1. Introduction
Personal space is the area around us perceived as safe, causing discomfort or flight reactions when violated by others. It is a human analog of territorial behavior in animals, also termed the immediate body-buffer zone [Reference Felipe and Sommer11,Reference Horowitz, Duff and Stratton21,Reference Knapp, Hall and Horgan24]. Personal space is critical in nonverbal social communication; it is mediated by social status, familiarity, and attachment style [Reference Hall, Coats and LeBeau15,Reference Hayduk16,Reference Knapp, Hall and Horgan24]. Regulation of personal space is dynamically linked to state and context variables such as size, gender, eye gaze, direction of approach and cultural background [Reference Knapp, Hall and Horgan24].
Personal space regulation as approach-avoidance behavior has received little attention in schizophrenia [Reference de la Asuncion, Docx, Sabbe, Morrens and de Bruijn8]. Tests of real interpersonal distance revealed increased values in schizophrenia [Reference de la Asuncion, Docx, Sabbe, Morrens and de Bruijn8,Reference Deus and Jokic-Begic9,Reference Horowitz, Duff and Stratton21,Reference Park, Ku, Kim, Jang, Kim and Kim30,Reference Sommer35]. In contrast, projective paper and pencil based measures of interpersonal distance yielded inconsistent results [Reference Duke and Mullens10,Reference Nechamkin, Salganik, Modai and Ponizovsky28]. Likewise, the association of personal space with clinical syndromes remains less clear [Reference Deus and Jokic-Begic9], with one study demonstrating increased interpersonal distance with more severe positive symptoms [Reference de la Asuncion, Docx, Sabbe, Morrens and de Bruijn8]. However, no study has investigated whether personal space regulation in schizophrenia is altered as a function of current paranoid threat. In fact, threat leads to increased personal space in healthy individuals [Reference Horowitz, Duff and Stratton21,Reference Knapp, Hall and Horgan24,Reference Sambo and Iannetti34]. Likewise, avoidance behavior including increased personal space is preferred in subjects with high levels of anxiety such as in social anxiety or post-traumatic stress disorders [Reference Bogovic, Mihanovic, Jokic-Begic and Svagelj4,Reference Wieser, Pauli, Grosseibl, Molzow and Muhlberger40].
Amygdala activity is critical for personal space regulation as it corresponds to personal space violations in close proximity in healthy subjects. In fact, a patient with complete bilateral amygdala lesions lacked interpersonal space regulation: she preferred the experimenter to be abnormally close [Reference Kennedy, Glascher, Tyszka and Adolphs23]. In addition, active avoidance behavior relies on a circuit between amygdala and nucleus accumbens [Reference Ramirez, Moscarello, LeDoux and Sears33]. In line with this, we detected increased structural connectivity between amygdala and nucleus accumbens in schizophrenia patients with increased paranoid symptoms which may trigger avoidance behavior [Reference Bracht, Horn, Strik, Federspiel, Razavi and Stegmayer6]. Structural and functional alterations of the limbic system, particularly amygdala and ventral striatum, have been associated with paranoid experience of threat, persecutory or referential delusions in schizophrenia [Reference Heinz and Schlagenhauf18,Reference Menon, Schmitz, Anderson, Graff, Korostil and Mamo26,Reference Orliac, Naveau, Joliot, Delcroix, Razafimandimby and Brazo29,Reference Pinkham, Liu, Lu, Kriegsman, Simpson and Tamminga32,Reference Stegmayer, Horn, Federspiel, Razavi, Bracht and Laimbock36].
The findings of increased interpersonal space in subjects experiencing threat, the association of amygdala activity in personal space regulation and the link between paranoid symptoms and aberrant function of the limbic system in schizophrenia suggest increased personal distance to be particularly pronounced in a subgroup of schizophrenia experiencing paranoid threat.
