An experimental investigation of emotional reactivity and delayed emotional recovery in borderline personality disorder: the role of shame

Available online at www.sciencedirect.com

Comprehensive Psychiatry 51 (2010) 275 – 285 www.elsevier.com/locate/comppsych

An experimental investigation of emotional reactivity and delayed emotional recovery in borderline personality disorder: the role of shame Kim L. Gratza,⁎, M. Zachary Rosenthalb , Matthew T. Tulla , C.W. Lejuezc , John G. Gundersond a

Department of Psychiatry and Human Behavior, University of Mississippi Medical Center, Jackson, MS 39216, USA b Duke University Medical Center, Durham, NC 27710, USA c Center for Addictions, Personality, and Emotion Research and the Department of Psychology, University of Maryland, College Park, MD 20742, USA d Department of Psychiatry, McLean Hospital and Harvard Medical School, Belmont, MA 02478, USA

Abstract Despite the emphasis on emotional reactivity and delayed emotional recovery in prominent theoretical accounts of borderline personality disorder (BPD), research in this area remains limited. This study sought to extend extant research by examining emotional reactivity (and recovery following emotional arousal) to 2 laboratory stressors (one general, and the other involving negative evaluation) and exploring the impact of these stressors on subjective responding across the specific emotions of anxiety, irritability, hostility, and shame. We hypothesized that outpatients with BPD (compared to outpatients without a personality disorder; non-PD) would demonstrate heightened subjective emotional reactivity to both stressors, as well as a delayed return to baseline levels of emotional arousal. Results provide evidence for context- and emotion-specific reactivity in BPD. Specifically, BPD participants (compared to non-PD participants) evidenced heightened reactivity to the negative evaluation but not the general stressor. Furthermore, results provide support for shame-specific reactivity in BPD, with BPD participants (vs non-PD participants) evidencing a significantly different pattern of change in shame (but not in reported anxiety, irritability, or hostility) across the course of the study. Specifically, not only did BPD participants report higher levels of shame in response to the negative evaluation, their levels of shame remained elevated following this stressor (through the post-recovery period at the end of the study). Findings suggest the importance of continuing to examine emotional reactivity in BPD within specific contexts and across distinct emotions, rather than at the general trait level. © 2010 Elsevier Inc. All rights reserved.

1. Introduction Problems with the experience and expression of emotions are considered to be central to borderline personality disorder (BPD) [1,2], with the higher-order trait of emotional dysfunction identified by many BPD researchers as a “core” personality trait underlying the disorder. Although several different lower-order emotion-related traits have been linked at a theoretical and empirical level to BPD, including affective instability [2], anxiousness [3], affect intensity [1], and anxiety sensitivity [4], 2 lower-order traits considered to be particularly relevant to BPD are emotional reactivity (ie, the tendency to have strong emotional

⁎ Corresponding author. Tel.: +1 601 815 6450. E-mail address: [email protected] (K.L. Gratz). 0010-440X/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.comppsych.2009.08.005

reactions to stimuli) and delayed recovery (ie, slow return to baseline levels of emotional arousal following the activation of an emotion) [1]. In fact, even the characteristic of affective instability (see references [2,5,6]) was defined in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) to emphasize emotional reactivity (see reference [7]). Yet, despite the emphasis on emotional reactivity and delayed recovery in prominent theoretical accounts of BPD [1], research in this area remains relatively limited and no studies have examined emotional reactivity and delayed recovery in regard to specific emotional states and contexts. Thus, the goal of the current study was to extend extant research by examining emotional reactivity (and recovery following emotional arousal) to 2 laboratory stressors and exploring the impact of different stressors on specific subjective emotional states.

276

K.L. Gratz et al. / Comprehensive Psychiatry 51 (2010) 275–285

1.1. Emotional reactivity and delayed recovery in BPD Emotional reactivity broadly refers to responses that occur within 1 or more systems of emotional responding (ie, subjective, physiological, motoric/expressive) following changes in the external or internal environment [8]. Furthermore, emotional reactivity can be differentiated from the related construct of affect intensity, with the latter denoting the general tendency to experience intense emotions across environmental contexts and the former referring to more discrete and proximal responses to specific environmental events. Despite the theorized importance of emotional reactivity per se to BPD, however, very few studies have examined this trait in particular, with most studies of self-reported emotional dysfunction in BPD collapsing across multiple lower-order traits. For example, although studies using the Affect Intensity Measure (AIM) [9] provide evidence for heightened levels of a generalized affect intensity/reactivity among individuals with BPD (compared to individuals with other personality disorders or bipolar II disorder) [10,11], none of these studies has examined the extent to which emotional reactivity per se is heightened among individuals with BPD; thus, it is possible that the observed betweengroup differences on the AIM are solely the result of heightened levels of self-reported affect intensity among individuals with BPD. Nonetheless, providing preliminary support for a relationship between emotional reactivity per se and BPD, a recent study found that BPD features among college students are positively associated with self-reported negative emotional reactivity [12]. Moreover, despite failing to provide consistent evidence for heightened psychophysiological reactivity in BPD [13-18], results of laboratory-based studies investigating emotional reactivity in BPD provide further support for heightened subjective emotional reactivity in BPD. Specifically, 2 studies examining emotional reactivity to emotionally evocative and neutral images found that individuals with BPD (compared to healthy controls) reported greater subjective reactivity to neutral images, despite demonstrating lower psychophysiological reactivity to all images [14,15]. Even less research has examined the relationship between delayed emotional recovery and BPD, with most literature in this area being theoretical in nature [1]. However, results of a recent study using an ecological momentary assessment approach provide suggestive evidence for delayed recovery following emotional arousal among individuals with BPD, finding that individuals with BPD reported a longer duration of emotional distress than healthy controls [19]. The preliminary evidence for heightened subjective emotional reactivity and delayed emotional recovery in BPD notwithstanding, further research is needed to examine the nature and extent of emotional reactivity and delayed recovery among individuals with BPD across specific emotional states. In particular, although evidence suggests

