Supplemental Material - Cambridge University Press



Supplemental Material TOC \o "1-3" Supplemental Material 1: Morphed Faces Task PAGEREF _Toc496452517 \h 2Supplemental Material 2: Social Goals Task PAGEREF _Toc496452518 \h 3Practice Vignettes PAGEREF _Toc496452519 \h 3Vignettes PAGEREF _Toc496452520 \h 3Internal Consistency PAGEREF _Toc496452521 \h 4Supplemental Material 3: Potential Confounding Factors PAGEREF _Toc496452522 \h 5Supplemental Material 4: Outliers and Categorical Data PAGEREF _Toc496452523 \h 5Mahalanobis Distance PAGEREF _Toc496452524 \h 5Influence of Categorical Data (i.e. residential/community or diagnostic group) PAGEREF _Toc496452525 \h 6Supplemental Material 5: Exploring different types of trauma PAGEREF _Toc496452526 \h 8Does Amygdala Fear Modulation interact with prior Abuse in predicting CU traits? PAGEREF _Toc496452527 \h 8Does Amygdala Fear Modulation interact with prior Neglect in predicting CU traits? PAGEREF _Toc496452528 \h 9Supplemental Material 6: Exploring Effects of level of PTSD and Anxiety PAGEREF _Toc496452529 \h 11Supplemental Material 7: Visualizing the moderated mediation PAGEREF _Toc496452530 \h 12Supplemental Material 1: Morphed Faces TaskThe experimental design was adapted from Marsh and colleagues ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"2l066pnhgr","properties":{"formattedCitation":"{\\rtf (Marsh {\\i{}et al.}, 2008)}","plainCitation":"(Marsh et al., 2008)"},"citationItems":[{"id":6191,"uris":[""],"uri":[""],"itemData":{"id":6191,"type":"article-journal","title":"Reduced amygdala response to fearful expressions in children and adolescents with callous-unemotional traits and disruptive behavior disorders","container-title":"Am J Psychiatry","page":"712-20","volume":"165","issue":"6","archive_location":"18281412","abstract":"OBJECTIVE: Extensive work implicates abnormal amygdala activation in emotional facial expression processing in adults with callous-unemotional traits. However, no research has examined amygdala response to emotional facial expressions in adolescents with disruptive behavior and callous-unemotional traits. Moreover, despite high comorbidity of callous-unemotional traits and attention deficit hyperactivity disorder (ADHD), no research has attempted to distinguish neural correlates of pediatric callous-unemotional traits and ADHD. METHOD: Participants were 36 children and adolescents (ages 10-17 years); 12 had callous-unemotional traits and either conduct disorder or oppositional defiant disorder, 12 had ADHD, and 12 were healthy comparison subjects. Functional MRI was used to assess amygdala activation patterns during processing of fearful facial expressions. Patterns in the callous-unemotional traits group were compared with those in the ADHD and comparison groups. RESULTS: In youths with callous-unemotional traits, amygdala activation was reduced relative to healthy comparison subjects and youths with ADHD while processing fearful expressions, but not neutral or angry expressions. Functional connectivity analyses demonstrated greater correlations between the amygdala and the ventromedial prefrontal cortex in comparison subjects and youths with ADHD relative to those with callous-unemotional traits. Symptom severity in the callous-unemotional traits groups was negatively correlated with connectivity between amygdala and ventromedial prefrontal cortex. CONCLUSIONS: This is the first study to demonstrate reduced amygdala responsiveness in youths with callous-unemotional traits. These findings support the contention that callous and unemotional personality traits are associated with reduced amygdala response to distress-based social cues.","DOI":"10.1176/appi.ajp.2007.07071145","ISSN":"1535-7228 (Electronic) 0002-953X (Linking)","shortTitle":"Reduced amygdala response to fearful expressions in children and adolescents with callous-unemotional traits and disruptive behavior disorders","journalAbbreviation":"The American journal of psychiatry","language":"eng","author":[{"family":"Marsh","given":"A. A."},{"family":"Finger","given":"E. C."},{"family":"Mitchell","given":"D. G."},{"family":"Reid","given":"M. E."},{"family":"Sims","given":"C."