The wandering mind in borderline personality disorder: Instability in self- and other-related thoughts Philipp Kanskea*, Lars Schulzeb, Isabel Dziobekc, Hannah Scheibnerc, Stefan Roepked#, Tania Singera# a Department of Social Neuroscience, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany b Department of Educational Sciences and Psychology, Free University Berlin, Germany c d Berlin School of Mind and Brain, Humboldt-Universität zu Berlin, Germany Charité - Universitätsmedizin Berlin, Campus Benjamin Franklin, Department of Psychiatry, Berlin, Germany # These authors contributed equally. Running title: Mind-wandering in borderline personality *Correspondence to Philipp Kanske, Department of Social Neuroscience, Max Planck Institute for Human Cognitive and Brain Sciences, Stephanstr. 1a, 04103 Leipzig, Germany, email: kanske@cbs.mpg.de, Tel: +49-(0)341 9940-2693; Fax: +49 (0)341 9940-2356 1 Abstract Diagnostic criteria for borderline personality disorder (BPD) include instability in identity and interpersonal relationships. Here, we probed whether instability is already present in BPD patients’ thoughts about themselves and others. We tested BPD patients (N=27) and healthy controls (N=25) with a mind-wandering task that assesses content and variability of stimulusindependent self-generated thoughts. Multi-level modeling revealed that while BPD patients and healthy controls mind-wander to a similar extent, BPD patients’ thoughts are colored predominantly negatively. Most importantly, although their thoughts concerned the self and others as much as in controls, they fluctuated more strongly in the degree to which their thoughts concerned themselves and others and also gave more extreme ratings. Self- and other related thoughts that were more extreme were also more negative in valence. The increased variability supports current conceptualizations of BPD and may account for the instability in identity and interpersonal relationships. Keywords: borderline personality disorders, mind-wandering, self-generated thought, identity, self-other representations, interpersonal relationships 2 1. Introduction Borderline personality disorder (BPD) is a severe and debilitating mental disorder that affects between 0.7 and 2.7% of the general population (Torgersen et al., 2001; Trull et al., 2010). Patients with BPD are characterized by impulsivity and affective dysregulation, but also by instability in identity and interpersonal relations (American Psychiatric Association, 2013; Gunderson, 2007; Sanislow et al., 2002). The DSM-5 alternative model for BPD further highlights self and interpersonal problems as crucial for the disorder. These may play out in dramatic shifts in self-image or sense of self and unstable, intense relationships. Across the theoretical spectrum, unstable self-other representations are integral parts of etiological models of BPD (Beck et al., 2003; Horowitz, 2004; Levy et al., 2006; Livesley and Jang, 2000) and have also been hypothesized to constitute the fundamental impairment that gives rise to BPD symptomatology, including affective dysregulation and impulsivity (Bender and Skodol, 2007). Empirical investigations of unstable self and other representations have mainly focused on self and observer evaluations (e.g., of interview or projective data; Porcelli et al., 2006; Tramantano et al., 2003) or specific stimulus-induced reactions of the patients (e.g., of others’ facial expressions; Lerner and St Peter, 1984; Westen et al., 1990). Clinicians’ ratings of their patients’ self-concept, for example, revealed that BPD patients have a more incoherent and inconsistent self-concept compared to other personality disorder patients and non-clinical control groups (Wilkinson-Ryan and Westen, 2000). BPD patients’ selfdescriptions in questionnaires are typically very negative (Klein et al., 2001; Rüsch et al., 2007; Sieswerda et al., 2005), and they also report lower self-concept clarity (Roepke et al., 2011), increased alexithymia (i.e., an inability to identify one’s own emotions New et al., 2012), and describe themselves more in terms of opposites than in terms of salient attributes in repertory grid tests (de Bonis et al., 1995). Using a card sorting task, Vater et al. (Vater et 3 al., 2015) recently showed that BPD patients have a more compartmentalized and negative self-concept, that is, they have a tendency to organize knowledge about the self into discrete, extremely negatively valenced categories (Showers, 1992). BPD patients’ representations of others also differ from healthy controls (HC) (Herpertz and Bertsch, 2014). Neutral or ambiguous facial expressions, for example, are evaluated as more negative (Domes et al., 2008; Domes et al., 2009; Dyck et al., 2009), as are film characters (Barnow et al., 2009; Sieswerda et al., 2013) and words, irrespective of comorbid depression (Kurtz and Morey, 1998). Moreover, their descriptions of others are characterized by multidimensional dichotomous thinking, a cognitive style that refers to the tendency to evaluate experiences in terms of mutually exclusive categories rather than falling along continua (Beck et al., 2001). Film characters, for example, are rated as either extremely negative or positive (Napolitano and Mckay, 2007; Veen and Arntz, 2000), as are real interaction partners (Arntz and ten Haaf, 2012). The studies reported above investigated BPD patients’ self-other representations at a given point in time and therefore represent a momentary snapshot. A direct test of instability would, however, require the assessment of self- and other-related thoughts at multiple points and explore how they change over time. Furthermore, if trait judgments are requested, momentary snapshots also reflect patients’ own integration of previous experience and, therefore, rely on meta-cognitive capabilities. As has been demonstrated for affective dysregulation, as well as dissociation and paranoid ideation, interpersonal disturbances and suicidality, sampling from the actual momentary experience of participants and integrating the data statistically may give a more realistic and valid characterization of BPD patients’ psychopathology (Ebner-Priemer et al., 2015; Ebner-Priemer et al., 2007; Santangelo et al., 2014). 4 Therefore, we utilized an established mind-wandering paradigm that, because of low task demands, induces stimulus-independent, self-generated thoughts and probed these at multiple points across the task (Ruby et al., 2013a). To increase the number of sampling points, the task was performed twice approximately ten days apart. In contrast to previous studies that tested participants’ responses to specific stimuli such as faces, the mindwandering paradigm gives an indication of patients’ self-generated mental contents. In the general population, such self-generated thought, mind-wandering, is highly prevalent, occurring in up to 50% of waking time (Kane et al., 2007; Killingsworth and Gilbert, 2010). While mind-wandering has been linked to negative mental health outcome (Killingsworth and Gilbert, 2010; Smallwood and Schooler, 2015), we know little about mind-wandering activity and specific mind-wandering content in psychopathology (Ottaviani et al., 2015; Smallwood, 2013). So far, no studies have investigated mind-wandering in BPD. Given the reported predominance of negative self and other evaluations, we hypothesized that patients in the BPD group would have more negative and less positive thoughts compared to healthy matched controls. Crucially, we also expected to find evidence for the hypothesized instability in self and other representations. This instability should be reflected in greater variation of the ratings of self- and other-related thoughts. Specifically, we explored two measures of variability, fluctuations in ratings between