Child attention to ain and ain tolerance are deendent uon anxiety and attention control: An eye-tracking study Running Head: Child anxiety, attention control, and ain Heathcote, L.C. 1, MSc, Lau, J.Y.F., PhD, Mueller, S.C. 4, PhD, Eccleston, C. 3, PhD, Fox, E. 1, PhD, Bosmans, M. 4, MSc & Vervoort, T. 4, PhD * 1 Deartment of Exerimental Psychology, Tinbergen Building, 9 South Parks Road, Oxford, OX1 3UD, United Kingdom. lauren.heathcote@sy.ox.ac.uk; elaine.fox@sy.ox.ac.uk King's College London, Institute of Psychiatry Psychology & Neuroscience, Deartment of Psychology, Box PO77, Henry Wellcome Building, De Cresigny Park, Denmark Hill, London, SE5 8AF, United Kingdom. jennifer.lau@kcl.ac.uk 3 Centre for Pain Research, 1 West Level 3, University of Bath, Bath, BA 7AY, United Kingdom. c.eccleston@bath.ac.uk 4 Deartment of Exerimental-Clinical and Health Psychology, Ghent University, Henri Dunantlaan, B-9000 Gent Belgium. Sven.Mueller@ugent.be; martijnbosmans@gmail.com ; tine.vervoort@ugent.be * Corresonding author: Tine Vervoort, Deartment of Exerimental-Clinical and Health Psychology, Ghent University, Henri Dunantlaan, B- 9000 Ghent, Belgium. Tel: +3 (0)9 64 91 08 Fax: +3 (0)9 64 64 71. Electronic mail may be sent to Tine.Vervoort@Ugent.be Original article This work was carried out at the Deartment of Exerimental Clinical and Health Psychology at the University of Ghent. Lauren Heathcote is a DPhil student at the University of Oxford, a Research Training Fellow for UK charity Action Medical Research for Children (Grant Reference: GN1), and a trainee member of Pain in Child Health, a research training initiative of the Canadian Institutes of Health Research. Elaine Fox is suorted by an ERC Advanced Investigator Award (Ref: 34176). The authors declare that they have no cometing financial interests. The funders had no role in this study. 1
What s already known about this toic? Attention biases towards ain are associated with oorer ain outcomes in adults. In adolescents, attention control seems to moderate the association between emotional resonse to ain and attention bias to ain. What does this study add? This is the first study to use eye tracking to study attention to ain cues in children. Attention control is an imortant moderator of anxiety on attention bias to ain and tolerance of cold ressor ain in youth.
ABSTRACT Background: Pain is common and can be debilitating in childhood. Theoretical models roose that attention to ain lays a key role in ain outcomes, however very little research has investigated this in youth. This study examined how anxiety-related variables and attention control interacted to redict children s attention to ain cues using eye-tracking methodology, and their ain tolerance on the cold ressor test (CPT). Methods: Children aged 8-17 years had their eye-gaze tracked whilst they viewed hotograhs of other children dislaying ainful facial exressions during the CPT, before comleting the CPT themselves. Children also comleted self-reort measures of anxiety and attention control. Results: Findings indicated that anxiety and attention control did not imact children s initial fixations on ain or neutral faces, but did imact how long they dwelled on ain versus neutral faces. For children reorting low levels of attention control, higher anxiety was associated with less dwell time on ain faces as oosed to neutral faces, and the oosite attern was observed for children with high attention control. Anxiety and attention control also interacted to redict ain outcomes. For children with low attention control, increasing anxiety was associated with anticiating more ain and tolerating ain for less time. Conclusions: This is the first study to examine children s attention to ain cues using eyetracking technology in the context of a salient ainful exerience. Data suggest attention control is an imortant moderator of anxiety on multile outcomes relevant to young eole s ain exeriences. 3
INTRODUCTION Affective-motivational models roose that ain imoses an attentional riority and motivates avoidance behaviour, articularly when one is anxious about ain (Eccleston and Crombez, 1999; Vlaeyen and Linton, 000). Relatedly, how one attends to ain, and subsequently engages in behaviour to avoid ain, is thought critical in the transition from acute to chronic ain. As ain often first becomes chronic in youth (Perquin et al., 000), understanding the relationshi between anxiety and ain attending, as well as avoidance behaviour, is articularly imortant in aediatric oulations (King et al., 011). Paediatric research in this area is only beginning, yet both aediatric and adult literatures show conflicting findings, with some studies showing that increased anxiety contributes to more ain attending and behavioural avoidance (Crombez et al., 1998; Heathcote et al., 015; Van Damme et al., 004; Verhoeven et al., 010), whilst others rovide null findings (Roelofs et al., 003; Van Ryckeghem et al., 01), or evidence counter to exectations (Boston and Share, 005; Vervoort et al., 01, 011). Understanding when anxiety contributes to these outcomes is imortant for identifying those at risk for adverse ain exeriences. Attention control, the ability to focus effortfully in the face of distraction and flexibly