},{"family":"Kosson","given":"D. S."},{"family":"Towbin","given":"K. E."},{"family":"Leibenluft","given":"E."},{"family":"Pine","given":"D. S."},{"family":"Blair","given":"R. J."}],"issued":{"date-parts":[["2008",6]]}}}],"schema":""} (Marsh et al., 2008). Specifically, participants were shown photographs of faces of 10 men and women from the Pictures of Facial Affect series ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"21jopou6vp","properties":{"formattedCitation":"{\\rtf (Marsh {\\i{}et al.}, 2008)}","plainCitation":"(Marsh et al., 2008)"},"citationItems":[{"id":6191,"uris":[""],"uri":[""],"itemData":{"id":6191,"type":"article-journal","title":"Reduced amygdala response to fearful expressions in children and adolescents with callous-unemotional traits and disruptive behavior disorders","container-title":"Am J Psychiatry","page":"712-20","volume":"165","issue":"6","archive_location":"18281412","abstract":"OBJECTIVE: Extensive work implicates abnormal amygdala activation in emotional facial expression processing in adults with callous-unemotional traits. However, no research has examined amygdala response to emotional facial expressions in adolescents with disruptive behavior and callous-unemotional traits. Moreover, despite high comorbidity of callous-unemotional traits and attention deficit hyperactivity disorder (ADHD), no research has attempted to distinguish neural correlates of pediatric callous-unemotional traits and ADHD. METHOD: Participants were 36 children and adolescents (ages 10-17 years); 12 had callous-unemotional traits and either conduct disorder or oppositional defiant disorder, 12 had ADHD, and 12 were healthy comparison subjects. Functional MRI was used to assess amygdala activation patterns during processing of fearful facial expressions. Patterns in the callous-unemotional traits group were compared with those in the ADHD and comparison groups. RESULTS: In youths with callous-unemotional traits, amygdala activation was reduced relative to healthy comparison subjects and youths with ADHD while processing fearful expressions, but not neutral or angry expressions. Functional connectivity analyses demonstrated greater correlations between the amygdala and the ventromedial prefrontal cortex in comparison subjects and youths with ADHD relative to those with callous-unemotional traits. Symptom severity in the callous-unemotional traits groups was negatively correlated with connectivity between amygdala and ventromedial prefrontal cortex. CONCLUSIONS: This is the first study to demonstrate reduced amygdala responsiveness in youths with callous-unemotional traits. These findings support the contention that callous and unemotional personality traits are associated with reduced amygdala response to distress-based social cues.","DOI":"10.1176/appi.ajp.2007.07071145","ISSN":"1535-7228 (Electronic) 0002-953X (Linking)","shortTitle":"Reduced amygdala response to fearful expressions in children and adolescents with callous-unemotional traits and disruptive behavior disorders","journalAbbreviation":"The American journal of psychiatry","language":"eng","author":[{"family":"Marsh","given":"A. A."},{"family":"Finger","given":"E. C."},{"family":"Mitchell","given":"D. G."},{"family":"Reid","given":"M. E."},{"family":"Sims","given":"C."},{"family":"Kosson","given":"D. S."},{"family":"Towbin","given":"K. E."},{"family":"Leibenluft","given":"E."},{"family":"Pine","given":"D. S."},{"family":"Blair","given":"R. J."}],"issued":{"date-parts":[["2008",6]]}}}],"schema":""} (Marsh et al., 2008) displaying either neutral expressions or fearful expressions of differing intensities (50, 100 or 150%). Neutral expressions (0% fear) were morphed into fearful expressions to create composites (50% or 100% intensity) or by extrapolating to create exaggerated expressions (150% intensity), see Figure 2. The neutral expressions were created by morphing neutral into happy expressions and taking the 25% composite, following previous indications that neutral expressions may appear threatening ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"10j18feq0v","properties":{"formattedCitation":"{\\rtf (Phillips {\\i{}et al.