Ra

Ja

b

c

d

a

ARRA

KCSVP

1

hiSrbf

eocflPfavm

KG

0h

Journal of Anxiety Disorders 27 (2013) 627– 634

Contents lists available at ScienceDirect

Journal of Anxiety Disorders

esponses to voluntary hyperventilation in children with separationnxiety disorder: Implications for the link to panic disorder

oe Kossowskya,b, Frank H. Wilhelmc, Silvia Schneiderd,∗

Department of Clinical Psychology & Psychotherapy, University of Basel, SwitzerlandDepartment of Anesthesiology, Perioperative, and Pain Medicine, Children’s Hospital Boston, Harvard Medical School, USADivision of Clinical Psychology, Psychotherapy, and Health Psychology, Department of Psychology, University of Salzburg, AustriaDepartment of Clinical Child and Adolescent Psychology, Faculty of Psychology, Ruhr University Bochum, Germany

r t i c l e i n f o

rticle history:eceived 3 January 2013eceived in revised form 15 July 2013ccepted 2 August 2013

eywords:hildhood anxietyeparation anxiety disorderoluntary hyperventilation

a b s t r a c t

Background: Biological theories on respiratory regulation have linked separation anxiety disorder (SAD) topanic disorder (PD). We tested if SAD children show similarly increased anxious and psychophysiologicalresponding to voluntary hyperventilation and compromised recovery thereafter as has been observed inPD patients.Methods: Participants were 49 children (5–14 years old) with SAD, 21 clinical controls with other anxietydisorders, and 39 healthy controls. We assessed cardiac sympathetic and parasympathetic, respiratory(including pCO2), electrodermal, electromyographic, and self-report variables during baseline, pacedhyperventilation, and recovery.

sychophysiology Results: SAD children did not react with increased anxiety or panic symptoms and did not show signsof slowed recovery. However, during hyperventilation they exhibited elevated reactivity in respiratoryvariability, heart rate, and musculus corrugator supercilii activity indicating difficulty with respiratoryregulation.Conclusions: Reactions to hyperventilation are much less pronounced in children with SAD than in PDpatients. SAD children showed voluntary breathing regulation deficits.

. Introduction

Separation anxiety disorder (SAD) is one of the major child-ood anxiety disorders. Lifetime prevalence rates of childhood SAD

n recent studies lay between 4.1% and 5.1% (Kessler et al., 2005;hear, Jin, Ruscio, Walters, & Kessler, 2006). The most frequently

eported symptoms are separation-related distress, avoidance ofeing alone/without an adult and sleeping away from caregivers orrom home (Allen, Lavallee, Herren, Ruhe, & Schneider, 2010).

Abbreviations: AX, anxiety; CC, clinical controls; ECG, electrocardiogram; EMG,lectromyography; ET, end task; HC, healthy controls; HF, high frequency rangef heart period variability; HR, heart rate; IBI, interbeat intervals; ICG, impedanceardiography; MAP, mean arterial pressure; NSFn, non-specific skin conductanceuctuations; pCO2, arterial partial pressure of carbon dioxide; PD, panic disorder;EP, preejection period; RL, restlessness; RMSSD, root mean squared successive dif-erence; RR, respiratory rate; RSA, respiratory sinus arrhythmia; SAD, separationnxiety disorder; SCL, skin conductance level; TPR, total peripheral resistance; VH,oluntary hyperventilation; Vmin, minute ventilation; Vt, tidal volume; VtV, rootean squared successive breath-by-breath differences of tidal volume.∗ Corresponding author at: Ruhr-Universität Bochum, Fakultät für Psychologie,linische Kinder- und Jugendpsychologie, Universitätsstr. 150, 44789 Bochum,ermany. Tel.: +49 234 322 3169; fax: +49 234 320 3169.

E-mail address: silvia.schneider@rub.de (S. Schneider).

887-6185/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.janxdis.2013.08.001

© 2013 Elsevier Ltd. All rights reserved.

Donald Klein established the “separation anxiety hypothesis ofpanic disorder” by pointing to similarities between children withseparation anxiety and adults with panic disorder (PD). The bio-logical mechanism linking the two disorders was expressed in hissuffocation false alarm hypothesis (Klein, 1993) which suggestedthat both PD and the consistent respiratory abnormalities seen inpanic patients may be due to hypersensitive, medullary CO2 detec-tors. Klein’s work stimulated much research, yielding ambiguousresults. Three types of research have tested the separation anxietyhypothesis: retrospective reports of childhood SAD in adults withPD (e.g. Battaglia et al., 1995; Lipsitz et al., 1994); top-down (e.g.Biederman et al., 2001; Unnewehr, Schneider, Florin, & Margraf,1998; Warner, Mufson, & Weissman, 1995) and bottom-up (Last,Perrin, Hersen, & Kazdin, 1996; Martin, Cabrol, Bouvard, Lepine, &Mouren-Simeoni, 1999) studies of familial aggregation of the twodisorders, as well as research on biological correlates of SAD andPD (Pine, Coplan, et al., 1998; Pine et al., 2000). A recent meta-analysis demonstrated that a childhood diagnosis of separationanxiety disorder significantly increases the risk of panic disorder

and any anxiety disorder (Kossowsky et al., 2013).

A large research literature exists on psychophysiological base-line functioning and response to threat stimuli in adult anxietydisorders such as PD and post-traumatic stress disorder (for a

6 nxiety

roirtptM2rasld&lPthadRs&(SdSCt(ti

toFisLsecw&baPaaiepabaseopd(c(wo

28 J. Kossowsky et al. / Journal of A

eview, see Craske et al., 2009) that has yielded mixed resultsn the specificity of psychophysiological indicators for enhanc-ng nosological differentiation of anxiety disorders. One area ofelatively strong findings in this regard are psychophysiologicalheories proposing a direct connection between changes in arterialartial pressure of carbon dioxide (pCO2) due to hyper- or hypoven-ilation and the experience of anxiety and panic in PD (e.g. Ehlers &

argraf, 1989; Meuret et al., 2011; Meuret, Wilhelm, Ritz, & Roth,008; for a critical review see Roth, Wilhelm, & Pettit, 2005). For thiseason, research on the biological correlates of childhood SAD anddult PD has relied on biological challenges aimed at the respiratoryystem. These challenges have increasingly been applied as both aaboratory model for evoking anxious responding, and as a potentialiagnostic marker for anxiety disorders, particularly PD (Wilhelm

Roth, 2001; Zvolensky & Eifert, 2001). Most research on venti-atory physiology in children with SAD (Pine, Coplan, et al., 1998;ine et al., 2000; Roberson-Nay et al., 2010) utilized CO2 challengeechniques, which are based on the association between PD andeightened response to CO2, in which patients with PD report morenxiety, dyspnea, and panic attacks than comparison participantsuring CO2 inhalation (e.g. Blechert, Wilhelm, Meuret, Wilhelm, &oth, 2010; Gorman et al., 1997; Klein, 1993). The heightened CO2ensitivity has been suggested to be a risk factor for PD (Papp, Klein,

Gorman, 1993) and has also been associated with childhood SADBattaglia et al., 2009; Pine et al., 2000; Roberson-Nay et al., 2010).ome studies investigating respiratory abnormalities have foundifferences between children with mixed anxiety groups, includingAD, and non-anxious children: enhanced respiratory rate duringO2 inhalation, as well as elevated minute ventilation, increasedidal volume, and lower end-tidal pCO2 during room-air breathingPine, Coplan, et al., 1998; Pine et al., 2000). However, these respira-ory differences during room-air breathing could not be replicatedn another study (Kossowsky, Wilhelm, Roth, & Schneider, 2012).