different, successive thought probes, and the extremity of the ratings. The reported negative self and other evaluations in BPD also gave rise to the hypothesis that self- and other-related thoughts and their variability during mind-wandering may be colored more negatively in the BPD group. We therefore related thought valence (negative, positive) to the ratings and variability of selfand other-related thoughts. Furthermore, we included additional ratings that asked for the temporal focus of the self-generated thoughts (past- or future-oriented). We had no explicit hypotheses regarding these ratings in BPD, but included them to more comprehensively 5 characterize the thought space (Smallwood and Schooler, 2015). To assess current mood, we also included mood ratings, hypothesizing more negative mood in BPD patients. Finally, in addition to testing differences between BPD patients and controls, we also tested whether the effects would be associated with symptom severity in BPD patients only. 6 2. Methods 2.1 Participants 2.1.1 BPD group Thirty patients with Borderline Personality Disorder according to DSM-IV-TR (American Psychiatric Association, 2000; Saß et al., 2003) were recruited at the Department of Psychiatry, Charité – Universitätsmedizin Berlin, Germany. All were inpatients at the Charité hospital and admitted for specialized BPD treatment from a waiting list; all BPD patients had outpatient status before admission, none of the patients were transferred from another institution to our hospital or admitted for acute care. Testing was performed in the first two weeks after admission in a laboratory at the hospital. Because of technical and timing difficulties (specifically, a software problem with the experimental computer), three patients could not complete the mind-wandering protocol, yielding a final sample of 27 patients (for demographics, see Table 1). 2.1.2 Control group Thirty-four healthy control participants were recruited and matched to the patients in age, gender, and education. Because of technical and timing difficulties (as for the patient group), data from nine control participants were missing, yielding a final sample of 25 healthy controls (see Table 1). The presence or absence of individual diagnoses in patients and controls was established with the German versions of the Mini International Neuropsychiatric Interview (Ackenheil et al., 1999; Lecrubier et al., 1997) and the Structured Clinical Interview II (First et al., 1997; Wittchen et al., 1997). All interviewers were psychiatrists or clinical psychologists, trained in the application of the SCID II and MINI interview and received supervision on the SCIDs and MINIs. In a previous study, interrater reliability of SCID-II BPD diagnoses by the same interviewers employed in the current study was good, κ=0.82 (Ritter et al., 2014). Exclusion 7 criteria for the patients included any psychotic disorder, current substance abuse/ dependency, mental retardation, epilepsy/ organic brain disease, and age younger than 18. For controls, the same exclusion criteria applied and the presence of any current mental disorder additionally led to exclusion from the study. All participants gave written informed consent prior to participation. The study was approved by the ethics committee of the Charité – Universitätsmedizin Berlin. 2.2 Task We used an established mind-wandering paradigm that probes self-generated thoughts during a choice reaction time task (Baird et al., 2012; Ruby et al., 2013a; Smallwood et al., 2013a). A series of black digits between 1 and 8 was presented. One sixth of the digits (randomly selected) was presented in red, signaling to participants that they should indicate via button press if this number was odd or even. Black digits were presented for 1000 ms and red digits for 2000 ms. Responses had to be made while the colored digits were still present on the screen. Stimuli were separated by a fixation cross of variable duration (2200–4400 ms). Thought probes were presented between four and nine times at random intervals during the choice reaction time task. At each probe time point, participants were asked to rate their current thoughts using nine-point Likert scales on several separate dimensions including (1) how positive, (2) how negative, (3) how self-related, (4) how other-related, (5) how pastoriented, (6) how future-oriented, and (7) how off task their thoughts at that point in time had been. The extent to which participants rated their thoughts as off task is interpreted as an indicator of mind-wandering. Additionally, participants rated their current mood (i.e., how positive and how negative they felt). (For the exact phrasing of all rating questions, please see Ruby et al., 2013). The entire task lasted approximately 20 min. Stimuli were presented using E-prime 2.0 (Psychology Software Tools, Inc., Sharpsburg, PA, USA). 8 The time of day for testing was matched for the two groups (mean experiment start time was 1:42 pm for BPD patients and 1:26 pm for control participants). In contrast to most previous mind-wandering studies, each participant completed the task twice in two identical sessions approximately 10 days apart (mean time difference 9.0 ± 8.2 days for BPD patients and 11.6 ± 9.4 days for healthy control participants; t(50) = -1.1, p > .25) to increase the number of thought probes (combined mean number of probes 11.2 ± 2.0 for BPD patients and 11.9 ± 2.5 for healthy controls; t(50) = -1.1, p > .25). 2.3 Questionnaire The German version of the Personality Assessment Inventory - Borderline Features (PAIBOR) was used to measure symptom severity in BPD patients (Groves and Engel, 2007; Hopwood et al., 2013; Morey, 1991). The PAI-BOR assesses core aspects of BPD, that is, affective instability, identity problems, negative relationships, and self-harm. Previous validation studies indicated good-to-excellent internal consistency and illustrated the usefulness of the PAI-BOR for determining features of BPD in clinical samples and the general population (e.g., (BellPringle et al., 1997; Gardner and Qualter, 2009; Stein et al., 2007). Internal consistency of the PAI-BOR in the present sample was excellent, with a Cronbach’s alpha of 0.95. 2.4 Data analysis All statistical analyses were performed using SPSS (IBM Corp. Released 2013. IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: IBM Corp.). 2.4.1 Performance We analyzed the reaction times and accuracy of participants’ responses in the choice reaction time task with independent samples t-tests. 2.4.2 Thought probes 2.4.2.1 Principal components analysis 9 To test for congruency with previous reports on the contents of self-generated thoughts in healthy participants, we first calculated a principal components analysis (as described in Ruby et al., 2013a), which yielded a three-component solution that conformed with the literature (see Supplementary Table 1; Ruby et al., 2013b; Smallwood et al., 2013b). Specifically, component 1 mainly includes positive and negative ratings, component 2 includes past and other ratings, and component 3 includes future and self ratings. For a more fine-grained characterization of self-generated thought in BPD, we here focused on analyzing the ratings individually. 