shift attention, is likely imortant in understanding the variable relationshi between ainrelated anxiety and ain attending. Indeed, we recently found that adolescents who catastrohize about ain, and have low attention control, showed high vigilance to ain facial exressions (Heathcote et al., 015). However, whilst attention control rovides one exlanation for mixed findings, our results are reliminary, and outstanding questions remain. Most studies have relied on aradigms that examine static measures of attention, indirectly via resonse times (e.g., dot-robe (Beck et al., 011; Boyer et al., 006; Vervoort et al., 013b)). Eye-tracking methods alternatively rovide continuous measures of attention over time. This may be articularly relevant in the context of effortful attention control, which 4
is likely to have more ronounced effects on attention maintenance as oosed to initial selection. Further, most studies have relied on linguistic stimuli (ain words), which have low ecological validity and are indirectly related to ain. A significant advance would be to use ictorial stimuli that relate to an imending ain exerience (e.g., see (Vervoort et al., 013b). There are also concetual considerations. Secifically, catastrohizing has been studied as the rimary index of ain-related anxiety. However, affective-motivational models (Asmundson et al., 01; Eccleston and Crombez, 1999; Van Damme et al., 010a; Vlaeyen and Linton, 000) osit that general anxiety symtoms when facing ain also influence attention. Moreover, it has recently been argued that catastrohizing in childhood may be best understood as anxiety about ain rather than catastrohic beliefs (Eccleston et al., 01). Therefore, examining broader anxiety symtoms, as well as catastrohizing, in children s attending to ain is imortant. Further, while reliminary findings suggest attention control is imortant in understanding variable relationshis between anxiety and attention, it remains unknown whether it has a similar moderating role on behavioural avoidance, which is considered key in exlaining ain outcomes. This study examines how anxiety-related variables and attention control interact to redict children s attention to ain cues, and their tolerance and exerience of ain on the cold ressor test (CPT). Children s eye movements were tracked whilst viewing hotograhs of other children dislaying ainful facial exressions during the CPT, before comleting the CPT themselves. We hyothesized that, articularly when attention control is low, high anxiety and ain catastrohizing would contribute to (1) increased attention towards ain cues, and () decreased tolerance of cold ressor ain, which can be thought of as an avoidance of further ain (Boston and Share, 005; Vervoort et al., 013b). 5
. METHODS.1 Particiants The majority of the articiants were recruited from a samle of schoolchildren and adolescents (hereafter referred to as children ) (N= 46) who articiated in a questionnaire study about ain exeriences that took lace aroximately 5 months reviously (unublished data). Only children who had rovided informed consent to be re-contacted were aroached (N = 136). Exclusion criteria for the current study were as follows: (1) recurrent or chronic ain, () develomental delay, (3) having insufficient knowledge of the Dutch language, and (4) not being between the ages of 8-17 years. From the total of 136 children, 11 were contacted via their arents (the remaining did not rovide contact details (N=10), or were younger or older than 8-17 years (N = 14)). Of those contacted, 96.4% (N =108) met the inclusion criteria and 41.07% (N=46) agreed to articiate. Thirty-two ercent (N=36) refused to articiate. The main reason for refusal to articiate was lack of time or other family resonsibilities. The remainder of the children (N=1) showed initial interest in the study but were not contactable beyond the first time of contact (i.e., did not resond to a subsequent email or voicemail). To account for otential data loss during eye-tracking and to increase samle size, an additional 9 children were recruited via social media advertisement (advertisement was aimed at arents). There were no differences in anxiety, ain catastrohizing, and attention control between articiants recruited via schools and those recruited via social media (all t (5) 1.63, ns). As one of the articiants recruited from the revious questionnaire study contributed as a ilot subject, the final samle of the resent study consisted of 54 children (range 8-17 years, M=1.1 years, SD=.39, 19 boys). The arents of these children articiated in an indeendent study directly following the current study. Procedures relevant to the arental study occurred following the currently described methodology and are thus not exected to interfere with current results. Particiants were 6