that the emotional reactivity and delayed recovery in BPD may be specific to negative emotions [19,20] (with studies demonstrating a general underresponsiveness and hyporeactivity for positive emotions in BPD; see references [21,22]), additional research is needed to examine these traits across distinct negative emotions. Indeed, research on the related trait of affective instability has found that the affective instability characteristic of BPD does not involve all emotions but is emotion-specific [11]. Specifically, research is needed to examine the particular emotions of anger, anxiety, and shame, all of which have been emphasized in the literature on BPD [3,11,23-28]. 1.2. Anger and anxiety in BPD Historically, theoretical and clinical literature on BPD emphasized the relevance of anger and anxiety to this disorder [23-25], noting the prominence of these emotions among individuals with BPD [24-26,29,30]. This literature is supported by empirical research suggesting the relevance of these emotions to BPD. For example, in the aforementioned study of affective instability in BPD, Koenigsberg and colleagues [11] found that the affective instability characteristic of patients with BPD is specific to anger and anxiety, with results indicating greater lability of anger and anxiety (but not depression and elation) among patients with BPD (vs those with other personality disorders), even when controlling for mood disorders, gender, and age. With regard to research on the relevance of anger in particular to BPD, studies have found heightened levels of anger, hostility, and irritability among individuals with BPD (compared to both healthy controls [31] and patients with a mood disorder [32]), consistent with the theorized centrality of anger in general [25,29,30], and hostility and irritability in particular [23,24,33], in BPD. As for research on the relevance of anxiety to BPD, in addition to evidence that patients with BPD exhibit elevated symptoms of anxiety (eg, see references [26,34]) and have high rates of co-occurring anxiety disorders [35-37], the trait of anxiousness has the highest loading on the emotion dysregulation factor of the Dimensional Assessment of Personality Pathology [38], proposed by Livesley and colleagues [3] to be the core feature of BPD. 1.3. Shame and BPD More recently, literature has begun to emphasize the central role of shame in BPD, with this emotion in particular thought to underlie some of the most troubling symptoms associated with this disorder (eg, suicidal behaviors, deliberate self-harm, and aggressive behaviors) [1,27,28]. Indeed, it has even been suggested that BPD may best be conceptualized as a chronic shame response [39]. The development of an overwhelming shame response (or shame-proneness) among individuals with BPD is thought to begin at an early age in response to chronic experiences of abuse, neglect, and/or invalidation [27,39].

K.L. Gratz et al. / Comprehensive Psychiatry 51 (2010) 275–285

Unlike guilt, which is associated with the belief “I have done something bad,” shame is associated with the belief “I am bad” [40]. Not surprisingly, guilt and shame have been found to be differentially related to psychopathology, with guilt being negatively associated with or unrelated to psychopathology and shame being positively associated with psychopathology [40]. Indeed, many of the behaviors found to be positively associated with shame are common among individuals with BPD, including interpersonal hostility [40] and suicidality [41,42]. Although limited, recent empirical research provides support for the theorized centrality of shame to BPD. For example, Rüsch et al [27] found that women with BPD reported higher levels of shame-proneness and state shame than women with social phobia and healthy controls, in addition to evidencing a more shame-prone self-concept on an implicit association test. Yet, although both theoretical and empirical literature alike suggest the relevance of shame to the emotional dysfunction of BPD, further research is needed to examine whether emotional reactivity and delayed recovery in BPD are differentially related to shame versus other emotional states. 1.4. The current study The overall goal of the current study was to extend past research on emotional reactivity and delayed recovery in BPD by examining the role of several discrete BPDrelevant negative emotions (specifically, shame, anxiety, irritability, and hostility), rather than just negative emotions in general. Further, this study explored the relevance of 2 different types of laboratory stressors to emotional dysfunction in BPD (one of which was general and the other which involved negative evaluation). Specifically, we examined reactivity to a computerized stressor shown to induce anxiety, irritability, and frustration (the Paced Auditory Serial Addition Task—Computerized Version [PASAT-C]; see references [43-45]),and then provided participants with negative feedback about their performance during this stressor (thus providing a negatively evaluative stressor). Self-reported emotional states (ie, anxiety, irritability, hostility, and shame) were assessed at several different times throughout the study. It was hypothesized that, compared to a clinical comparison group of outpatients without a personality disorder (non-PD), outpatients with BPD would demonstrate heightened subjective emotional reactivity to the laboratory stressors, reporting higher levels of emotions in response to both the general and negatively evaluative stressors. Furthermore, we hypothesized that outpatients with BPD (vs non-PD outpatients) would demonstrate a delayed return to baseline arousal, as evidenced by a failure to return to baseline levels of emotional arousal by the end of the final recovery period. Given that all of the emotions examined in this study are considered to be BPD-relevant, no specific hypotheses were made regarding emotion-specific reactivity or delayed recovery.

277

2. Method 2.1. Participants Two groups of outpatients (BPD and non-PD) were recruited through advertisements posted at a psychiatric hospital and on 2 websites. Participants were required to be between 18 and 60 years of age and currently receiving outpatient psychiatric treatment. After providing written informed consent, potential participants were interviewed using the Diagnostic Interview for DSM-IV Personality Disorders [46] and the Current Mood Episodes, Psychotic Symptoms, and Alcohol and Non-Alcohol Dependence sections of the Structured Clinical Interview for DSM-IV Axis I Disorders [47]. Inclusion criteria included either meeting 5 or more criteria for BPD (BPD group), or not meeting full criteria for any PD and not meeting more than 3 criteria for BPD (non-PD group). Exclusion criteria for both groups focused on the presence of Axis I psychopathology that could influence emotional responding to the stressors, including current manic, hypomanic, or depressive mood episodes (but not lifetime history of mood disorders), current substance dependence, and/or current diagnosis of a psychotic disorder. Thirty-six participants (18 per group) completed the experimental portion of the study. However, after completing the experiment, 1 participant in the BPD group revealed to the experimenter that he had been informed about the purpose of the study and use of deception before participation and had altered his responses accordingly; thus, this participant was excluded, resulting in a final sample size of 35. 2.1.1. Borderline personality disorder group Participants with BPD were predominantly female (88%), White (77%), single (77%), and highly educated (some college = 24%; college graduate = 35%; graduate school = 35%), and ranged in age from 18 to 52, with a mean age of 34.06 (SD = 11.15). In regard to their clinical characteristics, 35% of the participants reported at least one inpatient psychiatric hospitalization in the past year, and 35% reported past-year treatment in a partial hospitalization program. Participants received an average of 3.30 (SD = 3.23) hours of outpatient psychiatric treatment per week and were taking an average of 3.18 (SD = 1.74) psychiatric medications. With regard to the particular classes of psychiatric mediations being taken by participants, 65% reported taking antidepressant medications, 53% reported taking anxiolytics, 18% reported taking antipsychotic medications, 29% reported taking mood stabilizers, 47% reported taking sleep medications, and 12% reported taking stimulants. As with most BPD samples, participants had high rates of Axis II comorbidity (65%; specifically, avoidant, dependent, and obsessive-compulsive personality disorders), with 47% meeting criteria for 1 other personality disorder, 12% meeting criteria for 2 other personality disorders, and 6% meeting criteria for 3 additional personality disorders.

