Demographic and Clinical Characteristics
As shown in , each patient group was demographically comparable to their healthy comparison sample in terms of age, gender, ethnicity, and parental education level, except that prodromal patients were significantly younger than their comparison sample by about one year on average, F(1,24)=6.75, p=.02. Attempts were not made to match the groups on education level given the potential influence of illness on educational attainment; prodromal and first-episode schizophrenia patients reported fewer years of education relative to their comparison groups, F(1,24)=2.47, p=.03 and F(1,67)=3.29, p=.01, respectively.
Demographic and Clinical Characteristics of Participants
As expected, first-episode patients had a significantly shorter duration of illness than chronic schizophrenia patients, F(1,70)=62.3, p=.001. Dosage of antipsychotics also was significantly higher in chronic patients as compared to prodromal and first-episode patients, F(2,75)=10.02, p=.001. The three patient groups, however, did not differ significantly in global symptom levels as assessed by the SAPS, F(2,85)=.65, p=.53, or the SANS, F(2,84)=1.61, p=.21. Negative symptoms were higher than positive symptoms across prodromal individuals, F(1,24)=4.70, p=.04, first-episode patients, F(1,75)=39.22, p=.001, and chronic patients, F(1,70)=34.70, p=.001.
Affective Modulation as a Function of Valence
To determine if prodromal, first-episode, and chronic schizophrenia patients exhibited an overall pattern of affective modulation consistent with previous studies (Curtis et al., 1999
; Schlenker et al., 1995
; Volz et al., 2003
), responses to neutral images were compared to those elicited by unpleasant and pleasant pictures. To equate the number of content categories represented within each valence, a subset of unpleasant categories (i.e., contamination, mutilation, and animal attack) was selected for comparison with the pleasant categories (i.e., couples/families, food/nature, and mild erotica) to yield comparable levels of overall reported arousal (unpleasant: M
=5.02; pleasant: M
=5.08), as determined by standardized ratings of the IAPS (Lang, Bradley, & Cuthbert, 2005
Regarding subjective evaluations of the images, significant main effects for valence, F(2,272)=231.50, p=.001, and arousal, F(2,272)=42.09, p=.001, indicated that unpleasant images were rated as low in experienced pleasure and high in experienced arousal while pleasant pictures were rated as high in experienced pleasure and arousal relative to neutral pictures (see ). Although none of the clinical samples differed from the healthy comparison group, patients experiencing prodromal symptoms did rate pleasant pictures as less pleasurable than chronic schizophrenia patients and they found neutral images to be less pleasing than both groups of schizophrenia patients, as supported by a significant group-by-valence interaction effect, F(6,272)=3.56, p=.01. No significant main or interaction effects involving group were observed for ratings of arousal.
Evaluative and Physiological Responses to Pleasant, Neutral, and Unpleasant Pictures, Averaged Across Specific Picture Contents
As shown in , the prototypic valence effect on the eyeblink reflex was observed in the healthy comparison sample and in each of the patient groups, F(2,320)=13.71, p=.001. Unpleasant images were associated with significantly larger eyeblinks than pleasant images while eyeblinks to neutral images were intermediate. Neither the main effect for group, F(3,160)=1.40, p=.25, nor the group-by-valence interaction effect, F(6,320)=1.15, p=.33, was significant.
Heart rate response also was consistent with expected patterns: a sustained deceleration to unpleasant pictures and a triphasic response (deceleration-acceleration-deceleration) to pleasant images (see ). Thus, across the four groups, the best fitting function was quadratic when viewing unpleasant (r
=.98) scenes and cubic when viewing pleasant (r
=.91) or neutral (r
=.92) images (Bradley et al., 2001
). As shown in , the groups did not differ in the magnitude of their initial deceleration, F
=.18, and no group-by-valence interaction effect was observed, F
=.74. Although pleasant and unpleasant pictures tended to prompt greater initial deceleration relative to neutral images, the effect did not achieve statistical significance, F
=.18. Later during picture viewing, a group effect did emerge with prodromal patients showing greater peak acceleration relative to chronic schizophrenia patients who persisted in a pattern of deceleration across valence categories during the last 3 s of picture viewing, F
=.02. A main effect for valence, F
=.01, replicated prior research with unpleasant scenes resulting in the least peak acceleration followed by pleasant and then neutral images.
Figure 1 Averaged waveforms for heart rate change when viewing pleasant, neutral, and unpleasant images, with a sustained deceleratory pattern to unpleasant pictures and a triphasic (deceleratory, acceleratory, deceleratory) pattern to pleasant pictures, across (more ...)
For skin conductance responses (SCRs; see ), the main effect of valence was not statistically significant, F(2,266)= 1.04, p=.36, probably owing to reliance on lower but more comparable levels of arousal for analyses of hedonic valence. The typical pattern would be increases in sympathetic activation with more arousing stimuli. Unexpectedly, SCRs were significantly higher in patients experiencing prodromal symptoms than in the other three groups, F(1,133)=2.92, p=.04, irrespective of valence.
Consideration of the average effect sizes between each patient group and the healthy comparison sample reveals a similar pattern of findings. With the exception of higher SCRs in prodromal patients, the effect sizes detected were generally quite modest with an occasional moderate effect but no consistent pattern of discrimination between patients and healthy individuals.
Evaluation of the Defensive Motivational System
Although all unpleasant images were rated as low in experienced pleasure (see ), a significant main effect for picture content category, F(6,816)=31.40, p=.001, was modified by a significant group-by-content category interaction, F(18,816)=1.92, p=.02. Prodromal patients rated scenes of human attack as significantly less unpleasant than schizophrenia patients while pictures of accidents also were rated as less unpleasant when compared to judgments made by chronic patients. No significant differences were detected between healthy participants and the two schizophrenia patient groups.