2.4.2.2 Rating levels To analyze the ratings, we used multi-level models because they take correlated observations within individuals into account and perform well with missing data or unequal numbers of data points within individuals (Jahng et al., 2008). The models were specified to test for associations of the group factor (BPD group vs. control group, coded as 0 and 1, respectively) with the thought probe ratings. The ratings on each dimension (i.e., how other-related, selfrelated, negatively valenced, positively valenced, past-oriented, future-oriented and how off task the thoughts were) were entered as dependent variable to be predicted in separate models. The main predictor was the group factor. Significant effects of group would indicate that BPD is, for instance, associated with higher levels of negative thoughts. Additional covariates included the number of the particular sampling point within the session (e.g., sample count 5, indicating that this rating value was obtained during the fifth probe for this participant) to control for changes due to the repetition of the rating questions, and session (i.e., session 1 or 2) to control for changes due to the repetition of the whole procedure. 2.4.2.3 Fluctuations in ratings In addition to the rating level (e.g., to what extent a certain thought was negative), we investigated two indices of instability: fluctuations and extremity of ratings. To obtain a 10 measure of how much individuals fluctuate in their ratings from one thought probe to the next, we calculated squared successive differences, an established procedure in experience sampling studies (Ebner-Priemer et al., 2007; Jahng et al., 2008; Skirrow et al., 2014; Trull et al., 2008). As for the rating levels, we then calculated multi-level models with fluctuations as dependent variable and group (BPD group vs. control group) as predictor. Sample count and session were again included as covariates. 2.4.2.4 Extremity of ratings To obtain a measure of the extremity of the individual ratings, we calculated the squared difference of each rating from the mean for that variable. While this measure is naturally related to the fluctuations (e.g., very strong fluctuations will also lead to high extremity), it does not take into account the extent of these successive changes, but rather indicates the degree to which a certain rating differs from the “norm” (e.g., fluctuations may be low, but extremity still high if ratings are uniformly high during the first half and low during the second half of a session). As for the rating levels, we then calculated multi-level models with extremity as dependent variable and group (BPD group vs. control group) as predictor. Sample count and session were included as covariates. 2.4.2.5 Testing the valence of self- and other-related thoughts A new set of models tested for relations between the variables that showed significant differences between BPD patients and controls. For instance, we tested whether extremity of self-related ratings was related to negativity. To this end, rating levels, fluctuations, and extremity of self and other ratings were entered as dependent variable in separate models. Again, group (BPD group vs. control group) was used as a predictor, but the ratings on how positive and negative the thoughts were, were entered as additional predictors. Crucially, the interaction of group with the additional predictors was included to test, for instance, if the 11 association of extremity in self-relatedness and negativity of a thought was stronger/weaker in BPD patients. 2.4.3 Mood probes The ratings of how positive and how negative participants’ current mood was were analyzed in the same way as the ratings, fluctuations, and extremity of the thought probes. 2.4.4 Relation to symptom severity For models where group was a significant predictor, follow-up analyses were used to examine symptoms severity as a predictor in the BPD group only. We included the PAI-BOR scores as covariates in the models described above, only testing BPD patients in this analysis and replacing group (BPD group vs. control group) with the PAI-BOR scores as a predictor. All multi-level models included a random intercept and random effects for sample count and order. Maximum Likelihood (ML) was used as estimation method. 12 3. Results 3.1 Performance Independent samples t-tests of performance measures in the choice reaction time task showed no significant differences between BPD patients and HCs in accuracy (t(50) = -0.17, p > 0.85; BPD: 89.2 % ± 14.5, HC: 89.9 % ± 15.1) or reaction times (t(50) = 0.86, p > .35; BPD: 904.8 milliseconds (ms) (169.4), HC: 869.4 ms (120.5)). 3.2 Thought probes 3.2.1 Rating levels Multi-level models revealed that BPD patients did not differ from HCs in the extent of offtask thoughts, that is, in the amount of reported mind-wandering (see Fig.1A and Table 2). However, the content of the self-generated thoughts differed. BPD patients rated their thoughts to be more negative (Cohen’s d = 1.85) and less positive than HCs (Cohen’s d = 1.70). While HCs showed a positive bias (i.e., more positive than negative thoughts; t(24) = 8.2, p < 0.001), BPD patients showed a negative bias (i.e., more negative than positive thoughts; t(26) = 2.1, p < 0.05). Including current mood (see below) as a covariate (positive mood: b = 0.594, S.E. = 0.037, p < 0.001, negative mood: b = 0.483, S.E. = 0.032, p < 0.001) showed significant relations between mood and thought valence, but did not change the group differences in positive (b = -13.265, S.E. = 2.844, p < 0.001) and negative thoughts (b = 21.869, S.E. = 2.970, p < 0.001). There were no group differences regarding how self- and other-related or past- and future-oriented the rated thoughts were. 3.2.2 Fluctuations in ratings As a measure of how much individuals fluctuate in their ratings from one thought probe to the next, squared successive differences (Ebner-Priemer et al., 2007; Jahng et al., 2008; Skirrow et al., 2014; Trull et al., 2008) were subjected to multi-level models (see Fig. 1B and Table 2). The results show increased fluctuations in BPD patients compared to HCs in self- 13 and other-related thoughts (Cohen’s d = 0.89 and d = 0.45, respectively). There were no further differences between the groups. 3.2.3 Extremity of ratings As a measure of the extremity of the individual ratings, the squared differences of each rating from the mean were subjected to multi-level models (see Fig. 1C and Table 2). As for fluctuations, we also observed significantly increased extremity of self- and other-related thoughts in BPD patients compared to HCs (Cohen’s d = 2.15 and d = 1.79, respectively). Additionally, BPD patients showed more extreme ratings of off-task thoughts (Cohen’s d = 0.60). 3.2.4 Testing the valence of self- and other-related thoughts The previous analyses showed increased negative and decreased positive thoughts in BPD patients, as well as increased fluctuations and extremity of self- and other-related thoughts. Here we tested whether the valence of thoughts (negative, positive) was associated with the rating levels, fluctuations, and extremity of self- and other-related thoughts (Table 3). The results show that the more negative a thought, the more it was also self-related. This was true across both BPD patients and HCs, there was no interaction with group. There was no relation to positive thoughts or between other-related thoughts and the level of positive and negative thoughts. With regard to fluctuations in self- and other-related thoughts, there were no significant relations to the level of negative or positive thoughts. Regarding the extremity of self- and other-related thoughts, interactions of group and negative thoughts indicated greater extremity of how self- and other-related the negative thoughts were in BPD patients, that is, self- and other-related thoughts that were more extreme were also more negative in valence in BPD patients compared to HCs (see Fig. 2). The interaction of group and positive thoughts was not significant. 14 negative valence of thoughts positively predicted the extremity ratings of self and other relatedness, that is, self- and other-related thoughts that were more extreme were also more negative in valence. 