}, 2004)}","plainCitation":"(Phillips et al., 2004)"},"citationItems":[{"id":290,"uris":[""],"uri":[""],"itemData":{"id":290,"type":"article-journal","title":"Differential neural responses to overt and covert presentations of facial expressions of fear and disgust","container-title":"NeuroImage","page":"1484-1496","volume":"21","issue":"4","source":"PubMed","abstract":"There is debate in cognitive neuroscience whether conscious versus unconscious processing represents a categorical or a quantitative distinction. The purpose of the study was to explore this matter using functional magnetic resonance imaging (fMRI). We first established objective thresholds of the critical temporal parameters for overt and covert presentations of fear and disgust. Next we applied these stimulus parameters in an fMRI experiment to determine whether non-consciously perceived (covert) facial expressions of fear and disgust show the same double dissociation (amygdala response to fear, insula to disgust) observed with consciously perceived (overt) stimuli. A backward masking paradigm was used. In the psychophysics experiment, the following parameters were established: 30-ms target duration for the covert condition, and 170-ms target duration for the overt condition. Results of the block-design fMRI study indicated substantial differences underlying the perception of fearful and disgusted facial expressions, with significant effects of both emotion and target duration. Findings for the overt condition (170 ms) confirm previous evidence of amygdala activation to fearful faces, and insula activation to disgusted faces, and a double dissociation between these two emotions. In the covert condition (30 ms), the amygdala was not activated to fear, nor was the insula activated to disgust. Overall, findings demonstrate significant differences between the neural responses to fear and to disgust, and between the covert presentations of these two emotions. These results therefore suggest distinct neural correlates of conscious and unconscious emotion perception.","DOI":"10.1016/j.neuroimage.2003.12.013","ISSN":"1053-8119","note":"PMID: 15050573","journalAbbreviation":"Neuroimage","language":"eng","author":[{"family":"Phillips","given":"Mary L."},{"family":"Williams","given":"Leanne M."},{"family":"Heining","given":"Maike"},{"family":"Herba","given":"Catherine M."},{"family":"Russell","given":"Tamara"},{"family":"Andrew","given":"Christopher"},{"family":"Bullmore","given":"Ed T."},{"family":"Brammer","given":"Michael J."},{"family":"Williams","given":"Steven C. R."},{"family":"Morgan","given":"Michael"},{"family":"Young","given":"Andrew W."},{"family":"Gray","given":"Jeffrey A."}],"issued":{"date-parts":[["2004",4]]}}}],"schema":""} (Phillips et al., 2004). Each trial started with the presentation of a face (2000ms), after which a fixation cross was displayed (900ms). Participants were asked to identify the gender by button press. Each run consisted of 40 face trials randomly interspersed with 40 fixation trials (1000ms) in order to jitter stimulus presentation. The total task involved two runs, totaling ~6 minutes.Examples of morphed expressions. Shown are four examples of morphed expressions. Examples are taken from different actors with fear intensity increasing from left to right. Marsh, A. A., E. C. Finger, D. G. Mitchell, M. E. Reid, C. Sims, D. S. Kosson, K. E. Towbin, E. Leibenluft, D. S. Pine, and R. J. Blair. 2008. “Reduced Amygdala Response to Fearful Expressions in Children and Adolescents with Callous-Unemotional Traits and Disruptive Behavior Disorders.” Am J Psychiatry 165 (6):712–20. , Mary L., Leanne M. Williams, Maike Heining, Catherine M. Herba, Tamara Russell, Christopher Andrew, Ed T. Bullmore, et al. 2004. “Differential Neural Responses to Overt and Covert Presentations of Facial Expressions of Fear and Disgust.” NeuroImage 21 (4):1484–96. Material 2: Social Goals TaskPractice VignettesYou are in the library checking out books for a project. A kid grabs the book you need right before you are about to take it off the shelf.You are outside for recess.? A kid runs to the swing that you were about to use, and starts swinging on it instead. VignettesA new kid at your school is coming