Another type of respiratory biological challenge is volun-ary hyperventilation (VH), which involves breathing in excessf metabolic demand (Unnewehr, Schneider, Margraf, Jenkins, &lorin, 1998; Zvolensky & Eifert, 2001). VH is based on the find-ng that hyperventilation is critical in the production of somaticymptoms and the development of clinical anxiety conditions (Carr,ehrer, Hochron, & Jackson, 1996; Margraf, 1993). For example,tudies found that persons with PD reported greater anxiety andxperienced more panic attacks during VH compared to nonclini-al persons, persons with generalized anxiety disorder, and personsith social phobia (Gorman et al., 1988; Rapee, Brown, Antony,

Barlow, 1992). Ordinarily when hyperventilation is stopped,reathing patterns adjust spontaneously, normalizing pCO2 within

few minutes (Roth, 2005). This pCO2 recovery is compromised inD patients (Gorman et al., 1988; Maddock & Carter, 1991). After

3-min hyperventilation period, PD differed from social phobiand controls in having much slower symptomatic and physiolog-cal recovery (Wilhelm, Gerlach, & Roth, 2001). Group differenceffect sizes (Cohen’s d) in pCO2 at the end of the 10-min recoveryeriod in this study were in the order of 1.2, indicating consider-ble diagnostic specificity. Slow pCO2 recovery was accompaniedy slower recovery in heart rate (HR) and skin conductance levels,s well as increased reports of shortness of breath and anxiety. Inum, previous results have pointed to a greater anxiety-inducingffect of VH challenge procedures in PD patients, as compared withther anxiety disorders and controls, indicating its utility as a usefulrovocation method. Research has recently turned to specific car-iac autonomic indices regulating HR. Respiratory sinus arrhythmiaRSA), a measure of the magnitude of rhythmic fluctuations in HR

aused by respiration, is the preferred indicator of vagal activityBeauchaine, 2001; Berntson, Cacioppo, Quigley, & Fabro, 1994),hile cardiac preejection period (PEP) has been shown to be one

f the best non-invasive measures of cardiac sympathetic activity

Disorders 27 (2013) 627– 634

(Berntson et al., 1994). Hyperventilation in adults with PD led todecreased vagal and increased sympathetic activity (Sullivan et al.,2004).

The current study aims to expand research on biological corre-lates of SAD and their potential link to PD by examining if VH elicitssimilar symptoms in children with SAD as were found in adultswith PD. Further, we examine if children with SAD exhibit a slowphysiologic and symptomatic recovery from VH, as demonstratedby adults with PD. Our measures were selected to register a widearray of autonomic and respiratory correlates of anxiety (Pine et al.,2000; Wilhelm & Roth, 1998) in children diagnosed with SAD, chil-dren diagnosed with other anxiety disorders beside SAD (clinicalcontrols, CC), and children without current or past mental disorders(healthy controls, HC). We expected a disorder-specific reactionpattern across the three groups: First, that only SAD children wouldreport significantly more anxiety, physical symptoms, and anxietyrelated cognitions during VH and recovery than would the CC andHC children. Second, that VH and recovery would be associatedwith higher sympathetic activation and more vagal withdrawal inthe SAD compared to the other groups (Sullivan et al., 2004). Third,that compared to the HC and CC groups, children with SAD wouldexhibit a slower pCO2 recovery.

To test our predictions, we assessed heart rate, PEP, skin con-ductance level, and RSA, as the primary physiological measuresof autonomic dysregulation in anxiety (Wilhelm & Roth, 1998);end-tidal pCO2, respiratory rate, tidal volume, and minute venti-lation as indices of the quality and quantity of respiratory changes;and self-reported anxiety, as well as a set of symptoms relatedto panic and VH as psychological measures. In addition, we mea-sured facial musculus corrugator supercilii electromyography asan index of negatively valenced facial expression or mental effort(Tassinary, Cacioppo, & Vanman, 2007), as well as tidal volumevariability because elevations in this variable have been observedrepeatedly in PD patients during baseline and anxiety provoca-tions (Abelson, Weg, Nesse, & Curtis, 2001; Martinez et al., 1996;Papp et al., 1995; Stein, Millar, Larsen, & Kryger, 1995; Wilhelm,Trabert, & Roth, 2001a,b) and were also prevalent during recoveryfrom VH (Wilhelm, Gerlach, et al., 2001). In addition, this measureis sensitive to difficulty in following breathing instructions duringVH smoothly and thus provides a putative measure of voluntarybreathing regulation inefficiency (Khoo, 1999)

2. Method

2.1. Participants

The methods are described in detail in our previous publication(Kossowsky et al., 2012). In brief, the experimental groups consistedof 49 children with a primary diagnosis of SAD, 21 CC with a pri-mary diagnosis of anxiety disorder other than SAD, and 39 HC notmeeting criteria for any current diagnosis. The ages of the sam-ple can be found in Table 1. Seven participants in the SAD groupand four participants in the CC group were recruited through localchild and adolescent psychiatrists, psychologists, and pediatricians.The rest of the sample was recruited through newspaper advertise-ments and flyers. The diagnoses were assessed by trained doctoralstudents in clinical child psychology, blinded to group status, usingthe Diagnostic Interview for Mental Disorders in Children and Ado-lescents (Kinder-DIPS; Neuschwander, In-Albon, Adornetto, Roth,& Schneider, 2013; Schneider, Unnewehr, & Margraf, 2009). It is awell-validated structured interview for diagnosing DSM-IV disor-

ders in children and has alternate forms for children and parents.Inclusion criteria were knowledge of the local language, the chil-dren’s and their parent’s informed consent, and completion ofpsychological assessments. Children were excluded if they were

J. Kossowsky et al. / Journal of Anxiety Disorders 27 (2013) 627– 634 629

Table 1Demographic and psychometric characteristics of the study groups.