278

K.L. Gratz et al. / Comprehensive Psychiatry 51 (2010) 275–285

2.1.2. Clinical comparison group Non-PD participants were predominantly female (78%), White (100%), single (61%), and highly educated (some college = 33%; college graduate = 39%; graduate school = 28%), and ranged in age from 18 to 57, with a mean age of 37.33 (SD = 12.08). In regard to the clinical characteristics of this group, 17% of the participants reported at least 1 inpatient psychiatric hospitalization in the past year, and 22% reported past-year treatment in a partial hospitalization program. Participants received an average of 2.57 (SD = 3.18) hours of outpatient psychiatric treatment per week and were taking an average of 2.56 (SD = 2.62) psychiatric medications. With regard to the particular classes of psychiatric medications being taken by participants, 61% reported taking antidepressant medications, 33% reported taking anxiolytics, 11% reported taking antipsychotic medications, 22% reported taking mood stabilizers, 39% reported taking sleep medications, and 6% reported taking stimulants. Finally, it warrants mention that although participants in the clinical comparison group were required to not meet full criteria for a personality disorder, 44% of these participants qualified for a subthreshold personality disorder (ie, meeting all but 1 criteria for [at least] 1 of the personality disorders), indicative of at least some chronic difficulties. 2.2. Materials 2.2.1. Clinical Interviews The Current Mood Episodes, Psychotic Symptoms, and Alcohol and Non-Alcohol Dependence sections of the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I/P) [47] were used to assess for the exclusion criteria (including current mood episodes, current substance dependence, and current psychotic disorders). Interviews were conducted by masters-level clinicians, trained in the administration of the SCID-I/P. All interviews were reviewed by a PhD-level clinician (KLG); diagnostic agreement was 100%. The Diagnostic Interview for DSM-IV Personality Disorders (DIPD-IV) [46] was used to assess for the presence of Axis II personality disorders. The DIPD-IV has demonstrated good interrater and test-retest reliability [48], with interrater κ coefficients ranging from 0.68 to 0.73 (based on 84 pairs of raters independently rating 27 videotaped assessments) and a test-retest κ coefficient of 0.69 to 0.74. The DIPD-IV was administered by masterslevel clinicians trained in the administration of this interview. All interviews were reviewed by a PhD-level clinician (KLG). In the 3 cases for which a discrepancy in diagnosis was evident, areas of disagreement were discussed as a group and a consensus was reached. As a result of the collaborative, iterative process we used to diagnose PDs, no specific data on the reliability of the diagnostic assessments are available. 2.2.2. Affect Intensity Measure The Affect Intensity Measure (AIM) [9] is a 40-item measure of the characteristic (ie, trait) intensity and reactivity

of emotional responses, independent from the frequency and hedonic level of emotional responses. Research suggests that the AIM has high internal consistency and good test-retest reliability over a period of 2 years [9,49], as well as good construct validity [9,50-52]. Moreover, evidence for the convergent validity of the AIM is provided by findings of a significant association with borderline pathology among a sample of psychiatric patients [53]. Although Larsen and Diener [9] developed the AIM as a unidimensional measure of affect intensity, research suggests that the AIM is multidimensional, measuring both positive and negative affect intensity and emotional reactivity [54-57]. Given recent findings that the relationship between affect intensity/reactivity and BPD may be specific to negative emotions [20], as well as the focus of the current study on negative emotional reactivity, scores were computed for both the negative affect intensity and negative emotional reactivity variables (based on the criteria of Weinfurt et al [56]). Items were recoded so that higher scores in every case indicated greater negative affect intensity or reactivity. The AIM was included in the present study to provide an assessment of trait negative affect intensity and emotional reactivity, allowing us to examine the extent to which trait-based measures of emotional responding correspond to and/or differ from real-time assessments of emotional intensity and reactivity to specific laboratory stressors. Internal consistency in the current sample was α = .82 for the negative affect intensity subscale and α = .63 for the negative emotional reactivity subscale. 2.2.3. Subjective emotional response assessments Subjective emotional responses were assessed at 5 different time points throughout the experiment. In line with the methods used in previous studies assessing reactivity to the laboratory stressor used here (ie, the PASAT-C; see references [43,45]), participants were asked to report on their levels of anxiety, irritability, and hostility throughout the study. In addition, given its suggested centrality to BPD [28,39], participants also were asked to rate their levels of shame. Specifically, at baseline and again at 4 different points throughout the study (see Procedure), participants were asked to rate the extent to which they were currently (“right now”) experiencing anxiety, irritability, hostility, and shame on a scale from 1 (not at all) to 5 (extremely). 2.2.4. Laboratory stressors This study used a modified version of the PASAT-C, an empirically-supported laboratory stressor shown to induce emotional distress in the form of anxiety, frustration, and irritability [43-45]. During this task, numbers are sequentially flashed on a computer screen, and participants are instructed to sum the most recent number with the previous number (using the computer mouse to click on the correct answer). After providing

K.L. Gratz et al. / Comprehensive Psychiatry 51 (2010) 275–285

each sum, the participant must ignore the sum and add the following number to the most recently presented number. When a correct answer is provided, a point is obtained. If an incorrect answer is provided, or if the participant fails to provide an answer before the next number is presented, an “explosion” sound is played and the score does not change. The version of the PASAT-C used in this study consisted of 4 levels, the first 3 of which had increasingly shorter latencies between number presentations. Because the correct answer must be provided before the presentation of the next number to obtain a point, difficulty increases as latencies decrease. Level 1 (low difficulty) had a 3-second latency between number presentations, Level 2 (medium difficulty) had a 2-second latency, and Levels 3 and 4 (high difficulty) had a 1-second latency. The first level lasted 1 minute, the second level lasted 2 minutes, and the third level (which served as a prime for the final level) lasted 1 minute. Following a brief 60second resting period at the end of level 3 to complete the emotion ratings (to assess reactivity to the PASAT-C and prevent differing durations in the final level from influencing the emotion ratings; see references [43,44]), the final level began. The final level had the same latency between number presentations as level 3 (ie, 1 second) but lasted 7 minutes and included an option to terminate the task at any time. Immediately following completion of the PASAT-C, standardized negative feedback appeared on the participants' computer monitor. Specifically, participants were presented with the following message: “Your task performance was [below average]. Compared to others who have completed the task, your performance was in the [bottom 10%]. Based on your performance, you will receive [8 minutes], out of a possible 20 minutes, to solve the anagrams.” (For information on the anagrams task, see below.) 2.3. Procedure All methods used in this study received prior approval by the institution's institutional review board. After providing written informed consent, participants completed the screening interview and self-report questionnaire, for which they were reimbursed $25. Eligible participants were then scheduled for the experimental portion of the study. Upon arrival to the laboratory, participants were informed that the purpose of the study was to examine performance on 2 problem-solving tasks, 1 of which involved working memory (PASAT-C) and the other which assessed verbal problemsolving abilities (anagrams). Participants were provided with instructions for each of these tasks, and informed that their performance on the tasks would determine the amount of money they would receive as reimbursement for their participation, ranging from $15 to $25 (providing an incentive to perform well on the tasks). Following provision of the study instructions, participants were escorted to a study cubicle, separate from the experimenter. The experimenter