comensated 30 for articiating in this study. The study was aroved by the Faculty of Psychology and Educational Sciences ethics committee, Ghent University, Belgium.. Aaratus A cold ressor aaratus was used as an exerimental ain task to induce ain in articiants and to measure ain tolerance. Particiants were instructed to lower their right hand into the cold water u to just above the wrist. Particiants ket their hand in the cold ressor box u to their ain tolerance level or an uninformed ceiling of 4 minutes. The temerature of the water was maintained at 10 C (+/-1 C) and was circulated continuously by a um (see also (Vervoort et al., 013b)). The cold ressor aaratus is well suited for use with children and the ain exerienced is considered to rovide an analogue for various naturally occurring acute ains (Birnie et al., 014; Von Baeyer et al., 005). A second tank was also used, for which the water was maintained at room temerature (1 C; +/-1 C). To standardize skin temerature rior to cold ressor immersion, all articiants first immersed their hand in the room temerature water tank for a total of minutes (see also (Birnie et al., 014; Vervoort et al., 013b). This rocedure will henceforth be referred to as the Cold Pressor Test (CPT)..3 Pain tolerance Pain tolerance was defined as the articiant s total time of immersion in the cold water in seconds, measured by a stowatch. Prior to the cold ressor test, articiants were rovided with the following instructions Hold your hand in the cold water as long as you can endure the ain. When you think/feel I cannot endure the ain any longer I cannot handle more, I want you to say STOP and take your arm out of the cold water box. The maximum duration of exosure to the cold ressor test was 4 minutes. However, the articiants were not informed of this ceiling so that there would be no risk that they mistakenly thought they were exected to leave their arm in the water for this length of time. 7
.4 Self-reort measures Attention control Attention control was assessed with the Dutch version of the Attention Control Scale (ACS) (Muris et al., 008). The ACS consists of 0 items, and yields a total score that can range from 0 to 80, with higher scores indicating good attention control. The ACS has two subscale scores for attention focusing (e.g., My concentration is good even if there is music in the room around me ) and attention shifting (e.g., It is easy for me to alternate between two different tasks ). The ACS has shown both good reliability and redictive utility, redicting resistance to interference in Stroo-like satial conflict tasks as well as attentional disengagement from threat stimuli among highly anxious eole (Derryberry and Reed, 00). Attention control has also been measured with good reliability and validity in children (Heathcote et al., 015; Meesters et al., 007). Cronbach alha in this study was.67 for the total score. State anxiety State anxiety was assessed by the State version of the Dutch State-Trait Anxiety Inventory for Children (Bakker et al., 1989; Sielberger et al., 1973). The STAIC-state is a 0-item questionnaire designed to measure current anxious state in children. Particiants are resented with various emotions (e.g., calm, nervous) and asked to read each item and to decide how they feel right now. Each item consists of a statement beginning I feel., and ending with three ossible emotional states (e.g., very calm, calm or not calm ). Total scores can range from 0 to 40. The decision to measure state rather than trait anxiety was based on revious findings indicating that state anxiety is usually highly correlated with trait anxiety and is likely to be a more roximal redictor of articiants situational resonses (see also (Vervoort et al., 014)). Furthermore, we were interested in state anxiety symtoms in the face of an ucoming ainful exerience (the cold ressor test), making the measure 8
relevant within the context of ain. The STAIC-state has been shown to be a reliable and valid instrument in revious research (see(bakker et al., 1989; Sielberger et al., 1973)). Cronbach alha in this study was.74. Catastrohic thinking about ain A situation-secific measure based uon the original Pain Catastrohizing Scale for Children (PCS-C; (Crombez et al., 003)) was administered to measure the articiants catastrohizing thoughts about the ucoming ain task. Again, we were interested in state catastrohizing in the face of an ucoming ainful exerience (the cold ressor test), making the measure relevant for this healthy samle, and more relevant for the outcome of ain tolerance. The original PCS-C consists of 13 items describing different thoughts and feelings that young eole may exerience when they are in ain, and yields three subscale scores for rumination, magnification, and hellessness. The PCS-C has shown to be a reliable and valid instrument in young eole from 9 to 15 years (Crombez et al., 003). The use of a situationsecific measure of catastrohizing, i.e., assessing the level of catastrohizing thoughts about an ucoming ain task, is consistent with revious studies (Vervoort et al., 009). The situation-secific measure of the PCS-C consists of adated items from each subscale (Rumination: At this moment, to what extent do you kee thinking about how much ain you will exerience during the test? ; Magnification: At this moment, to what extent do you think that, because of the ain, something serious might haen to you? ; Hellessness: At this moment, to what extent do you think you will not be able to endure the test because of the ain? ). The 6 items of the situation-secific PCS-C were rated on an 11-oint numerical rating scale (0 = not at all ; 10 = very much ), and were comleted before the attention task (i.e., eye-tracking task). The total score on catastrohizing could range from 0 to 60 and was used as an index of the articiant s catastrohizing thoughts about anticiated ain. Cronbach alha in this study was.81. 9
