Research Master’s Thesis
Frontolimbic connectivity
during Emotion Regulation in
Borderline Personality Disorder
University of Amsterdam
Department of Psychology
Author:
Sascha Béla Duken
Student number:
11117478
Email:
sascha.duken@student.uva.nl
sascha_duken@web.de
Supervisor:
dr. Henk Cremers
2
ndAssessor:
prof. dr. Arnoud Arntz
1
Abstract
Borderline Personality Disorder (BPD) is a severe mental illness characterized as an instability of affects, relationships, and the self-image. One key problem in BPD seems to be emotion dysregulation. While emotion generation involves limbic regions (e.g. the amygdala), emotion regulation recruits prefrontal regions (e.g. the ventromedial prefrontal cortex). The current study investigated the role of aberrant communication between these regions (i.e. functional connectivity) during emotion regulation in BPD. Further, the study investigated whether aberrant functional connectivity is specific for BPD or whether it also occurs in Cluster-C Personality Disorders (ClCs), which are marked by persistent anxiety or fear. Participants viewed negative and neutral stimuli. Prior to each picture, they were instructed to look at the stimulus without regulating their emotions or to down-regulate their emotions by realizing that they were safe. Frontolimbic connectivity during emotion regulation was compared between BPD patients (n = 55), ClC patients (n = 24), and non-patients (n = 42) using psychophysiological interactions. Psychophysiological interactions allow to investigate task-dependent changes in functional connectivity between brain regions. BPD patients failed to recruit prefrontal regions to voluntarily down-regulate their emotions but showed normal frontolimbic connectivity during emotion regulation. ClC patients recruited prefrontal regions during emotion regulation but showed impaired frontolimbic connectivity. These results suggest that emotion dysregulation in BPD and ClC might be associated with distinct neural correlates.
2
1. Introduction
People who suffer from Borderline Personality Disorder (BPD) can experience extreme emotional heights only to skid into a devastating emotional valley a few moments later. BPD is characterized as an instability of affects, relationships, and self-image (DSM-V; American Psychiatric Association, 2013). It is associated with severe functional impairment, high suicidal risk, and pronounced psychological and emotional misery (Gratz & Roemer, 2008; Leichsenring, Leibing, Kruse, New, & Leweke, 2013). With a prevalence of 0.5 to 5.9 percent, BPD poses a significant burden not only to many individuals but to society as a whole (Leichsenring et al., 2013; van Asselt, Dirksen, Arntz, & Severens, 2007). Consequently, it is important to understand why BPD patients experience such intense, unstable emotions, and why they cannot regulate their emotional states. The current study investigated functional connectivity in the brain during emotion regulation to provide insight into the neural correlates of aberrant emotion processing in BPD. This might help to understand core features of BPD and how to address them in clinical interventions.
Emotion dysregulation seems to underlie not only BPD but also other psychiatric disorders (Zilverstand, Parvaz, & Goldstein, 2017) such as anxiety disorders (Cremers et al., 2015; Etkin & Wager, 2007), mood disorders (Chen et al., 2008; Johnstone, van Reekum, Urry, Kalin, & Davidson, 2007; Price & Drevets, 2012), and Obsessive Compulsive Disorder (Milad & Rauch, 2012; Taylor & Liberzon, 2007). However, emotions and mood fluctuate more suddenly in BPD than in other disorders. While the mood of BPD patients can change within a few hours (usually in response to environmental stimuli), dysregulated mood states in other disorders can last multiple days or weeks (American Psychiatric Association, 2013). Further, the emotional instability in BPD is particularly pronounced in social situations and often results in quick changes between elevation and demonization of the patients themselves and of significant others, which contributes to social problems (American Psychiatric Association, 2013; Arntz & ten Haaf, 2012; King-Casas et al., 2008). This instability of social evaluations might arise from an inability of BPD patients to mentalize, i.e. to imagine and understand mental states of themselves and others (Choi-Kain & Gunderson, 2008; Fonagy & Bateman, 2008). Consequently, while emotional instability in BPD has a certain similarity to other disorders, it is distinct in its rapid fluctuation and in the important role of social situations and relationships.
Even though the expression of psychopathology can be very different across disorders, theories that aim to explain specific disorders, especially on a biological level, are often quite similar and cannot account for the differences in psychopathology (Aldao & Nolen-Hoeksema, 2010; Taylor & Liberzon, 2007; Zilverstand et al., 2017). This study included a cluster-c personality disorder (ClC) control group to investigate whether similar or different neural correlates are associated with BPD and ClC (Aldao & Nolen-Hoeksema, 2010; Taylor & Liberzon, 2007; Zilverstand et al., 2017). ClCs include Avoidant Personality Disorder, Dependent Personality Disorder, and Obsessive Compulsive Disorder, which are characterized by exaggerated or generalized experiences of anxiety and fear, i.e. emotion dysregulation (American Psychiatric Association, 2013). This makes a comparison between neural correlates of BPD and ClC particularly interesting.
In order to investigate emotion dysregulation it is informative to differentiate between emotion generation and emotion regulation (Gross & Thompson, 2007). Emotion generation broadly consists of four steps (Gross & Thompson, 2007; Gross, Sheppes, & Urry, 2011). First, a situation elicits an affective or psychophysiological reaction in an individual. Second, attention is shifted towards or away from the situation. Third, the situation is cognitively appraised. Fourth, the individual responds behaviorally and physiologically to the situation. Emotion regulation can target any of the stages of the emotion processing pipeline (Gross & Thompson, 2007). One of the most effective emotion regulation strategies is reappraisal, which targets the cognitive appraisal of the situation (Buhle et al., 2014; Diekhof, Geier, Falkai, & Gruber, 2011). Specifically,
3
reappraisal aims at changing the interpretation of the situation or its consequences. Since BPD patients often fail to use reappraisal strategies to down-regulate their emotions (Koenigsberg, Siever, et al., 2009), clinical interventions aim at teaching such strategies (Arntz & van Genderen, 2009; Goodman et al., 2014). For example Schema-Focused Therapy, one of the most effective treatments available for BPD (Sempértegui, Karreman, Arntz, & Bekker, 2013), teaches patients to reappraise situations by realizing to be safe (Arntz & van Genderen, 2009).
Reappraisal strategies require active cognitive processing, whereas immediate emotional reactions are rapid and automatic (Buhle et al., 2014). Accordingly, biological models of emotion regulation suggest that cortical regions that are involved in higher cognitive functions (mainly located in the prefrontal cortex) regulate activity in limbic regions that are active during automatic emotional reactions (i.e. the amygdala, the striatum and the anterior insula; Ochsner, Silvers, & Buhle, 2012; Ochsner & Gross, 2005). The vmPFC might play a key role in emotion regulation because it not only directly regulates limbic activity but also mediates the effects of other higher-order brain regions on limbic activity (Diekhof et al., 2011; Motzkin, Philippi, Wolf, Baskaya, & Koenigs, 2015; Ochsner et al., 2012; Urry, 2006). However, some research suggests that activation of other cortical regions than the vmPFC might be more important for emotion regulation, e.g. the lPFC, the cingulate cortex, the parietal cortex, the insula, and the striatum (Buhle et al., 2014; Heatherton & Wagner, 2011; Kanske, Heissler, Schönfelder, Bongers, & Wessa, 2011; Morawetz, Bode, Derntl, & Heekeren, 2017). In sum, evidence for enhanced activation of prefrontal regions during emotion regulation is strong but the role of specific regions remains unsolved.
While activation of prefrontal regions might be necessary to successfully regulate emotions, it is not sufficient. The prefrontal cortex also has to effectively communicate with limbic regions. Activation of the prefrontal cortex without co-activation of limbic regions might reflect enhanced cognitive efforts to regulate emotions even though the regulation attempt fails to affect emotional experience. Studies of functional connectivity can reveal how signals in multiple areas of the brain correlate and thus, expand the understanding of frontolimbic emotion processing. Initial evidence suggests that successful emotion regulation depends on enhanced amygdala-frontal connectivity during reappraisal (Banks, Eddy, Angstadt, Nathan, & Luan Phan, 2007; Di, Huang, & Biswal, 2017; Morawetz, Bode, Baudewig, & Heekeren, 2016). Prefrontal regions that show enhanced coupling with limbic regions in studies of functional connectivity broadly overlap with regions that were found in studies of brain activation (e.g. mPFC, insula, cingulate cortex ; Di et al., 2017; Etkin, Egner, & Kalisch, 2011; Morawetz et al., 2017). Consequently, both, prefrontal activation and frontolimbic connectivity, might be necessary for successful emotion regulation.
