Processing of visual stimuli in borderline personality disorder: A combined behavioural and magnetoencephalographic study

International Journal of Psychophysiology 78 (2010) 257–264

Contents lists available at ScienceDirect

International Journal of Psychophysiology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / i j p s yc h o

Processing of visual stimuli in borderline personality disorder: A combined behavioural and magnetoencephalographic study Angela Merkl a,⁎, Nina Ammelburg a, Sabine Aust b, Stefan Roepke a, Hans Reinecker c, Lutz Trahms d, Isabella Heuser a, Tilmann Sander d a

Department of Psychiatry, Charité, Universitätsmedizin, Campus Benjamin Franklin, Eschenallee 3, 14050 Berlin, Germany Cluster of Excellence “Languages of Emotion” and Dahlem Institute for Neuroimaging of Emotion, Freie Universität Berlin, Habelschwerdter Allee 45, 10957 Berlin, Germany Otto-Friedrich-Universität Bamberg, Fakultät für Humanwissenschaften, Markusplatz 3, 96045 Bamberg, Germany d Physikalisch-Technische Bundesanstalt (PTB), Abbestrasse 2-12, 10587 Berlin, Germany b c

a r t i c l e

i n f o

Article history: Received 28 March 2010 Received in revised form 16 August 2010 Accepted 20 August 2010 Available online 25 August 2010 Keywords: MEG Face perception BPD Emotion Occipital face area M170

a b s t r a c t Background: Behavioral studies on facial emotion recognition yielded heterogeneous results in patients with Borderline Personality Disorder (BPD). Extrastriate cortex hyperactivation has been demonstrated in imaging studies in patients with BPD during face recognition, but electrophysiological studies are lacking. The aim was to investigate temporal processes following face perception in patients with BPD. Methods: Magnetoencephalography (MEG) was used in eleven non-medicated patients with BPD and nine age-matched healthy subjects. Behavioral responses to visual stimuli and an emotion discrimination task were evaluated. First, participants had to silently watch faces, houses and animals. Emotional expressive faces then had to be judged from two basic emotions in a two-alternative forced choice task. Regional field power (RFP) of MEG signals was obtained from two regions of interest: Temporal and occipital areas. Psychometric assessment was performed. Results: Patients with BPD had significantly reduced RFP amplitudes in the right posterior occipital region of interest, for the time window between 150 and 160 ms, irrespective of the type of visual stimulus or the emotional face category. Patients with BPD had significantly higher error rates for recognition of emotional expressive faces compared to healthy controls though they showed a higher accuracy in detecting fearful faces. Controls improved during face recognition, whereas patients showed no learning effect. Conclusion: This MEG study provides evidence for disturbances in cortical visual perception in BPD patients regardless of emotional salience of the stimulus. In line with previous studies subtle deficits in visual perception might be related to impairment in interpersonal communication in BPD. © 2010 Elsevier B.V. All rights reserved.

1. Introduction Borderline Personality disorder (BPD) is a serious psychiatric disorder and the most prominent features are emotional instability, interpersonal disturbances, chronic emptiness and chronic suicidal tendencies (Linehan et al., 1993). BPD patients have a pattern of high psychiatric service use. The debate regarding the neurobiological mechanisms underlying BPD is still ongoing, and therefore identification of biomarkers would be a major step toward improving diagnostic accuracy and identifying therapeutic targets (Stanley and Siever, 2010). A common way to study the emotional instability of BPD is the measurement of responses after presentation of emotional stimuli. Faces convey information that is essential for interpersonal interactions and emotional expressions play a central role, since they

⁎ Corresponding author. Tel.: + 49 30 8445 8604; fax: + 49 30 8445 8233. E-mail address: [email protected] (A. Merkl). 0167-8760/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.ijpsycho.2010.08.007

are crucial for inferring the observed person's feelings and intentions. A substantial number of behavioral studies have investigated emotional face recognition in patients with BPD, but they demonstrated heterogeneous results. So far, researchers used static images, i.e. Ekman faces and in these studies, BPD patients were able to correctly identify emotional expressions at times more accurately than healthy controls (Domes et al., 2008; Lynch et al., 2006; Wagner and Linehan, 1999). However, when facial emotion recognition tasks present more complex situations, by setting time limits for recognizing emotions in faces (Dyck et al., 2009), or with additional prosodic information (Minzenberg et al., 2006), patients with BPD showed increased error rates. The study by Lynch et al. (2006) used a morphing affect recognition paradigm with several emotion intensities and showed a “hyper-responsiveness” for patients with BPD especially towards fearful faces (Lynch et al., 2006). Imaging studies in BPD have shown fusiform gyrus, as well as amygdala, inferior and middle temporal cortical areas hyperactivity during emotional face expression tasks (Guitart-Masip et al., 2009; Herpertz et al., 2001).