Anticiated and exerienced ain intensity Anticiated and exerienced ain intensity during the cold ressor test were assessed using 11-oint 1-item scales with the anchors no ain (0) and a lot of ain (10). Particiants were requested to rovide written ratings for these measures immediately before and after comleting the cold ressor test. For the current study, the articiant s self-reorted anticiated and exerienced ain intensity were considered secondary outcome measures (ain tolerance was the rimary outcome of interest)..5 Stimulus materials The stimulus set for the viewing task consisted of 40 grey-scaled ictures of 10 different children (5 boys and 5 girls; age range 9-16 years) dislaying neutral and ain facial exressions. These ictures were reviously used in a dot-robe study in children and an eyetracking study in arents examining child and arental attention to child ain, resectively (see (Vervoort et al., 014, 013b)). These ictures were reviously reliably coded for occurrence and intensity of facial ain dislay by means of the Child Facial Coding System (CFCS) (Chambers et al., 1996). For each of the 10 children in the stimulus set, 4 ictures were chosen to reflect each of 4 categories of facial ain exression: one icture of a articular child deicted the child s neutral face (NFE); one icture deicted the child with a low ain facial exression (LFE); one showed a moderate ain facial exression (MFE), and one deicted a high ain facial exression (HFE). Using the 40-icture set, three tyes of icture airings were generated, resulting in 30 slides. Secifically, each slide consisted of two ictures of the same child resenting a neutral exression combined with the same child s low ain, moderate ain, or high ain exression. This allowed examination of whether articiants selective attention to ain faces varied with varying levels of facial ain exressiveness. Pairs were comiled twice such that the neutral exression aeared equally often on the left and right side (i.e., 60 slides were shown: 0 neutral-low ain; 0 neutral- 10
moderate ain, and 0 neutral-high ain). The validity of the resent stimulus set is suorted by revious findings that differential facial ain exressiveness (i.e., NFE, MFE, LFE, HFE) reflects differences in observers ain intensity ratings such that increasing levels of facial ain exressiveness corresonded with observers increasing ain ratings (see (Vervoort et al., 013a, 013b))..6 Eye movement measurement Eye movements were tracked with a 300 Hz Tobii (TX300) table-mounted eye tracker (Tobii Technology AB, Danderyd, Sweden). This system consists of a 17-inch comuter screen with a camera and infrared LED otics embedded beneath it. The system records eye movements based on the corneal reflection caused by the infrared light source. Particiants were seated comfortably 60cm away from the centre of the screen. In order to calibrate the eye-tracker, articiants were asked to focus on nine sequentially aearing red dots resented in random lacement on the screen. The viewing task commenced after valid calibration..7. Viewing (eye-tracking) task Particiants were informed that they would comlete a viewing task, in which their eye movements would be tracked. Particiants were informed that during the task, they would view hotograhs of other children on the screen. To make the stimuli salient and related to a ersonal ainful exerience, articiants were told that these hotograhs reresented children undergoing the cold ressor task (i.e., the ainful task that that they would subsequently erform). Particiants were shown an overview of one trial (on aer) in order to ensure familiarity with the exerimental set u. Particiants were instructed to first focus on a centrally-resented white fixation cross when it aeared on screen and then to simly view the faces that would subsequently aear on the screen. Each trial within the viewing task began with a 500 ms resentation of the white fixation cross. Then, a slide with the air 11
of facial stimuli against a grey background was resented for 3500 ms and articiants were free to visually exlore the slide. Following an inter-trial interval of 00 ms (grey screen), the next trial again began with the resentation of the fixation cross. In total, the exeriment consisted of 60 trials: each of the 30 slides was resented twice, once with the ain face on the left and once on the right side of the screen. Pictures were 16 cm high and 10 cm wide. Pictures were searated by 4.4 cm from their central oints. Slides were resented to articiants in a fully randomized order..8 Picture ain intensity ratings After comleting the viewing task, articiants were asked to rate each resented ain face for ain intensity using a 0-10 NRS (i.e., How much ain has the child dislayed on the icture; anchors; no ain at all ; a lot of ain ). Pictures were resented on a comuter screen using Office PowerPoint in a randomized order that was fixed across articiants. Particiants were instructed to make written ratings of ain intensity for all ictures and were encouraged to roceed as fast as