Emotion dysregulation in psychopathology has been studied by comparing prefrontal and limbic activity during emotion regulation between patients and non-patient controls (NPC; van Zutphen, Siep, Jacob, Goebel, & Arntz, 2015; Zilverstand et al., 2017). However, findings have been inconsistent and no region consistently showed hypo- or hyper-activation during emotion generation or emotion regulation in BPD patients (van Zutphen et al., 2015). Consequently, a lack of frontolimbic connectivity rather than aberrant brain activity might cause emotion dysregulation in BPD. Initial evidence suggests that functional connectivity is altered in BPD patients compared to non-patients (Dudas et al., 2016; Wolf et al., 2011; Xu et al., 2016). Critically, BPD patients show aberrant functional connectivity between the amygdala and prefrontal regions (Cullen et al., 2011; Krause-Utz et al., 2014; New et al., 2007). Aberrant functional connectivity in BPD seemed to be clinically relevant because it correlated positively with symptom severity (Krause-Utz et al., 2014; Wolf et al., 2011; Xu et al., 2016). Furthermore, one study showed that BPD patients could train to down-regulate emotional arousal in response to negative pictures using neurofeedback (Paret et al., 2016). The improvement in emotion regulation was not reflected in reduced amygdala activity but in increased amygdala-vmPFC connectivity. While these studies show the importance of frontolimbic connectivity in general,
4
most of them have not investigated functional connectivity during an emotion regulation task. Consequently, the present study investigated whether BPD patients show aberrant frontolimbic connectivity that is specific for emotion regulation.
In this study, we reanalyzed a dataset of BPD patients, ClC patients, and non-patients controls (NPC), who viewed emotional and neutral pictures after being instructed to either only look at the stimuli (emotion generation) or to down-regulate their emotions using cognitive reappraisal (emotion regulation). Results from the activation study are presented elsewhere (van Zutphen et al., in prep). In summary, when BPD patients looked at emotional stimuli without cognitive regulation, they showed hyper-activation of the anterior insula, the temporoparietal junction and the dlPFC compared to NPC. However, when BPD patients were instructed to down-regulate their emotions, they failed to recruit prefrontal regions such as the anterior cingulate cortex, the dlPFC, and the middle temporal gyrus. Another study with the same sample investigated amygdala-prefrontal connectivity in resting-state fMRI before and after the emotion regulation task (Baczkowski et al., 2016). They found that non-patients compared to BPD patients showed increased amygdala connectivity with the mPFC, the dlPFC, the vlPFC and the superior temporal gyrus after the task. BPD patients showed lower amygdala connectivity with the posterior cingulate cortex and higher connectivity with the superior parietal lobe after the task. These results provide initial evidence that frontolimbic connectivity is altered in BPD and might relate to emotion dysregulation. However, it remains unclear which experimental manipulation caused the difference between pre- and post-task resting state fMRI. The current study investigated frontolimbic connectivity during the emotion regulation task using psychophysiological interaction (PPI) analyses. PPIs allow to investigate how the connectivity between brain regions varies during a task, depending on the experimental manipulations of interest, i.e. emotion generation and emotion regulation.
To summarize, BPD is most commonly defined as a disorder of emotion dysregulation characterized by abnormal frontolimbic activity. We investigated whether emotion dysregulation in
BPD is associated with abnormal frontolimbic connectivity during emotion regulation. We further
tested whether aberrant frontolimbic connectivity is specific for BPD or whether it also occurs in ClC. We expected that BPD patients show lower amygdala - prefrontal connectivity during emotion regulation compared to non-patients and ClC patients. Since the role of specific cortical and subcortical regions in emotion regulation has not been fully established (Diekhof et al., 2011; Buhle et al., 2014), we explored functional connectivity beyond amygdala - prefrontal pathways in whole-brain analyses.
2. Methods
2.1. Operationalization
The current study reanalyzed an existing dataset to investigate frontolimbic connectivity during emotion regulation in BPD patients, ClC patients, and non-patients (Van Zutphen et al., in prep; Wetzelaer et al., 2014). Participants performed an emotion regulation task (see Figure 1)
similar to paradigms that have been used in previous studies (Diekhof et al., 2011; van Zutphen et al., 2015). Situated in a 3T scanner, participants viewed a series of emotional pictures and were instructed to maintain their emotional reactions to the picture or to down-regulate them. The task consisted of four runs with 24 trials, respectively. The stimuli consisted of 24 negative, positive, erotic, and neutral pictures, respectively (Jacob, Arntz, Domes, Reiss, & Siep, 2011; Lang, Bradley, & Cuthbert, 1997). All pictures depicted social situations in the sense that multiple persons were interacting. Social stimuli were used because BPD symptoms are usually most prevalent in social situations (Arntz & ten Haaf, 2012; Crowell, Beauchaine, & Linehan, 2009; Koenigsberg, Fan, et al., 2009). Prior to each picture, participants were instructed to either passively look at the stimulus (emotion generation condition) or to actively down-regulate their
5
effective reappraisal strategy taught in Schema-Focused Therapy (Arntz & van Genderen, 2009). After each picture, participants indicated how they felt using a visual analogue scale ranging from negative (-100 mm) to positive (100 mm). After the emotion regulation task, participants rated
each picture regarding arousal and valence using the Self-Assessment Manikin Scale (SAMS; Bradley & Lang, 1994).
Figure 1. Overview of a trial during scanning. The total trial duration was 19 to 20.5s. Figure
from van Zutphen et al. (in prep).
2.2. Sample characteristics
Complete preprocessed functional imaging data were available from 54 BPD patients, 25 ClC patients, and 43 non-patient controls. One participant with BPD had to be excluded from all left amygdala connectivity analyses because it was not possible to extract PPI values for the contrast of interest (nBPD – left amygdala PPI = 53). All participants were female and met the respective inclusion
criteria. BPD and ClC patients were diagnosed using the Structural Clinical Interview (SCID) II (First, Spitzer, Gibbon, Williams, & Benjamin, 1997) and I (First, Spitzer, Gibbon, & Williams 1994), respectively. Only BPD patients who scored higher than 20 on the BPD Severity Index were included in the study. BPD patients that showed comorbidity with narcissistic or antisocial personality disorder, full or sub-threshold, were excluded. ClC patients who met criteria for full or sub-threshold cluster-b personality disorder or who met more than two BPD criteria were excluded. Non-patients who met diagnostic criteria for any Axis I or Axis II disorder using the SCID-I and SCID-II were excluded. The three groups did not differ regarding age, handedness, or IQ, but differed significantly regarding symptom severity (van Zutphen, in prep). Some of the patients were taking medication for clinical reasons (mainly antidepressants but also antipsychotics, hypnotics and mood stabilizers). However, previous analyses of this dataset showed no effect of medication on brain activity during emotion generation and regulation (Zutphen et al., in prep).
The initial significance threshold was p ≤ .005 in a cluster of at least 10 adjoining voxels (k ≥ 10). At an uncorrected p-value of .005, the study had 80% power to detect a large difference between BPD patients and non-patients (Cohen’s d ≥ .71), a large difference between BPD patients and ClC patients (Cohen’s d ≥ .84), and a large difference between ClC patients and non-patients (Cohen’s d ≥ .88), respectively.
2.3. Image Acquisition
Data were acquired on 3T scanners in Maastricht, Freiburg, and Lübeck. T1-weighted structural images were acquired (echo time = 2.6ms, repetition time = 2250ms, flip angle = 9°, field of view = 2562mm, voxel size = 13mm). T2*-weighted functional images were acquired via
echo planar imaging (EPI; echo time = 27 ms, repetition time = 2000 ms, flip angle = 90°, field of view = 1922mm, voxel size = 33mm).