SAD groupMean ± SD

CC groupMean ± SD

HC groupMean ± SD

Statistic Post hoc significance

n in sample 49 21 39% female 51.0 66.7 48.7 �2(2) = 1.94Age (years) 8.3 ± 2.5 9.5 ± 2.4 9.9 ± 2.3 F(2,106) = 4.71* SAD < CC = HCAge range 5–13 5–14 6–14Hours sleep in last 24 h 9.76 ± 1.10 9.41 ± 0.87 9.55 ± 1.13 F(2,105) = .87Caffeine intake in last 24 ha 0.45 ± 1.20 0.78 ± 1.57 0.68 ± 1.47 F(2,102) = 50SAI-C 1.82 ± 0.91 1.44 ± 0.57 0.56 ± 0.61 F(2,83) = 23.41*** SAD = CC > HCSAI-P 2.35 ± 0.81 1.72 ± 0.84 0.56 ± 0.46 F(2,89) = 52.93*** SAD > CC > HCSASC-R 2.38 ± 0.58 2.84 ± 0.84 1.93 ± 0.63 F(2,57) = 7.02** SAD = CC > HCRCMAS-C 0.43 ± 0.16 0.43 ± 0.13 0.27 ± 0.14 F(2,96) = 12.72*** SAD = CC > HCRCMAS-P 0.41 ± 0.17 0.39 ± 0.16 0.27 ± 0.14 F(2,72) = 5.94** SAD = CC > HC

Note: SAD = separation anxiety disorder; CC = clinical controls; HC = healthy controls; SAI-C = Separation Anxiety Inventory for Children; SAI-P = Separation Anxiety Inventoryfor Parents; SASC-R = Social Anxiety Scale for Children-Revised; RCMAS-C = Revised Children’s Manifest Anxiety Scale-Child version; RCMAS-P = Revised Children’s ManifestAnxiety Scale-Parent version

* P < .05

tftrt

de2ag4wcmAsrAGidtc

2

Rt8abnrp(wlwnWdc

** P < .01.*** P < .001

a Number of drinks containing caffeine.

aking psychotropic medication. Children with SAD were offeredree diagnostic assessment and treatment for their participation inhe study, while children in the control groups received a monetaryeward. The local ethics committee for medical research approvedhe study.

Psychometric assessment of the study groups included theisorder-specific Child and Parent versions of the Separation Anxi-ty Inventory for Children (SAI-C/P; In-Albon, Meyer, & Schneider,013) in which parents and children assessed the degree of avoid-nce of separations in a variety of settings (e.g., “I/my child avoid/soing to sleep alone”) using a 5-point scale ranging from 0 (never) to

(always). The psychometric properties of the SAI-C/P were goodith a test-retest reliability of r = .84. Internal consistency of the

urrent sample was alpha = .88 (children), .93 (parents). Further-ore, to measure trait anxiety, the Revised Children’s Manifestnxiety Scale (RCMAS; Reynolds & Richmond, 1978; German ver-ion, Boehnke, Silbereisen, Reynolds, & Richmond, 1986) includingatings from both the children and the mothers was used. The Socialnxiety Scale for Children-Revised (SASC-R; La Greca & Stone, 1993;erman version, Melfsen & Florin, 1997) was included, as the orig-

nal study design was to recruit a group of children with a primaryiagnosis of social phobia as the clinical control group. Both ques-ionnaires, the RCMAS and the SASC-R, demonstrated good internalonsistency (Cronbach’s ˛: 0.82–0.92) in the current sample.

.2. Procedures

With regard to breathing duration, Meuret, Ritz, Wilhelm, andoth (2005) suggested a minimum length of 2 min for hyperven-ilation and a post-hyperventilation observation period of at least

min. Hornsveld, Garssen, and van Spiegel (1995) found that nearlyll symptoms appear after 3 min of VH. Levels of pCO2 fallingelow 35 mmHg typically indicate that breathing is in the hypocap-ic range (Oakes, 1996) and levels around 30 mmHg or lower onepeated occasion or across longer measurement periods have beenroposed as being indicative of “unequicoal chronic hypocapnia”Bass & Gardner, 1985). Meuret et al. (2005) indicate that a PD groupas physiologically distinguishable from controls when the pCO2

evel reached 20 mmHg. In our piloting with children of ages 5–14,e found that it is very difficult, if not impossible for a significant

umber of them, to keep pCO2 at 20 mmHg for several minutes.e therefore chose a target level of 25 mmHg, in order to stan-

ardize this important experimental variable and to better comparehildren and groups.

Two experimenters conducted the individual testing sessions,which took approximately 90 min and consisted of three standard-ized tasks in succession, starting with a separation task (Kossowskyet al., 2012), followed by a mild social stressor and the hyper-ventilation test presented in this paper, in a temperature- andsound-controlled room. Special precautions were taken to accom-modate the participating children to the testing situation andapparatuses (see Wilhelm, Schneider, & Friedman, 2005). Partici-pants were seated in a comfortable armchair at a 90◦ angle from themother, who was located in one corner of the room approximately3 meters from the child, so that eye contact for both parties waspossible. Electrodes and sensors were attached. The behaviors ofthe child and mother were monitored in an adjoining control roomthrough two discreetly placed cameras. Communication with theparticipant was via an intercom.

The VH paradigm began with a three-minute baseline. Dur-ing VH the participants paced their breathing at 20 breaths perminute by synchronizing it with a voice recording instructing theparticipants when to breathe in and out through their nose. Theexperimenters monitored the participants’ end-tidal pCO2, withthe goal of inducing each subject to quickly reach and maintaina pCO2 at or slightly lower than 25 mmHg. Children who didnot reach this threshold within a minute were instructed to takedeeper breaths. After this period, participants were instructed tosit still and resume normal breathing for nine minutes. Self-reportmeasures of anxiety and symptoms were collected by means ofquestionnaires immediately after the baseline and recovery period.On the second occasion, participants were asked to fill out oneset of questionnaires to describe how they felt at the end of therecovery period and then to fill out a second set to describe retro-spectively how they felt during the fast breathing. If necessary, theexperimenter assisted the child in filling out the questionnaires.

3. Measures

3.1. Self-report measures

Subjective ratings by the children were recorded during thephysiological testing. For this purpose a specially developedchild-friendly, 11-point Likert Scale questionnaire (0 = not at all,10 = extremely) was used (Wilhelm et al., 2005). In addition to a

number of items that were described in detail in an earlier reportmeasuring anxiety, restlessness, and panic related cognitions andsymptoms (Kossowsky et al., 2012), we also included a hyperventi-lation subscale, which specifically assessed symptoms related to

6 nxiety

hhScattNdlc

3

hGfoafe

3

iattrmXEr(tsrtibp

3

gmasgct

3

rNtrTfrpri

30 J. Kossowsky et al. / Journal of A

yperventilation (HVS, pressure in or tightening of chest, light-eadedness or fainting, things seem unreal, tingling sensations).ince answering the questionnaires occasionally proved to be diffi-ult for children younger than 8 years, a short version was created. Itssessed anxiety (AX), restlessness (RL), the four symptoms relatedo HV, and whether the child felt he/she would have liked to endhe task (ET). 13 children completed only the shortened version.o difference was found in the pattern of results between the chil-ren completing the shortened version and those completing the

ong version (p > .2). All questionnaires demonstrated good internalonsistency (Cronbach’s ˛: 0.83–0.94) for this sample.

