,35) = 1.22, ns). However, there was a significant group difference for self-reported social anhedonia, with patients reporting higher levels than healthy participants (F(1,35) = 4.12, p=.05; see Table 2).
Bivariate correlations between performance on the LEAS and anhedonia self-reports are shown in Table 3. Any correlations found to be significant were run also as partial correlations controlling for the potential influence of age, sex, or premorbid IQ; all correlations remained significant when controlling for these factors. Higher physical anhedonia levels were correlated with lower scores on the LEAS Self full scale (r=−.47, p<.05) and LEAS Self “complex” subscale (r = −.62, p=<.01) in healthy participants. Lower scores on the social anhedonia scale were also correlated with higher LEAS Self “complex” scores (r = −.44, p<.05) in healthy participants, although the significance level dropped to p = 0.218 when we controlled for sex, age, and WRAT scores. In the healthy group there was also a correlation between higher LEAS Self full scale scores and lower social anhedonia (r=−.48, p<.05). Although it did not reach a statistical significance, the relationship between higher LEAS Other “complex” scores and lower social anhedonia was in the moderate effect size range for healthy subjects (r=−.39, p<.05).
In contrast, individuals with SSD showed no significant relationship between scores on the LEAS scales and physical and social anhedonia scales, with the largest effect size at 0.16. It is noteworthy that all subscales of the LEAS and also their full scales had a stronger correlation with anhedonia subscales in healthy subjects than in SSD subjects, although it only reached statistical significance in the cases explained above (see Table 3).
In controls a higher frequency of a score of “3” in the LEAS was correlated to a lower level of social anhedonia (r=−.53; p=.01) and a higher frequency of a score of “0” in the LEAS was positively correlated with social anhedonia (r=.47; p<.05). This correlation remained significant in the LEAS Other subscale score distribution (for the score of “0” r=.48, p<.05; for the score of “3” r=−.46, p<.05) and only for the score of “3” in the LEAS Self subscale score distribution (r=.50, p<.05).
Better performance on the LEAS Self complex scale was correlated with higher scores on the PANSS general subscale (r = .47, p<.05), a measurement of general symptomatology in schizophrenia (i.e. anxiety, disorientation, poor attention, poor impulse controls, depression). This correlation was particularly driven by the specific item addressing somatic concerns (r=.44, p<.05); when considering the whole LEAS Self scale, positive correlations were found between emotional awareness for self and the items addressing somatic concerns (r=.46, p<.05) and anxiety (r=.46, p<.05) and a score adding all negative emotions in the PANSS General Scale including somatic concern, anxiety, guilt feelings, tension and depression (r=.49, p<.05). However, there were no other statistically significant relationships between scores on the LEAS self scales and clinical assessments. Interestingly, higher scores on the LEAS Other complex subscale were significantly correlated with higher HCQL intrapsychic scores indicating better quality of life (r = .49, p<.05), whereas the correlation between the LEAS Other simple subscale and the HCQL intrapsychic score did not quite reach significance (r = .44, p=.06). Scores on the HCQL intrapsychic scale were not significantly correlated with other LEAS subscales or full scales, nor with anhedonia measures in this sample. As expected, lower social anhedonia levels were associated with better functioning on the HCQL interpersonal scale (r = −.59, p<.01), but this HCQL scale did not correlate with any other anhedonia or LEAS scales or subscales. The remaining HCQL scales (common and instrumental) were not significantly correlated with any other measures of emotion functioning in this relatively small sample.
This study was conducted to compare schizophrenia-spectrum patients (SSD) to healthy subjects in their level of awareness of their emotional state in response to real-life social situations, and to examine whether reductions in this awareness was related to social functioning and psychopathology. Schizophrenia-spectrum patients demonstrated less complex understandings of others’ emotions than healthy subjects specifically in difficult social scenarios, even when factors potentially influencing descriptions of emotional reactions such as premorbid intelligence and sex were controlled for. In contrast, in analyses of simple scenarios, there was a clear relationship between higher premorbid intelligence and greater understanding of someone else’s emotion in the SSD group. The SSD group also reported that they perceived less pleasure in social situations, as measured by the self-reported social anhedonia subscale, than did healthy subjects. At the same time, no differences between groups were found for measures of awareness of subjects’ own emotional experiences in social situations, nor in measures of subjects’ ability to get pleasure from non-social stimuli.
In the current study, we found that the ability to experience pleasure, as assessed by the anhedonia scales, was independent of SSD subjects’ ability to provide complex descriptions of their own emotional experiences. Even when SSD subjects demonstrated complex understandings of their own emotions, this ability did not support greater experiences of pleasure from either the physical or social environment. The ability to experience pleasure and the ability to provide descriptions of emotional experiences in oneself are significantly related in healthy individuals; it would appear that each of these components of emotion processing are better integrated than in SSD subjects. Although both measures (levels of emotional awareness scale and anhedonia measures) look at emotion processing, they represent different aspects of the phenomenon: the anhedonia measures examined the subjects’ ability to experience one particular emotion (pleasure) in certain situations while the LEAS measured the ability to describe an emotional experience with a cognitive meta-representation in a developmental context (Lane and Garfield, 2005). It is likely that healthy individuals have a more fluid interaction between these two components of emotion processing and that this interaction is impaired in SSD subjects in whom different aspects of emotional processing (e.g., emotional experience, cognitive elaborations of emotional experience) are not well integrated. Finding other ways to measure affective-cognitive integration should be pursued to further establish the nature and implications of this deficit in schizophrenia or SSD. It should also be noted that the anhedonia measures used in this study also depend on the subjects’ report of their ability to experience pleasure from past experiences and memory of past positive experiences and there is evidence that these reports may be significantly dissociated from actual experiences of positive affect in schizophrenia (Herbener et al, 2007). It remains to be determined whether the ability to experience other emotions (such as fear) also has a similar pattern of correlation with the level of emotional awareness in healthy subjects versus individuals with schizophrenia or SSD.