+
Visual instruction 2 s Look -or- Safe‘How do you feel?’ NEG POS
8 s 4 s 5-6.5 s
Implement instruction
6
2.4. Preprocessing & statistical analyses
Data were preprocessed as described in Zutphen et al. (in prep). Briefly, the preprocessing steps of the functional images included slice time correction, motion correction for three translation and three rotation parameters, and high-pass filtering of two sines/cosines per run. Participants underwent two anatomical scans – one before the emotion regulation task and one before another task. These images were skull stripped and corrected for intensity inhomogeneities. If both scans were available for a subject, an average of the two anatomical scans was used for all analyses to achieve better resolution and contrast. The functional images per run were co-registered with the structural image. The data were smoothed using an isotropic Gaussian kernel of 6mm FWHM. Finally, the data were normalized to the Talairach and Tournoux space (1988). During image acquisition, there were differences across sites regarding the selection of the field of view. As a consequence, the overall coverage and group comparisons excluded some lower temporal areas as well as parts of the parietal cortex and motor areas. The preprocessed data were converted from brain voyager (Goebel, Esposito, & Formisano, 2006) to NIfTI for SPM12 (Wellcome Department of Cognitive Neurology, London; www.fil. ion.ucl.ac.uk/spm) and FSL (Smith et al., 2004). All subsequent steps of the data analysis were carried out in SPM12. FSLview was used for data visualization.
2.5. Activation analysis
Prior to the functional connectivity analysis, task dependent activation was investigated within and between groups using t-tests. First, a general linear model (GLM) was built for each subject to investigate the pattern of activation. The GLM included ten task regressors and six motion parameters: instruction, rating, negative/emotion-generation, positive/emotion-generation, erotic/emotion-generation, neutral/emotion-generation, negative/emotion-regulation, positive/emotion-regulation, erotic/emotion-regulation, neutral/emotion-regulation, and six motion parameters. Each predictor was convolved with a double gamma hemodynamic response function. Second, brain activation associated with emotion generation was investigated by comparing the negative-look with the neutral-look condition. Third, brain activation associated with emotion regulation was investigated by comparing the negative-safe with the negative-look condition. Each contrast was first tested separately within each group and then between groups (BPD vs NPC, BPD vs ClC, and ClC vs. NPC). Results are reported for each contrast corrected for multiple comparisons using a whole-brain cluster-wise false discovery rate (FDR), pFDR ≤ .05.
Uncorrected results and test statistics are reported in Appendix A.1-A.2, with an initial threshold
of p ≤ .005 and a minimum cluster size of ten voxels (k ≥ 10).
In order to further investigate differences in amygdala activity during emotion generation (negative-look vs. neutral-look) and during emotion regulation (negative-safe vs. negative-look), we extracted a mean value of the contrast estimates of all voxels within the amygdala (based on a 50% probability mask from the Harvard-Oxford subcortical probability atlas) to compare the mean activity between groups. In order to further investigate differences in prefrontal activity during emotion regulation (negative-safe vs negative-look), we extracted a mean value of the contrast estimates of all voxels within a meta-analytic mask of prefrontal emotion regulation regions from Diekhof et al. (2011). This mask includes the vmPFC, the dmPFC and the middle frontal gyrus. We tested the robustness of the effects based on the Diekhof mask by trying to replicate them using a meta-analytic mask of prefrontal emotion regulation regions from Neurosynth (key word: reappraisal, forward inference, search date: 13-07-2017; Yarkoni, Poldrack, Nichols, Van Essen, & Wager, 2011). Neurosynth allows automated large-scale meta-analyses based on specific terms (i.e. reappraisal). For the meta-analyses of mean contrast estimates, the significance threshold was set to p ≤ .05.
7
2.6. Functional connectivity analysis
We used a psychophysiological interaction (PPI) analysis to investigate task-dependent functional connectivity (Friston et al., 1997; Gitelman, Penny, Ashburner, & Friston, 2003; O’Reilly, Woolrich, Behrens, Smith, & Johansen-Berg, 2012). PPI analyses identify voxels in which activity is more related to activity in a seed region of interest (seed ROI) during a specific psychological context, i.e. emotion regulation, compared to another psychological context, i.e. emotion generation (O’Reilly et al., 2012). We investigated functional connectivity with the right and the left amygdala as separate seed regions. PPI is one of the most common methods to investigate task-dependent connectivity between two or more brain regions (McLaren, Ries, Xu, & Johnson, 2012). Specifically, a significant PPI shows that activity in a target region (e.g. the vmPFC) can be predicted not only by 1) a specific psychological context (e.g. emotion regulation vs. emotion generation) and 2) activity in another region (i.e. the physiological variable, e.g. the amygdala), but also by 3) the interaction of the psychological context (1) and the physiological variable (2). This means that the PPI explains variance in the signal in a target region over and above what is explained by the effect of the task (1) and the correlation with the seed region (2). In this study for example, a significant PPI indicates that two regions (e.g. the vmPFC and the amygdala) are stronger functionally connected during emotion regulation than during emotion generation. This emotion regulation-dependent connectivity was hypothesized to be impaired in BPD patients compared to non-patients.
For the PPI analysis, the GLM of the activation analysis was extended with the voxel-wise time-course of the amygdala (based on a 50% probability mask from the Harvard-Oxford subcortical probability atlas, provided in FSL) and the interaction of this time-course with the psychological predictors. Consequently, the design matrix included: 1) the psychological variables, i.e. the experimental manipulations (the ten predictors from the activation analysis); 2) the physiological variable, i.e. the mean time-course of activity in the amygdala; 3) the interaction of the psychological variables and the physiological variables, i.e. the generalized PPI (McLaren et al., 2012); 4) six motion parameters. A significant PPI effect in a cluster of voxels indicated that the respective cluster showed increased or decreased connectivity with the amygdala during emotion regulation compared to emotion generation. Since PPI effects only explain variance over and above what is explained by the task and the correlation with the seed-region, they tend to lack statistical power, and false negative findings could be expected (O’Reilly et al., 2012).
2.6.1. Amygdala - whole-brain connectivity
In a set of confirmatory analyses, we first tested PPIs within groups, i.e. whether amygdala - whole-brain connectivity is higher during emotion regulation than during emotion generation. Subsequently, we tested whether the PPI effects differed among groups. Specifically, we tested whether amygdala - whole-brain connectivity during emotion regulation is lower in BPD patients than in healthy subjects and whether this deficit is specific for BPD. We restricted the confirmatory PPI analyses to a comparison of emotion regulation and emotion generation during negative trials to facilitate interpretability (neutral, positive, and erotic stimuli were excluded from the contrast). Results are reported for each contrast corrected for multiple comparisons, pFDR ≤
.05. Uncorrected results and test statistics are reported in Appendix A.3-A.4, with an initial
threshold of p ≤ .005 and a minimum cluster size of ten voxels (k ≥ 10). 2.6.2. Amygdala – prefrontal connectivity
Based on the importance of prefrontal regions and statistical power considerations, we conducted another PPI analysis with a prefrontal emotion regulation target ROI. In this analysis, FDR multiple comparison correction was restricted to the target ROI. The ROI was the Diekhof mask (2011) that was also used for the activation analysis. In order to overcome the problem of low power of PPIs, we extracted the mean amygdala-connectivity of every voxel within the emotion regulation mask from Diekhof et al. (2011) and compared the resulting means between
8
groups using t-tests. In order to test for the robustness of these effects we tried to replicate them by extracting a mean value of connectivity of every voxel within the emotion regulation mask created with Neurosynth (Yarkoni, et al., 2011).
3. Results
3.1. Manipulation check
According to Zutphen et al. (in prep), all groups reported to feel more positive during the safe-condition than during the look-condition, which indicates that emotion regulation was at least partially successful. Further, BPD patients reported lower levels of well-being in negative and neutral trials than non-patients, which indicates that BPD patients experienced aberrant emotion processing. BPD patients did not differ from ClC patients, and ClC patients did not differ from non-patients.
3.2. Activation analysis
The uncorrected clusters that resulted from our activation analyses (Appendix A1 and A2)
were almost identical to the results from Zutphen et al (in prep), showing that the implementation of the GLM in brain voyager and SPM are comparable. However, we used different thresholds (puncorrected ≤ .005, cluster size ≥ 10) and multiple comparison corrections
(FDR) than Zutphen et al. (p ≤ .01, cluster size emotional sensitivity ≥ 15, cluster size emotion regulation ≥ 22, based on Monte Carlo simulation-based cluster-level correction). Consequently, our results after multiple comparison correction differed slightly.