down the hall from the other direction, and suddenly bumps into your shoulder hard, knocking your books to the floor.A kid laughs when you walk down the hallway and accidentally trip. ?You are throwing balls in the schoolyard with your classmates. One kid never throws the ball to you. During lunch, a kid pushes in front of you in the line waiting for food. You are talking to your classmates about a TV show that you really like and one of them says that it’s stupid.Internal ConsistencyCronbach’s αAvoid Conflict.739Reconciliation.846Dominance .804Revenge.740Forced Respect.792Supplemental Material 3: Potential Confounding Factors Response times correlated with age and gender and ICU scores correlated with IQ. In addition, a considerable proportion of the youth in this study was on psychotropic medication, as medications were not withheld at the time of the scan. Moreover, anonymous reviewers raised concerns that motion and/or DBD symptoms could have been driving our interaction finding. Therefore, we repeated our key analysis within our bilateral amygdala ROIs and included age, IQ, gender, psychotropic medication, average T-score on the CBCL DMS-IV oriented ODD and CD subscales and average motion per TR as additional nuisance regressors. This revealed that the fear intensity modulated BOLD response in the right amygdala maintained an interaction between ICU and CTQ (32 voxels, Z=2.94, Cohen’s d=0.35, xyz=19.2;-4.5;-25.0).Supplemental Material 4: Outliers and Categorical DataMahalanobis DistanceExtreme cases might have invalidated the case for a dimensional analytical approach to the data, thereby leading to a spurious interaction effect between ICU and CTQ. To address this issue, we performed a Mahalanobis Distance (MD) analysis in order to calculate how many standard deviations each youth is away from the mean of our sample distribution, given our variable space. We calculated the MD for each participant on four variables: the ICU, the CTQ, the CBCL Anxiety T-score and the CBCL PTSD T-score. We then calculated a p-value per participant, under the Χ2-distribution, to determine whether their MD value should be considered an outlier. 4 participants were identified at p<=.05. We reran our main analysis excluding these participants. This showed that the fear intensity modulated BOLD response in the right amygdala still showed an interaction between ICU and CTQ (28 voxels, Z=3.41, Cohen’s d=0.41, xyz=19.2;-2.5;18.0). Thus, it is unlikely that extreme cases, at least with respect to their ICU, CTQ, Anxiety and PTSD scores, could have determined our interaction finding.Influence of Categorical Data (i.e. residential/community or diagnostic group)Categorically different subgroups of participants might have invalidated the case for a dimensional analytical approach. Below, we display a scatterplot between the ICU and CTQ, with data points marked according to diagnostic group. We fitted a linear trend line separately for each diagnostic group. The equation for each trend line is shown, as well as the correlation coefficient. Fisher’s r-to-z transformations showed that none or the correlations differed significantly from each other (all p>.484). This suggests that the relationship between the ICU and the CTQ does not depend on diagnostic group. Next, we display scatterplots between ICU and CTQ separately for each placement group (residential vs. community) and fitted a linear trend line separately for each group. A Fisher’s r-to-z transformation showed that these correlations did not significantly differ from each other (p=.653). This suggests that the relationship between the ICU and the CTQ does not depend on participant source. Supplemental Material 5: Exploring different types of trauma All analyses were conducted using the total scores for the five trauma subscales of the CTQ. However, it is possible that type of trauma matters in the interaction with the ICU. To address this possibility, we repeated our key analysis twice, once using Neglect scores (Emotional Neglect + Physical Neglect) and once using Abuse scores (Emotional Abuse + Physical Abuse + Sexual Abuse) of the CTQ. This revealed comparable interactions between ICU and both