279

turned on the participants' computer monitor, left the cubicle, and instructed the participant to await further instructions. All study procedures were computerized, requiring limited interaction between the participants and experimenter; the only contact after the start of the experimental procedure occurred when the experimenter entered the cubicle to read aloud the experimental instructions that appeared on the participants' computer screen. Participants were first asked to rate their levels of anxiety, irritability, hostility, and shame, providing a baseline assessment of subjective emotional responses (Baseline). After a 5-minute resting period, participants completed the PASAT-C. As described above (see Laboratory stressors), participants were asked to rate their current emotional state between Levels 3 and 4 of the PASAT-C (PASAT-C Reactivity). Following completion of the PASAT-C and receipt of the negative feedback (which appeared on their computer screen immediately after completion of the PASAT-C), participants were once again instructed to rate their current emotional state across the 4 emotions of interest (Feedback Reactivity). Next, participants were provided with a list of 48 solvable anagrams, each of which contained 6 to 7 letters [58]. Examples include EMKONY (ie, monkey) and NORGEA (ie, orange). Participants were instructed to solve as many anagrams as they could in the time provided. All participants received 8 minutes to work on the anagrams. At the end of the 8 minutes, participants were instructed to stop working on the anagrams and were once again asked to report on their subjective emotional responses (Post-Anagrams). Next, participants were instructed to sit quietly for 5 minutes, after which they completed the final ratings of their current emotional state (Post-Recovery). Once the experimental procedure was completed, participants were guided through a brief relaxation exercise (designed to alleviate any lingering distress). After the relaxation exercise, participants were fully debriefed about the purpose of the study, the use of deception, and the rationale for providing the false feedback. Participants also were provided with a list of coping strategies for managing distress. All participants (regardless of their performance) were reimbursed $25 for participation in this part of the study. 3. Results 3.1. Preliminary analyses A series of t tests and χ2 analyses were conducted on relevant demographic (ie, age, gender, and racial/ethnic background) and clinical (ie, amount of therapy, number of psychiatric medications, rates of use of the different classes of psychiatric medications [ie, antidepressants, anxiolytics, antipsychotics, mood stabilizers, sleep medications, and stimulants], and past-year hospitalization rates) characteristics to determine equivalence across

280

K.L. Gratz et al. / Comprehensive Psychiatry 51 (2010) 275–285 1

groups. Results indicate no significant between-group differences on any of these variables (ts b 1.00, χ2 s b 1.37; η2p s b 0.05, contingency coefficients b 0.19; Ps N .10), with one exception: significantly more BPD participants than non-PD participants self-identified as nonWhite (χ2 = 4.78, P b .05). Given that none of the dependent variables was significantly associated with racial background, however, this variable was not included as a covariate in later analyses. Furthermore, preliminary analyses were conducted to examine between-group differences in performance on the laboratory tasks (expected to influence participants' emotions in response to and following the tasks). Results indicate that the groups did not differ significantly in performance on the PASAT-C (t = 1.06, P N .10) or anagrams task (t = 0.13, P N .10). 3.2. Analysis plan To examine between-group differences in trait negative emotional intensity and reactivity, a 1-way (BPD vs non-PD) multivariate analysis of variance (MANOVA) was conducted on the negative affect intensity and negative emotional reactivity variables from the AIM. Likewise, to examine between-group differences in subjective emotional responses at the various time points throughout the study, a 1-way (BPD vs non-PD) MANOVA was conducted with self-reported anxiety, irritability, hostility, and shame across the 5 assessment points (ie, Baseline, PASAT-C Reactivity, Feedback Reactivity, Post-Anagrams, and Post-Recovery) serving as the dependent variables. To examine between-group differences in the pattern of change in subjective emotional responses following exposure to the first laboratory stressor (ie, the PASAT-C) through postrecovery, a series of 2 (group) × 4 (assessment point) 1 Analyses were also conducted to explore more fully the effect of psychiatric medication use on the outcomes of interest. To this end, we conducted a series of exploratory analyses examining the associations between the use of different classes of medication (ie, antidepressants, anxiolytics, antipsychotics, mood stabilizers, sleep medication, and stimulants), as well as the number of medications, and self-reported emotional responses at each of the assessment points (ie, Baseline, PASAT-C Reactivity, Feedback Reactivity, Post-Anagrams, and PostRecovery). Findings indicated that the use of antidepressant, anxiolytic, antipsychotic, or sleep medications was not significantly associated with self-reported emotional responses at any assessment point. However, the use of stimulant medications was associated with higher levels of anxiety and shame at several assessment points (Fs N 6.12, Ps b .05); the use of mood stabilizers was associated with higher levels of hostility following the anagrams (F1,33 = 4.91, P b .05); and number of medications was positively associated with levels of shame at Post-Recovery (r = 0.41, P b .05) and levels of anxiety at baseline, post-recovery, and after the negative feedback (r's N 0.35, P's b .05). Given these findings, we examined whether the results of our primary analyses changed when controlling for the use of different classes of psychiatric medications or the number of medications used. The results of the repeated measures analyses of covariance (see Section 3.3.2) did not change when controlling for medication status across the different classes of medications or number of different psychiatric medications.

repeated measures analyses of covariance (ANCOVAs; controlling for baseline levels of that emotional response, as well as performance on the laboratory tasks) were conducted on anxiety, irritability, hostility, and shame. In all repeated measures ANCOVAs, the multivariate test statistic was used. For those analyses indicating a statistically significant group by assessment point interaction effect, the following post hoc analyses were conducted: (a) 1-way (BPD vs non-PD) ANCOVAs (controlling for baseline levels of the emotional response and task performance) were used to examine between-group differences in emotional responses at each assessment point; (b) 1-way (Baseline vs Post-Recovery) repeated measures ANCOVAs (controlling for performance on the laboratory tasks) conducted within each group separately were used to examine changes in levels of the emotional response from baseline to post-recovery; (c) 1-way within-group repeated measures ANCOVAs (controlling for baseline levels of that emotional response, as well as performance on the laboratory tasks) were used to examine changes in levels of the emotional response from its peak level to postrecovery; and (d) independent-sample t tests were conducted to examine between-group differences in the slopes of the lines from the highest point of arousal to both the next time point (to examine between-group differences in the rate of emotional recovery immediately following peak arousal) and post-recovery (to investigate between-group differences in the rate of recovery from peak arousal to the end of the study). For all analyses, α was set at .05, using the more conservative 2-tailed test. 3.3. Primary analyses 3.3.1. Trait negative emotional intensity and reactivity Results provide evidence for the differential relevance of trait negative emotional intensity and reactivity to BPD, suggesting the importance of examining these emotionrelated traits separately. Specifically, although the overall effect of group on trait negative emotional intensity and reactivity was significant (Wilks Λ = .52, F2,32 = 14.96, P b .001, multivariate η2p = 0.48), examination of the univariate effects revealed significant between-group differences in only negative emotional intensity, with BPD participants reporting significantly higher levels of negative emotional intensity than non-PD participants (BPD = 46.00 ± 4.95; non-PD = 35.33 ± 6.29; F1,33 = 30.85, η2p = 0.48; P b .01). The groups did not differ significantly in trait negative emotional reactivity (BPD = 25.94 ± 4.85; non-PD = 24.06 ± 3.96; F1,33 = 1.59, η2p = 0.05; P N .10). 3.3.2. Emotional reactivity and recovery in response to the laboratory stressors Means and standard deviations for self-reported emotional responses across all assessment points are presented in Table 1. Results of the MANOVA examining betweengroup differences in self-reported emotional responses revealed a significant overall effect of group (Wilks Λ = .19, F14,20 = 3.02, P b .05, multivariate η2p = 0.81). Consistent with

K.L. Gratz et al. / Comprehensive Psychiatry 51 (2010) 275–285 Table 1 Descriptive and inferential statistics for between-group differences in emotional responses BPD (n = 17)