3.2.1. Emotional Sensitivity
Brain activation associated with emotion generation, i.e. emotional sensitivity, was investigated by comparing the negative-look with the neutral-look condition. First, we investigated brain activity during emotion generation within groups (see Table 1). BPD patients showed enhanced
activation in the bilateral occipital cortex, in the right supramarginal gyrus, and in the right superior frontal gyrus and frontal pole. ClC patients showed enhanced activation in the bilateral inferior occipital gyrus and middle temporal gyrus, in the bilateral supramarginal and postcentral gyrus, and in the middle frontal gyrus. Further, ClC patients showed a tendency towards enhanced activity in the anterior cingulate gyrus, even though this difference did not survive multiple comparison correction. Non-patients showed enhanced activity in the bilateral inferior occipital cortex, in the right precentral gyrus, in the postcentral gyrus, the supramarginal gyrus and the parietal operculum cortex, in the anterior cingulate gyrus, and in a cluster stretching from the right thalamus to the right inferior frontal gyrus.
Second, we compared brain activity during emotion generation between groups. A whole-brain two (negative, neutral) by three (BPD, ClC, non-patients) F-test resulted in a significant the left occipital cortex and middle temporal gyrus (Fpeak = 81.67; pFDR < .001). However, t-tests on
the whole brain did not reveal any differences between groups (see Appendix A.1.2. for
uncorrected results).
Finally, a mean value of activation of the bilateral amygdala was extracted per group based on the Harvard-Oxford subcortical probability atlas. Figure 1A shows that the mean activation of
the amygdala did not differ between BPD patients and non-patients (t(95) = 1.14, p = .257), nor between ClC patients and non-patients (t(66) = 1.50, p = .137).
9
Cluster-level Peak level Coordinates Brain region pFDR k puncorr T puncorr x y z
BPD <.001 106 <.001 8.95 <.001 -48 -64 1 lateral occipital cortex/ middle temporal gyrus <.001 607 <.001 6.38 <.001 63 -22 31 supramarginal gyrus <.001 610 <.001 4.48 <.001 6 50 43 superior frontal gyrus/
frontal pole
ClC <.001 310 <.001 9.05 <.001 -48 -61 -2 lateral occipital cortex/ middle temporal gyrus .005 265 .001 5.81 <.001 -63 -28 28 supramarginal gyrus/
postcentral gyrus <.001 437 <.001 5.63 <.001 60 -19 31 supramarginal gyrus/
postcentral gyrus <.001 469 <.001 5.06 <.001 48 -4 40 precentral gyrus
.071 114 .015 4.39 <.001 0 17 28 anterior cingulate gyrus
NPC <.001 369 <.001 9.73 <.001 -48 -70 -2 lateral occipital cortex/ middle temporal gyrus <.001 198 <.001 5.96 <.001 3 -25 1 right thalamus to inferior
frontal gyrus .002 297 <.001 5.38 <.001 24 -10 49 precentral gyrus .002 290 <.001 5.16 <.001 -57 -4 37 precentral gyrus
.002 327 <.001 5.01 <.001 0 -4 37 anterior cingulate gyrus .066 101 .020 3.85 <.001 -36 -13 43 Precentral gyrus
Table 1. Emotional generation – within groups. Negative-look vs. neutral-look contrast. Only FDR
corrected p-values ≤ .10 are presented. Tables with all uncorrected clusters are presented in Appendix A1.1.
3.2.2. Emotion regulation
Brain activation associated with emotion regulation was investigated by comparing the negative-safe with the negative-look condition. First, we investigated brain activity during emotion regulation within groups. Using a whole-brain analysis, none of the groups differed in brain activation between the negative-safe and the negative-look condition (see Appendix A.2.1.
for uncorrected results).
Second, we investigated differences in brain activity during emotion regulation between groups. A whole-brain two (negative, neutral) by three (BPD, ClC, non-patients) F-test resulted in no significant cluster. However, t-tests between groups (see Table 2) revealed that BPD
patients showed less activation than non-patients in the superior temporal gyrus and the middle temporal gyrus, even though this trend only approached significance (pFDR = .098). Further, BPD
patients showed less activation than ClC patients in the right frontal pole and the right inferior frontal gyrus. ClC patients did not differ from non-patients.
Finally, the mean activation of prefrontal regions involved in emotion regulation was extracted based on the meta-analytic masks from Diekhof et al. (2011) and Neurosynth (Yarkoni, et al., 2011). Figure 1B illustrates that BPD patients showed less activation in prefrontal regions than
non-patients (Diekhof-mask: t(95) = -2.55, p = .012; Neurosynth-mask: t(95) = -2.17, p = .032). On the contrary, ClC patients did not show different activation compared to non-patients (Diekhof-mask: t(66) = -0.39, p = .696; Neurosynth-mask: t(66) = -0.30, p = .767). There were no differences in mean amygdala activation between BPD patients and nonpatients (t(95) = -0.62, p = .537), nor between ClC patients and non-patients(t(66) = 0.2871, p = .775).
10
pFDR k puncorr T puncorr x y z
NPC vs BPD
.098 215 .002 4.84 <.001 66 -13 -2 posterior superior and middle temporal gyrus
ClC vs BPD
<.001 809 <.001 5.46 <.001 48 38 1 frontal pole/
inferior frontal gyrus Table 2. Emotional regulation – between groups. Negative-safe vs. negative-look contrast. Only FDR
corrected p-values ≤ .10 are presented. Tables with all uncorrected clusters are presented in Appendix A2.2.
Figure 1. A) Amygdala activation during emotion generation based on a 50% probability mask from the
Harvard-Oxford subcortical probability atlas. B) Prefrontal activation during emotion regulation based on
meta-analytic mask from Diekhof et al (2011) and Neurosynth (Yarkoni, et al., 2011).
3.3. Functional connectivity analysis - PPI
Emotion regulation dependent connectivity of prefrontal regions with the amygdala was investigated by comparing the PPIs between the negative-safe and the negative-look condition. We investigated functional connectivity with the right and the left amygdala separately. First, we investigated amygdala connectivity of the whole brain within groups. However, none of the groups showed a significant difference in right amygdala or in left amygdala connectivity between the negative-safe and the negative-look condition (see Appendix A.3.1. and A.4.1. for
uncorrected results of right and left amygdala, respectively).
Second, we compared amygdala connectivity between groups. Investigating right amygdala connectivity, a whole-brain two (negative, neutral) by three (BPD, ClC, non-patients) F-test resulted in no significant cluster. Direct comparisons of right amygdala connectivity between groups revealed no difference between BPD patients and non-patients. However, ClC patients showed reduced right amygdala connectivity compared to BPD patients and non-patients in the parietal operculum cortex (see Table 3 and Figure 2). Furthermore, ClC patients showed
reduced right amygdala connectivity compared to non-patients in the right insular cortex. However, the differences between ClC patients and non-patients only approached significance (pFDR = .073). Investigating left amygdala connectivity, a whole-brain two (negative, neutral) by
three (BPD, ClC, non-patients) F-test resulted in no significant cluster. Direct comparisons between groups yielded no significant results (see Appendix A.4.2. for uncorrected results).
11
A) BPD patients vs. ClC patients
B) ClC patients vs. non-patients
Figure 2. Functional connectivity with the right amygdala in a whole-brain analysis. A) BPD patients vs.
ClC patients (x = 45, z = 16, respectively). B) ClC patients vs non-patients (x = 42, z = 4, respectively).
The maps show right amygdala connectivity with T ≥ 2.6178 (puncorrected = .005). Increased connectivity is
12
Cluster-level Peak level Coordinates Brain region pFDR k puncorr T puncorr x y z
BPD vs ClC
.001 172 <.001 5.15 <.001 45 -13 16 central opercular and parietal operculum cortex
NPC vs ClC .073 85 .001 4.53 <.001 42 -7 4 insular cortex .073 72 .003 3.96 <.001 -36 -43 19 parietal operculum cortex/ supramarginal gyrus Table 3. Right amygdala connectivity during emotion regulation – between groups. Negative-safe vs.
negative-look contrast. Only FDR corrected p-values ≤ .10 are presented. Tables with all uncorrected clusters are presented in Appendix A.3.1.
After investigating amygdala connectivity in a whole-brain analysis we investigated emotion regulation dependent connectivity of prefrontal regions applying a meta-analytic mask based on Diekhof et al. (2011). Neither within group comparisons, nor between group comparisons yielded significant results in the prefrontal ROI in left or right amygdala connectivity, respectively.