3.3 Mood probes Multi-level models revealed that BPD patients showed elevated levels of negative and decreased levels of positive mood (see Supplementary Fig. 1 and Table 2). Fluctuations and extremity of mood ratings were not significantly different between groups. 3.4 Relation to symptom severity To test for a relation of the observed effects to symptom severity in the BPD group, we included the PAI-BOR scores as covariates in the models (only testing BPD patients in this analysis). PAI-BOR was a predictor of increased levels of negative thoughts (b = 1.377, S.E. = 0.520, p = 0.014) and of extremity of other-related thoughts (b = 21.537, S.E. = 9.590, p = 0.028). Numerically, but not significantly, PAI-BOR also predicted reduced levels of positive thoughts (b = -0.896, S.E. = 0.439, p = 0.051), reduced positive mood (b = -0.861, S.E. = 0.467, p = 0.077), and fluctuations in self- (b = 23.941, S.E. = 14.196, p = 0.093) and otherrelated thoughts (b = 35.156, S.E. = 20.099, p = 0.095). All other effects were not significantly related to PAI-BOR scores (all p > 0.30). 15 4. Discussion The present study aimed at investigating the self-generated thoughts of patients with BPD and utilized a standard mind-wandering task that probes the amount and specific content of self-generated thoughts (Ruby et al., 2013a). The results reveal that BPD patients and HCs mind-wander to a similar extent. BPD patients also think of the past, the future, themselves, and others as often as healthy individuals; however, their thoughts are colored predominantly negatively, while the thoughts of HCs are colored positively. Crucially, BPD patients are more unstable in their self- and other-related thoughts, as indicated by increased fluctuations between successive thought probes and by more extreme values of these ratings. The more these ratings of self- and other-related thoughts differed from the mean, the more negative they were in BPD patients. The observation of more negative thoughts in BPD patients is in line with previous reports of a negative bias in emotional face recognition (Dyck et al., 2009) and in questionnaire evaluations of themselves and others (Klein et al., 2001; Rüsch et al., 2007; Sieswerda et al., 2013). Critically, it extends these findings to more negative self-generated mental content, as assessed with multiple probes of ongoing mind-wandering. As previously reported, BPD patients also showed more negative mood (Nisenbaum et al., 2010). The negative bias in thought contents remained, however, when controlling for current mood, which suggests some specificity of mood and thought contents. Given the negative biases in information processing across different psychopathologies (e.g., in major depressive disorder (Raes et al., 2006), generalized anxiety disorder (Mogg and Bradley, 2005), or social anxiety disorder (Joormann and Gotlib, 2006)), it is possible that the predominance of negative thoughts observed here in BPD patients is a more general characteristic of psychopathology, which should be tested in future studies (for a recent report in depression, see Hoffmann et al., 2016). Nevertheless, it may represent an important problem also in BPD. While studies 16 on the consequences of specific thought patterns are still rare, there is some indication that more negative thoughts increase the stress-related cortisol response to psycho-social stressors (Engert et al., 2014). Future studies should therefore elucidate if a negative bias constitutes a psychopathogenic factor and how it may be improved. BPD patients also showed more variability in self- and other-related thoughts. Variability was measured as within-subject fluctuations between successive thought probes and as the extremity of these ratings, the latter being related to particularly negative thoughts in BPD patients. The increased variability points towards instability in self- and other-related mental representations and supports accounts of BPD across the theoretical spectrum that posit these instabilities as central to the disorder (Beck et al., 1961; Horowitz, 2004; Levy et al., 2006; Livesley and Jang, 2000). However, it also extends these accounts by showing that BPD patients already differ in how much they vary in thinking of themselves and others, not only in whether their thoughts of themselves and others vary more strongly in valence. It is conceivable that this variability is related to other characteristics of BPD such as diminished self-concept clarity (Roepke et al., 2011). In contrast to previous empirical investigations of disturbed self (de Bonis et al., 1995; Roepke et al., 2011; Vater et al., 2015; Wilkinson-Ryan and Westen, 2000) and other representations in BPD (Arntz and ten Haaf, 2012; Napolitano and Mckay, 2007; Veen and Arntz, 2000), the present study tested (a) the contents of selfgenerated mental activity and (b) sampled ongoing mental activity at multiple time points. Testing at multiple time points may be more valid than single measurements, as has been demonstrated for affective dysregulation in BPD (Ebner-Priemer et al., 2015; Ebner-Priemer et al., 2007), and it also allows a direct test of instability because participants are not asked to integrate across experiences themselves in their judgments. A better understanding of the self-generated thoughts of BPD patients is also important because of the high prevalence of mind-wandering, occurring in 25 to 50% of 17 waking time (Kane et al., 2007; Killingsworth and Gilbert, 2010). Beyond impairment in current task performance, which is mainly affected by the extent of off-task thoughts (Smallwood et al., 2003; Smallwood et al., 2007), the content of this prevalent mental activity may also have far-reaching consequences, as indicated by studies on stress reactivity or on the effects of habitual rumination styles (Engert et al., 2014; Hertel, 1998; Schick et al., 2013). Investigating the specific consequences of instability in self- and other-related thoughts, as observed here for BPD patients, could elucidate their etiological role for the disorder, for example by linking unstable other-related thoughts to problems in social cognition and ultimately in interpersonal relationships. Neuroimaging studies have found that mind-wandering is associated with increased activity in the default-mode network of cortical regions that are active during resting state, where the amount of self-generated thought correlates with activation (Mason et al., 2007). Patients with BPD show increased “trait-level” functional connectivity within the defaultmode network at rest (Kluetsch et al., 2012; Wolf et al., 2011), and it would be interesting to test the association between this increased connectivity and the increased variability in thought content. During external stimulation, however, in particular when reasoning about the affective states of others, activity within this network is reduced in BPD (Dziobek et al., 2011; Mier et al., 2013). Future research could investigate whether the altered trait activity within this network, which may be due to instability in self- and other-related thoughts, hampers the functioning of this network under external demands, leading to impaired social cognition. While BPD patients’ self-generated thoughts showed a negative bias and increased variability in how self- and other-related they were, the extent of off-task thoughts and the degree to which thoughts were past- or future-oriented did not differ from healthy controls. First and foremost, this speaks for specificity of the observed alterations. Given that 18 borderline personality features have been linked to increased rumination (Baer and Sauer, 2011), increased overall off-task thoughts, that is, mind-wandering, with a past focus could have been expected in addition to a negative bias. As rumination in BPD has so far only been studied with trait questionnaires, the results may suggest that BPD patients momentary and