The positive syndrome in schizophrenia includes both, symptoms related to the paranoid experience of threat (e.g. paranoid delusions, hallucinations, hostility or suspiciousness), and symptoms with positive emotional valence (e.g. grandiosity or the paranoid experience of power). Persecutory delusions or paranoid ideation of threat have been shown to differ from grandiose delusions; the former are linked to negative emotions such as anxiety while the latter are not [Reference Ben-Zeev, Ellington, Swendsen and Granholm2,Reference Ben-Zeev, Morris, Swendsen and Granholm3,Reference Freeman, Dunn, Fowler, Bebbington, Kuipers and Emsley12,Reference Garety, Gittins, Jolley, Bebbington, Dunn and Kuipers14]. While delusions of paranoid threat would make individuals feel unsafe and anxious, paranoid experience of power would be associated with feelings of control and mastery. In order to disentangle the contribution of paranoid threat and paranoid experience of power on social approach-avoidance behavior, we tested the minimal tolerable interpersonal distance and subjective comfort ratings at fixed distances in schizophrenia patients currently experiencing paranoid threat, paranoid power or with neutral affect and healthy controls. We hypothesized that patients with paranoid threat would present with increased personal space and increased discomfort at fixed interpersonal distance compared to patients without paranoid threat and controls. We suspected that patients with neutral affect would differ from patients with paranoid threat, but performance would not differ from controls. Furthermore, we hypothesized that paranoid experience of power would lead to reduced avoidance behavior and thus reduced interpersonal distance and increased comfort at close proximity.
2. Method
2.1. Participants
In total, 64 patients with schizophrenia spectrum disorders (45 schizophrenia, 13 schizoaffective, 6 schizophreniform) and 24 healthy control subjects matched for age (39.7 vs 41.9 years) and gender (53% vs 54% male) were included in this study. Patients were recruited from the inpatient and outpatient departments of the University Hospital of Psychiatry, Bern, Switzerland. Healthy controls were recruited from the community via advertisement. Exclusion criteria included substance abuse or dependence (other than nicotine). Exclusion criteria for healthy controls were history of any psychiatric disorder and first-degree relatives with schizophrenia spectrum disorders. All participants were interviewed with the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I). In addition, we applied the Liebowitz Social Anxiety Scale (LSAS) [Reference Heimberg, Horner, Juster, Safren, Brown and Schneier17] to assess social phobia severity in all participants as this could contribute to personal space regulation, however, none scored above the 30-point cut-off. Duration of illness and number of episodes were obtained from the case files and the information given by the patients. All but four patients received antipsychotic medication at the time of assessments. Chlorpromazine equivalents were calculated according to Woods [Reference Woods41]. Clinical and demographic data are given in Table 1. After complete explanation of the procedure of the study, participants provided written informed consent. The protocol adhered to the declaration of Helsinki and was approved by the local ethics committee.
Table 1 Participants’ characteristics, mean (SD).
2.2. Procedures
For all participants, the order of assessments was: screening, psychopathological evaluation, personal space experiments, and if necessary retrieval of further information from case files. Two well-trained raters (GS and AK) performed psychopathological evaluations. In patients, psychopathology was assessed with the Positive And Negative Syndrome Scale (PANSS) [Reference Kay, Fiszbein and Opler22] and the Bern Psychopathology Scale (BPS) [Reference Strik, Wopfner, Horn, Koschorke, Razavi and Walther38]. The BPS is a research instrument to assess three psychotic symptom dimensions related to specific brain systems: motor behavior, language and paranoid affectivity. The global domain scores are rated after completion of the interview on a 7-point Likert scale ranging from severe inhibition (−3) to severe disinhibition (+3), while 0 represents no abnormality. On the global affectivity scale, the ratings range from severe paranoid threat to paranoid grandiosity (here referred to as paranoid power). The affectivity dimension takes into account subjective experience of paranoid threat or paranoid power as well as indirect signs such as the emotional content of hallucinations and delusions or unpleasant body sensations. The BPS has good internal consistency and external validity [Reference Bracht, Heidemeyer, Koschorke, Horn, Razavi and Wopfner5,36–Reference Strik, Wopfner, Horn, Koschorke, Razavi and Walther38]. Cluster analysis revealed clear distinction between inhibition and disinhibition of the BPS affectivity dimension, i.e. paranoid experiences of threat and power [Reference Lang, Stierlin, Stegmayer, Walther, Becker and Jäger25].