.2. Physiological measures

Physiological channels were recorded at 1000 Hz using BIOPACardware and AcqKnowledge software (BIOPAC Systems, Inc.,oleta, CA, USA). Data reduction and editing of artifacts were per-

ormed using ANSLAB software (Wilhelm & Peyk, 2005). Placementf electrodes/sensors, recording and data reduction were done inccordance with published guidelines and conventions establishedor psychophysiological research (e.g. Fowles et al., 1981; Sherwoodt al., 1990).

.2.1. Cardiovascular measuresR-waves from a standard Lead-II electrocardiogram (ECG) were

dentified automatically. Cubic spline interpolation and resamplingt 4 Hz were used to convert interbeat intervals (IBI) into an equidis-ant time series. Eight spot electrodes were attached pairwise tohe neck and thorax to obtain an impedance cardiogram (ICG). Theaw ICG dz/dt-signal was ensemble averaged over 1-min experi-ental intervals to allow for reliable identification of the B-, Z- and-points. Preejection period was calculated as the interval fromCG Q-point to ICG B-point. We examined the individual respi-atory rates to see if they lay outside the 9–30 cycles/min range0.15–0.5 Hz). Since this was never the case, RSA was quantified byhe natural logarithm of the summed Welch power spectral den-ity of IBI in this range, corresponding to the high frequency (HF)ange of heart period variability. The choice of the 0.5-Hz cutoff ashe higher bound for the HF band is also in line with prior studiesn children (e.g. Pine, Wasserman, et al., 1998). PEP and RSA are theest available non-invasive measures of cardiac sympathetic andarasympathetic efferent activity, respectively.

.2.2. Electrodermal measuresTwo Ag/AgCl Beckman electrodes filled with isotonic electrode

el were attached to the volar surfaces of the medial index andiddle fingers of the child’s non-dominant hand. A constant volt-

ge of 0.5 V between the electrodes afforded calculation of averagekin conductance level (SCL). The signal was scanned for risesreater than 0.02 �Siemens from a zero-slope baseline, which wereounted as the number of non-specific skin conductance fluctua-ions (NSFn).

.2.3. Respiratory measuresThe following respiratory variables were calculated from tho-

acic and abdominal pneumographic channels (James Long, Inc.,ew York) calibrated for each individual using the mean of

wo fixed volume breathing calibration procedures: respiratoryate (RR), tidal volume (Vt), minute ventilation (Vmin) (Wilhelm,rabert, & Roth, 2001a). The root mean squared successive dif-erence (RMSSD) of Vt (VtV) was the metric of breath-by-breath

espiratory variability for Vt. Expiratory pCO2 (end-tidal partialressure of expired CO2) was measured continuously using capnog-aphy (N-1000, Nellcor) and dual nostril prongs. Participants werenstructed to keep their mouth closed and breathe only through

Disorders 27 (2013) 627– 634

their nose. End-tidal pCO2 was defined as the level at which pCO2stopped rising at the end of an expiration.

3.2.4. Electromyographic measuresThe musculus corrugator supercilii electromyography (EMG)

raw signal from two electrodes at the left eye-brow was filtered,rectified, and smoothed using a 50 ms moving average window.

4. Data analyses

Data were visually inspected for normal distribution and out-liers. For each subject, a spreadsheet was assembled containingphysiological parameters for each test segment for review. Datapoints 2 SDs above or below the mean for a subject during an epoch,prompted reanalysis for validation without knowledge of the clini-cal status. If data within conditions were missing, the interpolationmethod described by Stemmler (1989) was applied. Percentage ofmissing data for single physiological variables ranged between 0%and 5%.

Three comparisons between parallel self-report and physiolog-ical data were made: baseline differences between the groups,differences in reactivity (HV phase minus baseline measurement),and differences in recovery (HV recovery phase minus HV phase).

To avoid alpha inflation, the selected physiological variableswere submitted to multivariate analyses of covariance (MANCO-VAs) controlling for age. Linearity was assessed from scatterplots. Ifthe MANCOVA was significant, the individual ANCOVAs controllingfor age were analyzed. If they were significant, group contrasts con-trolling for age are reported. This procedure was done separatelyfor baseline, reactivity, and recovery analyses.

Non-parametric Kruskal–Wallis ANOVA and Mann–WhitneyU tests were used to analyze the self-report data since theirdistribution was strongly left-skewed and could not be normal-ized by transformation. Our non-parametric tests were performedonce with the whole sample and once with an age-stratifiedsample (overlapping age range: 6–13). The patterns of resultswere equivalent, and only the analyses for the whole sample arereported. Significant Kruskal–Wallis ANOVA results were followedby Mann–Whitney U tests. Significance levels were set at p<.05 formeasures relevant for the reported hypotheses (HR, PEP, NSFn, RSA,pCO2, RR, Vt, Vmin, VtV, EMG) and p < 0.01 for other measures.All statistical tests were two-tailed. Effect sizes (Cohen’s d) werecalculated for the three pairwise group comparisons on primarymeasures. Group comparisons used uncorrected p-values.

5. Results

5.1. Sample characteristics

Table 1 compares the demographic and clinical characteristics ofthe SAD, CC, and HC groups. While the groups differed significantlyin age, these ages overlapped to a very high degree (overlap 6–13years). The groups did not differ in sex distribution. The groups didnot differ with regard to pre-challenge caffeine intake, substanceuse, sleep, as well as general medication usage (p’s > 0.5) in the last24 h. The SAD and CC groups reported higher mean scores than theHC group on all questionnaires administered during the diagnosticphase. As expected, the SAD group scored significantly higher on thedisorder-specific Separation Anxiety Inventory than the CC groupbut only on the parent form (SAI-P).

5.2. Self-report measures

Fig. 1 presents the group means. Group differences duringbaseline were found for anxiety, �2(2, N = 107) = 12.83, p = .002)

J. Kossowsky et al. / Journal of Anxiety Disorders 27 (2013) 627– 634 631

F f the sB 10.

abUitpt

pigbt

5

wt

FlfFccVnd

ig. 1. Self-report data (means and standard errors) of the principal measures o = baseline; HV = hyperventilation; R = recovery. Total potential range of scores: 0–

nd desire to end task, �2(2, N = 107) = 14.21, p = .001). Duringaseline, both the SAD U(46,37) = 686.0, p = .010, d = .58) and CC(37,19) = 211.5, p < .001, d = 1.17) group differed from the HC group

n anxiety. While no child chose to terminate the experiment, bothhe SAD U(46,37) = 657.0, p = .012, d=.57) and CC U(37,19) = 178.0,

< .001, d = 1.21) reported significantly more desire to end the taskhan the HC group during baseline.