In order to further investigate whether patients showed aberrant amygdala-prefrontal connectivity during emotion regulation, a mean value of amygdala-connectivity within prefrontal ROIs was extracted based on the meta-analytic masks from Diekhof et al. (2011) and Neurosynth (Yarkoni, et al., 2011). Figure 3A illustrates left amygdala - prefrontal connectivity. BPD patients
did not differ from non-patients (Diekhof-mask: t(49) = 1.21, p = .229; Neurosynth-mask: t(94) = 0.49, p = .622). Similarly, ClC patients did not differ from non-patients (Diekhof-mask: t(66) = 0.1923, p = .0.848; Neurosynth-mask: t(66) = 0.01, p = .0.993). Figure 3B illustrates right
amygdala - prefrontal connectivity during emotion regulation. BPD patients did not differ from non-patients (Diekhof-mask: t(95) = .465, p = .643; Neurosynth-mask: t(96) = -0.06, p = .953). Using the Diekhof mask, ClC patients did not differ from non-patients (t(66) = -1.35, p = .180). However, using the Neurosynth mask, ClC patients showed lower right amygdala-prefrontal connectivity than non-patients (t(66) = -2.05, p = .045).
A) Left amygdala B) Right amygdala
Figure 3. A) Left amygdala - prefrontal connectivity during emotion regulation. B) Right amygdala -
prefrontal connectivity during emotion regulation. Prefrontal regions were selected based on a meta-analytic mask from Diekhof et al (2011) and Neurosynth (Yarkoni, et al., 2011).
13
4. Discussion
Borderline Personality Disorder (BPD) patients often fail to down-regulate their emotions and suffer from extreme emotional instability (van Zutphen et al., 2015). Emotion generation involves limbic regions, whereas emotion regulation recruits prefrontal regions (Ochsner et al., 2012). The current study investigated whether emotion dysregulation in BPD is associated with aberrant frontolimbic activity and connectivity by comparing BPD patients with ClC patients and non-patients. We found that BPD patients failed to recruit prefrontal brain regions to voluntarily down-regulate their emotions. ClC patients recruited prefrontal regions during emotion regulation but showed decreased frontolimbic connectivity.
4.1. Emotional sensitivity
Limbic regions were not more active in BPD or ClC patients than in NPC patients when looking at negative compared to neutral pictures. Therefore, it seems that disturbed emotion processing in these disorders does not result from enhanced emotional sensitivity. It is possible that patients and non-patients differ in their unregulated reaction to emotional stimuli, not because the limbic reactivity is enhanced but because stimuli are evaluated and appraised differently (Arntz & ten Haaf, 2012; Choi-Kain & Gunderson, 2008). For example, BPD patients tend to differ from non-patients in their ability to mentalize when they evaluate social situations (Choi-Kain & Gunderson, 2008; Fonagy & Bateman, 2008; King-Casas et al., 2008). If a negative stimulus depicts a social scene, BPD patients might have a similar limbic response as non-patients. However, they might appraise the situation differently – possibly reflected in a unique pattern of prefrontal activation – and draw incorrect and excessively negative conclusions (Arntz & ten Haaf, 2012; Choi-Kain & Gunderson, 2008). However, we found no conclusive evidence regarding different brain activation during emotion generation between BPD, ClC, and non-patients. This finding is broadly in line with previous research that could not find consistent patterns of hyperactivity in BPD patients while viewing emotional stimuli (van Zutphen et al., 2015).
4.2. Emotion regulation
4.2.1. Activity
If emotional sensitivity in terms of limbic hyper-reactivity to emotional stimuli cannot explain emotion dysregulation in BPD, it is possible that aberrant emotion regulation might play a more important role. Unexpectedly, there were no differences in brain activation between the negative-safe and the negative-look condition within any group. However, between-group comparisons revealed that BPD patients showed lower brain activity than non-patients and ClC patients in multiple prefrontal brain regions when attempting to regulate their emotions. Specifically, BPD patients showed lower activation than non-patients in the posterior temporal gyrus and the middle temporal gyrus. Further, BPD patients showed lower activation than ClC patients in the right frontal pole and the right inferior frontal gyrus. ClC patients did not differ from non-patients in brain activity during emotion regulation. These results were confirmed when comparing mean activation in prefrontal emotion regulation regions between groups based on meta-analytic masks. Consequently, BPD patients failed to voluntarily recruit prefrontal regions to down-regulate limbic activity on demand. This deficit was specific for BPD patients. These results indicate that, on a neural basis, failed reappraisal of negative emotions might play a more important role in BPD than enhanced emotional sensitivity.
Alternatively, BPD patients might have automatically tried to down-regulate their emotions in both, the emotion generation and the emotion regulation condition, while ClC and non-patients were able to experience the emotions in the look condition without automatic regulation. As a consequence, ClC and non-patients would have shown enhanced prefrontal activity in the emotion regulation condition compared to BPD patients not because BPD patients failed to
14
recruit prefrontal regions in the emotion regulation condition but because they already recruited them in the emotion generation condition.
Our conclusion that BPD patients failed to recruit prefrontal regions during emotion regulation are broadly in line with previous research (van Zutphen et al., 2015). However, findings across studies are inconsistent regarding specific regions that showed decreased activation compared to healthy subjects (van Zutphen et al., 2015). One explanation for these inconsistencies might be that cognitive reappraisal is a category rather than a specific strategy. The term cognitive reappraisal refers to multiple regulation strategies that are similar but not necessarily identical. For example, while emotional distancing or realizing to be safe both reflect reappraisal strategies that require higher cognitive functions, they might nevertheless be qualitatively different and rely on distinct brain structures (Opialla et al., 2014; Yokum & Stice, 2013).
4.2.2. Connectivity
We investigated right and left amygdala connectivity using PPIs. Contrary to our expectations, the right amygdala PPI revealed that BPD patients did not differ from non-patients regarding frontolimbic connectivity during emotion regulation. However, ClC patients showed less right amygdala connectivity with the right central and parietal operculum cortex and the right insular cortex than BPD patients and non-patients, respectively. These regions are located in immediate proximity and most likely reflect one cluster, probably in the insular cortex.
The brain regions that showed lower right amygdala connectivity in ClC patients were not located within prefrontal emotion regulation regions according to meta-analytic results (Diekhof et al., 2011). Again, one reason might be that this study used a slightly different reappraisal strategy than previous research (i.e. realizing to be safe instead of distancing or reinterpreting). However, when comparing average right amygdala connectivity of prefrontal emotion regulation regions based on a meta-analytic mask from Neurosynth (Yarkoni, et al., 2011), we found a significant difference in frontolimbic connectivity between ClC patients and non-patients. ClC but, not BPD patients, showed lower right amygdala connectivity in prefrontal emotion regulation regions compared to non-patients. In conclusion, both a whole-brain analysis and an analysis of mean frontolimbic connectivity showed a pattern of reduced right amygdala connectivity in ClC patients during emotion regulation.
The results of the left amygdala PPI were not as conclusive as those from the right amygdala. Since our results from the left amygdala and from the right amygdala PPI differed, future studies using PPIs should distinguish between the hemispheres when choosing a seed region. The statistical power of PPIs is lower than of classical brain activation contrasts (O’Reilly et al., 2012). Therefore, it might be more important to account for subtle differences between the hemispheres when investigating task-dependent connectivity, than when investigating brain activation. Even though results from the right amygdala connectivity were not replicated with the left amygdala, they were relatively consistent across different comparisons using right amygdala PPI. Consequently, we conclude that among ClC patients, various brain regions, including the anterior insula and prefrontal regions, fail to communicate with the right amygdala to down-regulate emotions.
4.3. Limitations and future directions
This study was one of the first to investigate functional connectivity during an emotion regulation task (Cullen et al., 2015; Paret et al., 2016; van Zutphen et al., 2015; Xu et al., 2016). Further, this study included a ClC patient and a non-patient control group to investigate the specificity of biological markers of pathological emotion dysregulation. Nevertheless, some limitations should be taken into consideration when interpreting the results. First, the sample included only women, which reduces the generalizability to male or mixed populations, especially because previous research has shown differences between male and female patients in emotion
15
regulation and psychopathology (Johnson et al., 2003; Nolen-Hoeksema, 2012). However, the exclusion of male participants reduced undesired variance in the BOLD signal and therefore is an appropriate first step to investigate task-dependent functional connectivity with PPI. Future studies should address gender differences and the generalizability of the results to other populations.