258

A. Merkl et al. / International Journal of Psychophysiology 78 (2010) 257–264

To date, the temporal dynamics of brain responses to emotional faces in BPD still remain unclear (Vuilleumier and Pourtois, 2007). Electrophysiological investigations in the primary visual or extrastriate cortex concerning the fundamental, visual-perceptive process of patients with BPD are sparse. An opportunity to measure the distinctive temporal electrophysiological responses towards face processing in vivo and non-invasively exists in the form of magnetoencephalography (MEG), a technique known for its excellent temporal resolution. A face-specific, event-related magnetic field (ERF) peaking at 170 ms (M170) post stimulus and recorded at posterior sensors corresponds to the right inferior occipito-temporal cortex (Deffke et al., 2007; Halgren et al., 2000). An even faster cortical response to the coarse recognition of faces in healthy subjects has been found as early as 100 ms (M100) (Liu et al., 2002). The “fusiform face area” (FFA) has been defined as an extrastriate module for face perception (Kanwisher et al., 1997). Posterior to the FFA, the “occipital face area” (OFA) (Gauthier et al., 2000), has been observed to respond preferably to faces than objects. The rOFA has been shown to be mandatory for recognition of face parts and emotional expressions after temporary lesions were produced in this area by applying repetitive transcranial magnetic stimulation (rTMS) (Pitcher et al., 2008). Thus, the purpose of our study was to investigate if BPD patients, compared with healthy controls, show electrophysiological differences in the early visual processing pathways, represented as M170 in the occipito-temporal visual cortices areas while processing faces and other stimuli. The study is focused on the early visually evoked fields as they are known to yield stable MEG responses. The perception of facial expression implicates a large neural network (Adolphs et al., 2003) and the investigation of the ventral visual pathway in patients with BPD contributes to the understanding of an affect-sensitive network, which has not been assessed so far by MEG. The current study implemented two visual processing tasks using different kinds of response formats. In the first task, subjects had to silently watch three stimulus categories: faces, houses and animals. With this simple task, we investigated electrophysiological responses to basic visual stimuli. The second task was an emotion-specific task in which subjects had to judge emotions within a distinct time limit. On the electrophysiological level, we expected differences between the two groups but had a non-directional differential hypothesis due to lacking prior MEG data with BPD patients. As BPD patients seem to have problems predominantly with the perception of the emotions fear and anger (Koenigsberg et al., 2002; Levine et al., 1997), we hypothesized differences in the perception of those negative emotions in patients with BPD compared to healthy controls in the behavioral task. Further, to address the issue of patients with BPD being a heterogeneous diagnostic group, we controlled for comorbid diagnoses, especially for comorbid post-traumatic stress disorder (PTSD) and major depression. Finally, we controlled for dissociation by implementing a dissociation rating during the MEG measures. 2. Materials and methods 2.1. Participants Thirteen right-handed, female patients with BPD (range of age: 20 to 38 years) were recruited for the study from consecutively admitted inpatients to the Charité, Department of Psychiatry, Berlin University Medicine. Eleven right-handed, female healthy controls (range of age: 25 to 41 years) without any history of neurological or psychiatric disorders were carefully selected and age-matched (Oldfield, 1971). Healthy controls were interviewed by a brief, semi-structured psychiatric questionnaire to ensure that the subject had no former or current psychiatric history and were free of psychotropic medication. The healthy controls were recruited by newspaper advertisement and received compensation.

Patients fulfilled DSM-IV-TR criteria for BPD, assessed by the appropriate segment and German translation of the Structured Clinical Interview for DSM-IV-TR for Personality Disorders (SCID-II; Fydrich et al., 1997; Spitzer et al., 1992). At the department, the diagnosis of BPD was established by two independent clinicians using SCID-II (Spitzer et al., 1992), as well as a two week observation phase in the ward and an interview with a family member of the patient was completed. At the time of inclusion in the study, the patients were in the first week of a 12-week scheduled dialectic behavioural therapy (DBT) program (Bohus et al., 2004; Linehan et al., 1999). Axis I comorbidity was assessed by the Structured Clinical Interview for DSM-IV Axis I Disorders (MINI-SCID; (Ackenheil et al., 1999; Sheehan et al., 1998)). Exclusion criteria were schizophrenia, major depression (current not lifetime), substance abuse and neurological disorders in the previous 6 months and psychopharmacologic medication during the past two weeks. Patients had been free of psychotropic medication for a minimum of 2 weeks prior to MEG examination to ensure that medication was not a confounding effect on the electrophysiological and behavioral measures. Medication prior to study inclusion had been selective serotonine reuptake inhibitors (SSRIs) (30.8%, escitalopram), tricyclic antidepressants (15.4%, imipramine, 7.6%, doxepine) and atypical neuroleptics (15.4%, quetiapine). Four patients (30.8%) were drug naive. Five patients out of the 13 took a combination of SSRIs and atypical neuroleptics. Patients and controls were thoroughly informed before the study and gave written consent assent before screening. The investigation was carried out in accordance with the latest version of the Declaration of Helsinki. The Local Ethics Committee of the Charité approved the study. MEG results of two patients and two controls were discarded due to head movements and artifacts in the MEG scanner leading to 11 patients and 9 controls included for further analysis. 2.2. Behavioral tasks The study consisted of two experiments. In the first experiment, subjects had to silently watch a randomly presented sequence of single “pictures of facial affect”, selected from Ekman and Friesen's standardized sets (Ekman, 1993). The faces presented were of basic emotions (happiness, anger, fear, disgust, surprise, and sadness) and of neutral faces. Emotional expressive faces were presented in equal numbers. The face pictures were randomly intermixed with single pictures of two control conditions (houses and animals), houses previously described in a study (Lueschow et al., 2004). All pictures were presented for 300 ms each with an inter-stimulus-interval (ISI) of 800 ms and a 800 ms cross-hair pattern at the beginning of each trial. For each of the three stimulus categories 65 images were presented leading to a total of 195 images. Task instruction was given prior to the experiment outside the MEG room and consisted of silently watching the pictures in the first experiment. In the second experiment, a set of 112 faces comprising pictures of 16 male faces and 16 female faces with neutral or emotional expressions were presented. Each trial started with a cross-hair pattern, presented for 800 ms, followed by the face stimulus appearing for 300 ms, which was followed by an interval of 800 ms duration blank screen. Subjects were asked either to fixate the crosshair or to watch the face. On the subsequent screen, two words were presented for 2000 ms and subjects were asked to decide between two words (one correct word and one distracter word, i.e. anger, fear, happiness, surprise, disgust, sadness, and neutral) which described best the emotional expression and press a button. In order to record the response and control for task compliance, we asked subjects to use a mouse button which was placed at each side next to the hips. Subjects were asked to indicate the valence of the facial expression within these two seconds by pressing one of two buttons with the index finger of the right or left hand. The side for the button press for a