Although there was a significant increase in anxiety (Z = 2.64, = .008), restlessness (Z = 3.24, p = .001) and panic symptoms sever-ty (Z = 4.10, p < .001) during VH for all groups combined and forroups individually, no significant group differences could be foundetween the groups with regard to reactivity or recovery in any ofhe self-report measures.

.3. Physiological measures

Fig. 2 presents the group means. Since the baseline MANCOVAas not significant (Wilks � = .832; F(20,182) = .88, p=.61), no fur-

her analyses were undertaken.The reactivity MANCOVA was significant (Wilks � = .735;

(18,186) = 1.72, p = .039) and all groups reached an average pCO2evel of 25 mmHg. Significant group reactivity differences wereound in following measures: HR, F(2,105) = 3.35, p=.039, EMG,(2,105) = 8.87, p<.001, and VtV F(2,105) = 3.54, p = .033. Post hocontrasts showed significantly higher increases in the SAD group

ompared to HC in HR (p = .014, d = .58), EMG (p = .001, d = .79), andtV (p = .014, d = .56). Additionally, the SAD group displayed a sig-ificantly higher EMG increase than the CC (p = 001, d = .72). Noifferences were found between the HC and CC groups.

eparation anxiety group (SAD), clinical controls (CC), and healthy controls (HC).

The recovery MANCOVA was not significant (Wilks � = 781;F(18,182) = 1.32, p = .17), so that no further analyses were under-taken.

An ANOVA of pCO2 found no differences between the groups inany of the analyses (p’s ≥ 0.1).

6. Discussion

This study is the first to investigate autonomic and respiratoryresponses in children diagnosed with SAD, clinical controls, andhealthy controls during a voluntary hyperventilation challenge. Awide range of cardiac, respiratory, electrodermal, and experientialresponses was examined. Our use of paced breathing with pCO2feedback was successful in that respiratory rate was increased andpCO2 was lowered to equal levels in all three groups. Dropoutswere avoided by not requiring too extreme or prolonged respi-ratory efforts. However, our expectation that SAD children wouldreact similarly to the challenge as has been observed in PD couldonly partially be supported by the data.

During VH, SAD children exhibited higher reactivity in VtV andHR compared to HC and a higher reactivity in corrugator EMG com-pared to HC and CC. The exaggerated increase in VtV indicates thatthe SAD group had particular difficulty with breathing evenly dur-ing the instructed paced breathing with ongoing feedback to alterdepth of breathing in order to reach and maintain a set goal ofhypocapnia. According to Klein’s suffocation false alarm hypothe-

sis (Klein, 1993) increases in VtV may represent sighing or gaspingfor air. However, this feature should primarily appear during invol-untary breathing regulation typical for baseline or recovery periodswhere we did not find any abnormality in SAD.

632 J. Kossowsky et al. / Journal of Anxiety Disorders 27 (2013) 627– 634

Fig. 2. Principal physiological measures (means and standard errors) for the separation anxiety group (SAD), clinical controls (CC), and healthy controls (HC). B = baseline;HV = hyperventilation; R = recovery; EMG = musculus corrugator supercilii electromyography; lnHF = natural logarithm of high frequency spectral power of heart period vari-a tidal pm ute ve

piceprc(TctogtbfaAaopi

bility; NSFn = number of non-specific skin conductance fluctuations; pCO2 = end-ean squared successive breath-by-breath differences of tidal volume; Vmin = min

The absence of group differences in self-reported anxiety andanic symptom responses to voluntary hyperventilation, which is

n contrast to our hypothesis, indicates that HR and facial mus-le changes are most probably due to difficulty and increasedffort in voluntary breathing regulation during the challengingCO2 feedback-supported hyperventilation task rather than a cor-elate of anxious responding. Both psychophysiological variablesan be indicators of effort during physical activity and mental taskse.g. de Morree & Marcora, 2010; van Boxtel & Jessurun, 1993).ogether with the increased VtV during VH this suggests that SADhildren had more difficulty pacing their breathing according tohe ongoing pacing cues and smoothly following the concurrentccasional feedback to breathe more or less deeply, pointing toreater difficulty in voluntary breathing regulation. We reason thathis difficulty in voluntary breathing regulation in SAD, which haseen observed here for the first time and has not been reportedor PD probably due to a lack of research in this regard, may ben important physiological vulnerability feature of this disorder.lthough it was rather subtle in this study it could develop into

more pathological feature during mental stress episodes latern in life and contribute to the development of PD, a disorderhenomenologically characterized by quite prominent respiratory

rregularities (Roth, 2005; Wilhelm, Gevirtz, & Roth, 2001). Despite

artial pressure of expired CO2; RR = respiratory rate; Vt = tidal volume; VtV = rootntilation

this interesting finding, our results could not replicate the similari-ties in physiological (and particularly respiratory) reactivity foundin CO2 inhalation paradigms between childhood SAD and adult PD(Battaglia et al., 2009; Pine et al., 2000; Roberson-Nay et al., 2010)and also did not find the expected slow physiological recovery.

The only subtle perturbations we observed may be due to theVH method, which may not have been strong enough to elicit largeanxious responding. To prevent drop-outs, breathing rates, pCO2reduction goals, and length of VH were set to levels that shouldhave been attainable by most, if not all of the children, but actu-ally were not far away from levels that were successful in inducinganxiety and panic in adults with PD. Future research could investi-gate whether a lower pCO2 level would be attainable by children inthis age group for a longer duration and thus be more successful ineliciting substantial anxiety. In studies on adult PD directly compar-ing CO2 inhalation and VH, CO2 inhalation was typically the morepanicogenic challenge (Gorman et al., 1988; Papp et al., 1997). Thismay be explained by existing challenge research suggesting thatcontrollability, predictability, and safety signals serve a mediating

role between the onset of bodily arousal and anxious responding(Zvolensky & Eifert, 2001). This would imply lower anxiety in thecondition most under the participants’ control, namely the volun-tary hyperventilation, compared to that in the conditions where

nxiety

tnpl

eriOrabvplaliect

riawaafgKei(fiawdeseca(P

napwns

fStipafbfdtpj

J. Kossowsky et al. / Journal of A

hey have little or no direct control (e.g. when lying in a canopy),amely the CO2-inhalation. The presence of the mother and specialrecautions we took to accommodate participating children to the

aboratory (see Wilhelm et al., 2005) may also have played a role.Despite only a weak panicogenic paradigm, children with SAD

xhibited subtle voluntary breathing regulation deficits. Theseespiratory perturbations were not associated with anxiety orncreased symptom awareness as in the case of adult panic patients.ur data is in line with an explanation of PD that suggests that

espiratory hypersensitivity in panic may be secondary to cognitivebnormalities (Clark, 1993). Rather than differing in any specificiological sensitivity, patients with PD may differ from healthy indi-iduals in a cognitive set that interprets many bodily sensations asotentially harmful. According to this model, respiratory dysregu-

ation may be primarily a consequence of dysregulated cognitivend emotional processes, rather than generative of them. The lowevel of hyperventilation symptoms and panic cognitions reportedn our SAD children indicates that VH either does not elicit strongnough physical changes or that these changes are not yet asso-iated with panic cognitions. The absence of these factors may inurn explain the absence of anxiety.