As shown in , arousal ratings differed between specific categories, F(6,816)=36.17, p=.001, but followed the expected ordering across unpleasant contents as supported by a significant linear trend, F(1,136)=138.39, p=.001. Using the Bonferroni adjusted level of significance, scenes of human attack, animal attack, mutilations, and accidents received the highest ratings and differed significantly from pictures depicting contamination, loss/pollution, and illness, which did not differ statistically from each other. Human attack was rated significantly higher in subjective arousal than mutilations or accidents, while animal attack was viewed as more arousing than accidents. Analyses investigating group differences did not reveal a significant main effect, F(3,136)=1.91, p=.13, or interaction, F(18, 816)=1.09, p=.36.
Further supporting the possibility that the defensive motivational system is not compromised by schizophrenia or its prodromal state, a significant main effect of unpleasant picture contents on startle reflex magnitude was obtained in healthy comparison subjects and across all three patient groups, F(6,960)=6.43, p=.001. The left side of illustrates the predicted linear association between increasing levels of self-reported arousal and potentiation of the startle eyeblink reflex in each group, as indicated by a significant effect for a linear trend, F(1,160)=17.45, p=.001. There were no significant effects involving group, F(3,160)=.31, p=.82, or the group-by-content category interaction, F(18,960)=1.04, p=.41. There also were no significant associations between total or subscale scores of the SAPS and SANS and magnitude of the startle reflex when picture contents were ordered by mean arousal ratings.
Figure 2 Mean startle blink magnitude and skin conductance changes when viewing specific pleasant, neutral, and unpleasant picture contents, across healthy comparison subjects, prodromal patients, first-episode schizophrenia patients, and chronic schizophrenia (more ...)
Analyses of heart rate response revealed significant effects for group, F(3,146)=6.83, p=.001, and group-by-content category, F(18,876)=2.05, p=.01, during initial deceleration, and for content category, F(6,876)=3.78, p=.01, and group-by-content category, F(18,876)=2.55, p=.01, during peak acceleration (see ). Although rated as less unpleasant by prodromal patients, pictures of human attack, animal attack, and accidents resulted in significantly greater initial deceleration in the high-risk group than in healthy individuals or schizophrenia patients. Among prodromal patients, a larger initial deceleration also was associated with scenes of loss/pollution when compared with chronic patients. Images of human attack, animal attack, and loss/pollution continued to differentiate prodromal patients, who exhibited less acceleration towards the end of picture viewing than the other three groups. In contrast, peak acceleration to pictures of contamination was greater in prodromal individuals than in chronic patients.
The effect of unpleasant contents on SCRs differed by group, as indicated by a significant interaction effect, F
=.01, and significant group-by-content category trend effects involving linear, F
=.02, and quadratic, F
=.03, relationships (see right side of ). Following Bradley et al. (2001)
, a main effect of picture content was obtained in healthy comparison subjects, F
=.01, reflecting a predominately linear association, F
=.001, and a small quadratic contribution, F
=.04. Comparisons between each patient group and the healthy sample revealed that first-episode and chronic schizophrenia patients were no different from healthy individuals, showing larger SCRs as arousal levels increased with unpleasant contents. Prodromal patients failed to exhibit this relationship, however, and provided a relatively undifferentiated pattern of SCRs across unpleasant picture contents that tended to be larger overall than in the other three groups, F
Evaluation of the Appetitive Motivational System
As shown in , all three pleasant content categories were rated as similarly high in experienced pleasure. A significant group effect, F(3,136)=2.80, p=.05, indicated that prodromal patients rated the images as less pleasant than healthy subjects and chronic schizophrenia patients. The expected linear effect of content category on subjective arousal ratings was obtained, F(1,136)=30.66, p=.001, such that pictures of mild erotica were rated as more arousing than images of food/nature and couples/families, F(2,272)=11.97, p=.001.
As further evidence that underlying motivational states showed no indication of abnormality in schizophrenia, startle eyeblink magnitude varied with pleasant picture contents, F(2,320)=6.91, p=.01, again reflecting an overall linear relationship, F(1,160)=11.39, p=.01. Mild erotica generally elicited greater inhibition or smaller eyeblinks than the other two pleasant categories (see left side of ), in the absence of a significant group effect, F(1,160)=.25, p=.86, or group-by-content category interaction, F(6,320)=1.88, p=.09. Neither total nor subscale scores of the SAPS and SANS corresponded with the effects of picture contents on the startle reflex.
When viewing different pleasant picture contents, the groups did not differ during initial heart rate deceleration, F(3,144)=.20, p=.90, or later peak acceleration, F(3,144)=2.07, p=.11, (see ). The magnitude of these deceleratory and acceleratory responses were not significantly different between content categories, F(2,288)=1.79, p=.17 and F(2,288)=1.81, p=.17, respectively, suggesting comparable levels of cardiovascular activation across the pleasant images.
Consistent with arousal ratings and the eyeblink reflex, SCRs varied with pleasant picture contents, F(2,246)=4.35, p=.02, and as was best accounted for by a linear trend (see right side of ), F(1,123)=6.16, p=.02. Pictures of mild erotica prompted significantly larger SCRs than scenes of couples/families while food/nature images were associated with moderate responses. Prodromal patients exhibited overall larger SCRs to pleasant images as compared to the other groups, F(3,123)=5.06, p=.01, but the group-by-content category interaction was not significant, F(6,246)=1.20, p=.31.