Anxiety Baseline PASAT-C Reactivity Feedback Reactivity Post-Anagrams Post-Recovery Irritability Baseline PASAT-C Reactivity Feedback Reactivity Post-Anagrams Post-Recovery Hostility Baseline PASAT-C Reactivity Feedback Reactivity Post-Anagrams Post-Recovery Shame Baseline PASAT-C Reactivity Feedback Reactivity Post-Anagrams Post-Recovery

Non-PD (n = 18)

Test statistic F

η2p

Mean

SD

Mean

SD

2.94 3.09 2.97 2.71 2.41

0.93 1.18 1.39 1.06 0.96

2.08 2.14 2.06 2.25 1.89

1.02 1.12 1.36 1.19 1.08

6.73⁎ 5.98⁎ 3.89 1.42 2.29

0.17 0.15 0.11 0.04 0.07

2.24 3.00 3.29 2.88 2.82

1.20 1.28 1.11 1.36 1.07

1.33 2.22 2.06 2.28 1.94

0.59 0.88 1.35 1.02 0.87

8.08⁎⁎ 4.46⁎ 8.77⁎⁎ 2.23 7.08⁎

0.20 0.12 0.21 0.06 0.18

1.35 2.18 2.24 2.06 2.06

0.61 1.29 1.20 1.03 1.03

1.00 1.44 1.28 1.67 1.22

0.34 0.62 0.75 0.91 0.43

4.56⁎ 4.70⁎ 8.10⁎⁎ 1.43 10.07⁎⁎

0.12 0.13 0.20 0.04 0.23

2.24 2.94 3.53 3.41 3.29

1.20 1.34 1.55 1.46 1.45

1.06 1.67 1.78 1.83 1.56

0.24 1.19 1.31 1.10 1.04

16.73⁎⁎ 8.85⁎⁎ 13.14⁎⁎ 13.16⁎⁎ 16.78⁎⁎

0.34 0.21 0.29 0.29 0.34

⁎ P b .05. ⁎⁎ P b .01.

findings of higher levels of trait negative affect intensity among BPD participants, examination of the univariate effects revealed that BPD (vs non-PD) participants reported significantly higher levels of all emotional responses at baseline. Thus, all primary repeated measures analyses reported below will control for baseline levels of the emotional response. Results indicate no between-group differences in the pattern of change in anxiety, hostility, or irritability over the course of the study, as the group × assessment point interaction was not significant for anxiety (F3,28 = 0.84, η2p = 0.08, P N .10), hostility (F3,28 = 1.95, η2p = 0.17, P N .10), or irritability (F3,28 = 2.28, η2p = 0.20, P N .10). Interestingly, however, analyses did provide evidence for between-group differences in emotional reactivity and recovery for the emotion of shame, indicating a significant group × assessment point interaction effect (F3,28 = 3.91, η2p = 0.30, P b .05) (see Fig. 1).2 Post hoc ANCOVAs (controlling for baseline levels of shame and performance on the PASAT-C) revealed that although there was no 2 Given evidence of gender differences in emotional intensity and reactivity [9,51,52], as well as the fact that most of the sample was female, the same analysis was conducted among just the female participants (to ensure that the findings are applicable to women in particular). Findings among the subsample of female participants (n = 29) were the same as those obtained among the larger mixed-gender sample, indicating a significant group × assessment point interaction effect for shame (F3, 22 = 3.16, η2p = 0.30, P b .05).

281

significant between-group difference in shame in response to the PASAT-C (PASAT-C Reactivity; F1,31 = .43, η2p = 0.01, P N .10), BPD participants reported significantly higher levels of shame after the negative feedback (F1,31 = 5.64, η2p = 0.15, P b .05). Furthermore, additional post hoc ANCOVAs (controlling for baseline levels of shame and performance on both the PASAT-C and anagrams task) revealed that their levels of shame remained significantly higher than the non-PD group at all subsequent time points (ie, post-anagrams and post-recovery) (F's1,30 N 6.71, η2ps ≥ 0.18, Ps b .05), and did not decrease significantly over the remaining time points (F2,12 = 1.59, η2p = 0.21, P N .10). Finally, although both groups reported higher levels of shame at the end of the recovery period (compared to baseline), this difference was statistically significant only for the BPD participants (for BPD: F1,14 = 6.02, η2p = 0.30, P b .05; for non-PD: F1,15 = 3.50, η2p = .19; P N .08). Interestingly, however, post hoc analyses examining between-group differences in the slopes of the lines from participants' peak levels of shame (ie, PASAT-C Reactivity for non-PD participants and Feedback Reactivity for BPD participants; see Fig. 1) to both the following period and post-recovery revealed no significant differences in the rate of emotional recovery across the groups. Specifically, findings indicated no significant between-group differences in the slopes of the lines from participants' peak levels of shame to either the following period (t33 = 1.30, P = .20) or post-recovery (t33 = 0.38, P = .71). 4. Discussion The purpose of this study was to extend extant research on emotional reactivity and delayed emotional recovery in BPD

Fig. 1. Group by assessment point interaction for shame (controlling for baseline levels of shame and performance on the laboratory tasks).

282

K.L. Gratz et al. / Comprehensive Psychiatry 51 (2010) 275–285

by examining the precise nature and extent of these traits in BPD, especially with regard to specific emotional states and the contexts in which these emotions occur. To this end, this study examined emotional reactivity (and recovery following emotional arousal) to 2 different laboratory stressors, exploring the impact of these stressors on subjective responding across the specific emotions of anxiety, irritability, hostility, and shame. Consistent with prominent theoretical accounts of BPD [1], findings provide some evidence for subjective emotional reactivity in BPD. However, rather than suggesting the presence of a generalized emotional reactivity in BPD, results provide evidence for context-dependent and emotion-specific reactivity in BPD. Indeed, findings indicated no group differences in trait negative emotional reactivity. Furthermore, whereas BPD participants (compared to non-PD participants) evidenced heightened reactivity in response to the negative evaluation, BPD and non-PD participants did not differ in their reactivity to the general laboratory stressor (which has been found in previous research to induce emotional distress in the form of anxiety, frustration, and irritability) [43-45]. As such, results suggest that the heightened emotional reactivity within BPD may be context-dependent, and more likely to occur in response to stressors such as negative evaluation than more general stressors. Likewise, evidence suggests that the emotional reactivity and delayed recovery in BPD may be specific to particular emotional responses. That is, whereas BPD participants did not evidence a different pattern of change in subjective anxiety, irritability, or hostility across the course of the study than non-PD participants, they did evidence a significantly different pattern of change in shame. Specifically, not only did BPD participants report higher levels of shame in response to the negative evaluation, their levels of shame remained elevated following this stressor (through the postrecovery period at the end of the study). Furthermore, only BPD participants reported significantly higher levels of shame at the end of the study (relative to their baseline levels). Importantly, however, post hoc analyses indicated no differences in the rate of emotional recovery across groups, suggesting that although BPD individuals may demonstrate a delay in the return to baseline levels of shame (relative to non-PD individuals), this is likely the result of the magnitude of their emotional reaction rather than the maintenance of emotional arousal once activated. Thus, although findings provided evidence for shame-specific delayed recovery among participants with BPD, this delayed recovery was due to the greater intensity of their shame response rather than a slower rate of emotional recovery. Altogether, these findings suggest the importance of continuing to examine emotional reactivity and recovery in BPD across specific contexts and emotions, rather than at the more general trait level (which may obscure the complexity and intricacies of these characteristics in BPD). Findings also provide evidence for the differential relations of emotional intensity and emotional reactivity to