Contrary to our hypotheses, we found no evidence of slowecovery after VH or any respiratory or autonomic abnormalitiesn any of the groups during recovery. Small group sizes do notccount for this finding because the study was sufficiently poweredith a relatively large sample size for pathophysiological research

nd the corresponding effect sizes are negligible (d’s < .1). Neitherre the measures and analysis techniques unreliable because weound significant reactivity to VH and also differences between theroups during the preceding separation challenge (Brand, Wilhelm,ossowsky, Holsboer-Trachsler, & Schneider, 2011; Kossowskyt al., 2012) and previous research has also emphasized the reliabil-ty of the autonomic and respiratory measurement methodologyBlechert, Lajtman, Michael, Margraf, & Wilhelm, 2006). This nullnding during recovery from VH has also been found in one recentdult study on PD (Wollburg, Meuret, Conrad, Roth, & Kim, 2008),here the suggestion was offered that response to the VH challengeepends on the subtype of PD into which the individual belongs, forxample, a respiratory or non-respiratory subtype. Further, recenttudies have emphasized the role of PD in parents (Roberson-Nayt al., 2010) and of genetic determinants as the major underlyingause of the continuity of childhood SAD into adult PD and thessociation of both disorders with heightened sensitivity to CO2Battaglia et al., 2009). Due to our small sample size of parents withD, we were not able to take these variables into account.

A distinct limitation in the current study was the small but sig-ificant age difference between the groups, which we addressed bydding age as a covariate where possible. We performed the non-arametric analyses twice, once with the whole sample and onceith an age stratified sample. Only after determining that age hado significant effect, did we consider the results from the wholeample to be valid.

Our results indicate that hyperreactivity to and slow recoveryrom lowered pCO2 levels do not play a central role in children withAD during VH. However, they appear to display difficulty in volun-ary breathing regulation. This difficulty in childhood could developnto a more pathological feature later on in life and contribute to theossible development of PD and might explain some of the associ-tions between the two disorders reported in the literature. Whileuture studies are warranted to confirm our results on voluntaryreathing regulation, a worthwhile point of research would be tourther investigate why children with SAD exhibit anxious respon-

ing to CO2 inhalation but not to VH, with particular emphasis onhe mediating role of psychological factors such as controllability,redictability, and awareness and appraisal of bodily symptoms,

ust to name a few that are likely to play some role. At current, both

Disorders 27 (2013) 627– 634 633

the psychopathological and treatment models of SAD have yet toincorporate the role and findings of psychophysiology. Given thefact that we found a response concordance between physiologicaland subjective measures using a different paradigm (Kossowskyet al., 2012), further studies are needed to clarify the discordancefound in the current study and its implications for psychopatho-logical and treatment models of SAD. Answering these questionsshould help us better understand the pathophysiology of SAD andits association to PD and thus lay the basis for successful prevention.

Acknowledgments

This study was supported by grant project (PP001-68701;105314-116517) awarded to Silvia Schneider by the Swiss NationalScience Foundation. We thank Dr. Peter Peyk for his researchassistance. The authors report no biomedical financial interests orpotential conflicts of interest.

References

Abelson, J. L., Weg, J. G., Nesse, R. M., & Curtis, G. C. (2001). Persistent respiratoryirregularity in patients with panic disorder. Biological Psychiatry, 49(7), 588–595.

Allen, J., Lavallee, K., Herren, C., Ruhe, K., & Schneider, S. (2010). DSM-IV criteriafor childhood separation anxiety disorder: informant, age, and sex differences.Journal of Anxiety Disorders, 24(8), 946–952.

Bass, C., & Gardner, W. N. (1985). Respiratory and psychiatric abnormalitiesin chronic symptomatic hyperventilation. British Medical Journal, 290(6479),1387–1390.

Battaglia, M., Bertella, S., Politi, E., Bernardeschi, L., Perna, G., Gabriele, A., et al. (1995).Age at onset of panic disorder: influence of familial liability to the disease andof childhood separation anxiety disorder. American Journal of Psychiatry, 152(9),1362–1364.

Battaglia, M., Pesenti-Gritti, P., Medland, S. E., Ogliari, A., Tambs, K., & Spatola, C. A.(2009). A genetically informed study of the association between childhood sep-aration anxiety, sensitivity to CO(2), panic disorder, and the effect of childhoodparental loss. Archives of General Psychiatry, 66(1), 64–71.

Beauchaine, T. (2001). Vagal tone, development, and Gray’s motivational theory:toward an integrated model of autonomic nervous system functioning in psy-chopathology. Development and Psychopathology, 13, 183–214.

Berntson, G. G., Cacioppo, J. T., Quigley, K. S., & Fabro, V. T. (1994). Autonomic spaceand psychophysiological response. Psychophysiology, 31(1), 44–61.

Biederman, J., Faraone, S. V., Hirshfeld-Becker, D. R., Friedman, D., Robin, J. A., &Rosenbaum, J. F. (2001). Patterns of psychopathology and dysfunction in high-risk children of parents with panic disorder and major depression. AmericanJournal of Psychiatry, 158(1), 49–57.

Blechert, J., Lajtman, M., Michael, T., Margraf, J., & Wilhelm, F. H. (2006). Identify-ing anxiety states using broad sampling and advanced processing of peripheralphysiological information. Biomedical Sciences Instrumentation, 42, 136–141.

Blechert, J., Wilhelm, F. H., Meuret, A. E., Wilhelm, E. M., & Roth, W. T. (2010). Respira-tory, autonomic, and experiential responses to repeated inhalations of 20% CO2

enriched air in panic disorder, social phobia, and healthy controls. BiologicalPsychology, 84(1), 104–111.

Boehnke, K., Silbereisen, R. K., Reynolds, C. R., & Richmond, B. O. (1986). What I thinkand feel – German experience with the revised form of the Children’s ManifestAnxiety Scale. Personality and Individual Differences, 7(4), 553–560.

Brand, S., Wilhelm, F. H., Kossowsky, J., Holsboer-Trachsler, E., & Schneider, S. (2011).Children suffering from separation anxiety disorder (SAD) show increased HPAaxis activity compared to healthy controls. Journal of Psychiatric Research, 45(4),452–459.

Carr, R. E., Lehrer, P. M., Hochron, S. M., & Jackson, A. (1996). Effect of psychologi-cal stress on airway impedance in individuals with asthma and panic disorder.Journal of Abnormal Psychology, 105(1), 137–141.

Clark, D. M. (1993). Cognitive mediation of panic attacks induced by biological chal-lenge tests. Advances in Behaviour Research and Therapy, 15(1), 75–84.