2.3. Measures
2.3.1. Affect scores
Based on the ratings of BPS global affect score, patients were stratified into three groups: patients with paranoid threat (−3, −2, −1), paranoid power (1, 2, 3) and neutral affectivity (0). This classification is based on the current state at assessment, clearly separating subjects with paranoid delusions and grandiose delusions. However, at different time points, the same subjects may suffer from different types of delusions. The three patient groups differed at trend levels in gender distribution, education, duration of illness and PANSS positive and general psychopathology scores, as the group with paranoid threat included more women, subjects with longer education, shorter duration of illness, and more pronounced symptoms (Table 1).
2.3.2. Experimental paradigms
The stop-distance paradigm is a reliable, frequently used measure of the minimal tolerable interpersonal distance [Reference Hayduk16]. Participants indicate the minimal tolerable interpersonal distance to an experimenter while the interpersonal distance is varied. In the beginning of the procedure, the participant was positioned at one end of the room facing the experimenter from a distance of 7 m. Participants were instructed to tell the experimenter to stop at a distance, where they would start feeling less comfortable in the passive approach condition and stop moving themselves in the active approach condition. We applied a modified version of the stop-distance paradigm that would also allow testing the effects of eye contact and type of approach (active vs passive). The test included four different conditions: participant approach (active) versus experimenter approach (passive) as well as approach with versus without eye contact. Considering the dynamic nature of personal space preferences, we tested each condition five times in randomized order. Once the correct distance has been determined, a laser distance sensor (Leica Disto D2, Leica Geosystems, Heerbrugg, Switzerland) was used to measure the distance between the bodies considering the line from the medial aspect of the clavicula in both persons. The outcome variable was the minimally tolerable interpersonal distance in meter.
The fixed-distance paradigm assessed subjective evaluation of comfort at given interpersonal distances of 0.5, 1.0, 1.5, 2.0 and 2.5 m between experimenter and participant with eye contact based on a Visual Analogue Scale (VAS) ranging from 0 mm (maximum discomfort) to 100 mm (maximum comfort). The outcome measure was the VAS rating of comfort.
Due to gender effects on interpersonal distance [Reference Holt, Cassidy, Yue, Rauch, Boeke and Nasr19], we chose to conduct both experiments exclusively with a male experimenter maintaining a neutral facial expression. The experimenter was completely unfamiliar to the participants and blind to their clinical status. The experimenter had a height of 1.80 m.
2.4. Statistical analyses
Differences between groups in clinical and demographic parameters were explored with ANOVAs and χ2-tests respectively. Repeated measures ANCOVAs were applied to test group differences in the stop-distance paradigm including the minimum tolerated interpersonal distance and the factors “approach” (active vs passive), “eye contact” (with vs without), and “trial” (1–5). Repeated measures ANCOVAs of the fixed-distance paradigm included the VAS ratings of comfort and the factor “distance” (0.5 m, 1.0 m, 1.5 m, 2.0 m, 2.5 m). For each paradigm, we tested 3 ANCOVAs:
all patients vs all controls;
four groups (patients with threat, power, and neutral affect and controls);
three groups (patients only).
In all tests, gender was added as covariate, because gender effects have been reported [Reference Holt, Cassidy, Yue, Rauch, Boeke and Nasr19] and groups differed at trend level in gender. In order to demonstrate a specific effect of paranoia on personal space regulation beyond the reported effect of positive symptoms or general psychopathology [Reference de la Asuncion, Docx, Sabbe, Morrens and de Bruijn8], the corresponding PANSS subscales were included as additional covariates in the ANCOVAs. In all repeated measures, ANCOVAs Greenhouse-Geisser corrections were applied as well as post-hoc Sidak tests. For the stop-distance paradigm condition with active approach and eye contact, we calculated receiver-operating curves (ROC) to determine the specificity and sensitivity of detecting paranoia with interpersonal distance. Active approach and eye contact were chosen due to the clinical relevance, as patients more often actively approach mental health professionals maintaining eye contact. Further associations between clinical characteristics and personal space were tested using linear regression models; as we had to limit the number of these analyses, we chose conditions of high clinical relevance and calculated regression models for mean minimum interpersonal distance with active approach and eye contact or comfort ratings at 1.0 m fixed interpersonal distance. We tested the impact of PANSS subscales, CPZ and BPS affect ratings on the personal space measures. Furthermore, we explored the correlations of the LSAS with personal space measures and paranoia ratings (PANSS and BPS).