Second, PPI analyses lack statistical power which inflates the likelihood of false negative findings (O’Reilly et al., 2012). Since PPI analyses explain variance over and above what is explained by the main effect of the task and of the correlation with the seed regions, the effect has to be large on a within subject level. Furthermore, our sample was only able to detect large effects across subjects. Consequently, it is possible that the current study did not detect some meaningful patterns of functional frontolimbic connectivity due to a lack of power. One way to increase statistical power is to extract mean values of connectivity from anatomically or meta-analytically defined brain regions. This method circumvents the problem of multiple comparison correction in PPI analyses. In line with this solution, we extracted a mean value of frontolimbic connectivity based on meta-analytic results on emotion regulation. This analysis showed significantly lower right amygdala connectivity among ClC patients compared to non-patients. However, our results differed between the right and the left amygdala connectivity analysis. Future studies need to address whether the right and the left amygdala truly differ in connectivity patterns or whether the current study lacked statistical power to detect the same patterns of connectivity for both hemispheres.
Third, the coverage of the preprocessed images excluded some lower temporal areas as well as parts of the parietal cortex and motor areas due to inconsistencies during image acquisition across sites. While future studies should include these brain regions to investigate differences across the entire brain, previous studies found that other limbic and prefrontal regions were more important for emotion processing. Consequently, the brain coverage in this study was acceptable to investigate the most relevant regions involved in emotion generation and regulation.
Fourth, emotion generation and emotion regulation cannot always be distinguished as clearly as the model of cognitive control of emotions suggests (Gross & Barrett, 2011; Gross, Sheppes, & Urry, 2011) and it is difficult to ensure whether and how participants implement the safe and look instructions. Participants might attempt to regulate their emotions even when they are instructed not to. However, there is a multitude of neuroimaging studies that investigated emotion regulation and emotion generation with similar manipulations both in patient and non-patient samples (Buhle et al., 2014; van Zutphen et al., 2015; Zilverstand et al., 2017). These studies reliably detected differences in brain activation between emotion regulation and generation instructions. Complementary studies showed that physiological and behavioural emotional responses decreased in participants who were instructed to reappraise, which indicates that these instructions indeed lead to emotion regulation (Gross, 1998). Consequently, emotion generation vs. regulation manipulations as implemented in this study are useful to investigate aberrant emotional sensitivity and emotion regulation.
At this stage, it is not possible to draw strong conclusions about how impaired functional connectivity relates to behavioral and subjective differences in emotion dysregulation or psychopathology. Future research should investigate behavioral and self-reported emotionality as correlates of aberrant functional connectivity to identify how it affects emotion regulation. Most importantly, it will be necessary to investigate whether aberrant frontolimbic connectivity is associated with different problems in emotion regulation than aberrant frontolimbic activity.
5. Conclusion
The current study investigated neural markers of emotion dysregulation in BPD and ClC patients compared to non-patients. Our results indicated that patients did not show enhanced emotional sensitivity to negative stimuli, reflected in limbic hyper-reactivity. If anything, BPD patients, ClC patients, and non-patients showed unique patterns of activation in multiple regions
16
other than limbic regions such as the left occipital gyrus and the middle temporal gyrus. However, results on aberrant emotional sensitivity in BPD or ClC were not conclusive and should be interpreted with caution. Patients differed more profoundly from non-patients in emotion regulation. Therefore, aberrant emotion regulation seems to play a more important role in BPD and CIC than emotional sensitivity. BPD patients and ClC patients showed specific patterns of aberrant neural activity or connectivity during emotion regulation. On the one hand, BPD patients showed normal frontolimbic connectivity, but lower activity in prefrontal regions compared to ClC patients and non-patients. On the other hand, ClC patients showed normal activity in prefrontal regions, but reduced right amygdala connectivity with the insular cortex and prefrontal emotion regulation regions compared to BPD patients and non-patients.
In sum, emotion dysregulation in BPD patients seemed to result from a failure to recruit prefrontal regions to down-regulate limbic activity, whereas emotion dysregulation in ClC seemed to result from failed communication between prefrontal and limbic regions. These results suggest that future research needs to investigate neural markers of psychopathology beyond mere activation of brain regions and compare patients not only to non-patients but also to other clinical groups.
17
6. References
Aldao, A., & Nolen-Hoeksema, S. (2010). Specificity of cognitive emotion regulation strategies: A transdiagnostic examination. Behaviour Research and Therapy, 48(10), 974–983.
http://doi.org/10.1016/j.brat.2010.06.002
American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders. (5.
Edition, Ed.). Arlington VA: American Psychiatric Association.
Arntz, A., & ten Haaf, J. (2012). Social cognition in borderline personality disorder: Evidence for dichotomous thinking but no evidence for less complex attributions. Behaviour Research and Therapy, 50(11), 707–718. http://doi.org/10.1016/j.brat.2012.07.002
Arntz, A., & van Genderen, H. (2009). Schema therapy for borderline personality disorder. Chichester:
Wiley-Blackwell. Retrieved from http://www.azmonyar.com/DownloadPDF/20987818.pdf Baczkowski, B. M., van Zutphen, L., Siep, N., Jacob, G. A., Domes, G., Maier, S., … van de Ven,
V. (2016). Deficient amygdala–prefrontal intrinsic connectivity after effortful emotion regulation in borderline personality disorder. European Archives of Psychiatry and Clinical Neuroscience, 1–15. http://doi.org/10.1007/s00406-016-0760-z
Banks, S. J., Eddy, K. T., Angstadt, M., Nathan, P. J., & Luan Phan, K. (2007). Amygdala-frontal connectivity during emotion regulation. Social Cognitive and Affective Neuroscience, 2(4), 303–312.
http://doi.org/10.1093/scan/nsm029
Bradley, M., & Lang, P. J. (1994). Measuring Emotion: The Self-Assessment Semantic
Differential Manikin and the. Journal of Behavior Therapy and Experimental Psychiatry, 25(I), 49–59.
http://doi.org/10.1016/0005-7916(94)90063-9
Buhle, J. T., Silvers, J. A., Wage, T. D., Lopez, R., Onyemekwu, C., Kober, H., … Ochsner, K. N. (2014). Cognitive reappraisal of emotion: A meta-analysis of human neuroimaging studies.
Cerebral Cortex, 24(11), 2981–2990. http://doi.org/10.1093/cercor/bht154
Chen, C., Suckling, J., Ooi, C., Fu, C., Williams, S., Walsh, N., … Bullmore, E. (2008). Functional coupling of the amygdala in depressed patients treated with antidepressant medication.
Neuropsychopharmacology, 33(8), 1909–1918. http://doi.org/10.1038/sj.npp.1301593
Choi-Kain, L. W., & Gunderson, J. G. (2008). Mentalization: ontogeny, assessment, and application in the treatment of borderline personality disorder. Am J Psychiatry, 165(9), 1127–
1135. http://doi.org/10.1176/appi.ajp.2008.07081360
Cremers, H. R., Veer, I. M., Spinhoven, P., Rombouts, S. A. R. B., Yarkoni, T., Wager, T. D., & Roelofs, K. (2015). Altered cortical-amygdala coupling in social anxiety disorder during the anticipation of giving a public speech. Psychological Medicine, 45(7), 1521–1529.
http://doi.org/10.1017/S0033291714002657
Crowell, S. E., Beauchaine, T. P., & Linehan, M. M. (2009). A biosocial developmental model of borderline personality: Elaborating and extending Linehan’s theory. Psychological Bulletin, 135(3),
495–510. http://doi.org/10.1037/a0015616
Cullen, K. R., LaRiviere, L. L., Vizueta, N., Thomas, K. M., Hunt, R. H., Miller, M. J., … Schulz, S. C. (2015). Brain activation in response to overt and covert fear and happy faces in women with borderline personality disorder. Brain Imaging and Behavior, 319–331.
http://doi.org/10.1007/s11682-015-9406-4
Cullen, K. R., Vizueta, N., Thomas, K. M., Han, G. J., Lim, K. O., Camchong, J., … Schulz, S. C. (2011). Amygdala Functional Connectivity in Young Women with Borderline Personality Disorder. Brain Connectivity, 1(1), 61–71. http://doi.org/10.1089/brain.2010.0001
Di, X., Huang, J., & Biswal, B. B. (2017). Task modulated brain connectivity of the amygdala: a meta-analysis of psychophysiological interactions. Brain Structure and Function, 222(1), 619–634.