We found that higher levels of intrapsychic functioning (reflected in indices of motivation, anhedonia, curiosity, and empathy) were positively associated with greater ability to provide complex descriptions of the emotional experiences of others (LEAS Other complex). Perhaps the ability to effectively interpret emotional cues in complex social environments increases interest in engagement in the larger social environment, resulting in better intrapsychic functioning. Our findings also supported a previous relationship between increased ability to experience pleasure in social situations and social functioning (Blanchard et al, 1998; Herbener et al, 2005).
There was a significant positive correlation between more complex understanding of one’s own emotional state and higher PANSS General scores but no correlations were found between the ability to describe one’s own emotions and positive symptoms; this differs from Serper and Berenbaum’s report (2008) who found that lower levels of attention to emotion and lower emotion clarity (measured with TAS-20) were correlated with more severe hallucination ratings. With the TAS-20 being a self-report measure, it is possible that this relationship is biased and only captured patients that were able to notice their emotional deficits, while with a measure not based on self-report, like the LEAS, we were able to capture a wider range of patients and this relationship with positive symptoms disappeared when more patients with limited understanding of their emotions were included.
One hypothesis to explain the relationship between general symptoms and emotional awareness is that increased self-awareness could increase awareness of negative emotions generating this association with general symptoms (i.e.: somatic concern, anxiety, tension, depression). Further, there is a significant literature indicating that higher levels of insight in individuals with schizophrenia, particularly in the first years following initial onset of illness, are associated with higher levels of mood symptoms and suicidality (Crumlish et al, 2005; Kim et al, 2003; Siris, 2000; Amador et al, 1996). However, in the current study, no significant relationship was found between levels of emotional awareness and scores on the PANSS Positive and Negative subscales or the Hamilton Depression Rating Scale. In non-clinical samples, increasing awareness of one’s emotions by specifying them can reduce their intensity (Philippot et al, 2006). This certainly indicates that increased insight or emotional awareness might lead to different outcomes in individuals with schizophrenia, and raises the important issue of what combination of psychological, interpersonal, or contextual factors are needed in order to simultaneously support emotional awareness, insight, and long-term positive outcome. Data from research on other chronic illnesses, such as cancer, suggest that it is possible to come to terms with the fact of illness and its effects, and develop hopefulness about the future (Rosenbaum et al, 2004). Although psychiatric illnesses pose a particularly difficult problem because abnormalities in cognitive processes can impact beliefs about illness and the future, we do not know whether interventions and supports used in other populations with chronic medical illness might be similarly beneficial for individuals with schizophrenia. Further, it appears that the schizophrenic individuals at greatest risk for suicide are those who are most insightful (Amador et al, 1996); this particular subgroup may be particularly able to take advantage of cognitive interventions focusing on integration of illness into one’s life story, and ability to use hopefulness to move forward in their lives.
Studies focusing on theory of mind assess similar constructs than the ones analyzed in this study. These data thus point to the importance of clarifying differences in task methodology and difficulty, and how these characteristics may effect conclusions about the specific nature of deficits in emotional processing in individuals with SSD. Although many tasks might overlap on what they measure, each task might point to a different functional deficit and this may relate to specific clinical characteristics or deficits.
The current analyses focused on results from a relatively small sample of schizophrenia spectrum and healthy subjects. At the same time, even in this small sample, the current measures identified significant differences between groups that have important implications in future studies. Our understanding of emotional and social processing in schizophrenia is at a very early stage; past research in this area has focused primarily on deficits in affect recognition and theory of mind. It will be important to assess how performance on the LEAS is associated with these well-characterized deficits. At the same time, research on social and emotional deficits is relatively young, and there may be many additional aspects of social and emotional functioning which have not yet been addressed in empirical studies, which may similarly influence our understanding of social and emotional functioning in schizophrenia. Thus, although we acknowledge that this study on a previously untested social cognitive characteristic in individuals with schizophrenia has limited impact, we adhere to the belief that advances and innovations are the product of the cumulative evidence from such studies. Future studies should also evaluate if the level of emotional awareness can be modified with treatment interventions or if it changes with the course of the illness over time and if and how those changes relate to domains of clinical attention (symptoms, functioning, quality of life).
In the current study, we attempted to integrate essential concepts from cognitive, affective, and social neuroscience to assess behavior in SSD and healthy samples in their interpretation of social scenarios. The LEAS is a measure that helps explain how patients cognitively elaborate their emotional experiences and indirectly measures the level of their understanding of social situations. These results add another dimension to our understanding of emotional deficits in individuals with schizophrenia previously demonstrated in affect recognition and theory of mind tasks, by measuring cognitive-emotional interactions in a social context. The demonstrated clinical relevance of these findings suggests that we need to carefully examine in which individuals emotional awareness can improve quality of life, and in which individuals it may increase the risk of complications.
Grant support: This work was supported by a NIH award (MH 67223) to E. Herbener.