In order to replicate the findings with the well-known PANSS, we divided the patients into groups of high and low paranoia according to the ratings in item P6 – suspiciousness/persecution (low: < 4 and high: ≥ 4) [Reference Pinkham, Liu, Lu, Kriegsman, Simpson and Tamminga32]. Again, we ran the ANCOVAs for the stop-distance and fixed-distance paradigms with three groups (low paranoia, high paranoia and controls), covarying for gender. All statistics were performed with SPSS 22® (IBM Corporation, New York, USA).
3. Results
3.1. Stop-distance paradigm
Comparing all patients with controls, we detected a trend to a group effect with increased interpersonal distance in patients (Table 2). In contrast, we found clear effects investigating current paranoid experience on the minimum tolerated interpersonal distance (Fig. 1 and Table 2), which remained stable when correcting for gender, PANSS positive and PANSS general psychopathology (Supplementary figure S1). Patients with paranoid threat increased their minimum tolerated interpersonal distance by a factor of > 2 compared to all other patients. There was an interaction of eye contact and group, but post-hoc tests did not identify group differences. No further significant interactions were observed. Thus, irrespective of the variations in the stop-distance paradigm, patients with paranoid threat maintained increased interpersonal distance.
Fig. 1 Stop-distance paradigm. Left panel: active approach of the participants (right) with eye contact. Interpersonal distances are scaled to body height of the experimenter (1.8 m). Persons in black indicate mean values and the grey scales indicate minimum and maximum values. All distances were corrected for gender. Right panel: stop-distance paradigm all conditions. Bars represent means ± 95% CI, distances corrected for gender.
Table 2 Minimum tolerated interpersonal distance ANCOVAs.
To investigate the discriminative value of the results, we applied ROC statistics to the minimal tolerated interpersonal distance in the active approach with eye contact in patients, classifying them to paranoid threat versus non-threat (neutral and paranoid power). The test had excellent discriminative value (ROC area under the curve = 0.93; P < 0.001, Supplementary figure S2). At an interpersonal distance of 1.10 m, the test had 93% sensitivity and 83% specificity.
3.2. Fixed-distance paradigm
When considering all patients versus controls, we detected a trend towards an effect of distance on VAS comfort ratings, and an effect of group but no distance-group interaction (Table 3). Again, we found clear effects current paranoid experience on the VAS comfort ratings in the fixed-distance paradigm, which remained stable after controlling for gender, PANSS positive and PANSS general psychopathology (Table 3, Fig. 2 and supplementary figure S1). In controls, we noted a ceiling effect of comfort ratings at distances > 1.5 m. In patients with neutral affect or paranoid threat, comfort ratings steadily increased with larger interpersonal distance. Instead, patients with paranoid power displayed little variation of their generally enhanced comfort.
Fig. 2 Fixed-distance paradigm. Lines indicate Means ± 95% CI, adjusted for gender.
Table 3 VAS comfort ratings ANCOVAs.
3.3. Personal space and clinical variables
To explore the association with personal space regulation and clinical factors (PANSS positive, PANSS negative, CPZ and BPS affect rating), we entered the mean minimum tolerated interpersonal distance at active approach with eye contact and the VAS comfort ratings at 1.0 m interpersonal distance in regression analyses. Stepwise linear regression analysis for mean minimum tolerated interpersonal distance indicated best fit for a model with two predictors explaining R2corr = 53% of the variance (F = 36.84, df = 2, P < .001). In this model, the factor BPS affect rating explained 50% of the variance (ΔR2 = .50, β = −0.65, T = −7.36, P < 001), while PANSS positive added another 5% of the variance (ΔR2 = 0.05, β = 0.22, T = 2.52, P = 0.01). Stepwise linear models of comfort ratings at 1.0 m interpersonal distance indicated best fit for a model that only contained BPS affect ratings (R2 = 0.26, F = 21.25, df = 1, P < 0.001, β = 0.51; T = 4.61, P < 0.001). PANSS negative and CPZ had no significant effect on any of the measures of interpersonal space. Gender effects on interpersonal distance in the stop-distance paradigm and in the fixed distance paradigm did not reach significance, nor were the interactions of gender and group significant. Higher LSAS total scores correlated with discomfort at 1.0 m in controls (r = −0.43, P = .038), but not in patients. LSAS scores were also unrelated to paranoia ratings (PANSS positive, BPS affectivity dimension).