Neglect and Abuse in right amygdala on the fear intensity modulated BOLD response as those between ICU and CTQ total score (though only the cluster showing an interaction with Abuse scores survived an FWE correction at αuse s). For details see Supplemental Material 3a (Abuse) and 3b (Neglect).Does Amygdala Fear Modulation interact with prior Abuse in predicting CU traits?Panel A) The fear intensity modulated BOLD response in right amygdala (24 voxels, Z=3.00, Cohen’s d=0.35, xyz=19.2;-4.5;-25.0) showed an interaction between ICU and Abuse scores (Abuse scores were obtained by summing the Emotional Abuse, Physical Abuse and Sexual Abuse subscales of the CTQ) at p<.05 uncorrected, also surviving small volume correction (FWE α=0.05). Panel B) A moderation model was tested using the PROCESS macro for SPSS ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"iTkcNWaT","properties":{"formattedCitation":"(Hayes, 2013)","plainCitation":"(Hayes, 2013)"},"citationItems":[{"id":10858,"uris":[""],"uri":[""],"itemData":{"id":10858,"type":"book","title":"Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach","publisher":"The Guilford Press","publisher-place":"New York","event-place":"New York","abstract":"Explaining the fundamentals of mediation and moderation analysis, this engaging book also shows how to integrate the two using an innovative strategy known as conditional process analysis. Procedures are described for testing hypotheses about the...","URL":"","shortTitle":"Introduction to Mediation, Moderation, and Conditional Process Analysis","author":[{"family":"Hayes","given":"A. F."}],"issued":{"date-parts":[["2013"]]},"accessed":{"date-parts":[["2017",2,15]]}}}],"schema":""} (Hayes, 2013), similar to the model represented in Figure 3 in the main paper. Shown is the linear regression equation for predicted ICU scores, where FM stands for Fear Modulated Amygdala Response. The table contains Beta estimates (β), standard errors (SE), t-statistic (T), p-value (p) and Lower (LLCI) and Upper Limit Confidence Intervals (ULCI) for the coefficients of the linear regression model from panel B; i.e. a1=constant, b1=main effect of amygdala fear modulation on ICU score, b2=main effect of Abuse score on ICU score, b3=moderation of amygdala fear modulation on ICU score by Abuse score. Panel C) Trend lines and associated confidence intervals for conditional effect of fear intensity modulated BOLD response on ICU at values (-1 SD; mean; +1 SD) of Abuse. The Johnson-Neyman Technique shows that higher fear intensity modulated amygdala responses predict lower ICU scores for relatively low levels of Abuse (Abuse<=18.8). Fear intensity modulated amygdala responses were not significantly correlated with ICU scores at medium to high levels of prior trauma (Abuse>18.8).Does Amygdala Fear Modulation interact with prior Neglect in predicting CU traits?Panel A) The fear intensity modulated BOLD response in right amygdala (15 voxels, Z=2.46, Cohen’s d=0.29, xyz=19.2;2.5-18.0) showed an interaction between ICU and CTQ Neglect scores (Neglect scores were obtained by summing the Emotional Neglect and Physical Neglect of the CTQ) at p<.05 uncorrected, also surviving small volume correction (FWE α=0.05). Panel B) A moderation model was tested using the PROCESS macro for SPSS ADDIN ZOTERO_ITEM CSL_CITATION {"citationID":"iTkcNWaT","properties":{"formattedCitation":"(Hayes, 2013)","plainCitation":"(Hayes, 2013)"},"citationItems":[{"id":10858,"uris":[""],"uri":[""],"itemData":{"id":10858,"type":"book","title":"Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach","publisher":"The Guilford Press","publisher-place":"New York","event-place":"New York","abstract":"Explaining the fundamentals of mediation and moderation analysis, this engaging book also shows how to integrate the two using an innovative strategy known as conditional process analysis. Procedures are described for testing hypotheses about the...","URL":"","shortTitle":"Introduction to Mediation, Moderation, and Conditional Process Analysis","author":[{"family":"Hayes","given":"A. F."