trait self-judgments diverge, which should be explicitly tested in future investigations. Because of the methodological difficulties associated with studying the content of stimulus-independent mental activity, the investigation of mind-wandering has only really gained momentum in the last decade (Smallwood and Schooler, 2015). Regarding psychopathology, this advancement offers the chance to gain better understanding not only of the amount of mind-wandering, as has been attempted in depression (Watts et al., 1988), ADHD (Shaw and Giambra, 1993), and schizophrenia (Elua et al., 2012), but also of the specific content of self-generated thoughts, which has so far not been investigated. The present study demonstrates the feasibility of using mind-wandering tasks as a tool in clinical psychological research. Furthermore, while the mind-wandering literature has so far focused on the extent of off-task thoughts and increasingly also its content (Smallwood, 2013), the present study shows that investigating the variability of mind-wandering across time yields an additional interesting characteristic of self-generated thoughts and allows insight into the (in)stability of mental representations. There are some limitations to the present study. We could not include a matched clinical control group. Therefore, the specificity of the observed results to BPD could not be tested. The correlation of the observed effects to symptom severity, as measured with the PAI-BOR, is some indication of a relation to BPD, but future studies should probe whether the effects remain specific in comparison to other clinical groups. Furthermore, a larger sample would have been advantageous, in particular with regard to analyzing subgroups with different comorbidities or medication. As most BPD patients were treated with psychotropic 19 medication, influence of medication on the data could not be evaluated. Similarly, the subgroups with or without different comorbidities (e.g., current or past depression, PTSD, eating disorder) were very small. Future studies should include larger samples that allow comparison of these subgroups, because comorbidity or medication may impact the effects, for instance the observed negativity bias. Additionally, the small number of men in the present sample hampers generalizations of the results to all BPD patients irrespective of gender. The present results have several clinical implications. An increasing body of research investigates the effects of mindfulness practices for stabilizing the mind (Lutz et al., 2007), and mindfulness exercises are also implemented in one of the most effective psychotherapeutic treatments for BPD, dialectical behavioral therapy (Linehan, 1987; Linehan et al., 2006). Given that mindfulness training has been shown to reduce mindwandering (Mrazek et al., 2013), the present results suggest that mindfulness exercises are an important component of BPD treatment, possibly through their effect on mind-wandering. Mentalization-based therapy for BPD, in contrast, focuses on elaborating the capacity to implicitly and explicitly interpret the actions of oneself and others (Bateman and Fonagy, 2004, 2010). As an element of increasing the patient’s mentalization capacity, it may necessary to also stabilize the thoughts related to self and others. It may also prove fruitful to directly target the negative bias in self-generated thoughts observed in BPD, because a negative cognitive bias has been consistently linked to negative outcome and targeted interventions are available, such as “cognitive bias modification” or mental imagery based interventions (see e.g. Beck et al., 2003; Hallion and Ruscio, 2011; Lang et al., 2012). To conclude, using a mind-wandering paradigm we observed more negatively colored thoughts and more instability in self- and other-related thoughts in patients with BPD. In BPD patients, more extreme ratings of self- and other-related thoughts were also more 20 negative. Given the high prevalence of such self-generated thoughts, these alterations may have consequences, for example on patients’ stress responsivity and interpersonal relationships, and may thereby play an important etiological role for BPD. 21 Acknowledgments We wish to thank Lisa Deunert and Tom Dreßler for their help in data collection. 22 References Ackenheil, M., Stotz, G., Dietz-Bauer, R., Vossen, A., 1999. Mini International Neuropsychiatric Interview. German version 5.0.0. Psychiatrische Universitätsklinik München, Munich, Germany. American Psychiatric Association, 2000. Diagnostic and Statistical Manual of Mental Disorders (4th ed.) - Text Revision. American Psychiatric Publishing, Inc., Washington, D.C. American Psychiatric Association, 2013. Diagnostic and Statistical Manual of Mental Disorders (5th ed.). American Psychiatric Publishing, Inc., Washington, D.C. Arntz, A., ten Haaf, J., 2012. Social cognition in borderline personality disorder: evidence for dichotomous thinking but no evidence for less complex attributions. Behav Res Ther 50 (11), 707-718. Baer, R.A., Sauer, S.E., 2011. Relationships Between Depressive Rumination, Anger Rumination, and Borderline Personality Features. Personality Disorders-Theory Research and Treatment 2 (2), 142-150. Baird, B., Smallwood, J., Mrazek, M.D., Kam, J.W., Franklin, M.S., Schooler, J.W., 2012. Inspired by distraction: mind wandering facilitates creative incubation. Psychol Sci 23 (10), 1117-1122. Barnow, S., Stopsack, M., Grabe, H.J., Meinke, C., Spitzer, C., Kronmuller, K., Sieswerda, S., 2009. Interpersonal evaluation bias in borderline personality disorder. Behav Res Ther 47 (5), 359-365. Bateman, A.W., Fonagy, P., 2004. Mentalization-based treatment of BPD. J Pers Disord 18 (1), 36-51. Bateman, A.W., Fonagy, P., 2010. Mentalization based treatment for borderline personality disorder. World Psychiatry 9 (1), 11-15. Beck, A.T., Butler, A.C., Brown, G.K., Dahlsgaard, K.K., Newman, C.F., Beck, J.S., 2001. Dysfunctional beliefs discriminate personality disorders. Behav Res Ther 39 (10), 1213-1225. Beck, A.T., Freeman, A., Davis, D.D., 2003. Cognitive Therapy of Personality Disorders (2nd ed.). Guilford Press, New York. Beck, A.T., Ward, C.H., Mendelson, M., Mock, J., Erbaugh, J., 1961. An inventory for measuring depression. Arch Gen Psychiatry 4, 561-571. BellPringle, V.J., Pate, J.L., Brown, R.C., 1997. Assessment of borderline personality disorder using the MMPI-2 and the personality assessment inventory. Assessment 4 (2), 131139. Bender, D.S., Skodol, A.E., 2007. Borderline personality as a self-other representational disturbance. J Pers Disord 21 (5), 500-517. 23 de Bonis, M., De Boeck, P., Lida-Pulik, H., Feline, A., 1995. Identity disturbances and selfother differentiation in schizophrenics, borderlines, and normal controls. Compr Psychiatry 36 (5), 362-366. Domes, G., Czieschnek, D., Weidler, F., Berger, C., Fast, K., Herpertz, S.C., 2008. Recognition of facial affect in borderline personality disorder. J Pers Disord 22 (2), 135-147. Domes, G., Schulze, L., Herpertz, S.C., 2009. Emotion recognition in borderline personality disorder-a review of the literature. J Pers Disord 23 (1), 6-19. Dyck, M., Habel, U., Slodczyk, J., Schlummer, J., Backes, V., Schneider, F., Reske, M., 2009. Negative bias in fast emotion discrimination in borderline personality disorder. Psychol Med 39 (5), 855-864. Dziobek, I., Preissler, S., Grozdanovic, Z., Heuser, I., Heekeren, H.R., Roepke, S., 2011. Neuronal correlates of altered empathy and social cognition in borderline personality disorder. Neuroimage 57 (2), 539-548. Ebner-Priemer, U.W., Houben, M., Santangelo, P., Kleindienst, N., Tuerlinckx, F., Oravecz, Z., Verleysen, G., Van Deun, K., Bohus, M., Kuppens, P., 2015. Unraveling affective dysregulation in borderline personality disorder: a theoretical model