BPD, highlighting the importance of investigating these traits in BPD separately (rather than collapsing across these traits when examining emotional dysfunction in BPD). Indeed, in contrast to evidence that the emotional reactivity in BPD may be emotion- and context-specific (rather than generalized), result suggests the presence of a generalized heightened emotional intensity in BPD. Specifically, not only did BPD participants report significantly higher levels of trait negative emotional intensity, they reported significantly higher levels of all emotional responses at baseline. These findings are consistent with theories implicating an underlying emotional intensity in the pathogenesis of BPD [1] and suggest that there may be important differences in the nature and extent of emotional intensity (vs emotional reactivity) in BPD, with intensity being a more generalized vulnerability than reactivity. Although promising, the results of the present study are preliminary and must be evaluated in light of the study's limitations. First and foremost, this study used a small and homogenous sample of participants, potentially limiting our statistical power and ability to detect between-group differences, as well as the generalizability of the results. Moreover, the decision to exclude participants with a current mood episode (due to the fact that the presence of such an episode could influence task performance) may further limit the generalizability of the findings, especially with regard to BPD, as outpatients with BPD commonly present with cooccurring mood disorders [2]. Thus, replication of these findings in larger, more diverse, and more representative samples is needed. Another sample-related limitation concerns our comparison group. Although the use of a clinical comparison group is an asset of this study (providing a more stringent control group than a normal or nonpsychiatric control group), the specific Axis I conditions present within this group were not assessed and significant Axis II psychopathology was explicitly excluded. Thus, the specificity of our findings to BPD (vs PDs in general) is unclear. Future studies are needed to explore whether the pattern of context- and emotion-specific reactivity found here is specific to BPD or associated with PDs more broadly. Only by controlling for the presence of other PD symptoms or specifically comparing individuals with BPD to those with other PDs will future research be able to speak to the specificity of these findings to BPD in particular. For example, although clinical and empirical literature emphasizing the centrality of shame to BPD [27,28,39] suggests that our findings may be BPD-specific, it is also possible that the experience of shame (and heightened shame reactivity) is associated with the presence of another PD (in particular, narcissistic personality disorder; see reference [59]). However, given that none of the BPD participants met criteria for narcissistic personality disorder, it is unlikely that findings are solely related to the presence of co-occurring narcissistic personality pathology among the BPD participants.

K.L. Gratz et al. / Comprehensive Psychiatry 51 (2010) 275–285

Data on the specific forms of Axis I psychopathology present within this sample also would provide valuable information on the specificity of these findings to BPD, and the ways in which emotional reactivity in BPD compares to emotional reactivity within other specific clinical groups. In particular, the absence of data on the anxiety disorder diagnoses of participants limits our ability to determine the extent to which findings of heightened shame reactivity are related to BPD specifically (rather than co-occurring anxiety disorders within the BPD group, such as social anxiety disorder). However, it warrants mention that comparable rates of participants in both groups were receiving treatment of anxiety disorder-related difficulties, and results did not change when anxiety-disorder focused treatment status was included in the analyses as a covariate.3 Nonetheless, to further address the specificity of these findings to BPD, future studies should include a more thorough assessment of Axis I psychopathology in general (and anxiety disorders in particular) and control for the presence of specific anxiety disorders across groups. Moreover, given that context-specific emotional reactivity is thought to play a central role in certain anxiety disorders as well (eg, PTSD and GAD; see references [60,61]), future research should examine whether the context-specific emotional reactivity observed here is more pronounced in BPD than in these other disorders. Likewise, the absence of data on general psychiatric symptom severity and/or global impairment limits our ability to speak to the effects of psychopathology on our findings. Nonetheless, the comparableness of these groups with respect to relevant clinical characteristics (including amount of treatment, number and type of psychiatric medications, and past year hospitalization rates) suggests that our findings reflect more than simply severity of psychopathology. To further clarify this issue, however, future studies should

3 Although the inclusion of baseline levels of emotional responses as a covariate in analyses examining the pattern of change in emotional responses over the course of the study limits concerns regarding the potential confounding influence of co-occurring anxiety disorders on baseline emotional responses, exploratory analyses were conducted to examine the potential influence of anxiety-related difficulties on the results. First, a χ2 analysis was conducted to examine between-group differences in the rate of participants currently receiving some form of treatment for anxiety disorder-related difficulties (including individual psychotherapy, group therapy, and/or pharmacotherapy). Findings indicate that rates of anxiety disorder-focused treatment did not differ between groups (with 53% of the BPD participants and 39% of the non-PD participants receiving some form of treatment focused on anxiety disorder–related difficulties; χ2 = .70, P = .40). Next, we conducted a series of analyses examining the associations between anxiety disorder–focused treatment status (yes vs no) and self-reported emotional responses at each of the assessment points and exploring whether the results of our primary analyses changed when controlling for this variable. Results indicate that anxiety disorder–focused treatment status was not significantly associated with self-reported emotional responses at any assessment point (P's N .10). Furthermore, results did not change when anxiety disorder–focused treatment status was included as a covariate in the repeated measures ANCOVAs.

283

include a more thorough assessment of general symptom severity/impairment. Limitations with regard to the experimental procedures also warrant mention. For example, the examination of only 4 BPD-relevant emotions may have limited our ability to detect emotion-specific reactivity in BPD. Although research provides support for the relevance of these emotions in particular to BPD [11,26,27,31,32,37], future studies should examine other emotions as well, such as dysphoria, loneliness, and/or worthlessness (see reference [62]). Furthermore, although the computerized nature of this study limited interactions between experimenters and participants (thereby reducing the extent to which experimenters may have inadvertently influenced participants' performance), masked assessments would have further decreased the potential for experimenter biases and should be used in future studies. Finally, due to our experimental design, this study examined reactivity to negative evaluation only after exposure to a general stressor. As a result, findings may not generalize to emotional reactivity to negative feedback in the absence of a prior stressor. Future studies should examine if exposure to a previous stressor and/or the presence of general emotional distress influences emotional responding to negative feedback among individuals with BPD. Future studies also would benefit from the use of physiological measures of emotional responding (such as heart rate variability and electrodermal response), as well as the inclusion of biological measures of emotional reactivity (eg, fluctuations in cortisol levels throughout the experimental procedure). In providing a more objective assessment of emotional responding, the use of such measures would assist in determining whether differences in emotional reactivity and recovery among individuals with BPD (vs those without BPD) are the result of actual physiological or biological differences in emotional responding or simply differences in how individuals with BPD respond to and evaluate their emotional experiences. Moreover, future research should examine the influence of participants' emotion regulation strategies during the laboratory session on emotional responding. For example, the use of dissociation or other emotionally avoidant regulation strategies in response to the laboratory stressors (likely more common among the BPD participants; see references [2,63,64]) may have resulted in lower emotional reactivity to these stressors, limiting our power to detect between-group differences in emotional reactivity and delayed recovery. Future studies should specifically assess and control for participants' reported regulation strategies during the stressors on emotional responding throughout the session. Furthermore, although neither rates of use nor number of psychiatric medications differed between groups and results did not change when controlling for psychiatric medications, medications likely influence emotional arousal and reactivity is a variety of ways. Thus, the wide variability in the use of medications within the groups