3.4. Paranoia according to PANSS item P6
When subdividing patients into low and high paranoia groups according to the P6 item of the PANSS, the group with high paranoia included more women (62% vs 40%). Therefore, gender was added as a covariate in further analyses. In the stop-distance paradigm, repeated measures ANCOVA indicated an effect of group (F = 20.08, df = 2, P < 0.001), which was driven by increased personal space in the high paranoia group (P < 0.001) (Supplementary figure S4). There was no effect of approach, eye contact or trial. A trend to an interaction was noted for eye-contact group (F = 2.70, df = 2, P = 0.07). In the fixed-distance paradigm, repeated measures ANCOVA of comfort ratings indicated an effect of distance (F = 5.29, df = 1.9, P = 0.007), an effect of group (F = 23.56, df = 2, P < 0.001) as well as a distance-group interaction (F = 6.74, df = 3.8, P < 0.001) (Supplementary figure S5). Post-hoc tests demonstrated reduced comfort ratings in high paranoia patients compared to low paranoia (P < 0.001) and controls (P < 0.001), but also for low paranoia patients versus controls (P = 0.04).
4. Discussion
This is the first study reporting paranoid threat to be associated with abnormal personal space regulation in two paradigms, indicated by increased chosen interpersonal distance and reduced comfort in fixed settings of interpersonal distance. The opposite emotional valence, i.e. the paranoid experience of power, was associated with increased comfort ratings independent of the interpersonal distance, though the chosen interpersonal distance was unaltered. Patients with neutral affect were not different from controls in interpersonal distance, but experienced less comfort. These findings held true after correction for gender, PANSS positive and PANSS general psychopathology ratings. The stop-distance paradigm with active approach and eye contact proved to be highly sensitive and specific to detect paranoid threat among patients. Finally, when we divided patients according to the PANSS item P6 (suspiciousness/persecution) into high and low paranoia groups, we replicated the findings regarding paranoid threat.
Other studies demonstrated marked differences in the stop-distance paradigm between patients and controls, which were only at trend level in our study. Some studies reported slightly shorter interpersonal distances particularly for controls [Reference de la Asuncion, Docx, Sabbe, Morrens and de Bruijn8,Reference Holt, Cassidy, Yue, Rauch, Boeke and Nasr19], which may result from younger participants, cultural context or female experimenters. Still, the minimum tolerable distances determined in controls and in the schizophrenia patients with neutral affectivity was in the range of other studies [Reference Deus and Jokic-Begic9,Reference Kennedy, Glascher, Tyszka and Adolphs23].
Our study supports the association of increased personal space with positive symptoms [Reference de la Asuncion, Docx, Sabbe, Morrens and de Bruijn8], but extends previous knowledge by focusing on both ends of the paranoid dimension. More than two-fold increases of interpersonal distance were found in schizophrenia patients experiencing paranoid threat. Paranoid threat as defined by the global affect rating of the BPS includes persecutory delusions, paranoid perception of existential threat, and flight or fight reactions (i.e. avoidance or aggression) [Reference Strik, Wopfner, Horn, Koschorke, Razavi and Walther38]. Increased personal space in paranoid patients can be explained as safety behavior: the interpersonal distance offers a safe territory and provides time to react with appropriate defense or escape in case of sudden unwanted approach by others [Reference Felipe and Sommer11,Reference Horowitz, Duff and Stratton21,Reference Knapp, Hall and Horgan24,Reference Sambo and Iannetti34]. In fact, persecutory delusions, which are frequently associated with paranoid threat lead to increased safety seeking behaviors, such as avoidance, escape, aggression or help seeking [Reference Freeman, Garety, Kuipers, Fowler, Bebbington and Dunn13]. In this study, the experimenters maintained neutral facial expressions, which may have been misinterpreted as threatening by paranoid patients as demonstrated with other nonverbal stimuli including faces, words, prosody, or gestures [Reference Bach, Buxtorf, Grandjean and Strik1,Reference Bucci, Startup, Wynn, Baker and Lewin7,Reference Holt, Titone, Long, Goff, Cather and Rauch20,Reference Pinkham, Brensinger, Kohler, Gur and Gur31,Reference Walther, Stegmayer, Sulzbacher, Vanbellingen, Muri and Strik39]. In line with this, we noted an eye-contact group interaction, indicating that paranoid patients benefit from averted gaze during approach (Fig. 1). In clinical practice, it is advisable to maintain increased personal space when dealing with paranoid patients. In fact, the ROC results suggest that spontaneously increased interpersonal distance ≥ 1 m would indicate paranoia in schizophrenia.