and empirical evidence. J Abnorm Psychol. Ebner-Priemer, U.W., Kuo, J., Kleindienst, N., Welch, S.S., Reisch, T., Reinhard, I., Lieb, K., Linehan, M.M., Bohus, M., 2007. State affective instability in borderline personality disorder assessed by ambulatory monitoring. Psychol Med 37 (7), 961-970. Elua, I., Laws, K.R., Kvavilashvili, L., 2012. From mind-pops to hallucinations? A study of involuntary semantic memories in schizophrenia. Psychiatry Research 196 (2-3), 165-170. Engert, V., Smallwood, J., Singer, T., 2014. Mind your thoughts: associations between selfgenerated thoughts and stress-induced and baseline levels of cortisol and alpha-amylase. Biol Psychol 103, 283-291. First, M.B., Gibbon, M., Spitzer, R.L., Williams, J.B.W., Benjamin, L., 1997. Structured Clinical Interview for DSM-IV Axis II Personality Disorders (SCID-II). American Psychiatric Publishing, Inc., Arlington. Gardner, K., Qualter, P., 2009. Reliability and validity of three screening measures of borderline personality disorder in a nonclinical population. Personality and Individual Differences 46 (5-6), 636-641. Groves, J.A., Engel, R.R., 2007. The German Adaptation and standardization of the Personality Assessment Inventory (PAI). J Pers Assess 88 (1), 49-56. Gunderson, J.G., 2007. Disturbed relationships as a phenotype for borderline personality disorder. Am J Psychiatry 164 (11), 1637-1640. Hallion, L.S., Ruscio, A.M., 2011. A meta-analysis of the effect of cognitive bias modification on anxiety and depression. Psychol Bull 137 (6), 940-958. 24 Herpertz, S.C., Bertsch, K., 2014. The social-cognitive basis of personality disorders. Curr Opin Psychiatry 27 (1), 73-77. Hertel, P.T., 1998. Relation between rumination and impaired memory in dysphoric moods. J Abnorm Psychol 107 (1), 166-172. Hoffmann, F., Banzhaf, C., Kanske, P., Bermpohl, F., Singer, T., 2016. Where the depressed mind wanders: Self-generated thought patterns as assessed through experience sampling as a state marker of depression. J Affect Disord 198, 127-134. Hopwood, C.J., Wright, A.G., Krueger, R.F., Schade, N., Markon, K.E., Morey, L.C., 2013. DSM-5 pathological personality traits and the personality assessment inventory. Assessment 20 (3), 269-285. Horowitz, L.M., 2004. Interpersonal Foundations of Psychopathology. American Psychological Association, Washington, D.C. Jahng, S., Wood, P.K., Trull, T.J., 2008. Analysis of affective instability in ecological momentary assessment: Indices using successive difference and group comparison via multilevel modeling. Psychol Methods 13 (4), 354-375. Joormann, J., Gotlib, I.H., 2006. Is this happiness I see? Biases in the identification of emotional facial expressions in depression and social phobia. J Abnorm Psychol 115 (4), 705714. Kane, M.J., Brown, L.H., McVay, J.C., Silvia, P.J., Myin-Germeys, I., Kwapil, T.R., 2007. For whom the mind wanders, and when - An experience-sampling study of working memory and executive control in daily life. Psychol Sci 18 (7), 614-621. Killingsworth, M.A., Gilbert, D.T., 2010. A wandering mind is an unhappy mind. Science 330 (6006), 932. Klein, M.H., Wonderlich, S.A., Crosby, R., 2001. Self-concept correlates of the personality disorders. J Pers Disord 15 (2), 150-156. Kluetsch, R.C., Schmahl, C., Niedtfeld, I., et al., 2012. Alterations in default mode network connectivity during pain processing in borderline personality disorder. Archives of General Psychiatry 69 (10), 993-1002. Kurtz, J.E., Morey, L.C., 1998. Negativism in evaluative judgments of words among depressed outpatients with borderline personality disorder. J Pers Disord 12 (4), 351-361. Lang, T.J., Blackwell, S.E., Harmer, C.J., Davison, P., Holmes, E.A., 2012. Cognitive Bias Modification Using Mental Imagery for Depression: Developing a Novel Computerized Intervention to Change Negative Thinking Styles. European Journal of Personality 26 (2), 145-157. Lecrubier, Y., Sheehan, D.V., Weiller, E., Amorim, P., Bonora, I., Sheehan, K., Janavs, J., Dunbar, G., 1997. The mini international neuropsychiatric interview (MINI). A short diagnostic structured interview: reliability and validity according to CIDI. European Psychiatry 12, 224-231. 25 Lerner, H.D., St Peter, S., 1984. Patterns of object relations in neurotic, borderline and schizophrenic patients. Psychiatry 47 (1), 77-92. Levy, K.N., Clarkin, J.F., Yeomans, F.E., Scott, L.N., Wasserman, R.H., Kernberg, O.F., 2006. The mechanisms of change in the treatment of borderline personality disorder with transference focused psychotherapy. J Clin Psychol 62 (4), 481-501. Linehan, M.M., 1987. Dialectical behavior therapy for borderline personality disorder: Theory and method. Bulletin of the Menninger Clinic 51, 261-276. Linehan, M.M., Comtois, K., Murray, A.M., et al., 2006. TWo-year randomized controlled trial and follow-up of dialectical behavior therapy vs therapy by experts for suicidal behaviors and borderline personality disorder. Archives of General Psychiatry 63 (7), 757766. Livesley, W.J., Jang, K.L., 2000. Toward an empirically based classification of personality disorder. J Pers Disord 14 (2), 137-151. Lutz, A., Dunne, J.D., Davidson, R.J., 2007. Meditation and the neuroscience of consciousness, in: Zelazo, P., Moscovitch, M., Thompson, E. (Eds.), Cambridge Handbook of Consciousness. Cambridge University Press, Campbridge, UK. Mason, M.F., Norton, M.I., Van Horn, J.D., Wegner, D.M., Grafton, S.T., Macrae, C.N., 2007. Wandering minds: The default network and stimulus-independent thought. Science 315 (5810), 393-395. Mier, D., Lis, S., Esslinger, C., Sauer, C., Hagenhoff, M., Ulferts, J., Gallhofer, B., Kirsch, P., 2013. Neuronal correlates of social cognition in borderline personality disorder. Soc Cogn Affect Neurosci 8 (5), 531-537. Mogg, K., Bradley, B.P., 2005. Attentional bias in generalized anxiety disorder versus depressive disorder. Cognitive Therapy and Research 29 (1), 29-45. Morey, L.C., 1991. The Personality Assessment Inventory: Professional Manual. Psychological Assessment Resources, Lutz, FL. Mrazek, M.D., Franklin, M.S., Phillips, D.T., Baird, B., Schooler, J.W., 2013. Mindfulness Training Improves Working Memory Capacity and GRE Performance While Reducing Mind Wandering. Psychol Sci 24 (5), 776-781. Napolitano, L.A., Mckay, D., 2007. Dichotomous thinking in borderline personality disorder. Cognitive Therapy and Research 31 (6), 717-726. New, A.S., aan het Rot, M., Ripoll, L.H., Perez-Rodriguez, M.M., Lazarus, S., Zipursky, E., Weinstein, S.R., Koenigsberg, H.W., Hazlett, E.A., Goodman, M., Siever, L.J., 2012. Empathy and alexithymia in borderline personality disorder: clinical and laboratory measures. J Pers Disord 26 (5), 660-675. Nisenbaum, R., Links, P.S., Eynan, R., Heisel, M.J., 2010. Variability and predictors of negative mood intensity in patients with borderline personality disorder and recurrent suicidal behavior: Multilevel analyses applied to experience sampling methodology. J Abnorm Psychol 119 (2), 433-439. 