284

K.L. Gratz et al. / Comprehensive Psychiatry 51 (2010) 275–285

complicates interpretation of our results. Research is needed to explore more rigorously the impact of medication use on emotional reactivity and recovery among individuals with and without BPD. Despite limitations, the results of this study suggest that individuals with BPD may be emotionally reactive (and, as a result of the magnitude of their emotional reactions, evidence a delay in their return to baseline levels of emotional arousal) in the context of specific emotions (eg, shame) and specific cues (eg, negative evaluation). Findings in the extant literature pertaining to emotional reactivity in BPD may be obscured by the reliance on broad-based examinations of general emotional reactivity without consideration of the contexts that differentially elicit emotional responses and the specific emotional states that characterize emotional reactivity. Our findings of emotion-specific and context-dependent reactivity demonstrate the importance of continuing to examine reactivity across specific emotions and in response to particular stressors to better understand the nature and extent of subjective emotional reactivity in BPD. Results of this study also highlight the relevance of shame to BPD, providing preliminary evidence for shame-specific reactivity in BPD and suggesting the importance of assessing for the presence of this emotion among BPD patients in clinical settings. Furthermore, these findings have potential clinical implications, suggesting the potential utility of targeted interventions for BPD aimed at decreasing shame [28].

[9] [10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

Acknowledgment This research was supported by a grant from the Psychosocial Foundation of McLean Hospital awarded to the first author. The authors wish to thank Donna Lacroce for her assistance in data collection.

[19]

[20]

References [1] Linehan MM. Cognitive-behavioral treatment of borderline personality disorder. New York, NY: The Guilford Press; 1993. [2] Skodol AE, Gunderson JG, Pfohl B, Widiger TA, Livesley WJ, Siever JJ, et al. The borderline diagnosis I: Psychopathology, comorbidity, and personality structure. Biol Psychiatry 2002;51:936-50. [3] Livesley WJ, Jang KL, Vernon PA. Phenotypic and genetic structure of traits delineating personality disorder. Arch Gen Psychiatry 1998;55: 941-8. [4] Gratz KL, Tull MT, Gunderson JG. Preliminary data on the relationship between anxiety sensitivity and borderline personality disorder: the role of experiential avoidance. J Psychiatr Res 2008;42: 550-9. [5] Siever LJ, Davis KL. A psychobiological perspective on the personality disorders. Am J Psychiatry 1991;148:1647-58. [6] Siever LJ, Torgersen S, Gunderson JG, Livesley WJ, Kendler KS. The borderline diagnosis III: Identifying endophenotypes for genetic studies. Biol Psychiatry 2002;51:964-8. [7] Gunderson JG, Zanarini MC, Kisiel C. Borderline personality disorder. In: Widiger TA, editor. DSM-IV sourcebook, Vol. 2. Washington (DC): American Psychiatric Association Press; 1996. p. 717-33. [8] Rothbart MK, Derryberry D. Development of individual differences in temperament. In: Lamb ME, Brown AL, editors. Advances in

[21]

[22]

[23]

[24] [25] [26] [27]

[28]

[29]

developmental psychology, Vol. 1. Hillsdale (NJ): Erlbaum; 1981. p. 37-86. Larsen RJ, Diener E. Affect intensity as an individual difference characteristic: a review. J Res Pers 1987;21:1-39. Henry C, Mitropoulou V, New AS, Koenigsberg HW, Silverman J, Siever LJ. Affective instability and impulsivity in borderline personality and bipolar II disorders: similarities and differences. J Psychiatr Res 2001;35:307-12. Koenigsberg HW, Harvey PD, Mitropoulou V, Schmeidler J, New AS, Goodman M, et al. Characterizing affective instability in borderline personality disorder. Am J Psychiatry 2002;159:784-8. Cheavens JS, Rosenthal MZ, Daughters SB, Nowak J, Kosson D, Lynch TR, et al. An analogue investigation of the relationships among perceived parental criticism, negative affect, and borderline personality disorder features: the role of thought suppression. Behav Res Ther 2005;43:257-68. Ebner-Priemer UW, Badeck S, Beckmann C, Wagner A, Feige B, Weiss I, et al. Affective dysregulation and dissociative experience in female patients with borderline personality disorder: a startle response study. J Psychiatr Res 2005;39:85-92. Herpertz SC, Kunert HJ, Schwenger UB, Sass H. Affective responsiveness in borderline personality disorder: a psychophysiological approach. Am J Psychiatry 1999;156:1550-6. Herpertz SC, Schwenger UB, Kunert HJ, Lukas G, Gretzer U, Nutzmann K, et al. Emotional responses in patients with borderline as compared with avoidant personality disorder. J Personal Disord 2000; 14:339-51. Herpertz SC, Werth U, Lucas G, Qunaibi M, Schuerkens A, Kunert H, et al. Emotion in criminal offenders with psychopathy and borderline personality disorders. Arch Gen Psychiatry 2001;58:737-45. Herpertz SC, Koetting K. Startle response in inpatients with borderline personality disorder vs. healthy controls. J Neural Transm 2005;112: 1435-63. Schmahl CG, Elzinga BM, Ebner UW, Simms T, Sanislow C, Vermetten E, et al. Psychophysiological reactivity to traumatic and abandonment scripts in borderline personality and posttraumatic stress disorders: a preliminary report. Psychiatry Res 2004;126:33-42. Stiglmayr CE, Grathwol T, Linehan MM, Ihorst G, Fahrenberg J, Bohus M. Aversive tension in patients with borderline personality disorder: a computer-based controlled field study. Acta Psychiatr Scand 2005;111:372-9. Levine D, Marziali E, Hood J. Emotion processing in borderline personality disorders. J Nerv Ment Dis 1997;185:240-6. Lenzenweger MF, Clarkin JF, Fertuck EA, Kernberg OF. Executive neurocognitive functioning and neurobehavioral systems indicators in borderline personality disorder: a preliminary study. J Personal Disord 2004;18:421-38. Renneberg B, Heyn K, Gebhard R, Bachmann S. Facial expression of emotions in borderline personality disorder and depression. J Behav Ther Exp Psychiatry 2005;36:183-96. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed. Washington (DC): American Psychiatric Association; 1994. Gunderson JG, Singer MT. Defining borderline patients: an overview. Am J Psychiatry 1975;132:1-10. Kernberg OF. Borderline personality organization. J Am Psychoanal Assoc 1967;15:641-85. Snyder S, Pitts WM. Characterizing anxiety in the DSM-III borderline personality disorder. J Personal Disord 1988;2:93-101. Rüsch N, Lieb K, Gottler I, Hermann C, Schramm E, Richter H, et al. Shame and implicit self-concept in women with borderline personality disorder. Am J Psychiatry 2007;164:500-8. Rizvi SL, Linehan MM. The treatment of maladaptive shame in borderline personality disorder: A pilot study of “opposite action.”. Cogn Behav Pract 2005;12:437-47. Grinker R, Werble B, Drye R. The borderline syndrome: a behavioral study of ego functions. New York (NY): Basic Books; 1968.