18
Diekhof, E. K., Geier, K., Falkai, P., & Gruber, O. (2011). Fear is only as deep as the mind allows. A coordinate-based meta-analysis of neuroimaging studies on the regulation of
negative affect. NeuroImage, 58(1), 275–285. http://doi.org/10.1016/j.neuroimage.2011.05.073
Dudas, R. B., Mole, T. B., Morris, L. S., Denman, C., Hill, E., Szalma, B., … Voon, V. (2016). Amygdala and dlPFC abnormalities, with aberrant connectivity and habituation in response to emotional stimuli in females with BPD. Journal of Affective Disorders, 208(June 2016), 460–466.
http://doi.org/10.1016/j.jad.2016.10.043
Etkin, A., Egner, T., & Kalisch, R. (2011). Emotional processing in anterior cingulate and medial prefrontal cortex. Trends in Cognitive Sciences, 15(2), 85–93.
http://doi.org/10.1016/j.tics.2010.11.004
Etkin, A., & Wager, T. D. (2007). Functional neuroimaging of anxiety: a meta-analysis of
emotional processing in PTSD, social anxiety disorder, and specific phobia. American Journal of Psychiatry, 164(10), 1476–1488. http://doi.org/10.1176/appi.ajp.2007.07030504
First M. B., Spitzer R. L., Gibbon M., Williams J. B. W. (1994). Structured clinical interview for DSM- IV axis I disorders (SCID-I). New York: Biometric Research Department.
First M. B., Spitzer R. L., Gibbon M., Williams J. B. W., Benjamin L. (1997). Structured clinical interview for DSM-IV axis II personality disorders (SCID-II). New York: Biometric Research
Department.
Fonagy, P., & Bateman, A. (2008). The development of borderline personality disorder - a mentalizing model. Journal of Personality Disorders, 22(1), 4–21.
http://doi.org/10.1521/pedi.2008.22.1.4
Friston, K. J., Buechel, C., Fink, G. R., Morris, J., Rolls, E., & Dolan, R. J. (1997).
Psychophysiological and modulatory interactions in neuroimaging. Neuroimage, 6(3), 218–229.
http://doi.org/10.1006/nimg.1997.0291
Gitelman, D. R., Penny, W. D., Ashburner, J., & Friston, K. J. (2003). Modeling regional and psychophysiologic interactions in fMRI: The importance of hemodynamic deconvolution.
NeuroImage, 19(1), 200–207. http://doi.org/10.1016/S1053-8119(03)00058-2
Goebel, R., Esposito, F., & Formisano, E. (2006). Analysis of Functional Image Analysis Contest (FIAC) data with BrainVoyager QX: From single-subject to cortically aligned group General Linear Model analysis and self-organizing group Independent Component Analysis. Human Brain Mapping, 27(5), 392–401. http://doi.org/10.1002/hbm.20249
Goodman, M., Carpenter, D., Tang, C. Y., Goldstein, K. E., Avedon, J., Fernandez, N., … Hazlett, E. A. (2014). Dialectical behavior therapy alters emotion regulation and amygdala activity in patients with borderline personality disorder. Journal of Psychiatric Research, 57(1),
108–116. http://doi.org/10.1016/j.jpsychires.2014.06.020
Gratz, K. L., & Roemer, L. (2008). The relationship between emotion dysregulation and
deliberate self-harm among female undergraduate students at an urban commuter university.
Cognitive Behaviour Therapy, 37(1), 14–25. http://doi.org/10.1080/16506070701819524
Gross, J. J. (1998). Antecedent- and Response-Focused Emotion Regulation: Divergent Consequences for Experience, Expression, and Physiology. Journal of Personality and Social Psychology, 74(1), 224–237. http://doi.org/10.1037/0022-3514.74.1.224
Gross, J. J., & Barrett, L. F. (2011). Emotion Generation and Emotion Regulation: One or Two Depends on Your Point of View. Emotion Review, 3(1), 8–16.
http://doi.org/10.1177/1754073910380974.Emotion
Gross, J. J., Sheppes, G., & Urry, H. L. (2011). Cognition and Emotion Lecture at the 2010 SPSP Emotion Preconference. Cognition & Emotion, 25(5), 765–781.
http://doi.org/10.1080/02699931.2011.555753
Gross, J. J., Sheppes, G., & Urry, H. L. (2011). Cognition and Emotion Lecture at the 2010 SPSP Emotion Preconference. Emotion generation and emotion regulation: A distinction we should make (carefully). Cognition & Emotion, 25(5), 765–781.
19
Gross, J. J., & Thompson, R. A. (2007). Emotion Regulation: Conceptual Foundations. In
Handbook of Emotion Regulation (pp. 3–24). New York: Guilford Press. Retrieved from
http://med.stanford.edu/nbc/articles/4 - Emotion Regulation - Conceptual Foundations.pdf Heatherton, T. F., & Wagner, D. D. (2011). Cognitive neuroscience of self-regulation failure.
Trends in Cognitive Sciences, 15(3), 132–139. http://doi.org/10.1016/j.tics.2010.12.005
Jacob, G. A., Arntz, A., Domes, G., Reiss, N., & Siep, N. (2011). Positive erotic picture stimuli for emotion research in heterosexual females. Psychiatry Research, 190(2–3), 348–351.
http://doi.org/10.1016/j.psychres.2011.05.044
Johnson, D. M., Shea, M. T., Yen, S., Battle, C. L., Zlotnick, C., Sanislow, C. A., … Zanarini, M. C. (2003). Gender differences in borderline personality disorder: Findings from the
collaborative longitudinal personality disorders study. Comprehensive Psychiatry, 44(4), 284–292.
http://doi.org/10.1016/s0010-440x(03)0090-7
Johnstone, T., van Reekum, C. M., Urry, H. L., Kalin, N. H., & Davidson, R. J. (2007). Failure to regulate: counterproductive recruitment of top-down prefrontal-subcortical circuitry in major depression. The Journal of Neuroscience : The Official Journal of the Society for Neuroscience, 27(33),
8877–8884. http://doi.org/10.1523/JNEUROSCI.2063-07.2007
Kanske, P., Heissler, J., Schönfelder, S., Bongers, A., & Wessa, M. (2011). How to regulate emotion? Neural networks for reappraisal and distraction. Cerebral Cortex, 21(6), 1379–1388.
http://doi.org/10.1093/cercor/bhq216
King-Casas, B., Sharp, C., Lomax-Bream, L., Lohrenz, T., Fonagy, P., & Montague, P. R. (2008). The Rupture and Repair of Cooperation in Borderline Personality Disorder. Science, 321(5890),
806–810. http://doi.org/10.1016/j.pestbp.2011.02.012.Investigations
Koenigsberg, H. W., Fan, J., Ochsner, K. N., Liu, X., Guise, K. G., Pizzarello, S., … Siever, L. J. (2009). Neural Correlates of the Use of Psychological Distancing to Regulate Responses to Negative Social Cues: A Study of Patients with Borderline Personality Disorder. Biological Psychiatry, 66(9), 854–863. http://doi.org/10.1016/j.biopsych.2009.06.010
Koenigsberg, H. W., Siever, L. J., Lee, H., Pizzarello, S., New, A. S., Goodman, M., … Prohovnik, I. (2009). Neural correlates of emotion processing in borderline personality disorder. Psychiatry Research - Neuroimaging, 172(3), 192–199.
http://doi.org/10.1016/j.pscychresns.2008.07.010
Krause-Utz, A., Veer, I. M., Rombouts, S. A. R. B., Bohus, M., Schmahl, C., & Elzinga, B. M. (2014). Amygdala and anterior cingulate resting-state functional connectivity in borderline personality disorder patients with a history of interpersonal trauma. Psychological Medicine, 44(13), 2889–901. http://doi.org/10.1017/S0033291714000324
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1997). International Affective Picture System (IAPS): Technical Manual and Affective Ratings. NIMH Center for the Study of Emotion and Attention, 39–58. http://doi.org/10.1027/0269-8803/a000147
Leichsenring, F., Leibing, E., Kruse, J., New, A. S., & Leweke, F. (2013). Borderline personality disorder. Lancet, 377, 74–84. http://doi.org/10.1016/S0140-6736(10)61422-5
McLaren, D. G., Ries, M. L., Xu, G., & Johnson, S. C. (2012). A generalized form of context-dependent psychophysiological interactions (gPPI): A comparison to standard approaches.