26 Ottaviani, C., Shahabi, L., Tarvainen, M., Cook, I., Abrams, M., Shapiro, D., 2015. Cognitive, behavioral, and autonomic correlates of mind wandering and perseverative cognition in major depression. Frontiers in Neuroscience 8. Porcelli, J.H., Shahar, G., Blatt, S.J., Ford, R.Q., Mezza, J.A., Greenlee, L.M., 2006. Social cognition and object relations scale: Convergent validity and changes following intense inpatient treatment. Personality and Individual Differences 41, 407-417. Raes, F., Hermans, D., Williams, J.M.G., 2006. Negative bias in the perception of others' facial emotional expressions in major depression - The role of depressive rumination. Journal of Nervous and Mental Disease 194 (10), 796-799. Ritter, K., Vater, A., Rusch, N., Schroder-Abe, M., Schutz, A., Fydrich, T., Lammers, C.H., Roepke, S., 2014. Shame in patients with narcissistic personality disorder. Psychiatry Research 215 (2), 429-437. Roepke, S., Schroder-Abe, M., Schutz, A., Jacob, G., Dams, A., Vater, A., Ruter, A., Merkl, A., Heuser, I., Lammers, C.H., 2011. Dialectic behavioural therapy has an impact on selfconcept clarity and facets of self-esteem in women with borderline personality disorder. Clin Psychol Psychother 18 (2), 148-158. Ruby, F.J.M., Smallwood, J., Engen, H., Singer, T., 2013a. How self-generated thought shapes mood--the relation between mind-wandering and mood depends on the socio-temporal content of thoughts. PLoS One 8 (10), e77554. Ruby, F.J.M., Smallwood, J., Sackur, J., Singer, T., 2013b. Is self-generated thought a means of social problem solving? Front Psychol 4. Rüsch, N., Lieb, K., Gottler, I., Hermann, C., Schramm, E., Richter, H., Jacob, G.A., Corrigan, P.W., Bohus, M., 2007. Shame and implicit self-concept in women with borderline personality disorder. Am J Psychiatry 164 (3), 500-508. Sanislow, C.A., Morey, L.C., Grilo, C.M., Gunderson, J.G., Shea, M.T., Skodol, A.E., Stout, R.L., Zanarini, M.C., McGlashan, T.H., 2002. Confirmatory factor analysis of DSM-IV borderline, schizotypal, avoidant and obsessive-compulsive personality disorders: findings from the Collaborative Longitudinal Personality Disorders Study. Acta Psychiatr Scand 105 (1), 28-36. Santangelo, P., Bohus, M., Ebner-Priemer, U.W., 2014. Ecological momentary assessment in borderline personality disorder: a review of recent findings and methodological challenges. J Pers Disord 28 (4), 555-576. Saß, H., Wittchen, H.-U., Zaudig, M., Houben, I., 2003. Diagnostisches und Statistisches Manual Psychischer Störungen – Textrevision – DSM-IV-TR Hogrefe, Göttingen, Germany. Schick, A., Wessa, M., Vollmayr, B., Kuehner, C., Kanske, P., 2013. Indirect assessment of an interpretation bias in humans: neurophysiological and behavioral correlates. Front Hum Neurosci 7, 272. Shaw, G.A., Giambra, L., 1993. Task-unrelated thoughts of college-students diagnosed as hyperactive in childhood. Dev Neuropsychol 9 (1), 17-30. 27 Showers, C., 1992. Compartmentalization of positive and negative self-knowledge: keeping bad apples out of the bunch. J Pers Soc Psychol 62 (6), 1036-1049. Sieswerda, S., Arntz, A., Wolfis, M., 2005. Evaluations of emotional noninterpersonal situations by patients with borderline personality disorder. Journal of Behavior Therapy and Experimental Psychiatry 36 (3), 209-225. Sieswerda, S., Barnow, S., Verheul, R., Arntz, A., 2013. Neither dichotomous nor split, but schema-related negative interpersonal evaluations characterize borderline patients. J Pers Disord 27 (1), 36-52. Skirrow, C., Ebner-Priemer, U., Reinhard, I., Malliaris, Y., Kuntsi, J., Asherson, P., 2014. Everyday emotional experience of adults with attention deficit hyperactivity disorder: evidence for reactive and endogenous emotional lability. Psychol Med 44 (16), 3571-3583. Smallwood, J., 2013. Distinguishing how from why the mind wanders: a process-occurrence framework for self-generated mental activity. Psychol Bull 139 (3), 519-535. Smallwood, J., Baracaia, S.F., Lowe, M., Obonsawin, M., 2003. Task unrelated thought whilst encoding information. Conscious Cogn 12 (3), 452-484. Smallwood, J., Fishman, D.J., Schooler, J.W., 2007. Counting the cost of an absent mind: Mind wandering as an underrecognized influence on educational performance. Psychonomic Bulletin & Review 14 (2), 230-236. Smallwood, J., Ruby, F.J.M., Singer, T., 2013a. Letting go of the present: mind-wandering is associated with reduced delay discounting. Conscious Cogn 22 (1), 1-7. Smallwood, J., Ruby, F.J.M., Singer, T., 2013b. Letting go of the present: Mind-wandering is associated with reduced delay discounting. Conscious Cogn 22 (1), 1-7. Smallwood, J., Schooler, J.W., 2015. The science of mind wandering: empirically navigating the stream of consciousness. Annu Rev Psychol 66, 487-518. Stein, M.B., Pinsker-Aspen, J.H., Hilsenroth, M.J., 2007. Borderline pathology and the personality assessment inventory (PAI): An evaluation of criterion and concurrent validity. J Pers Assess 88 (1), 81-89. Torgersen, S., Kringlen, E., Cramer, V., 2001. The prevalence of personality disorders in a community sample. Archives of General Psychiatry 58 (6), 590-596. Tramantano, G., Javier, R.A., Colon, M., 2003. Discriminating among subgroups of borderline personality disorder: an assessment of object representations. Am J Psychoanal 63 (2), 149-175. Trull, T.J., Jahng, S., Tomko, R.L., Wood, P.K., Sher, K.J., 2010. Revised NESARC personality disorder diagnoses: gender, prevalence, and comorbidity with substance dependence disorders. J Pers Disord 24 (4), 412-426. Trull, T.J., Solhan, M.B., Tragesser, S.L., Jahng, S., Wood, P.K., Piasecki, T.M., Watson, D., 2008. Affective instability: measuring a core feature of borderline personality disorder with ecological momentary assessment. J Abnorm Psychol 117 (3), 647-661. 28 Vater, A., Schroder-Abe, M., Weissgerber, S., Roepke, S., Schutz, A., 2015. Self-concept structure and borderline personality disorder: evidence for negative compartmentalization. J Behav Ther Exp Psychiatry 46, 50-58. Veen, G., Arntz, A., 2000. Multidimensional dichotomous thinking characterizes borderline personality disorder. Cognitive Therapy and Research 24 (1), 23-45. Watts, F.N., Macleod, A.K., Morris, L., 1988. Associations between phenomenal and objective aspects of concentration problems in depressed-patients. British Journal of Psychology 79, 241-250. Westen, D., Ludolph, P., Silk, K., Kellam, A., Gold, L., Lohr, N., 1990. Object relations in borderline adolescents and adults: developmental differences. Adolesc Psychiatry 17, 360384. Wilkinson-Ryan, T., Westen, D., 2000. Identity disturbance in borderline personality disorder: an empirical investigation. Am J Psychiatry 157 (4), 528-541. Wittchen, H.-U., Zaudig, M., Fydrich, T., 1997. SKID-II. Strukturiertes Klinisches Interview für DSM-IV [Structural Clinical Interview for DSM-IV Axis II Disorders]. Hogrefe, Göttingen. 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. J Psychiatry Neurosci 36 (6), 402-411. 29 Figure captions Figure 1: (A) Level of reported off-task thoughts, other- and self-related, positive and negative, and future- and past-oriented thoughts (rating scale numbers without unit varying between 0 and 100). (B) Fluctuations between thought probes (squared successive differences). (C) Extremity of ratings (squared deviations from mean ratings). * p < 0.05; # p < 0.10 Figure 2: Scatterplot depicting the relation of the level of negative thoughts to the extremity (squared deviations from mean ratings) of (A) self-related and (B) other-related thoughts, separately for BPD patients and HCs. Intra-individual standard errors are displayed in grey and model predictions from the multi-level model as lines in the respective color. 