A. Merkl et al. / International Journal of Psychophysiology 78 (2010) 257–264

correct answer was randomized across subjects to control for lateralization effects. The task was designed as a two-alternative forced choice task (“respond as quick as possible”) and subjects received a 2000 ms lasting feedback on accuracy: “right” or “wrong”. In total, there were 6 blocks with 672 decisions per subject and the second experiment lasted 38.3 minutes in total. Block and trial order was randomized, but each participant evaluated the same set of Ekman faces. Scores were calculated as the total number of correct discrimination for all items. After the first experiment a pause of 5 min was done. Hence, both experiments lasted approximately 55 min. The whole experiment lasted about 2 h including pauses, electrode placements and ratings. Subjects were laying supine with their head inside the MEG helmet viewing images projected onto a glass panel (approximately 9 in. screen) at 20 centimeters (cm) distance. The size of the presented stimuli on the screen was 1.9 × 27 in. 2.3. Dissociation self-rating during MEG recording A visual analogue scale to rate dissociation was used for the patients with BPD during the two experiments in order to control for attention; the scale indicated “no tension” as 0% and “extreme tension” as 100% (“Could you please indicate a number between 0 and 100 to describe the level of your tension?”). The dissociation ratings were applied to the patients in order to ensure that the level of arousal did not interfere with their task performance. Patients spoke via a microphone with the experimenter in the control room outside the MEG room. If a self-rating tension/dissociation score was over 70%, the experiment was interrupted. BPD patients enrolled in the DBT program were used to regularly rate their inner tension on the ward. The dissociation rating was done after the first experiment. Healthy controls were asked if they were ready to continue after the five minute pause between experiment one and two. 2.4. Psychometric instruments Depressive symptoms were assessed with the Hamilton Depression Rating Scale, 17-item Version (HAMD-17) (Hamilton, 1960). Traumatic experiences were assessed with the German version of the Posttraumatic Distress Scale (PDS) with 17 items (Foa and Tolin, 2000). We applied the Borderline Symptom List, a self-rating instrument to specifically quantify borderline-typical symptoms (BSL) (Bohus et al., 2001). The HAMD-17 and PDS questionnaires were administered to control for confounding variables. The tests were undertaken before the MEG session in order to identify patients with severe depressive or posttraumatic symptoms. 2.5. MEG acquisition MEG was recorded in patients and healthy controls using a 93channel whole-head MEG system ET160 (Eagle Technology, Japan). Participants would lie supine, facing a semitransparent glass panel serving as the projection target of a computer beamer (Sanyo XF 12, Japan) outside of the magnetically shielded room. In addition to the MEG channels we sampled at 500 Hz, further electrocardiographic (ECG) and two electroocculographic (EOG) channels were recorded using a low-noise biosignal amplifier (Custom made, PTB) (Scheer et al., 2006). The physiological arousal (heart rate) as an objective criterion of distress and tension was recorded in both groups. Response times (RT) and performance (error percentage) were registered on the stimulation PC running Presentation Version 11.3 (Presentation, Neurobehavioral Systems, Inc.). The offline data analysis started with bandpass filtering from 0.5 to 45 Hz. The independent component analysis (ICA) was then applied to remove artifacts such as heart beat and eye movements (Sander et al., 2007). The remaining artifacts were spike-like and were scanned using a

259

threshold of 5% of the epochs for rejection. Two heart beat and two eye movement ICA components were removed similarly in the patient and healthy control group and the groups did not differ significantly in the number of rejected epochs with spike-like artifacts. Subsequently, the evoked epochs were scanned for spike like artifacts and contaminated epochs were rejected. The event related fields (ERFs) were calculated only for epochs where a correct answer was given. The ERFs were baseline corrected using an interval of −200 ms to 0 ms relative to the start of image presentation. A regional field power (RFP) analysis was performed calculating the root-mean-square of the ERFs as a function of time for regions of interest (ROI), consisting of groups of 6 to 15 MEG channels. The ROIs represent occipital left/right and temporal left/right regions of the sensor area (Fig. 1). This choice was motivated by the visual M100 located occipitally and the slightly later face selective M170 located more temporally. For MEG, the fixed geometry of the sensor relative to the variable individual anatomy leads to subject-dependent distances between the MEG sensor and the cortical layer. To remove variability, the RFP curves for each condition and each subject were rescaled to have a mean of unity in the window from 80 to 120 ms post stimulus. Before calculating these relative RFP curves, the existence of a M100 response was assured. In performing normalization in the window mentioned above, we were removing any inter-subject as well as inter-condition variability in the M100. Rescaling to unity for a certain time window meant here that we were looking at relative changes in signal strength at times before and after the time window chosen. 2.6. Statistical analysis All statistical analyses were performed using SPSS version 17.0 (SPSS Inc., Chicago, IL, USA). The clinical and behavioral data were analyzed for 13 patients and 11 controls. The student's t-test for continuous data with a 2-tailed significance at p b 0.05 was used. A one-factorial (Group) MANOVA with age as covariate for the variables “correct answers” and “percentage of wrong answers” was calculated. T-tests for differences between response times of each single emotion were performed. Differences between groups for error rates for each single emotion were calculated with a Mann–Whitney U-Test, since the data was not normally distributed. Spearman's coefficients of correlation were calculated for relations between clinical characteristics and behavioral data. MEG analysis was done with 11 patients and 9 controls. The RFP curves were segmented into adjacent 10 ms time windows and a mean RFP value was calculated for each curve within this window. Using these segmented data within an exploratory analysis we used pairwise t-tests comparing the groups in the time windows ranging from 120 to 220 ms. The groups were compared by stimulus category (face, house, animal) in the first experiment and by emotion category in the second experiment. The windows for t b 120 ms were excluded as the group variance of the relative RFP curves was by definition

Fig. 1. Definition of the channel groups for the regional field power analysis. The solid outline is a top view on a two dimensional rendering of the sensor area, nose at the top as indicated and neck at the bottom. The occipital groups are most sensitive to the primary visual areas (M100), the temporal groups to later processing stages (M170).