}],"issued":{"date-parts":[["2013"]]},"accessed":{"date-parts":[["2017",2,15]]}}}],"schema":""} (Hayes, 2013), similar to the model represented in Figure 3 in the main paper. Shown is the linear regression equation for predicted ICU scores, where FM stands for Fear Modulated Amygdala Response. The table contains Beta estimates (β), standard errors (SE), t-statistic (T), p-value (p) and Lower (LLCI) and Upper Limit Confidence Intervals (ULCI) for the coefficients of the linear regression model from panel B; i.e. a1=constant, b1=main effect of amygdala fear modulation on ICU score, b2=main effect of Neglect score on ICU score, b3=moderation of amygdala fear modulation on ICU score by Neglect score. Panel C) Trend lines and associated confidence intervals for conditional effect of fear intensity modulated BOLD response on ICU at values (-1 SD; mean; +1 SD) of Neglect. The Johnson-Neyman Technique shows that higher fear intensity modulated amygdala responses predict lower ICU scores for relatively low levels of Neglect (Neglect<=14.3) and higher fear intensity modulated amygdala responses predict higher ICU scores for relatively high levels of Neglect (Neglect>=34.3). Fear intensity modulated amygdala responses were not significantly correlated with ICU scores at medium levels of prior trauma (14.3<CTQ<34.3).Supplemental Material 6: Exploring Effects of level of PTSD and AnxietyThe goal of the current study was to determine whether level of prior trauma exposure moderated the relationship between CU traits and amygdala responsiveness to fearful expressions. However, prior trauma is related to the development of PTSD specifically and anxiety more generally. As such, one might expect that level of PTSD and anxiety symptoms would similarly moderate the relationship of CU traits with amygdala responsiveness as prior trauma exposure. To test this, we ran two linear regression models (AFNI’s 3dttest++) on the fear intensity beta coefficients within our amygdala ROIs: (i) using ICU, CBCL Anxiety T-scores and the interaction between ICU and CBCL Anxiety T-scores and (ii) using the ICU, CBCL PTSD T-scores and the interaction between ICU and CBCL PTSD T-scores as predictor variables. Both level of anxiety (8 voxels, Z=2.51, Cohen’s d=0.30, xyz=19.2;-8.0;-18.0) and particularly level of PTSD (24 voxels, Z=3.02, Cohen’s d=0.36, xyz=22.8;-8;-18.0) interacted with the ICU on the fear intensity modulated amygdala response at p<.05 uncorrected. Indeed, only the symptomatology particularly related to prior trauma exposure, PTSD symptomatology, survived an FWE correction at α=0.05. Supplemental Material 7: Visualizing the moderated mediationWe examined whether the indirect effect of fear intensity modulation in the amygdala on SGI through CU traits was moderated by prior trauma (Figure 5, panel A). To this end, we expressed SGI as a function of fear intensity modulated amygdala responses and prior trauma (see Methods section and Figure 5, panel B). Model estimates are summarized in Figure 5 (panel C). The bootstrapped model revealed a significant moderated mediation: the indirect effect of fear intensity modulation in the amygdala on SGI through CU traits is moderated by prior trauma (Index of moderated mediation = -.0455; SE = .0254). The conditional indirect effect of fear intensity modulated responses on SGI at values of CTQ are listed in panel D. This shows that there is only a significant negative association between SGI and fear intensity modulated amygdala responses at low levels of prior trauma (CTQ=25, but not at medium or high levels of trauma). This effect is visualized in the graph below. To visualize this indirect effect, we used the estimates from Figure 5, panel D and the linear regression model for SGI to plot trend lines for the association between SGI and fear intensity modulated amygdala responses at low, medium and high levels of prior trauma. Trend lines for the association between Social Goal Importance and fear intensity modulated amygdala responses at low, medium and high levels of prior trauma. * denotes significant association. Derived from beta estimates provided by PROCESS at the following values for CTQ: Low [-1 SD CTQ score], Medium [mean CTQ score] and High [+1 SD CTQ score] ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download