3.1 Differences between Diagnostic Groups on Rankings of Response to Emotional Stimuli during initial exposure
Self-reported responses to emotional stimuli were highly similar in schizophrenia and healthy participants across the full spectrum of valence and arousal levels. As shown in , rankings of both valence and arousal in response to the 131 images by the two groups were highly correlated (r = 0.98 for valence ratings; r = 0.95 for arousal ratings).
3.2 Differences between Groups in Ratings of Stimuli
Repeated measures analyses were used to test for group differences (between-group factor) in ratings of the five valence levels and for the five arousal levels (within-group factors). There was a main effect of valence level on subjects’ ratings of their emotional response to the stimuli (F(4,268) = 332.15, p < .001), but there was no significant group difference (F(1,67) = 0.09, p > .20) or group by image valence interaction (F(4,268) = 1.17, p > .20) (). In the analyses of arousal ratings, again, there was a large effect of normative image arousal level classification (F(4,244) = 148.09, p < .001) (), but, as with valence effects, there was no significant group difference (F(1,61) = 0.43, p > .20), or group by image arousal level interaction (F(4,244) = 0.23, p > .20). Overall the data indicate that both healthy and schizophrenia patients report very similar emotional valence and elicited arousal when viewing IAPS images.
Average ratings of valence and arousal categories of IAPS images by schizophrenia and healthy participants
3.3 Differences between Diagnostic Groups on Other Measures of Emotional Functioning
Multivariate analysis of variance (MANOVA) was used to assess group differences on the three measures of facial emotion perception (Emotion Acuity Test; Emotion Differentiation Test – Sad condition; Emotion Differentiation Test – Happy condition), and on the two measures of anhedonia. In these analyses, premorbid IQ was included as a covariate, as premorbid IQ measures showed some statistically significant relationships with anhedonia self-reports and performance on the emotion perception measures. There was no significant difference between groups in performance on the Penn facial emotion perception tests (F(3,62) = 2.01, p > .10.). However, there was a significant group difference for self-reported anhedonia (F(2,65) = 3.79, p < .05), with schizophrenia subjects reporting higher levels of both physical and social anhedonia than healthy subjects (Physical Anhedonia schizophrenia Mean (SD) = 16.38 (8.11); Healthy 11.91 (6.44); Social Anhedonia Schizophrenia mean 12.65 (7.04); Healthy = 8.91 (5.06)).
3.4 Relationship between anhedonia ratings and emotional response reports
There were significant correlations between anhedonia levels and valence ratings of the stimuli in both healthy and schizophrenia groups, with these relationships strongest at the more extreme positive and negative valence levels (correlation between physical anhedonia and rating of the most positive stimuli = −0.35, p < .05 for healthy subjects, r = − 0.49, p < .01 for schizophrenia subjects; correlation between physical anhedonia and rating of most negative stimuli = 0.39, p< .03 for healthy subjects, r = 0.41, p < .02 for schizophrenia subjects) (Figure 3). Thus, although anhedonia ratings were significantly higher in the schizophrenia patients, the healthy and schizophrenia groups demonstrated similar relationships between anhedonia ratings and valence ratings of immediate affective response to IAPS images.
3.5 Relationship between Emotional Experience and Measures of Other Aspects of Emotional Functioning
Partial correlations between performance on emotional response, anhedonia self-report, emotion perception, negative symptoms, and quality of life measures, controlling for estimated premorbid intellectual functioning are shown in . Premorbid intellectual functioning level was included as a covariate because it was found to correlate significantly with several of the measures in both groups. Given the number of correlations computed, only those which were significant at the p < .01 level are discussed. An initial review of indicates that there are many more nonsignificant than significant relationships in this table – indicating that there are multiple, independent dysfunctions in emotional processing in schizophrenia, rather than a single “emotional processing” deficit. The majority of correlations significant at the p < .01 are from approaches using a common methodology, and thus these significant correlations may be artifactually elevated by shared methods variance. Partial correlations significant at the p < .01 level using different measures were 1) higher physical anhedonia scores and more negative ratings of positively valenced IAPS stimuli; 2) higher social anhedonia and higher levels of negative symptoms and poorer performance on the Heinrichs-Carpenter Interpersonal Scale; 3) higher levels of negative symptoms and poorer performance on the Heinrichs-Carpenter Interpersonal and Heinrichs-Carpenter Intrapsychic Scales.
Table 2 Partial Correlations between Performance on Tasks Assessing Emotional Experience, Emotion Perception, Self-Reported Anhedonia, Negative Symptoms, and Quality Of Life Scales in Participants with Schizophrenia controlling for Premorbid Intellectual Functioning (more ...)