260

A. Merkl et al. / International Journal of Psychophysiology 78 (2010) 257–264

nervosa and three patients reported alcohol abuse in the past. Eight patients had a comorbid Axis II. Cluster C personality disorder in seven patients (avoidant) and one patient had a comorbid Cluster B personality disorder (narcissistic).

Table 1 Demographic characteristics and clinical characteristics.

Demographic data Age, yrs Female sex (%) Education (years) Clinical data Comorbid diagnoses Axis-I BSL HAMD-17 PDS

BPD (n = 13)

Healthy controls (n = 11)

26.2 ± 5.2 13 (100) 10.5 ± 1.76

30.3 ± 5.6 11 (100) 10.8 ± 4.85

2.23 ± 0.83

–

236.3 ± 51.7 7.85 ± 1.81 31.0 ± 14.3

23.6 ± 11.3 2.27 ± 1.11 0.55 ± 1.21

T

df

P-value

1.87

22

0.07

1.81

22

0.86

−14.9 −8.86 −7.65

22 22 22

3.2. Behavioural results: accuracy and response times

P b 0.001 P b 0.001 P b 0.001

Data are presented as mean and standard deviation (SD) if not otherwise specified. Abbreviations: BPD = borderline personality disorder patients; BSL = borderline symptom list; HAMD-17 = Hamilton depression rating scale-17 item version; PDS = posttraumatic distress scale (global score subscale), df = degrees of freedom.

reduced in comparison to the variance in the later time windows. For the time window t = 150–160 ms and the occipital right ROI, the exploratory t-test yielded a group effect for all categories in the first experiment and for most emotions in the second experiment. Therefore this window was chosen for further analysis, using an ANOVA with repeated measures. For the first experiment, repeated measures were “object category” (Face, House, and Animal) with factor “subject group” (Patient/Healthy Control). The repeated measures for the second experiment were the seven categories “emotion” and again the factor “subject group”. ANOVAs were calculated for the four combinations of right and left hemisphere and temporal and occipital ROI. The statistical inclusion of “hemisphere” is not considered as a repeated measure nor a classical factor linked to the experimental design (hemispheres are not independently stimulated from the each other). The spatial resolution of MEG allows separating hemispheric specific effects. Therefore a separate omnibus ANOVA with factors “subject group” and “hemisphere” and the aforementioned repeated measures was calculated for the first and second experiment data for temporal and occipital ROI. This omnibus ANOVA tests for complex interactions between hemispheres in relation to the patient and healthy control groups. We calculated effect size and power estimate. 3. Results 3.1. Demographics Demographics and clinical characteristics are shown in Table 1. Three patients fulfilled co-existing Axis I disorder criteria for lifetime major depressive disorder, two patients fulfilled criteria for bulimia

Patients compared to controls yielded a significant higher total error rate in the key responses to emotional expressive faces when analyzing all emotion decisions and correct answers (Table 2). Comparing each emotion separately, patients with BPD rated fearful faces significantly more accurately (Z = −2.058, p = 0.04) exhibiting fewer errors than healthy controls (Fig. 2). In the single emotion group comparisons no significant difference for error percentage was observed for the recognition of the emotion anger (Z = −0.985, p = 0.325), sadness (Z = −0.754, p = 0.451), surprise (Z = −0.812, p = 0.417), happiness (Z = −0.811, p = 0.417), disgust (Z = −0.435, p = 0.664) and neutral faces (Z = −0.985, p = 0.325). We did not observe a significant difference in response time (RT), as well as error percentages of late responses (defined as button press after 2000 ms) between patients with BPD and controls (Table 2). The mean number of errors between run 5 and run 1 (experiment 2) did not differ in patients with BPD (Z = −1.381, p = 0.167) but in healthy controls (Z = −2.848, p = 0.004). Age included as a covariate did not influence correct decisions of facial expressions (F (df) = 0.54 (1); p = 0.47). 3.3. Psychometric results and correlations Differences between groups in BSL, HAMD-17 and PDS are summarized in Table 1. Scores of correct answers during the recognition of emotional expressive faces correlated significantly with disease severity grade in the BSL questionnaire (r = −0.60, p = 0.002). 3.4. MEG results The relative RFP amplitudes for the occipital right ROI are shown in Fig. 3 for the first experiment. An ANOVA calculated for these occipital ROI right hemisphere data yielded a highly significant main effect for “group” (F1, 18 = 7.642, P = 0.001, η2 = 0.453, 1 − β = 0.95) with large effect size and power. Lower amplitude for patients with BPD was consistently observed for all three stimulus categories (Fig. 3) in support of the main effect between “groups”. For the temporal ROIs no significant effect was found. The omnibus ANOVA for the occipital ROI data with repeated measure “object category” showed a significant main effect between “groups” (F1, 36 = 5.161, p =0.029, η2 = 0.125, 1 − β= 0.60), but not between “hemispheres” (F1, 36 = 0.616, p = 0.437, η2 =0.017, 1 − β=0.12). The interaction effect between “group” and

Table 2 Accuracy of decisions and response times of the face perception task (Experiment 2) together with a statistical evaluation.

