key: cord-0262676-erebu22a authors: Farkas, K.; Pesthy, O.; Guttengeber, A.; Weigl, A. S.; Veres, A.; Szekely, A.; Komoroczy, E.; Szuromi, B.; Janacsek, K.; Rethelyi, J. M.; Nemeth, D. title: Altered interpersonal distance regulation in autism spectrum disorder date: 2021-08-07 journal: nan DOI: 10.1101/2021.08.06.21261686 sha: 5dd611705ba507a5fb67844b13a702f01a187634 doc_id: 262676 cord_uid: erebu22a Interpersonal distance regulation is an essential element of social communication. Its impairment in autism spectrum disorder (ASD) is widely acknowledged among practitioners, but only a handful of studies reported empirical research in real-life settings focusing only on children. However, these studies did not measure the alterations of vegetative functions related to interpersonal distance. Here, we introduced a new experimental design to systematically measure interpersonal distance along with heart rate variability (HRV) in adults with ASD and tested the modulatory effect of intentionality, eye contact, moving activity, and attribution. Twenty-two adults diagnosed with ASD and 21 matched neurotypical controls participated in our study from 2019 October to 2020 February. Our new experimental design combined the modified version of the stop distance paradigm with HRV measurement controlling for eye contact between the experimenter and the participant to measure interpersonal distance in incidental and intentional conditions. Our results showed greater preferred distance in ASD in the intentional but not in the incidental condition. These results were altered with eye contact and the participant's role (active vs. passive) in the stop distance task. Moreover, we found lower baseline HRV and reduced HRV reactivity in ASD; however, these vegetative measurements could not predict preferred interpersonal distance. Our study highlights the importance of interpersonal space regulation in ASD and the need for sophisticated experimental designs to grasp the complexity and underlying factors of distance regulation in typical and atypical populations. Autism spectrum disorder (ASD) is a neurodevelopmental condition characterized by persistent difficulties in social communication and social interaction across multiple contexts, such as abnormal social approach or failure to initiate or respond to social interactions; and restricted, repetitive patterns of behaviour, interests, or activities (American Psychiatric Association, 2013) . On the neural level, cortical (Ecker, 2017; Herringshaw et al., 2016 ), subcortical (Hadjikhani et al., 2017 Langen et al., 2014) , and autonomic (Gomez & Flores, 2020; Kushki et al., 2014) neural alterations can be observed, including developmental, structural and functional differences (Philip et al., 2012; Thye et al., 2018; Wood et al., 2021) , in parallel to the pervasive cognitive (Demetriou et al., 2019; Hill, 2004) , behavioural and physiological disturbances in ASD. However, one of the key components of social behahavior, namely interpersonal distance regulation, has received relatively less attention in ASD research (see exceptions: Candini et al., 2020; Lough et al., 2015; Massaccesi et al., 2021; Simões et al., 2020) even though its impairment in ASD is widely acknowledged among practitioners. Our study aims to measure the interpersonal distance regulation and a related physiological vegetative parameter (heart rate variability, HRV) during a social interaction task and test the modulatory effect of higher-level cognitive factors (intentionality, eye contact, active role, and attribution) in autism. Finding the appropriate social distance can be seen as the first step of physical social interactions. It is widely believed among practitioners that people with autism keep a greater or abnormal distance (Perry et al., 2015) . However, it is challenging to measure this phenomenon experimentally with high ecological validity. In autistic participants, preference for both closer (Asada et al., 2016; Ingram et al., 2007; Lough et al., 2015; Pedersen & Schelde, 1997) or farther distance (Candini et al., 2017; Gessaroli et al., 2013) can be found in the literature. Among these studies only one measured interpersonal distance in adult ASD; they found no difference between study groups (Kennedy & Adolphs, 2014) . Studies applying electrophysiological or imaging methods usually present video recordings of an approaching individual (Massaccesi et al., 2021) or use virtual reality displays (Mul et al., 2019; Simões et al., 2020) to measure interpersonal distance regulation in autism. The results both from real-life-and virtual settings remain inconclusive. In the present study, we intend to fill this gap by measuring the interpersonal distance among adult participants with ASD in an experimental setting with personal presence as close as possible to real-life situations. Despite the relevance, empirical studies on interpersonal distance regulation of participants with ASD were conducted only in the past few years. The nomenclature of the concept is still not unified. Personal space (Asada et al., 2016; Candini et al., 2017; Gessaroli et al., 2013) , physical distance (Kennedy & Adolphs, 2014) , and interpersonal space anddistance (Candini et al., 2020; Gessaroli et al., 2013; Massaccesi et al., 2021) are all commonly used. When measuring the physical distance between two people in one dimension, we use the term interpersonal distance. The Stop Distance Paradigm (Kennedy et al., 2009 ) is the most commonly used method for measuring interpersonal distance regulation. This method is considered to be an ecological measure of permeability and flexibility of interpersonal space regulation (Candini et al., 2020) . It can be combined with various factors. Previous studies suggested that the presence of eye contact (Asada et al., 2016) , the role played during the experiment (active approaching or passive indicating) (Candini et al., 2017) and social interactions as experimental interventions (Gessaroli et al., 2013; Massaccesi et al., 2021) might influence interpersonal distance. Most of the studies examined children or adolescents with autism (Asada et al., 2016; Candini et al., 2017; Gessaroli et al., 2013) and found larger preferred interpersonal distance in ASD, except one study (Asada et al., 2016) . The only study using the Stop Distance Paradigm among adults found no difference between groups in terms of interpersonal distance preferences (Kennedy & Adolphs, 2014) . Since different results were gained in different conditions, the question arises as to what extent these factors contribute to the outcome