NeuroImage, 61(4), 1277–1286. http://doi.org/10.1016/j.neuroimage.2012.03.068
Milad, M. R., & Rauch, S. L. (2012). Obsessive-compulsive disorder: Beyond segregated cortico-striatal pathways. Trends in Cognitive Sciences, 16(1), 43–51.
http://doi.org/10.1016/j.tics.2011.11.003
Morawetz, C., Bode, S., Baudewig, J., & Heekeren, H. R. (2016). Effective amygdala-prefrontal connectivity predicts individual differences in successful emotion regulation. Social Cognitive and Affective Neuroscience, 569–585. http://doi.org/10.1093/scan/nsw169
Morawetz, C., Bode, S., Derntl, B., & Heekeren, H. R. (2017). The effect of strategies, goals and stimulus material on the neural mechanisms of emotion regulation: A meta-analysis of fMRI
20
studies. Neuroscience and Biobehavioral Reviews, 72, 111–128.
http://doi.org/10.1016/j.neubiorev.2016.11.014
Motzkin, J. C., Philippi, C. L., Wolf, R. C., Baskaya, M. K., & Koenigs, M. (2015). Ventromedial prefrontal cortex is critical for the regulation of amygdala activity in humans. Biological
Psychiatry, 77(3), 276–284. http://doi.org/10.3816/CLM.2009.n.003.Novel
New, A. S., Hazlett, E. A., Buchsbaum, M. S., Goodman, M., Mitelman, S. A., Newmark, R., … Siever, L. J. (2007). Amygdala-prefrontal disconnection in borderline personality disorder.
Neuropsychopharmacology, 32(7), 1629–1640. http://doi.org/10.1038/sj.npp.1301283
Nolen-Hoeksema, S. (2012). Emotion Regulation and Psychopathology: The Role of Gender.
Annual Review of Clinical Psychology, 8(1), 161–187.
http://doi.org/10.1146/annurev-clinpsy-032511-143109
O’Reilly, J. X., Woolrich, M. W., Behrens, T. E. J., Smith, S. M., & Johansen-Berg, H. (2012). Tools of the trade: Psychophysiological interactions and functional connectivity. Social Cognitive and Affective Neuroscience, 7(5), 604–609. http://doi.org/10.1093/scan/nss055
Ochsner, K. N., & Gross, J. J. (2005). The cognitive control of emotion. Trends in Cognitive Sciences, 9(5), 242–249. http://doi.org/10.1016/j.tics.2005.03.010
Ochsner, K. N., Silvers, J. A., & Buhle, J. T. (2012). Functional imaging studies of emotion regulation: a synthetic review and evolving model of the cognitive control of emotion. Annals of the New York Academy of Sciences, 1251, E1-24.
http://doi.org/10.1111/j.1749-6632.2012.06751.x
Opialla, S., Lutz, J., Scherpiet, S., Hittmeyer, A., J??ncke, L., Rufer, M., … Br??hl, A. B. (2014). Neural circuits of emotion regulation: a comparison of mindfulness-based and cognitive reappraisal strategies. European Archives of Psychiatry and Clinical Neuroscience, 265(1), 45–55.
http://doi.org/10.1007/s00406-014-0510-z
Paret, C., Kluetsch, R., Zaehringer, J., Ruf, M., Demirakca, T., Bohus, M., … Schmahl, C. (2016). Alterations of amygdala-prefrontal connectivity with real-time fMRI neurofeedback in BPD patients. Social Cognitive and Affective Neuroscience, 11(6), 952–960.
http://doi.org/10.1093/scan/nsw016
Price, J. L., & Drevets, W. C. (2012). Neural circuits underlying the pathophysiology of mood disorders. Trends in Cognitive Sciences, 16(1), 61–71. http://doi.org/10.1016/j.tics.2011.12.011
Sempértegui, G. A., Karreman, A., Arntz, A., & Bekker, M. H. J. (2013). Schema therapy for borderline personality disorder: A comprehensive review of its empirical foundations, effectiveness and implementation possibilities. Clinical Psychology Review, 33(3), 426–447.
http://doi.org/10.1016/j.cpr.2012.11.006
Smith, S. M., Jenkinson, M., Woolrich, M. W., Beckmann, C. F., Behrens, T. E. J., Johansen-Berg, H., … Matthews, P. M. (2004). Advances in functional and structural MR image analysis and implementation as FSL. NeuroImage, 23(SUPPL. 1), 208–219.
http://doi.org/10.1016/j.neuroimage.2004.07.051
Talairach, J., & Tournoux, P. (1988). Co-planar stereotaxic atlas of the human brain. Stuttgart: Thieme
Taylor, S. F., & Liberzon, I. (2007). Neural correlates of emotion regulation in psychopathology.
Trends in Cognitive Sciences, 11(10), 413–418. http://doi.org/10.1016/j.tics.2007.08.006
Urry, H. L. (2006). Amygdala and Ventromedial Prefrontal Cortex Are Inversely Coupled during Regulation of Negative Affect and Predict the Diurnal Pattern of Cortisol Secretion among Older Adults. Journal of Neuroscience, 26(16), 4415–4425.
http://doi.org/10.1523/JNEUROSCI.3215-05.2006
van Asselt, A. D. I., Dirksen, C. D., Arntz, A., & Severens, J. L. (2007). The cost of borderline personality disorder: societal cost of illness in BPD-patients. European Psychiatry, 22(6), 354–
361. http://doi.org/10.1016/j.eurpsy.2007.04.001
Van Zutphen, L., Siep, N., Jacob, G. A., Domes, G., Sprenger, A., Willenborg, B., … Arntz, A. (in prep). Neural correlates of emotional instability in BPD: BPD specificity Neural correlates
21
of emotional instability in borderline personality disorder: stimulus category specificity and diagnosis specificity. Authors’ Manuscript.
van Zutphen, L., Siep, N., Jacob, G. A., Goebel, R., & Arntz, A. (2015). Emotional sensitivity, emotion regulation and impulsivity in borderline personality disorder: A critical review of fMRI studies. Neuroscience and Biobehavioral Reviews, 51, 64–76.
http://doi.org/10.1016/j.neubiorev.2015.01.001
Wetzelaer, P., Farrell, J., Evers, S. M. A. A., Jacob, G. A., Lee, C. W., Brand, O., … Arntz, A. (2014). Design of an international multicentre RCT on group schema therapy for borderline personality disorder. BMC Psychiatry, 14(319). http://doi.org/10.1186/s12888-014-0319-3
Wolf, R. C., Sambataro, F., Vasic, N., Schmid, M., Thomann, P. A., Bienentreu, S. D., & Wolf, N. D. (2011). Aberrant connectivity of resting-state networks in borderline personality disorder. Journal of Psychiatry and Neuroscience, 36(6), 402–411.
http://doi.org/10.1503/jpn.100150
Xu, T., Cullen, K. R., Mueller, B., Schreiner, M. W., Lim, K. O., Schulz, S. C., & Parhi, K. K. (2016). Network analysis of functional brain connectivity in borderline personality disorder using resting-state fMRI. NeuroImage: Clinical, 11, 302–315.
http://doi.org/10.1016/j.nicl.2016.02.006
Yarkoni, T., Poldrack, R. A., Nichols, T. E., van Essen, D. C., & Wager, T. D. (2011). Large-scale automated synthesis of human functional neuroimaging data. Nature Methods, 8(8), 665–670.
http://doi.org/10.1038/nmeth.1635.Large-scale
Yarkoni, T., Poldrack, R. A., Nichols, T. E., Van Essen, D. C., & Wager, T. D. (2011). Large-scale automated synthesis of human functional neuroimaging data. Nature Methods, 8(8), 665–
870. http://doi.org/10.1038/nmeth.1635
Yokum, S., & Stice, E. (2013). Cognitive regulation of food craving: Effects of three congitive reappraisal strategies on neural response to palatable foods. International Journal of Obesity, 37(12), 1565–1570. http://doi.org/10.1038/ijo.2013.39.Cognitive
Zilverstand, A., Parvaz, M. A., & Goldstein, R. Z. (2017). Neuroimaging cognitive reappraisal in clinical populations to define neural targets for enhancing emotion regulation. A systematic review. NeuroImage, 151(December 2015), 105–116.