Total decisions/counts Correct answers/counts Incorrect answers/counts Delayed answers/counts RT total/ms RT anger/ms RT disgust/ms RT fear/ms RT happiness/ms RT sadness/ms RT surprise/ms RT neutral/ms

Patients with BPD (n = 13)

Healthy controls (n = 11)

T

df

P-value

585.9 ± 113.4 483.7 ± 102.5 91.00 ± 80.08 11.23 ± 22.90 815.7 ± 145.9 870.8 ± 156.8 854.8 ± 176.8 912.7 ± 191.1 690.9 ± 119.2 845.1 ± 140.2 763.5 ± 184.2 772.4 ± 126.3

661.8 ± 33.77 611.8 ± 37.51 47.36 ± 5.27 2.64 ± 1.63 793.4 ± 100.8 833.9 ± 136.3 895.8 ± 104.2 927.6 ± 124.3 663.9 ± 82.5 787.1 ± 133.5 700.5 ± 132.0 749.5 ± 96.4

2.29 4.18 −1.92 −1.35 −0.43 −0.61 −0.68 −0.22 −6.35 −1.03 -0.95 −0.92

22 22 22 22 22 22 22 22 22 22 22 22

0.03 0.001 0.07 0.20 0.67 0.55 0.51 0.83 0.53 0.63 0.31 0.36

Abbreviations: BPD = borderline personality disorder patients; RT = response times; SD = standard deviation; df = degrees of freedom; n = number; bold P-values = significant.

A. Merkl et al. / International Journal of Psychophysiology 78 (2010) 257–264

261

Fig. 2. Behavioral data: Accuracy of emotion recognition in BPD as a function of error percentages of rated emotional expressive faces and neutral faces. BPD patients showed significantly (p b 0.05) higher accuracy in detecting fearful faces than controls. (black bars = controls; white bars = BPD).

“hemisphere” was close to significance (F1, 36 = 3.749, p = 0.061, η2 =0.094, 1 −β =0.47). For the second experiment, the relative RFP amplitudes for the occipital right ROI are shown in Fig. 4. As in the first experiment, lower amplitudes were observed for the patients compared to the controls in the second experiment. An ANOVA calculated for the occipital ROI data of the right hemisphere yielded a significant main effect for “group” (F1, 18 = 4.653, p = 0.045, η2 = 0.205, 1 − β = 0.53). There was no effect found for category “emotion” or the “group” and “emotion” interaction in the right occipital ROI data. No significant effects were found for the left hemisphere occipital ROI data and the right and left temporal ROI data of the second experiment. The omnibus ANOVA for the occipital ROI data with repeated measure “emotion category” and factors “group” (F1, 36 = 2.163, p = 0.15, η2 = 0.057, 1 − β = 0.29) and “hemisphere” (F1, 36 = .773, p = 0.39, η2 = 0.021, 1 − β = 0.14) showed no effect.

Fig. 4. MEG results: Normalized regional field power curves as a function of time for emotional expressive faces for the second experiment. Significant group amplitude differences were observed at a latency of 150 ms to 160 ms. Calculations are for sensors occipital right. Dashed line = controls; black line = patients; RFP = regional field power, a./u. = arbitrary units.

3.5. ECG data Fig. 3. MEG results: Normalised regional field power curves as a function of time for face, house and animal. The curves show an increase at 70 ms in MEG activity relative to the baseline RFP levels. Around 100 ms the curves intersect as a natural consequence of the normalization. After 120 ms the curve of the controls increases in contrast to the patient's curve. A maximum difference was reached around t = 155 ms. A significant group amplitude difference is found at a latency of 150 ms to 160 ms. Calculations are for sensors occipital right. Dashed line = controls; black line = patients. RFP = regional field power, a./u. = arbitrary units.

Mean heart rate in patients during the first experiment was 75.5 Hz (SD ±10.23) and 68.5 Hz (SD ±7.45) in controls. There was no significant difference between groups found (t (17) = −1.673, p = 0.113). During the second experiment patients had a mean heart rate of 76.40 Hz (SD ±10.31) and of 70.22 Hz (SD ±7.52) in controls, respectively, which was not statistically significant (t (17) = −1.475, p = 0.158).