30 Figure 1 31 Figure 2 32 Tables Table 1: Demographics and clinical characteristics of BPD patients and matched healthy control participants (means and standard deviations or absolute number and percent are given without and with parentheses, respectively). N Gender (female/male) Age Years of education PAI-BOR MDE Lifetime Current Current dysthymia Substance use disorder Any anxiety disorder Current PTSD Any eating disorder Any cluster A PD Any other Cluster B PD Any cluster C PD No psychotropic medication Current medication Antidepressants Atypical antipsychotics Mood stabilizer BPD 27 25/2 32.1 (9.7) 11.3 (2.3) 78.7 (7.5) HC 25 21/4 31.2 (10.1) 12.5 (2.6) 44.4 (10.9) statistic 21 (77.8) 5 (18.5) 1 (3.7) 9 (33.3) 2 (7.4) 9 (33.3) 9 (33.3) 0 1 (3.7) 4 (14.8) 7 (25.9) - - 15 (55.6) 6 (22.2) 2 (7.4) - - Χ2(1) = .33, p > 0.40 t(50) = .35, p > 0.70 t(50) = -1.82, p > 0.07 t(50) = 13.2, p < 0.001 33 Table 2: Differences between BPD patients and HCs in rating levels, fluctuations in ratings, and extremity of ratings as estimated with multi-level modeling. Model parameters for the group predictor are displayed. For full model information, see Supplementary Table 2. off-task other self negative positive past future negative mood positive mood rating levels b S.E. 2.794 6.016 -3.080 5.094 4.693 4.962 32.633 5.063 -27.821 4.704 5.151 4.714 -0.938 5.729 21.536 5.590 -24.178 5.648 p-value 0.644 0.548 0.349 0.000 0.000 0.280 0.871 0.000 0.000 fluctuations in ratings b S.E. p-value 115.281 164.766 0.485 402.132 152.093 0.008 571.535 124.799 0.000 22.618 99.668 0.821 143.081 156.439 0.361 204.707 152.534 0.186 232.985 261.637 0.373 -15.397 117.913 0.896 191.759 104.121 0.070 extremity of ratings b S.E. p-value 168.021 79.384 0.035 320.553 93.109 0.001 468.330 95.780 0.000 320.778 166.038 0.059 67.784 101.421 0.504 211.564 140.037 0.137 82.129 139.681 0.559 -59.828 120.937 0.622 -18.824 173.275 0.914 34 Table 3: Differences between BPD patients and HCs in the relation of rating levels, fluctuations, and extremity of self- and otherrelated ratings to the negativity of thoughts estimated as interactions in mutli-level models. Model parameters for valence predictors and their interaction with group are displayed. For full model information, see Supplementary Table 3. self other self other Predictors negative negative*Group negative negative*Group positive positive*Group positive positive*Group b rating levels S.E. p-value fluctuations in ratings b S.E. p-value extremity of ratings b S.E. p-value 0.143 0.065 0.028 4.958 4.124 0.230 -5.686 2.157 0.009 0.061 0.091 0.503 -6.791 5.172 0.190 8.184 2.706 0.003 -0.035 0.080 0.657 -0.705 5.336 0.895 -5.173 2.277 0.023 0.128 0.104 0.221 -1.389 6.799 0.838 10.491 2.873 0.000 -0.069 0.073 0.344 -6.355 3.974 0.110 1.661 2.004 0.408 0.029 0.093 0.751 8.820 5.128 0.086 -4.954 2.602 0.057 -0.017 0.081 0.836 -8.515 5.524 0.124 1.973 2.264 0.384 0.013 0.104 0.897 11.606 7.037 0.100 -5.681 2.908 0.051 35 Supplement Supplementary Table 1: Principal components analysis showing the component loadings for each rating question. Three components were observed: The affect component positively weighted on positive ratings and negatively on negative ratings, the socio-temporal past other component weighted positivelyon other and past ratings, and the socio-temporal future self component weighted positively on self and future. off-task other self negative positive past future F1 affect 0,273 0,019 -0,073 -0,920 0,940 -0,147 0,022 F2 past-other -0,622 0,797 0,067 0,159 -0,096 0,754 0,101 F3 future-self -0,326 0,284 0,727 0,071 0,001 -0,207 0,866 36 Supplementary Table 2: Differences between BPD patients and HCs in ratings levels, fluctuations in ratings and extremity in ratings as estimated with multi-level modelling off task other self negative positive past Model parameters rating levels Predictors b S.E. Intercept 52.694 5.662 Group 2.794 6.016 Session 10.736 3.384 Sample -1.586 0.425 Intercept 40.773 4.875 Group -3.080 5.094 Session -6,436 3.316 Sample 1.123 0.417 Intercept 31.156 4.309 Group 4.693 4.962 Session 5.672 3.667 Sample 1.056 0.383 Intercept 22.456 3.366 Group 32.633 5.063 Session -2.524 2.804 Sample 0.226 0.299 Intercept 70.455 3.805 Group -27.821 4.704 Session -0.370 2.561 Sample -0.372 0.303 Intercept 23.931 3.916 Group 5.151 4.714 Session -2.105 3.462 Sample 0.592 0.369 p-value 0.000 0.644 0.003 0.000 0.000 0.548 0.057 0.007 0.000 0.349 0.128 0.006 0.000 0.000 0.372 0.450 0.000 0.000 0.886 0.221 0.000 0.280 0.546 0.109 fluctuations in ratings b S.E. 799.759 212.261 115.281 164.766 -61.119 180.354 -9.323 33.358 499.961 191.175 402.132 152.093 -1.858 152.772 58.301 37.207 313.413 156.867 571.535 124.799 212.037 125.356 5.434 30.530 436.059 125.278 22.618 99.668 137.093 100.113 -16.314 24.382 460.652 161.947 143.081 156.439 -177.552 122.879 15.588 25.556 547.808 165.580 204.707 152.534 27.019 144.032 3.573 28.403 p-value 0.000 0.485 0.735 0.780 0.009 0.008 0.990 0.118 0.046 0.000 0.091 0.859 0.001 0.821 0.171 0.504 0.005 0.361 0.152 0.542 0.001 0.186 0.852 0.900 extremity of ratings b S.E. 500.623 100.018 168.021 79.384 258.725 79.688 -9.796 17.333 902.185 84.858 320.553 93.109 -261.131 61.667 12.894 18.744 757.377 81.636 468.330 95.780 -151.509 75.389 -4.759 18.250 866.964 132.635 320.778 166.038 -158.732 56.232 -2.854 12.447 898.414 80.079 67.784 101.421 -206.513 90.510 -2.580 18.857 719.115 127.268 211.564 140.037 -47.705 117.659 2.618 12.985 p-value 0.000 0.035 0.001 0.572 0.000 0.001 0.000 0.493 0.000 0.000 0.047 0.795 0.000 0.059 0.005 0.819 0.000 0.504 0.026 0.891 0.000 0.137 0.687 0.840 37 future negative mood positive mood Intercept Group Session Sample Intercept Group Session Sample Intercept Group Session Sample 43.085 -0.938 -2.680 0.941 27.779 21.536 0.796 1.063 70.052 -24.178 -3.418 -1.059 5.202 5.729 3.461 0.426 4.603 5.590 2.621 0.264 4.694 5.648 2.611 0.236 0.000 0.871 0.442 0.027 0.000 0.000 0.762 0.000 0.000 0.000 0.197 0.000 903.833 232.985 294.698 -36.194 411.628 -15.397 72.666 -8.574 88.677 191.759 129.848 18.717 258.846 261.637 192.771 36.399 141.951 117.913 105.024 23.917 94.822 104.121 69.235 25.114 0.000 0.373 0.127 0.320 0.004 0.896 0.492 0.720 0.351 0.070 0.072 0.458 1068.024 82.129 -58.920 -3.418 1146.016 -59.828 -215.965 -4.396 959.761 -18.824 -242.025 4.291 128.088 139.681 120.316 13.240 93.681 120.937 167.026 9.926 139.883 173.275 97.640 10.328 0.000 0.559 0.626 0.796 0.000 0.622 0.196 0.658 0.000 0.914 0.016 0.678 38 Supplementary Table 3: Differences between BPD patients and HCs in the relation of rating levels, fluctuations and extremity of selfand other-related ratings to the negativity of thoughts as estimates as interactions in mutli-level models Model parameters self other self other Predictors Intercept Group Session Sample negative negative*Group Intercept Group Session Sample negative negative*Group Intercept Group Session Sample positive positive*Group Intercept Group Session Sample positive positive*Group rating levels b S.E. p-value fluctuations in ratings b S.E. p-value extremity of ratings b S.E. p-value 27.570 3.870 0.000 211.055 178.000 0.236 818.476 89.658 0.000 -3.086 6.882 0.658 785.220 235.179 0.001 282.939 132.881 0.034 6.656 4.617 0.155 192.240 126.606 0.129 -123.647 84.836 0.153 1.020 0.367 0.006 6.196 30.585 0.840 -0.006 12.993 1.000 0.143 0.065 0.028 4.958 4.124 0.230 -5.686 2.157 0.009 0.061 0.091 0.503 -6.791 5.172 0.190 8.184 2.706 0.003 41.200 4.812 0.000 501.601 239.494 0.036 1003.133 96.490 0.000 -8.474 6.348 0.185 511.525 344.074 0.137 -81.833 141.267 0.563 -5.994 3.427 0.086 -13.458 156.009 0.931 -231.401 61.598 0.000 1.092 0.418 0.009 66.339 34.433 0.054 11.820 18.357 0.521 -0.035 0.080 0.657 -0.705 5.336 0.895 -5.173 2.277 0.023 0.128 0.104 0.221 -1.389 6.799 0.838 10.491 2.873 0.000 36.277 6.803 0.000 771.579 327.092 0.019 638.940 164.971 0.000 1.463 6.911 0.832 33.401 328.531 0.919 713.431 186.147 0.000 5.523 2.810 0.049 188.896 125.816 0.134 -141.482 84.208 0.098 1.004 0.378 0.008 3.701 30.555 0.904 -5.392 16.864 0.749 -0.069 0.073 0.344 -6.355 3.974 0.110 1.661 2.004 0.408 0.029 0.093 0.751 8.820 5.128 0.086 -4.954 2.602 0.057 41.894 7.775 0.000 1097.332 454.484 0.016 759.051 183.415 0.000 -3.685 8.227 0.654 -370.597 488.916 0.449 610.644 199.183 0.002 -6.561 3.825 0.094 17.801 161.792 0.913 -254.415 61.527 0.000 1.117 0.407 0.006 71.014 44.879 0.117 13.209 18.706 0.481 -0.017 0.081 0.836 -8.515 5.524 0.124 1.973 2.264 0.384 0.013 0.104 0.897 11.606 7.037 0.100 -5.681 2.908 0.051 39 Supplementary Figure 1: Figure 3: Level (A) of reported positive and negative mood, as well as (B) fluctuations between probes (squared successive differences) and (C) extremity of ratings (squared deviations from mean ratings). * p < 0.05; # p < 0.10 40