Craske, M. G., Rauch, S. L., Ursano, R., Prenoveau, J., Pine, D. S., & Zinbarg, R. E. (2009).What is an anxiety disorder? Depression and Anxiety, 26(12), 1066–1085.

de Morree, H. M., & Marcora, S. M. (2010). The face of effort: frowning muscle activityreflects effort during a physical task. Biological Psychology, 85(3), 377–382.

Ehlers, A., & Margraf, J. (1989). The psychophysiological model of panic attacks. In: P.M. G. Emmelkamp, W. T. A. M. Everaerd, F. Kraaimaat, & M. J. M. Van Son (Eds.),Fresh perspectives on anxiety disorders. Amsterdam: Swets & Zeitlinger.

Fowles, D. C., Christie, M. J., Edelberg, R., Grings, W. W., Lykken, D. T., & Venables, P.H. (1981). Committee report. Publication recommendations for electrodermalmeasurements. Psychophysiology, 18(3), 232–239.

Gorman, J. M., Browne, S. T., Papp, L. A., Martinez, J., Welkowitz, L., Coplan, J. D., et al.

(1997). Effect of antipanic treatment on response to carbon dioxide. BiologicalPsychiatry, 42(11), 982–991.

Gorman, J. M., Fyer, M. R., Goetz, R., Askanazi, J., Liebowitz, M. R., Fyer, A. J., et al.(1988). Ventilatory physiology of patients with panic disorder. Archives of Gen-eral Psychiatry, 45(1), 31–39.

6 nxiety

H

I

K

K

K

K

K

L

L

L

M

M

M

M

M

M

M

M

N

O

P

P

P

P

P

P

logical reactions to two levels of voluntary hyperventilation in panic disorder.

34 J. Kossowsky et al. / Journal of A

ornsveld, H., Garssen, B., & van Spiegel, P. (1995). Voluntary hyperventilation: theinfluence of duration and depth on the development of symptoms. BiologicalPsychology, 40(3), 299–312.

n-Albon, T., Meyer, A. H., & Schneider, S. (2013). Separation Anxiety AvoidanceInventory-Child and Parent Version: Psychometric Properties and Clinical Utilityin a Clinical and School Sample. Child Psychiatry & Human Development, 1–9.

essler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Walters, E. E.(2005). Lifetime prevalence and age-of-onset distributions of DSM-IV disordersin the national comorbidity survey replication. Archives of General Psychiatry, 62,593–603.

hoo, M. C. K. (1999). Periodic breathing and central apnea. In: M. D. Altose, &Y. Kawakami (Eds.), Control of Breathing in Health and Disease (pp. 203–250).New York: Marcel Dekker.

lein, D. F. (1993). False suffocation alarms, spontaneous panics, and relatedconditions: an integrative hypothesis. Archives of General Psychiatry, 50(4),306–317.

ossowsky, J., Pfaltz, M. C., Schneider, S., Taeymans, J., Locher, C., & Gaab, J. (2013).The separation anxiety hypothesis of panic disorder revisited. American Journalof Psychiatry, http://dx.doi.org/10.1176/appi.ajp.2012.12070893

ossowsky, J., Wilhelm, F. H., Roth, W. T., & Schneider, S. (2012). Separation anxi-ety disorder in children: disorder-specific responses to experimental separationfrom the mother. Journal of Child Psychology and Psychiatry, 53(2,), 178–187.

a Greca, A. M., & Stone, W. L. (1993). Social anxiety scale for children-revised:factor structure and concurrent validity. Journal of Clinical Child and AdolescentPsychology, 22(1), 17–27.

ast, C. G., Perrin, S., Hersen, M., & Kazdin, A. E. (1996). A prospective study of child-hood anxiety disorders. Journal of the American Academy of Child and AdolescentPsychiatry, 35(11), 1502–1510.

ipsitz, J. D., Martin, L. Y., Mannuzza, S., Chapman, T. F., Liebowitz, M. R., Klein, D.F., et al. (1994). Childhood separation anxiety disorder in patients with adultanxiety disorders. American Journal of Psychiatry, 151(6), 927–929.

addock, R. J., & Carter, C. S. (1991). Hyperventilation-induced panic attacks in panicdisorder with agoraphobia. Biological Psychiatry, 29(9), 843–854.

argraf, J. (1993). Hyperventilation and panic disorder: a psychophysiological con-nection. Advances in Behaviour Research and Therapy, 15(1), 49–74.

artin, C., Cabrol, S., Bouvard, M. P., Lepine, J. P., & Mouren-Simeoni, M. C. (1999).Anxiety and depressive disorders in fathers and mothers of anxious school-refusing children. Journal of the American Academy of Child and AdolescentPsychiatry, 38(7), 916–922.

artinez, J. M., Papp, L. A., Coplan, J. D., Anderson, D. E., Mueller, C. M., Klein, D.F., et al. (1996). Ambulatory monitoring of respiration in anxiety. Anxiety, 2(6),296–302.

elfsen, S., & Florin, I. (1997). Ein Fragebogen zur Erfassung sozialer Angst beiKindern (SASC-R-D). Kindheit und Entwicklung, 6, 224–229.

euret, A. E., Ritz, T., Wilhelm, F. H., & Roth, W. T. (2005). Voluntary hyperventi-lation in the treatment of panic disorder – functions of hyperventilation, theirimplications for breathing training, and recommendations for standardization.Clinical Psychology Review, 25(3), 285–306.

euret, A. E., Rosenfield, D., Wilhelm, F. H., Zhou, E., Conrad, A., Ritz, T., et al. (2011).Do unexpected panic attacks occur spontaneously? Biological Psychiatry, 70(10),985–991.

euret, A. E., Wilhelm, F. H., Ritz, T., & Roth, W. T. (2008). Feedback of end-tidalpCO2 as a therapeutic approach for panic disorder. Journal of Psychiatric Research,42(7), 560–568.

euschwander, M., In-Albon, T., Adornetto, C., Roth, B., & Schneider, S. (2013).Interrater-Reliabilität des Diagnostischen Interviews bei psychischen Störun-gen im Kindes- und Jugendalter. Zeitschrift fur Kinder- und Jugendpsychiatrie undPsychotherapie (in press).

akes, D. F. (1996). Clinical practitioner’s pocket guide to respiratory care (4th ed.).Old Town, MN: Health Educator Publications Inc.

app, L. A., Klein, D. F., & Gorman, J. M. (1993). Carbon dioxide hypersensitivity,hyperventilation, and panic disorder. American Journal of Psychiatry, 150(8),1149–1157.

app, L. A., Martinez, J. M., Klein, D. F., Coplan, J. D., Norman, R. G., Cole, R., et al. (1997).Respiratory psychophysiology of panic disorder: three respiratory challenges in98 subjects. American Journal of Psychiatry, 154(11), 1557–1565.

app, L. A., Welkowitz, L. A., Martinez, J. M., Klein, D. F., Browne, S., & Gorman, J.M. (1995). Instructional set does not alter outcome of respiratory challenges inpanic disorder. Biological Psychiatry, 38(12), 826–830.

ine, D. S., Coplan, J. D., Papp, L. A., Klein, R. G., Martinez, J. M., Kovalenko, P., et al.(1998). Ventilatory physiology of children and adolescents with anxiety disor-ders. Archives of General Psychiatry, 55(2), 123–129.

ine, D. S., Klein, R. G., Coplan, J. D., Papp, L. A., Hoven, C. W., Martinez, J., et al. (2000).