262

A. Merkl et al. / International Journal of Psychophysiology 78 (2010) 257–264

3.6. Dissociation self-rating during MEG recording 20% of the BPD patients reported increasing dissociation (up to 70% during the experiment, 60% percent indicated a decreasing dissociation and 20% reported a stable state throughout the experiment. Healthy controls reported no tension whatsoever. 4. Discussion This is the first study to our knowledge that investigated the course of the M170, the face-specific event-related magnetic field peaking at 170 ms, in patients with BPD. There were the following main findings in the study. Patients with BPD showed consistently lower relative M170 amplitudes than controls in all three stimulus categories in the first experiment leading to a significant group difference. The pattern of reduced activity in the relative M170 amplitude in patients with BPD was found irrespective of emotions in the second experiment. A significant group difference between patients and controls was observed. On the behavioral side, compared to healthy subjects, patients with BPD made less correct decisions compared to the total amount of decisions concerning the emotional expression of faces. This was related to the severity of BPD symptomatology in terms of subjective strain in such a way that high score on the BSL correlated with less correct decisions. Patient with BPD also had a subtle bias towards fearful faces, recognizing those more accurately than healthy controls did. Patients and healthy controls responded equally fast in terms of response times. Patients compared to healthy controls showed no improvement for the recognition of emotional expression in faces over the course of the first to the fifth and last session of the experiment. 4.1. MEG data BPD patients showed attenuation in relative amplitude of the M170 in all three stimulus categories. We did not find a distinct emotion specific effect or a modulation of the M170 by the emotionally competent stimuli in BPD patients compared to healthy subjects. To date, emotional modulations on extrastriatal visual areas in healthy subjects (Surguladze et al., 2003; Vuilleumier and Pourtois, 2007) and borderline patients (Herpertz et al., 2001) have only been shown with fMRI. Our results are in line with EEG/ERP studies that have shown the sensitivity of the N170 amplitude to stimulus content (Eimer and Holmes, 2007; Halgren et al., 2000). This suggests that patients with BPD may have subtle deficits in the processing of social and environmental stimuli, irrespective of the range and valence of emotional expressions. In addition, no MEG studies are available addressing this issue in patients with BPD. Interestingly, an MEG study with schizophrenic patients demonstrated a diminished activation in the right posterior fusiform area while face processing (Streit et al., 2001). It could be speculated that MEG alterations are associated with psychiatric diseases in an unspecific way; however future research is needed to investigate larger samples. In line with previous electrophysiological results and a study in a patient with right occipital face area lesion (rOFA) (Eimer and Holmes, 2007; Rossion et al., 2003a,b; Watanabe et al., 1999), group differences were significant in the right hemisphere only. This suggests that patients with BPD demonstrate subtle alterations in the processing of faces and other socially important stimuli at a very early perceptual processing level. In turn, MEG studies in schizophrenic patients have shown a right lateralized sustained activity in the amygdala while watching emotional faces (Ioannides et al., 2004), however the MEG method used there and ROI was different to our study. Borderline patients showed decreased M170 amplitude not only in the face category but in the two other categories tested, houses and animals, compared to those of controls. This is in line with a number of studies, which propose a domain-specific hypothesis arguing that

mechanisms engaged by faces are unspecific and represent a process which can be triggered by different stimuli classes. In a study with healthy subjects using MEG (Hadjikhani et al., 2009), it is found that objects incidentally perceived as faces evoked an early (165 ms) activation in the ventral fusiform cortex, at a time and location similar to that evoked by faces. In addition, activation in “face responsive” areas by non-facial stimuli like birds or cars has been described for “familiarized experts” in the respective category (Gauthier et al., 2000). Probably most important for the present study, it was shown that perceptual load manipulation modulates the N170 up to the degree of abolishing face selectivity (Mohamed et al., 2009). This domain specific hypothesis challenges the longstanding face-specificity hypothesis, which proposes that face perception is a highly specialized process dedicated to face perception per se and lesion studies show a double dissociation between the recognition of faces and objects (Kanwisher and Yovel, 2006). Clearly, we cannot completely rule out that MEG findings were due to an impaired vigilance in the BPD patients. However, the M170 amplitude difference between groups occurred in both experiments, while subjects were silently watching a face and while subjects were performing a forced-choice task which guided attention towards the emotional expression in faces. Furthermore, we controlled for attentional or motivational confounds by performing a tension rating during the experiment. In addition, heart rate as an indicator for physiological arousal and distress (Ebner-Priemer et al., 2008) was not decreased in patients. Of note is, patients did not respond significantly slower or faster than healthy controls to the stimuli suggesting that they were focused on both tasks. 4.2. Behavioural data Borderline patients were less accurate for all the emotional and neutral faces but demonstrated a bias towards fearful faces more than did controls. This is in line with previous studies indicating that Borderline patients are hyperresponsive to fearful faces (Lynch et al., 2006). However, contradictory to a study that investigated the “Reading the Mind in the Eyes” test (Fertuck et al., 2009), out patients were generally less accurate in all other emotion decisions. BPD patients did not improve in performance over the course of the first to the last session compared to healthy subjects in the recognition of emotional expressive faces. Given that our paradigm was not a learning experiment per se, further studies on emotional learning in patients with BPD would be mandatory to prove emotional learning deficits. A possible explanation could be that patients with BPD show decreased emotional control while responding to faces that is mediated by prefrontal brain structures. In line with this, decreased ventromedial prefrontal activity, compared to healthy subjects, in the context of negative emotions has been found in this patient group earlier (Silbersweig et al., 2007). 4.3. Methodological considerations Some confounding factors might limit the degree to which these results can be generalized. Sample size was small and the main result of reduced relative M170 amplitude has to be replicated. However, studies addressing electrophysiological changes in BPD with unmedicated patients are scarce. Moreover, patients with BPD had comorbid lifetime recurrent depressive episodes. This might increase the possibility that individual scores had a disproportionate influence, decreasing the detectability of group differences. We accounted for this potential influence by controlling the severity of current depressive symptoms with the HAMD-rating and rating of the traumatic symptoms before we included the patients in the study. The rate of comorbid Axis I disorders observed in the present study was comparable to that found in other clinical samples (Zanarini et al., 1998; Donegan et al., 2003) and BPD samples without additional Axis I