behaviour. None of the previous studies addressed the effect of eye contact, the active or passive role, and intentionality at the same time in interpersonal distance regulation. Here, we introduce a new experimental design in order to both systematically measure the interpersonal distance and determine the modulating factors in autism. Social communication and interactions are tremendously complex processes that can be altered in autism at cognitive, behavioural, and physiological levels. At the highest level, the theory of mind (mentalizing others) difficulties can be observed in autism (Baron-Cohen et al., 1985; Baron-Cohen, 1991) . To capture this phenomenon during the interpersonal distance measurement in a simplified way, we suggest the introduction of an additional condition: the participant's own or the experimenter's idea about the comfortable distance between them (attribution dimension). The processing of facial expressions, particularly that of the eye region, is highly relevant in the regulation of social behaviour, including interpersonal distance. Facial emotion processing and emotion recognition is altered in autism (Black et al., 2017; Monteiro et al., 2017; Philip et al., 2010; Teh et al., 2018; Wieckowski et al., 2020) . At the subcortical level, increased amygdala response for both threatening and neutral faces has been found in ASD, as compared to neurotypicals (Monk et al., 2010; Weng et al., 2011) . Furthermore, constraining eye contact led to an exaggerated increase of amygdala activation, while decreased eye contact was associated with diminished amygdala response to faces in ASD (Dalton et al., 2005; Hadjikhani et al., 2017; Tanaka & Sung, 2016; Tottenham et al., 2014) . These results suggest that altered amygdala functioning, including the regulation of eye contact and emotion processing, might have a substantial role in the disturbances of several aspects of social behaviour, such as in personal proximity or interpersonal space regulation as well (Frank et al., 2014; Kennedy et al., 2009; LeDoux, 2007) . Physiological, vegetative response to sensory, social, and emotional stimuli is suggested to be altered in autism in general, however, the methodology used is highly variable and the results are inconsistent (Lydon et al., 2016) . Since the classic electrophysiological experiment of Hutt et al. showed hyperarousal in children with autism (Hutt et al., 1965) , the majority of studies that measured autonomic regulation (pupillometry, skin conductance, or cardiac measures) found atypical resting-state functions indicating hyper-or hypoarousal in autism according to a recent review (Arora et al., 2021) . Recent studies aim to overcome these limitations by measuring heart rate variability (HRV). Heart rate is affected by both sympathetic and parasympathetic modulatory effects; thus, its variability might be a good marker of vegetative regulation, as higher HRV reflects the parasympathetic activity (Laborde et al., 2017) . Furthermore, a study found an association between HRV and cognitive flexibility in healthy individuals (Colzato et al., 2018) . A recent meta-analysis showed that heart rate variability is reduced in ASD: baseline HRV and HRV reactivity during social stress were significantly lower in participants with ASD, but HRV reactivity performing cognitive tasks did not differ (Cheng et al., 2020) . The reduced variability in the heart rate indicates an altered parasympathetic-sympathetic balance in ASD, suggesting the predominance of sympathetic activity and less flexible switching between vegetative states in ASD compared to neurotypicals. For these reasons, in our study, we measured interpersonal distance along with heart rate variability to examine their putative alterations and their relationship in autism. In this study, our main goal was to establish a comprehensive design to measure interpersonal distance and vegetative functions in ASD. First, we measured the interpersonal distance in adults with ASD. Second, we compiled a study design to examine the interaction of previously separately measured factors. Furthermore, in real life interpersonal distance regulation is more of an implicit process. To the best of our knowledge, no research compared the incidental and the intentional dimension of interpersonal space regulation. Here we studied the effect of eye contact, active or passive role, attribution, and the intentionality of social distance regulation in one experiment. Finally, the relationship between vegetative functions and all the factors listed above need to be investigated experimentally. We recorded heart rate variability to test if we could find decreased baseline HRV and reduced HRV reactivity during the social distance task in ASD. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 In total, forty-five adults participated in our research. Two neurotypical participants were excluded due to errors during data collection. The final sample consisted of forty-three participants, 22 were diagnosed with ASD, and 21 were neurotypical controls. The two groups did not differ in age, gender, and education (Table 1 ). All participants with ASD were diagnosed by trained clinicians, the diagnoses were confirmed with Autism Diagnostic Interview-Revised (ADI-R) and Autism Diagnostic Observation Schedule, IV-module (ADOS-IV.) (Lord et al., 1989 (Lord et al., , 1994 . (Table 1. ) Twelve participants had one or more comorbid disorders (attention deficit hyperactivity disorder (5), obsessive-compulsive disorder (3), generalized anxiety disorder (2), bipolar disorder (1), depression (1), and schizophrenia (1)). Participants with autism spectrum disorder were recruited from the outpatient unit of the Department of Psychiatry and Psychotherapy, Semmelweis University. Neurotypical control participants were recruited by advertisement. Participants (and legal guardians if applicable) provided written informed consent and did not receive financial compensation for their attendance. The study was conducted in accordance with the Declaration of Helsinki, and it was approved by the Regional and Institutional In our study, we developed an experimental design measuring social distance regulation along four dimensions. We intended to distinguish intentionally controlled and incidental, automatic behaviour. Therefore, in the first part of the experiment, in the implicit distance task participants did not get direct instruction related to social distance and performed a distractor task while distance data was recorded in two consecutive conditions. ( Figure 1A ). Next, in the second part, they underwent an explicit distance measurement, a modified version of the stop-distance paradigm (Kennedy et al., 2009) . In all conditions, the participant and the experimenter started from the opposite endpoints of the tape measure (five . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint meters) stuck on the ground. They were asked to consciously focus on comfortable social distance overall eight times with the following order. First, (1) participants were approaching actively and were asked to stop where it was still comfortable for them. Next, (2) participants were approaching actively and were asked to stop where they thought it was still comfortable for the experimenter. Then (3) participants stood passively and were asked to stop the experimenter where it was still comfortable for them, finally, (4) participants stood passively and were asked to stop the experimenter where they thought it was still comfortable for the experimenter. Participants repeated this procedure twice, with and without eye contact: either the experimenter was looking at the participant (eye contact condition), or at the papers she was holding (no eye contact condition). The order of these two conditions was randomised across participants. ( Figure 1B ). Experimental setting Panel A: Floor plan of the experimental room. Implicit condition. First, (1) participants placed a chair from its initial place anywhere comfortable in the room, then (2) they were allowed to move to complete a verbal task. Distance between the experimenter and participants were registered. Panel B: Experimental design of the explicit condition. The modified version of the stop distance Paradigm. First, (1) participants were approaching actively and were asked to stop where it was still comfortable for them. Next (2) participants were approaching actively and were asked to stop where they thought it was still comfortable for the experimenter. Then (3) participants stood passively and were asked to stop the experimenter where it was still comfortable for them; finally (4) participants stood passively and were asked to stop the experimenter where they thought it was still . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint comfortable for the experimenter. Participants repeated this procedure twice with and without eye contact; the order of these two conditions was randomised across participants. A wearable device Polar H10 was placed on participants' chests which recorded heart rate (HR) during the whole experiment. We measured cardiac interbeat intervals (RR intervals) using Polar H10 heart rate monitor chest strap (Saario, 2019) (Polar Electro Oy, Kempele, Finland), which has been shown to be a valid device to measure RR interval signals (Gilgen-Ammann et al., 2019). The HR monitor was connected to a Samsung Galaxy Tablet via Bluetooth. We used the Elite HRV application to export the recorded RR intervals as .txt files. We measured heart rate variability (HRV) at different conditions for a duration of 60 seconds: 1) at baseline (before participants entering the room), 2) during the incidental distance task (1 minute after entering the room), and 3) during the intentional interpersonal distance task (1 minute after starting the distance task) using the Root Mean Square of Successive RR interval Differences (RMSSD) method (Shaffer & Ginsberg, 2017) . Additionally, we calculated RMSSD at the preceding ten-second time window of trigger points set by researchers. These triggers corresponded to the time when participants arrived at their final location of each condition (see Data Pre-processing and analysis). In order to synchronize the distance data with the HRV data, both the experimenter and the participant wore a distance measuring device. The Obimon Prox (Veres, 2019) measures the distance and the relative orientation between two wearable devices in real-time. Two-Way Ranging (SDS-TWR) method (Neirynck, 2016) to both determine the distance between each other by emitting very short and low power radio transmissions and measure the so-called time-of-flight (ToF) with very high precision between transmission and . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint reception. The resolution of the measurement is in the range of a few centimeters, while the absolute precision is approximately 10 centimeters. The relative orientation is defined as the difference between the angles between the two devices taking the Earth's magnetic field as a reference. For increased precision, the device uses sensor fusion involving magnetometer, accelerometer, and gyroscope sensors. The results of the measurements are collected over Bluetooth LE wireless technology to a laptop computer and evaluated in real-time. Participants wore a Polar H10 and Obimon Prox device during the whole experiment. They placed the wearable device on themselves before the experiment started, then they waited five minutes while calibrating and registering 60 seconds of resting heart rate and HRV. First, after entering the study room, we measured the distance the participant prefers naturally (that is, without explicit instruction regarding their personal space). The experimenter asked them to sit anywhere in the room where it felt comfortable for them. To avoid the option of taking a seat wherever the chair had been already placed, we used a folding chair which the participants opened and placed for themselves. The experimenter remained at the same position during this procedure (see Figure 1A ). As a cover task, participants were instructed to count their heartbeats for one minute. Next, they were informed that they could move their chair closer to the experimenter in order to hear the instructions properly, and they completed a verbal fluency task. In the second part of the experiment, participants completed the intentional version of the social distance regulation task. They were asked to set up a comfortable distance along a marked measuring tape fixed on the floor (detailed description above: experimental paradigm, and Figure 1B) . Then, after a short break, they completed a computerized neurocognitive test battery -measuring working memory, executive functions, attention, inhibition, implicit learning, faux pas, etc. These results are not reported in this paper. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint Finally, they completed computerized versions of self-report questionnaires (AQ: Autism-Spectrum Quotient, MZQ: Mentalization Questionnaire, AAS: Adult Attachment Scale, ASRS: Adult ADHD Self-Report Scale, STAI-T: State-Trait Anxiety Inventory -Trait; see Supplementary material). To avoid the sensory over-reactivity effect, experimenters did not wear any jewelry, perfume, and had been asked not to eat spicy food before the experiment, they wore simple, casual, non-coloured clothes (jeans and black T-shirt). The room was curtained and artificially evenly lit. Preparation of HRV data was carried out using Python 3.7 with NumPy 1.20.1 (Harris et al., 2020), pandas 1.2.3 (Reback et al., 2020), and SciPy 1.6.1 (Virtanen et al., 2020) data processing packages. Since the samples were measured at a different rate for the Polar H10 (one sample per second) and the Obimon prox (one sample per milliseconds) devices, we resampled the Obimon data by taking the median for each second. Missing data were dropped from the analysis. To synchronize HRV with the proximity data we needed to obtain the timestamps for each file containing the RR intervals. The first timestamp was obtained from the name of the file which indicated the start time of the recording. Since the exported files only contained the RR intervals without a timestamp for each sample, the interval values themselves were used to create the time elapsed since the first sample. As RR intervals annotate the time between two successive heartbeats, it was possible to append the value of the RR interval to the time of the previous sample. After obtaining the timestamps, data points were replaced with the median if they indicated RR of 1200 milliseconds (ms) or above, or if their absolute Z score was higher than 2. Triggers added to the distance data (see Distance measuring) were adjusted manually if needed. HRV was estimated as the root mean square of successive RR interval differences (RMSSD) since this measurement is relatively . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint resistant to by-products caused by breathing (Hill & Siebenbrock, 2009) , and can be obtained for a shorter (10 seconds) period of time (Salahuddin et al., 2007) . Calculations were done by the following formula (1): Baseline HR and HRV were measured and calculated for 60 seconds (s), one minute before entering the experimental room (from -60 s to zero), reactive HR and HRV were measured for 60 seconds during implicit condition one minute after entering the room (from +60 s to +120 s) and during explicit interpersonal distance task one minute after starting the explicit paradigm (from +60 s to 120 s). The exact timing was indicated by the trigger points described above. Furthermore, we calculated RMSSD around time points where interpersonal distance data were reported. For implicit conditions, trigger points were used. To calculate RMSSD for each explicit condition, eight local minimums of the distance data were determined from data recorded by Obimon Prox. These eight time points indicate the shortest distances between the participant and the experimenter, corresponding to the time point when the reported distance was reached. RMSSD was calculated for an interval starting 10 seconds prior to reaching the reported distance. Statistical analysis was accomplished using R Version 3.6.3 (R Core Team, 2020), RStudio Version 1.2.1335 (RStudio Team, 2020), and JASP Version 0.14.0.0 (JASP Team, 2020). First, to measure if the two study groups do differ regarding age, gender, education, and scores on questionnaires, we conducted nonparametric Mann-Whitney U (Wilcoxon rank-sum test) tests and Chi-square test. To measure the effect of different conditions and . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint study groups on the distance and HRV data, repeated measures ANOVA tests were applied, while in the case of significant interaction effects post hoc tests with Bonferroni correction were used. Associations between distance, HRV, and scores of psychometric questionnaires and diagnostic tests were analysed with Spearman's rank-order correlations. Analyses were . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint Surprisingly, we did not find a statistically significant difference between groups in incidental social distance conditions (W inc1 =232, p=0.990, r=0.004, . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. In our sample, we measured larger interpersonal distance in autism during the intentional task. For the descriptive statistics of all conditions see supplementary material (Table S1. ). To test if the three conditions had different effects in the two study groups, we used two-way We found no significant Group × Condition interaction in any condition, which suggests that neither of the three manipulated conditions by themselves had significantly different effects in the ASD group. In the further analysis, we tested if the conditions influence each other's effect by applying two-way repeated-measures ANOVA. Eye-contact × Move interaction had a significant main effect (F(1,41)=6.150, p=0.017, 2 =0.003), and . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint Figure 3 Interpersonal distance in different intentional conditions. Participants with autism: green, neurotypical controls: blue. Conditions: Eye-contact: solid line, no eye-contact: dotted line. Active moving: darker color, passive: lighter color. Attribution of the other and self is marked on the x-axis. Participants with ASD kept a greater distance in conditions with eye contact when they moved actively. In general, participants with ASD had a slightly higher heart rate than NTP participants (Figure 4) , however, these differences were not significant. Measuring the effect of the conditions we used repeated-measures ANOVA. As caffeine intake could influence the heart rate, we included actual caffeine use as a covariate. The Group main effect was not statistically significant (F(1,30)=1.181, p=0.286, 2 =0.035), the main effect of Conditions was significant (F(2,60)=8.887, p<0.001, 2 =0.012), but the Condition × Group interaction was not (F(2,60)=0.582, p=0.562, 2 =0.001). It means, that in both groups we measured the highest HR during the intentional interpersonal distance task, and it was significantly higher than baseline (t ASD =4.543, p<0.001, 95% CI [1. 925, 9.853 CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint values were adjusted by using Bonferroni-correction. Levene's test showed that the variances were equal. ( Figure 4A ) Heart rate variability (HRV) was higher in the NTP group and at baseline, but not during the incidental and intentional conditions. Again, measuring the different effects of the conditions we used mixed-design ANOVA, with actual caffeine intake as a covariate. The group main effect (F(1,30)=2.880, p=0.100, 2 =0.072) and the Group × Condition interaction was not significant, but we found a significant Condition main effect . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint Heart rate and heart rate variability. Panel A: Baseline and reactive (incidental and intentional conditions) heart rate in beat per minute (bpm). Panel B: Baseline and reactive (incidental and intentional distance conditions) heart rate variability (RMSSD). Green line: participants with autism spectrum disorder, Blue line: neurotypical participants. Error bars: standard error of the mean. HRV (measured before the time point reported distance was reached) was numerically higher in the NTP group, but the difference was not significant between groups. For descriptive statistics see Supplementary material (Table S2 ). When we tested the effect of the conditions and their interactions, neither the main effect of eye contact, activity or attribution, and their interaction with the group, nor the group effect (F(1,29)=2.523, p=0.123, 2 =0.036) were significant using repeated-measures ANOVA. (Figure 5D -F). Vegetative functioning might be influenced by smoking, exercise, regular caffeine consumption, or the actual caffeine intake before the experiment. There was no difference between groups (see Table 1 ), however, including these variables as covariates did not change the results. The interpersonal distance (Panels A-C) and heart rate variability data measured by the RMSSD method (Panels D-F) are presented in Figure 5 to introduce their characteristics in different conditions in the two study groups. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint Is there any correlation between HRV, distance, and psychometric data? Correlation analysis was highly exploratory, due to the small sample size, but it might be suitable for further hypothesis generation. We found a relatively high consistency of distance and HRV (Figure 6 ), suggesting both the measured distance and HRV data indicate reliably a similar construct. Intentional interpersonal distance and HRV data show a weak negative correlation in NTP participants, suggesting that higher heart rate variability fosters prosocial activity and neurotypical participants with greater vegetative regulation could approach the experimenter more ( Figure 6 , upper triangle). In the case of participants with . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 autism, however, we found a positive association between distance and HRV results (except baseline HRV). ASD participants with higher HRV preferred longer distances (Figure 6 , lower triangle) Figure 6 Correlations between interpersonal distance and heart rate variability at the baseline and during the intentional interpersonal distance conditions. Dist = distance, HRV = heart rate variability, preHRV = 10s RMSSD, Eye = eye contact, No eye = no eye contact, Active = active moving, Passive = standing, Self = attribution to self, Other = attribution to the other conditions. Upper triangle: neurotypical participants, lower triangle: participants with autism. Warm colors refer to positive, cold colors refer to negative Spearman rank correlation rho values, grey asterisk marks the significant p values after (fdr) correcting for multiple comparisons. Results of psychometric questionnaires showed weak or no association with distance and HRV results ( Figure 7) ; however, the association with psychometric questionnaires in ASD showed a different pattern than in NTP. High trait anxiety level, poor mentalization, and attachment were weakly associated with greater interpersonal distance in the intentional . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 condition in autism (Figure 7 , lower triangle, second column), but neither of these correlations was significant after correction for multiple comparisons. Correlations between interpersonal distance, heart rate variability at the baseline and during the intentional interpersonal distance conditions, and psychometric data. Dist = distance, HRV = heart rate variability, AQ = Autism-spectrum Quotient, AAS = Adult Attachment Scale, MZQ = Mentalization Questionnaire, STAI-T = State-Trait Anxiety Inventory, Trait, ADI = Autism Diagnostic Interview-Revised, ADOS = Autism Diagnostic Observation Schedule. Upper triangle: neurotypical participants, lower triangle: participants with autism. Warm colors refer to positive, cold colors refer to negative Spearman rank correlation rho values, grey asterisk marks the significant p values after (fdr) correcting for multiple comparisons. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint In our study, we aimed at investigating interpersonal space regulation and the underlying vegetative response regulation in autism. To this end, we introduced an innovative methodology combining interpersonal distance measurement and vegetative parameter registration in incidental (automatic), and intentional (reflective) experimental settings in groups of adult participants with autism spectrum disorder and their matched neurotypical controls. We found increased interpersonal distance measured by the intentional interpersonal distance paradigm in autism, but there was no difference in the incidental setting. The difference was the most pronounced when participants approached actively and maintained eye contact with the experimenter during the intentional interpersonal distance task. Moreover, we found decreased baseline heart rate variability and decreased HRV reactivity in autism. The preferred interpersonal distance was similar among neurotypicals and participants with autism when they set their distance incidentally. In incidental conditions, both neurotypical and autistic participants took place at nearly three times greater distance than during the intentional conditions. When participants indicated the preferred distance at the beginning of the experiment, they mostly chose a distance within the zone of usual social distance (far phase 210 -370 cm, close phase 120 -210 cm). The social space (the zone between personal and social distance) is reserved for strangers or new acquaintances according to Hall's proxemic rules (Hall, 1966) ; all our participants behaved accordingly and chose a place automatically within this zone. Individual variability was high in both study groups, but they did not differ from each other. None of the incidentally chosen distances were associated with clinical variables. We argue that the incidental setting is suitable to measure the far phase of usual social distance, while by applying the intentional setting we estimate the personal distance. The interpersonal distance was larger in ASD during the . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint intentional interpersonal distance task that is a modified version of the stop distance paradigm. Participants were directly instructed to define a still comfortable position in relation to the position of the experimenter. Usually, in a stop distance paradigm, the participant and the experimenter are facing each other at the endpoints of a 3 to 6 meters long line along which the participants set their preferred interpersonal distance. In our experiment, we have chosen 5 meters as the initial distance. During this task, participants set closer distances than in the incidental setting, on average within the personal zone (far phase ~75 -120 cm, close phase ~45 -75 cm, the zone between the intimate distance and personal distance). Participants with autism set significantly greater distances: at the far region of the personal space zone. Prior to our experiment, five studies examined interpersonal space regulation in ASD, four of which studied children or adolescent populations. The only adult study observed no differences between participants with ASD and neurotypical controls (Kennedy & Adolphs, 2014) . However, the findings suggest that interpersonal space regulation is altered in ASD in childhood, autistic children preferred significantly larger interpersonal distance than neurotypical control participants (Candini et al., 2017; Gessaroli et al., 2013) . A study examining adolescents with ASD also concluded that their space regulation was altered. Interestingly, this conclusion was derived from opposing results: adolescents with ASD preferred shorter interpersonal distance than neurotypical controls (Asada et al., 2016) . Although, in this study from Japan, neurotypical participants preferred longer personal distances (cca. 130 -150 cm depending on condition) than in other cohorts, and the distances preferred by ASD participants were comparable to our and Kennedy et al.'s results (cca. 70 -100 cm) (Kennedy, 2014) . This raises the possibility of cultural differences in social rules and customs, including personal space arrangement. By applying an incidental condition in our study, we tried to capture the effect of the implicitly acquired social norms on interpersonal . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 distance regulation. Initially, when enough space was available, participants' behaviour was automatic, and the experimenter was a new acquaintance or a stranger; each group set a larger, in fact similarly large distance. Our results suggested that adult individuals with autism prefer greater interpersonal distance than neurotypicals only if they were aware of the situation. Several factors can influence the interpersonal distance between the experimenter and the participants (Candini, 2017; Candini 2020; Asada, 2016) . Being in either an active or a passive role during space regulation also influences the length of preferred interpersonal distance; regardless of group membership, participants tended to set a shorter distance when they were asked to approach the standing experimenter and stop when reaching optimal proximity (active role). The authors suggested that people compensate for the unpredictability of someone else approaching them (passive role) by keeping them further away from themselves; this effect was observed among ASD and neurotypical children as well previously (Candini et al., 2017) . We found an opposite effect: adult ASD and NTP participants stop at a greater distance in the active roles. Reflecting on this incongruity, we suggest that it may be due to age differences, or a cultural phenomenon; adult participants might be more polite and find it harder to stop the other. We argue that self-conscious determination of interpersonal distance is an important aspect of socially acceptable but assertive behaviour, which should be part of the social communication training in autism. Eye contact and active/passive role have been shown to affect the preferred interpersonal distance of ASD and neurotypical adolescents (Asada et al., 2016; Candini et al., 2020) . Only one of the five aforementioned studies manipulated the factor of eye contact between trials: in eye contact conditions when participants held passive roles, they preferred larger interpersonal distance, regardless of the group membership, and this effect did not emerge when holding active roles (Asada et al., 2016) . In contrast, in our study participants set a larger . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 interpersonal distance during active approaching with eye contact than without it. Within the conditions with maintained eye contact, we measured larger interpersonal distance during active, than passive roles. Furthermore, participants with ASD set greater interpersonal distance while actively approaching with maintained eye contact than NTPs in any condition, and than ASD participants themselves in other conditions, except passive roles without eye contact. We can speculate that eye contact might cause distress and evoke avoidance in autism, instead of an appetitive or comforting positive reinforcement, in line with the "eye avoidance" hypothesis (Tanaka & Sung, 2016; Tottenham et al., 2014) . Future studies directly examining this phenomenon and testing whether and how eye contact plays a role, seem warranted. Reciprocal social interactions are impaired in autism, it is often challenging to adapt the other person's perspective. In our study, besides the effect of eye contact and activity in approaching, we introduced attribution of self (participant) or the other (experimenter) as an additional variable to test whether our participants could consider their partners' hypothesized intentions. Considering the theory of mind difficulties in autism, we hypothesized that participants with ASD will show less difference along the attribution dimension than NTPs. We found no difference in ASD accordingly. Neurotypicals estimated the experimenter's preferred distance slightly greater than their own, however, this difference was not statistically significant. These results might suggest that participants with autism have decreased capacity to modify their behaviour according to others' aspect, in line with the well-known deficits in theory of mind in autism (Frith et al., 1991; Livingston et al., 2019) , but to test this hypothesis further studies are needed. An invention in our experimental design is that we combined interpersonal distance measurement with heart rate registration. In social behaviour parasympathetic regulation, the flexibility of vagal tone plays an important role according to Porges' polyvagal theory . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; (Muscatello et al., 2021; Porges, 2003 Porges, , 2009 . Resting high-frequency HRV was found to be associated with cooperative behaviour, using less disengagement and more socially adaptive emotion regulation strategies among healthy adults (Beffara et al., 2016; Geisler et al., 2013) . Variables associated with parasympathetic regulation (e.g. respiratory sinus arrhythmia) are associated with emotion recognition and symptom severity in autism (Bal et al., 2010) . Embodied mentalization theory refers to an inferential brain process of detecting, identifying, and regulating internal bodily signals (Debbané et al., 2016; Luyten & Fonagy, 2015) . Mentalizing required in social situations, however, can be described along four dimensions: automatic-controlled, internal-external, self-other, cognitive-affective mentalization (Fonagy & Luyten, 2009 ). High levels of stress and hyperarousal can disturb the balance at any dimension, and adaptive hypoactivity of the hypothalamic-pituitary-adrenal axis, parasympathetic rather than further sympathetic activation, might help regain it. We found reduced baseline heart rate variability in participants with autism spectrum disorder, in line with the results of previous studies corroborating altered autonomic nervous system functioning in ASD (Arora et al., 2021; Cheng et al., 2020; Thapa et al., 2019) . Furthermore, the HRV decrease induced by participating in a social situation was lower, assuming a decreased regulatory capacity in the ASD group (Cheng et al., 2020) . The average heart rate (reflecting sympathetic activity) was slightly higher in the ASD group than in neurotypicals, but this difference was not significant. We used the RMSSD method to measure HRV which captures the parasympathetic regulation rather than sympathetic arousal (Shaffer & Ginsberg, 2017) . Parasympathetic regulation is more relevant during social interactions, such as in situations where interpersonal distance regulation is needed (Porges, 2003) . We calculated HRV for 60 seconds at the baseline and during the two (incidental and intentional) task conditions. Psychotropic medication, caffeine intake, smoking, and exercising habits might influence autonomic regulation (Alvares et al., 2016; Benjamim et al., . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint 2020; Hautala et al., 2009; Middlekauff et al., 2014) , but study groups in this study did not differ in actual or regular caffeine intake, smoking and exercising habits (Table 1) . There were no significant differences between participants who took medication and those who did not. We found no significant difference, neither in heart rate nor in HRV considering these variables. Additionally, we calculated HRV during the social distance regulation task in 10 s time periods, applying ultra-short-term analysis (Salahuddin; 2007) exactly at the time point when participants arrived at the reported distance in order to take a closer look at the relationship of interpersonal distance and vegetative regulation. These results should be considered as hypothesis-generating findings and demands further testing in larger samples. To experimentally test more nuanced aspects of interpersonal distance regulation, we assessed the modulatory effect of eye contact in active or passive roles, and attribution. Eye contact was previously described as a factor unconsciously avoided in autism resulting in further difficulties in reading socially important signals (Joseph et al., 2008; Madipakkam et al., 2017) . Although we did not find a significant difference between study groups regarding the 10 s HRV metrics, we can speculate that, however, the reactive HRV values during the social distance task were somewhat reduced compared to the baseline, due to the baseline difference between groups, this reactive decrement was smaller in ASD and this might have prevented the further fine-tuning in autism. The polyvagal theory also provides a model of how prosocial behaviour, self-soothing, and calming are associated with the activity of parasympathetic regulatory function of the myelinated vagus nerve by inhibition of sympathetic-adrenal influences (active avoidance) or preventing freezing (passive avoidance) mediated by the unmyelinated vagus nerve (Porges, 2003) . This raises the possibility that disturbances in social interactions in autism are to some extent the consequence of neurophysiological states facilitating defensive rather than prosocial behaviours (Patriquin et . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint al., 2019; Porges et al., 2013) . Reduced regulatory capacity to which reduced baseline HRV refers, combined with elevated amygdala reactivity may lead to early exhaustion even during minimal social interaction. We can speculate that the larger interpersonal distance is the consequence of this early exhaustion of regulatory capacity and by keeping the distance they might avoid a more severe vegetative disturbance in social situations. To further test this hypothesis in real-life situations, applying widely available wearable devices might be useful, the experience gained in this way could later be used for example in the development of biofeedback tools in social communication training in autism. Despite the most careful planning, every study has its limitations. In this study, we examined adult participants with autism trying to fill a gap in this research area and measure interpersonal distance and vegetative regulation simultaneously. We recruited participants with average or above-average intellectual abilities, which increases the likelihood of adaptive skill acquisition. To overcome this limitation inclusion of a broader spectrum of autistic participants is needed in future studies. We found greater interpersonal distance in autism measured by the modified version of the stop distance paradigm, but there was no difference between study groups regarding heart rate variability during that part of the experiment. In subsequent research HRV differences should be measured at fixed distances as well, e.g. closer than the comfortable range for neurotypicals, farther than the comfortable range for participants with autism, and at intermediate distances, as well. Subjective rating of the level of comfort (both by participants and by the experimenter) might help to gain a better insight of how correctly the experimenter's perspective can be estimated by the participants. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint Unfortunately, we have had to finish the data collection due to the beginning of the COVIDpandemic. As the introduced social distancing measures deeply affect our senses of interpersonal distance, we cannot continue with the original study design. We plan to investigate the effect of this unforeseen but global intervention by repeating our experiment when the pandemic subsides, and the social aspects of the situation are consolidated. Additional conditions with and without wearing face masks might be considered, too. These further studies will be able to show us whether autistic people have been affected differently than neurotypicals by social distancing measures. Interpersonal distance regulation is a relevant nonverbal part of social communication. It reflects the individual needs for personal space and the ability to read others' intentions. Together with other biomarkers of vegetative functions, this might express how demanding a simple social interaction can be for people with autism. In this study, we introduced a new experimental design to measure these factors together, along with four dimensions relevant in autism and in social interactions. We found similar social distance in an automatic, incidental setting among the two study groups, but participants with autism preferred larger interpersonal distance during the intentional conditions. The difference was most pronounced when they had to approach the experimenter actively with maintained eye contact. Participants with autism had lower baseline heart rate variability and decreased heart rate variability reactivity than neurotypicals; however, their HRV and its changes in different conditions during the social distance task did not differ significantly. This raises the possibility that regulatory capacities were exhausted sooner, at a farther distance and it might be the background of a compensatory avoidant behaviour in ASD. We believe that the application of this experimental design could also be beneficial in the study of other . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint psychiatric conditions, such as borderline personality disorder, social phobia, or psychosis. These results could further expand our understanding of interpersonal distance regulation in ASD. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 7, 2021. ; https://doi.org/10.1101/2021.08.06.21261686 doi: medRxiv preprint Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) Is autonomic function during resting-state atypical in Autism: A systematic review of evidence Reduced Personal Space in Individuals with Emotion Recognition in Children with Autism Spectrum Disorders: Relations to Eye Gaze and Autonomic State The development of a theory of mind in autism: Deviance and delay? The Psychiatric Clinics of North America Does the autistic child have a 'theory of mind'? The Autism-Spectrum Quotient (AQ): Evidence from Asperger Syndrome/High-Functioning Autism, Malesand Females, Scientists and Mathematicians Resting high frequency heart rate variability selectively predicts cooperative behavior The plasticity of the interpersonal space in autism spectrum disorder Personal space regulation in childhood autism: Effects of social interaction and person's perspective Heart rate variability in individuals with autism spectrum disorders: A meta-analysis Working models of attachment: Implications for explanation, emotion, and behavior Adult attachment, working models, and relationship quality in dating couples Variable heart rate and a flexible mind: Higher resting-state heart rate variability predicts better taskswitching Attachment, Neurobiology, and Mentalizing along the Psychosis Continuum A developmental, mentalization-based approach to the understanding and treatment of borderline personality disorder Cardiac vagal tone is associated with social engagement and self-regulation Personal Space Regulation in Childhood Autism Spectrum Disorders RR interval signal quality of a heart rate monitor and an ECG Holter at rest and during exercise The Hidden Dimension Is a self-rated instrument appropriate to assess mentalization in patients with mental disorders? A behavioural and electroencephalographic study of autistic children Assessing children with autism, mental retardation, and typical development using the Playground Observation Checklist Affective response to eye contact and face recognition ability in children with ASD Violations of Personal Space by Individuals with Personal Space Regulation by the Human Amygdala Heart Rate Variability and Cardiac Vagal Tone in Psychophysiological Research -Recommendations for Experiment Planning, Data Analysis, and Data Reporting Autism diagnostic observation schedule: A standardized observation of communicative and social behavior Autism Diagnostic Interview-Revised: A revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders Violations of Personal Space in Young People with Autism Spectrum Disorders and Williams Syndrome: Insights from the Social Responsiveness Scale The neurobiology of mentalizing A systematic review of physiological reactivity to stimuli in autism Unconscious avoidance of eye contact in autism spectrum disorder Neural Correlates of Interpersonal Space Permeability and Flexibility in Autism Spectrum Disorder. Cerebral Cortex, bhaa404 Altered bodily self-consciousness and peripersonal space in autism Evidence for decreased parasympathetic response to a novel peer interaction in older children with autism spectrum disorder: A case-control study Reliability and validity of the Mentalization Questionnaire (MZQ) in forensic care Are ultra-short heart rate variability features good surrogates of short-term ones? State-of-the-art review and recommendations Behavioral aspects of infantile autism: An ethological description Interpersonal distance and social anxiety in autistic spectrum disorders: A behavioral and ERP study The Polyvagal Theory: Phylogenetic contributions to social behavior The polyvagal theory: New insights into adaptive reactions of the autonomic nervous system Polar Electro Oy. Suncom Systems An Overview of Heart Rate Variability Metrics and Norms. Frontiers in Public Health A Critical Review of Ultra-Short-Term Heart Rate Variability Norms Research Validation of the Expanded Versions of the Adult ADHD Self-Report Scale v1.1 Symptom Checklist and the Adult ADHD Investigator Symptom Rating Scale Virtual Reality Immersion Rescales Regulation of Interpersonal Distance in Controls but not in Autism Spectrum Disorder Manual for the State-Trait Anxiety Inventory Reduced heart rate variability in adults with autism spectrum disorder