Differential carbon dioxide sensitivity in childhood anxiety disorders and nonillcomparison group. Archives of General Psychiatry, 57(10), 960–967.

ine, D. S., Wasserman, G. A., Miller, L., Coplan, J. D., Bagiella, E., Kovelenku, P., et al.(1998). Heart period variability and psychopathology in urban boys at risk fordelinquency. Psychophysiology, 35(5), 521–529.

Disorders 27 (2013) 627– 634

Rapee, R. M., Brown, T. A., Antony, M. M., & Barlow, D. H. (1992). Response tohyperventilation and inhalation of 5.5% carbon dioxide-enriched air acrossthe DSM-III-R anxiety disorders. Journal of Abnormal Psychology, 101(3),538–552.

Reynolds, C. R., & Richmond, B. O. (1978). What I think and feel: a revised mea-sure of children’s manifest anxiety. Journal of Abnormal Child Psychology, 6(2),271–280.

Roberson-Nay, R., Klein, D. F., Klein, R. G., Mannuzza, S., Moulton, J. L.,Guardino, M., 3rd, et al. (2010). Carbon dioxide hypersensitivity in separation-anxious offspring of parents with panic disorder. Biological Psychiatry, 67(12),1171–1177.

Roth, W. T. (2005). Physiological markers for anxiety: panic disorder and phobias.International Journal of Psychophysiology, 58(2–3), 190–198.

Roth, W. T., Wilhelm, F. H., & Pettit, D. (2005). Are current theories of panic disorderfalsifiable? Psychological Bulletin, 131, 171–192.

Schneider, S., Unnewehr, S., & Margraf, J. (2009). Diagnostisches interview psychischerStörungen im Kindes- und Jugendalter (Kinder-DIPS). Heidelberg: Springer.

Shear, K., Jin, R., Ruscio, A. M., Walters, E. E., & Kessler, R. C. (2006). Prevalence andcorrelates of estimated DSM-IV child and adult separation anxiety disorder in theNational Comorbidity Survey Replication. American Journal of Psychiatry, 163(6),1074–1083.

Sherwood, A., Allen, M. T., Fahrenberg, J., Kelsey, R. M., Lovallo, W. R., & van Doornen,L. J. (1990). Methodological guidelines for impedance cardiography. Psychophys-iology, 27, 1–23.

Stein, M. B., Millar, T. W., Larsen, D. K., & Kryger, M. H. (1995). Irregular breath-ing during sleep in patients with panic disorder. American Journal of Psychiatry,152(8), 1168–1173.

Stemmler, G. (1989). The autonomic differentiation of emotions revisited: conver-gent and discriminant validation. Psychophysiology, 26(6), 617–632.

Sullivan, G. M., Kent, J. M., Kleber, M., Martinez, J. M., Yeragani, V. K., & Gorman, J.M. (2004). Effects of hyperventilation on heart rate and QT variability in panicdisorder pre- and post-treatment. Psychiatry Research, 125(1), 29–39.

Tassinary, L. G., Cacioppo, J. T., & Vanman, E. J. (2007). The skeletomotor system: sur-face electromyography. In: J. T. Cacioppo, L. G. Tassinary, & G. G. Berntson (Eds.),Handbook of psychophysiology (3rd ed., pp. 267–302). New York: CambridgeUniversity Press.

Unnewehr, S., Schneider, S., Florin, I., & Margraf, J. (1998). Psychopathology in chil-dren of patients with panic disorder or animal phobia. Psychopathology, 31,69–84.

Unnewehr, S., Schneider, S., Margraf, J., Jenkins, M., & Florin, I. (1998). Expo-sure to internal and external stimuli: reactions in children of patientswith panic disorder or animal phobia. Journal of Anxiety Disorders, 10(6),489–508.

van Boxtel, A., & Jessurun, M. (1993). Amplitude and bilateral coherency of facialand jaw-elevator EMG activity as an index of effort during a two-choice serialreaction task. Psychophysiology, 30, 589–604.

Warner, V., Mufson, L., & Weissman, M. M. (1995). Offspring at high and low riskfor depression and anxiety: mechanisms of psychiatric disorder. Journal of theAmerican Academy of Child and Adolescent Psychiatry, 34, 786–797.

Wilhelm, F. H., Gerlach, A. L., & Roth, W. T. (2001). Slow recovery from vol-untary hyperventilation in panic disorder. Psychosomatic Medicine, 63(4),638–649.

Wilhelm, F. H., Gevirtz, R., & Roth, W. T. (2001). Respiratory dysregulation in anx-iety, functional cardiac, and pain disorders: assessment, phenomenology, andtreatment. Behavior Modification, 25(4), 513–545.

Wilhelm, F. H., & Peyk, P. (2005). ANSLAB 4.0: autonomic nervous system labora-tory (Version 4.0). Basel., available from: http://sprweb.org/repository/index.cfm[last accessed 1 February 2012]

Wilhelm, F. H., & Roth, W. T. (1998). Taking the laboratory to the skies: ambulatoryassessment of self-report, autonomic, and respiratory responses in flying phobia.Psychophysiology, 35(5), 596–606.

Wilhelm, F. H., & Roth, W. T. (2001). The somatic symptom paradox in DSM-IV anx-iety disorders: suggestions for a clinical focus in psychophysiology. BiologicalPsychology, 57, 105–140.

Wilhelm, F. H., Schneider, S., & Friedman, B. (2005). Psychophysiological assessment.In: M. Hersen (Ed.), Comprehensive handbook of behavioral assessment. Volume II:child assessment (pp. 200–231). New York: Wiley.

Wilhelm, F. H., Trabert, W., & Roth, W. T. (2001a). Physiologic instability in panicdisorder and generalized anxiety disorder. Biological Psychiatry, 49(7), 596–605.

Wilhelm, F. H., Trabert, W., & Roth, W. T. (2001b). Characteristics of sighing in panicdisorder. Biological Psychiatry, 49(7), 606–614.

Wollburg, E., Meuret, A. E., Conrad, A., Roth, W. T., & Kim, S. (2008). Psychophysio-

Journal of Anxiety Disorders, 22(5), 886–898.Zvolensky, M. J., & Eifert, G. H. (2001). A review of psychological factors/processes

affecting anxious responding during voluntary hyperventilation and inhalationsof carbon dioxide-enriched air. Clinical Psychology Review, 21(3), 375–400.