A. Merkl et al. / International Journal of Psychophysiology 78 (2010) 257–264

disorders are highly uncharacteristic, and cannot be considered representative of BPD patients (Skodol et al., 2002). Age was close to significance between our groups and it was included as a covariate in the group analysis of the face recognition experiment, but no influence of age on the result was found. Finally, future research with MEG should involve a psychiatric control group and biomagnetic source reconstruction, to further delineate the cerebral sources of this effect. 5. Conclusion In conclusion, our MEG study provides evidence for disturbances in early visual perception in patients with BPD. This effect seems to be lateralized to the right hemisphere and it is observed irrespective of emotion category and extends to other stimulus categories. Impaired perception of facial expression and a heightened sensitivity towards fearful faces suggests a basis for the social difficulties in BPD. Although the number of participants was fairly low the main result of reduced relative M170 amplitude might indicate a future direction of a biological assessment for BPD patients. Finally, the results demonstrate the potential of MEG for investigating psychiatric diseases. Conflict of interest statement None declared. Acknowledgements Supported by Bundesministerium für Bildung und Forschung (BMBF) Grants 01 GO 0208 and 01 GO 0503, Berlin Neuroimaging Center (BNIC) to T.S. References Ackenheil, M., S.-I. G., Dietz-Bauer, A., Vossen, A., 1999. German Version of the MiniInternational Neuropsychiatric Interview (M.I.N.I.): The Development and Validation of a Structured Diagnostic Psychiatric Interview for DSM-IV and ICD-10. Psychiatrische Universitätsklinik, München. Adolphs, R., Tranel, D., Damasio, A.R., 2003. Dissociable neural systems for recognizing emotions. Brain Cogn. 52 (1), 61–69. Bohus, M., Limberger, M.F., Frank, U., Sender, I., Gratwohl, T., Stieglitz, R.D., 2001. Development of the borderline symptom list. Psychother. Psychosom. Med. Psychol. 51 (5), 201–211. Bohus, M., Haaf, B., Simms, T., Limberger, M.F., Schmahl, C., Unckel, C., et al., 2004. Effectiveness of inpatient dialectical behavioral therapy for borderline personality disorder: a controlled trial. Behav. Res. Ther. 42 (5), 487–499. Deffke, I., Sander, T., Heidenreich, J., Sommer, W., Curio, G., Trahms, L., et al., 2007. MEG/ EEG sources of the 170-ms response to faces are co-localized in the fusiform gyrus. Neuroimage 35 (4), 1495–1501. 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. Donegan, N.H., Sanislow, C.A., Blumberg, H.P., Fulbright, R.K., Lacadie, C., Skudlarski, P., Gore, J.C., Olson, I.R., McGlashan, T.H., Wexler, B.E., 2003. Amygdala hyperreactivity in borderline personality disorder: implications for emotional dysregulation. Biol. Psychiatry 54 (11), 1284–1293. Dyck, M., Habel, U., Slodczyk, J., Schlummer, J., Backes, V., Schneider, F., et al., 2009. Negative bias in fast emotion discrimination in borderline personality disorder. Psychol. Med. 39 (5), 855–864. Ebner-Priemer, U.W., Kuo, J., Schlotz, W., Kleindienst, N., Rosenthal, M.Z., Detterer, L., et al., 2008. Distress and affective dysregulation in patients with borderline personality disorder: a psychophysiological ambulatory monitoring study. J. Nerv. Ment. Dis. 196 (4), 314–320. Eimer, M., Holmes, A., 2007. Event-related brain potential correlates of emotional face processing. Neuropsychologia 45 (1), 15–31. Ekman, P., 1993. Facial expression and emotion. Am. Psychol. 48 (4), 384–392. Fertuck, E.A., Jekal, A., Song, I., Wyman, B., Morris, M.C., Wilson, S.T., et al., 2009. Enhanced ‘Reading the Mind in the Eyes’ in borderline personality disorder compared to healthy controls. Psychol. Med. 1–10. Foa, E.B., Tolin, D.F., 2000. Comparison of the PTSD symptom scale-interview version and the clinician-administered PTSD scale. J. Trauma. Stress 13 (2), 181–191. Fydrich, T., R. B., Schmitz, B., Wittchen, H.U., 1997. Strukturiertes Klinisches Interview für DSM-IV. Achse II. Persönlichkeitsstörungen (SKID-II). Hogrefe, Göttingen.