K.L. Gratz et al. / Comprehensive Psychiatry 51 (2010) 275–285 [30] Zetzel ER. A developmental approach to the borderline patient. Am J Psychiatry 1971;127:867-71. [31] Gardner DL, Leibenluft E, O'Leary KM, Cowdry RW. Self-ratings of anger and hostility in borderline personality disorder. J Nerv Ment Dis 1991;179:157-61. [32] Wilson ST, Stanley B, Oquendo MA, Goldberg P, Zalsman G, Mann JJ. Comparing impulsiveness, hostility, and depression in borderline personality disorder and bipolar II disorder. J Clin Psychiatry 2007;68: 1533-9. [33] Houck JH. The intractable female patient. Am J Psychiatry 1972;129: 27-31. [34] Domes G, Czieschnek D, Weidler F, Berger C, Fast K, Herpertz SC. Recognition of facial affect in borderline personality disorder. J Personal Disord 2008;22:135-47. [35] McGlashan TH, Grilo CM, Skodol AE, Gunderson JG, Shea MT, Morey LC, et al. The collaborative longitudinal personality disorders study: baseline axis I/II and II/II diagnostic co-occurrence. Acta Psychiatr Scand 2000;102:256-64. [36] Skodol AE, Oldham JM, Hyler SE, Stein DJ, Hollander E, Gallagher PE, et al. Patterns of anxiety and personality disorder comorbidity. J Psychiatr Res 1995;29:361-74. [37] Zanarini MC, Frankenburg FR, Dubo ED, Sickel AE, Trikha A, Levin A, et al. Axis I comorbidity of borderline personality disorder. Am J Psychiatry 1998;155:1733-9. [38] Livesley WJ, Jackson DN. Manual for the dimensional assessment of personality problems—basic questionnaire. Port Huron (MI): Sigma; 2009. [39] Crowe M. Never good enough—part 1: shame or borderline personality disorder? J Psychiatr Ment Health Nur 2004;11:327-34. [40] Tangney JP, Dearing RL. Shame and guilt. New York (NY): The Guilford Press; 2002. [41] Hastings ME, Northman LM, Tangney JP. Shame, guilt, and suicide. In: Joiner T, Rudd MD, editors. Suicide science: expanding the boundaries. Norwell (Mass): Kluwer Academic Publishers; 2000. p. 67-80. [42] Lester D. The association of shame and guilt with suicidality. J Soc Psychol 1998;138:535-6. [43] Brown RA, Lejuez CW, Kahler CW, Strong D. Distress tolerance and duration of past smoking cessation attempts. J Abnorm Psychol 2002; 111:180-5. [44] Lejuez CW, Kahler CW, Brown RA. A modified computer version of the Paced Auditory Serial Addition Task (PASAT) as a laboratorybased stressor. Behav Ther 2003;26:290-3. [45] Daughters SB, Lejuez CW, Kahler CW, Strong DR, Brown RA. Psychological distress tolerance and duration of most recent abstinence attempt among residential treatment-seeking substance abusers. Psychol Addict Behav 2005;19:208-11. [46] Zanarini MC, Frankenburg FR, Sickel AE, Young L. Diagnostic interview for DSM-IV personality disorders. Boston (Mass): McLean Hospital; 1996. [47] First MB, Spitzer RL, Gibbon M, Williams JBW. Structured clinical interview for DSM-IV Axis I disorders—patient edition (SCID-I/P,

[48]

[49] [50] [51]

[52]

[53]

[54]

[55]

[56]

[57] [58]

[59] [60]

[61]

[62]

[63] [64]

285

Version 2.0). New York (NY): New York State Psychiatric Institute; 1996. Zanarini MC, Skodol AE, Bender D, Dolan R, Sanislow C, Schaefer E, et al. The Collaborative Longitudinal Personality Disorders Study: II. reliability of Axis I and II diagnoses. J Personal Disord 2000;14: 291-9. Larsen RJ, Diener E, Emmons RA. Affect intensity and reactions to daily life events. J Abnorm Soc Psychol 1986;51:803-14. Flett GL, Boase P, McAndrews MP, Pliner P, Blankstein KR. Affect intensity and the appraisal of emotion. J Res Pers 1986;20:447-59. Fujita F, Diener E, Sandvik E. Gender differences in negative affect and well-being: the case for emotional intensity. J Pers Soc Psychol 1991;61:427-34. Larsen RJ, Diener E. A multitrait-multimethod examination of affect structure: hedonic level and emotional intensity. Pers Individ Differ 1985;6:631-6. Flett GL, Hewitt PL. Criterion validity and psychometric properties of the Affect Intensity Measure in a psychiatric sample. Pers Individ Differ 1995;19:585-91. Bryant FB, Yarnold PR, Grimm LG. Toward a measurement model of the Affect Intensity Measure: a three-factor structure. J Res Pers 1996; 30:223-47. Simonsson-Sarnecki M, Lundh L, Törestad B. Factor structure and validity of the Affect Intensity Measure in a Swedish sample. Pers Individ Differ 2000;29:337-50. Weinfurt KP, Bryant FP, Yarnold PR. The factor structure of the Affect Intensity Measure: in search of a measurement model. J Res Pers 1994; 28:314-31. Williams DG. Neuroticism and extraversion in different factors of the Affect Intensity Measure. Pers Individ Differ 1989;10:1095-100. Baumeister RF, Bratslavsky E, Muraven M, Tice DM. Ego depletion: is the active self a limited resource? J Pers Soc Psychol 1998;74: 1252-65. Ronningstam EF. Identifying and understanding the narcissistic personality. New York (NY): Oxford University Press; 2005. Orsillo SM, Batten SV, Plumb JC, Luterek JA, Roessner BM. An experimental study of emotional responding in women with posttraumatic stress disorder related to interpersonal violence. J Trauma Stress 2004;17:241-8. Mennin DS, Heimberg RG, Turk CL, Fresco DM. Preliminary evidence for an emotion dysregulation model of generalized anxiety disorder. Behav Res Ther 2005;43:1281-310. Zanarini MC, Frankenburg FR, DeLuca CJ, Hennen J, Khera GS, Gunderson JG. The pain of being borderline: dysphoric states specific to borderline personality disorder. Harv Rev Psychiatry 1998;6:201-7. Bijttebier P, Vertommen H. Coping strategies in relation to personality disorders. Pers Individ Differ 1999;26:847-56. Rosenthal MZ, Cheavens JS, Lejuez CW, Lynch TR. Thought suppression mediates the relationship between negative affect and borderline personality disorder symptoms. Beh Res Ther 2005;43: 1173-85.