The current experience of paranoid power, e.g. in grandiose delusions, was hypothesized to lead to perceived safety or control. While the assumption of reduced personal space proved wrong in this group, the prediction of increased comfort at fixed interpersonal distance was met. Most of the variance in the stop-distance paradigm was explained by the BPS global affectivity rating, while the PANSS positive subscale added less than 5% of the variance. This could be attributed to the mix of symptoms within the PANSS-positive syndrome, since different types of delusions are distinct regarding cognitive bias and negative emotions [Reference Garety, Gittins, Jolley, Bebbington, Dunn and Kuipers14]. Moreover, prospective experience sampling methods indicated anxiety to predict persecutory delusions, but not other types of delusion [Reference Ben-Zeev, Ellington, Swendsen and Granholm2,Reference Ben-Zeev, Morris, Swendsen and Granholm3]. Likewise, persecutory delusions have been repeatedly shown to correlate with high levels of anxiety [Reference Freeman, Dunn, Fowler, Bebbington, Kuipers and Emsley12], while grandiose delusions did not [Reference Garety, Gittins, Jolley, Bebbington, Dunn and Kuipers14]. Our results argue for distinct effects of the positive syndrome in schizophrenia on aberrant personal space regulation.
In addition to the important role of amygdala, personal space regulation has been associated with the interaction of the ventral premotor cortex and intraparietal sulcus in healthy subjects [Reference Holt, Cassidy, Yue, Rauch, Boeke and Nasr19,Reference Kennedy, Glascher, Tyszka and Adolphs23]. To date, however, there are no studies exploring the neural correlates of abnormal personal space regulation in schizophrenia. We would expect that the ventral striatum was involved in abnormal personal space regulation as it seems critical for aberrant salience and delusion formation [Reference Heinz and Schlagenhauf18,Reference Menon, Schmitz, Anderson, Graff, Korostil and Mamo26,Reference Mishara and Fusar-Poli27,Reference Stegmayer, Horn, Federspiel, Razavi, Bracht and Laimbock36].
The present study examined psychopathological state in a group of patients with schizophrenia spectrum disorders. The allocation to groups was based on the global affectivity rating of the BPS, a rather new psychopathology scale. However, findings regarding paranoid threat were replicated by subtyping patients according to the severity of the PANSS item suspiciousness (P6). The test procedure limited the experimenters to males with body height around 1.8 m; therefore, gender effects cannot be excluded [Reference Hayduk16,Reference Holt, Titone, Long, Goff, Cather and Rauch20,Reference Knapp, Hall and Horgan24]. The number of patients with paranoid power was small. Most patients were medicated, which, in principle, can alter the interpersonal behavior. Still, personal space regulation was not related to antipsychotic dosage, but this specific medication effects cannot be excluded. Our findings cannot be simply transferred to bipolar disorders, since distinctive symptoms for schizophrenia such as psychotic suspiciousness or grandiosity, appear to be relevant for personal space regulation. Based on these considerations, however, we would predict no increase of personal space in non-psychotic depression, but increased comfort at low distance for the non-paranoid elated state in mania.
In conclusion, the present study demonstrated that interpersonal distance is not generally increased in schizophrenia, but only in a subgroup of patients experiencing paranoid threat. Personal space measures detected this subgroup with excellent sensitivity and specificity. Paranoid experience of power or grandiosity instead alters the evaluation of close proximity. The findings argue for distinct contributions of positive symptoms to personal space regulation in schizophrenia.
Disclosure of interest
Dr. Walther received honoraria from Eli Lilly, Sandoz and Janssen-Cilag, he is advisory board member to Lundbeck/Otsuka. All other authors declare that they have no competing interest.
Acknowledgements
We thank the subjects for their participation in the study. The BPS can be downloaded at https://www.puk.unibe.ch/bps.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.eurpsy.2015.10.001.
Comments
No Comments have been published for this article.