263

Gauthier, I., Skudlarski, P., Gore, J.C., Anderson, A.W., 2000. Expertise for cars and birds recruits brain areas involved in face recognition. Nat. Neurosci. 3 (2), 191–197. Guitart-Masip, M., Pascual, J.C., Carmona, S., Hoekzema, E., Berge, D., Perez, V., et al., 2009. Neural correlates of impaired emotional discrimination in borderline personality disorder: an fMRI study. Prog. Neuropsychopharmacol. Biol. Psychiatry 33 (8), 1537–1545. Hadjikhani, N., Kveraga, K., Naik, P., Ahlfors, S.P., 2009. Early (M170) activation of facespecific cortex by face-like objects. NeuroReport 20 (4), 403–407. Halgren, E., Raij, T., Marinkovic, K., Jousmaki, V., Hari, R., 2000. Cognitive response profile of the human fusiform face area as determined by MEG. Cereb. Cortex 10 (1), 69–81. Hamilton, M., 1960. A rating scale for depression. J. Neurol. Neurosurg. Psychiatry 23, 56–62. Herpertz, S.C., Dietrich, T.M., Wenning, B., Krings, T., Erberich, S.G., Willmes, K., et al., 2001. Evidence of abnormal amygdala functioning in borderline personality disorder: a functional MRI study. Biol. Psychiatry 50 (4), 292–298. Ioannides, A.A., Poghosyan, V., Dammers, J., Streit, M., 2004. Real-time neural activity and connectivity in healthy individuals and schizophrenia patients. Neuroimage 23 (2), 473–482. Kanwisher, N., Yovel, G., 2006. The fusiform face area: a cortical region specialized for the perception of faces. Philos. Trans. R. Soc. Lond. B Biol. Sci. 361 (1476), 2109–2128. Kanwisher, N., McDermott, J., Chun, M.M., 1997. The fusiform face area: a module in human extrastriate cortex specialized for face perception. J. Neurosci. 17 (11), 4302–4311. Koenigsberg, H.W., Harvey, P.D., Mitropoulou, V., Schmeidler, J., New, A.S., Goodman, M., et al., 2002. Characterizing affective instability in borderline personality disorder. Am. J. Psychiatry 159 (5), 784–788. Levine, D., Marziali, E., Hood, J., 1997. Emotion processing in borderline personality disorders. J. Nerv. Ment. Dis. 185 (4), 240–246. Linehan, M.M., Heard, H.L., Armstrong, H.E., 1993. Naturalistic follow-up of a behavioral treatment for chronically parasuicidal borderline patients. Arch. Gen. Psychiatry 50 (12), 971–974. Linehan, M.M., Schmidt III, H., Dimeff, L.A., Craft, J.C., Kanter, J., Comtois, K.A., 1999. Dialectical behavior therapy for patients with borderline personality disorder and drug-dependence. Am. J. Addict. 8 (4), 279–292. Liu, J., Harris, A., Kanwisher, N., 2002. Stages of processing in face perception: an MEG study. Nat. Neurosci. 5 (9), 910–916. Lueschow, A., Sander, T., Boehm, S.G., Nolte, G., Trahms, L., Curio, G., 2004. Looking for faces: attention modulates early occipitotemporal object processing. Psychophysiology 41 (3), 350–360. Lynch, T.R., Rosenthal, M.Z., Kosson, D.S., Cheavens, J.S., Lejuez, C.W., Blair, R.J., 2006. Heightened sensitivity to facial expressions of emotion in borderline personality disorder. Emotion 6 (4), 647–655. Minzenberg, M.J., Poole, J.H., Vinogradov, S., 2006. Social-emotion recognition in borderline personality disorder. Compr. Psychiatry 47 (6), 468–474. Mohamed, T.N., Neumann, M.F., Schweinberg, S.R., 2009. Perceptual load manipulation reveals sensitivity of the face-selective N170 to attention. NeuroReport 20 (8), 782–787. Oldfield, R.C., 1971. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 9 (1), 97–113. Pitcher, D., Garrido, L., Walsh, V., Duchaine, B.C., 2008. Transcranial magnetic stimulation disrupts the perception and embodiment of facial expressions. J. Neurosci. 28 (36), 8929–8933. Rossion, B., Caldara, R., Seghier, M., Schuller, A.M., Lazeyras, F., Mayer, E., 2003a. A network of occipito-temporal face-sensitive areas besides the right middle fusiform gyrus is necessary for normal face processing. Brain 126 (Pt 11), 2381–2395. Rossion, B., Joyce, C.A., Cottrell, G.W., Tarr, M.J., 2003b. Early lateralization and orientation tuning for face, word, and object processing in the visual cortex. Neuroimage 20 (3), 1609–1624. Sander, T.H., Burghoff, M., Van Leeuwen, P., Trahms, L., 2007. Application of decorrelationindependent component analysis to biomagnetic multi-channel measurements. Biomed. Technol. (Berl.) 52 (1), 130–136. Scheer, H.J., Sander, T., Trahms, L., 2006. The influence of amplifier, interface and biological noise on signal quality in high-resolution EEG recordings. Physiol. Meas. 27 (2), 109–117. Sheehan, D.V., Lecrubier, Y., Sheehan, K.H., Amorim, P., Janavs, J., Weiller, E., et al., 1998. The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J. Clin. Psychiatry 59 (Suppl 20), 22–33 quiz 34-57. Silbersweig, D., Clarkin, J.F., Goldstein, M., Kernberg, O.F., Tuescher, O., Levy, K.N., et al., 2007. Failure of frontolimbic inhibitory function in the context of negative emotion in borderline personality disorder. Am. J. Psychiatry 164 (12), 1832–1841. Skodol, A.E., Gunderson, J.G., Pfohl, B., Widiger, T.A., Livesley, W.J., Siever, L.J., 2002. The borderline diagnosis I: psychopathology, comorbidity, and personality structure. Biol. Psychiatry 51 (12), 936–950. Spitzer, R.L., Williams, J.B., Gibbon, M., First, M.B., 1992. The Structured Clinical Interview for DSM-III-R (SCID). I: History, rationale, and description. Arch. Gen. Psychiatry 49 (8), 624–629. Stanley, B., Siever, L.J., 2010. The interpersonal dimension of borderline personality disorder: toward a neuropeptide model. Am. J. Psychiatry 167 (1), 24–39. Streit, M., Ioannides, A., Sinnemann, T., Wolwer, W., Dammers, J., Zilles, K., et al., 2001. Disturbed facial affect recognition in patients with schizophrenia associated with hypoactivity in distributed brain regions: a magnetoencephalographic study. Am. J. Psychiatry 158 (9), 1429–1436.

264

A. Merkl et al. / International Journal of Psychophysiology 78 (2010) 257–264

Surguladze, S.A., Brammer, M.J., Young, A.W., Andrew, C., Travis, M.J., Williams, S.C., et al., 2003. A preferential increase in the extrastriate response to signals of danger. Neuroimage 19 (4), 1317–1328. Vuilleumier, P., Pourtois, G., 2007. Distributed and interactive brain mechanisms during emotion face perception: evidence from functional neuroimaging. Neuropsychologia 45 (1), 174–194.

Wagner, A.W., Linehan, M.M., 1999. Facial expression recognition ability among women with borderline personality disorder: implications for emotion regulation? J. Pers. Disord. 13 (4), 329–344. Zanarini, M.C., Frankenburg, F.R., Dubo, E.D., Sickel, A.E., Trikha, A., Levin, A., Reynolds, V., 1998. Axis I comorbidity of borderline personality disorder. Am. J. Psychiatry 155 (12), 1733–1739.