ossible. Picture ratings were averaged for each level of facial ain dislay (NFE, LFE, MFE, HFE) resulting in 4 mean ain intensity ratings ranging from 0-10. This allowed examining whether differences in facial exressiveness of the ictures matched differences in the articiant ain intensity ratings of the ictures..9 Procedure All articiants were invited by telehone and received standardized information about the study. When arents and their children rovided consent, they were invited to the laboratory at Ghent University where the study was conducted (In July 014 and August/Setember 014). Uon arrival at the laboratory, one of two exerimenters (both female) accomanied the arent and their child to the test-room. They were were told that the child would erform a comuterized eye-tracking task (i.e., the viewing task), and comlete a number of questionnaires. Furthermore, the cold ressor aaratus was shown to the arent 1
and their child, and they were told that the child could remove their arm from the water at any time, for any reason, and that the child could withdraw from articiation at any time. After obtaining written arental and child consent, exerimenter 1 stayed with the articiant in the test-room while exerimenter accomanied the arent to an adjacent room. Before erforming the viewing task, articiants were requested to fill out measures on attention control, state anxiety, and state ain catastrohizing. Particiants comleted these measures before the eye-tracking task as we were rimarily interested in their baseline emotional resonse when entering the lab and beginning the study, and because we aimed to examine the influence of these baseline measures on subsequent attention. Particiants then comleted the eye-tracking task. As noted above, to make the stimuli salient and related to a ersonal ainful exerience, articiants were told that these hotograhs reresented children undergoing the cold ressor task, the ainful task that they would subsequently erform. Following the viewing task, the articiant s ain tolerance was assessed during the cold ressor test. Particiants anticiated and exerienced ain for and during the cold ressor test was assessed before and after they comleted the task, resectively. Particiants and arents could not see or hear each other throughout the duration of the study rocedure. In addition, to revent any contact between the exerimenter and the articiant, the exerimenter was seated silently behind a screen during erformance of the cold ressor test. After comletion of the entire study rocedure, arent-child dyads were reunited in the testroom and were fully debriefed about the urose of the study..10 Eye movement arameters and data analysis Gaze behaviour was analyzed off-line with the Tobii software analysis ackage (Tobii Studio 3.) using the Identification Velocity-Threshold (I-VT) filter (classifier: 30 /s; Velocity calculator window length: 0 ms). The minimum fixation duration was set at 100 ms. The I-VT fixation classifier is based on the Velocity-Threshold identification fixation filter as 13
described by Salvucci and Goldberg (Salvucci et al., 000) and Komogortsev et al (Komogortsev et al., 010). Within the resent study, the two target ictures were defined as areas of interest (AOIs) within which eye movements would be monitored. Two eye movement arameters were calculated: (1) robability of first fixation, and () total fixation duration. Probability of first fixation refers to whether or not articiants first fixated their attention on the ain face or the neutral face. A first fixation bias refers to a higher robability of the articiant to first fixate on the ain face rather than the neutral face. Accordingly, first fixation bias was calculated by comuting the number of trials when the first fixation was directed on the ain face as a roortion of all trials in which the first fixation was made to either face (roortion scores >0.5 reflect an orienting bias toward ain faces; 0.5 = no bias; <0.5 reflect an orienting bias toward neutral faces). This was done searately for each facial ain exression category (LFE, MFE, HFE), resulting in 3 robability indices of first fixation bias. Total fixation duration for each facial exression was generated by averaging the total time articiants sent looking at faces for each facial exression category searately (NFE, LFE, MFE, HFE). A total fixation bias refers to when a articiant sends longer looking at ain faces rather than neutral faces. In the current study (as in revious studies), total fixation duration may also be referred to as dwell time. To examine the imact of articiant anxiety, ain catastrohizing, and attention control uon attention to ain, a series of reeated measures analyses of covariance (ANCOVA) were erformed for each of the eye movement indices (i.e., robability of first fixation and total fixation duration) with facial exressiveness as a within subject factor and attention control and anxiety or ain catastrohizing entered as covariates, resectively, in searate analyses. To examine the imact of anxiety, ain catastrohizing, and attention control uon ain tolerance, univariate analyses of covariance (ANCOVAs) were erformed with ain tolerance as the deendent variable and attention control and anxiety or ain 14
catastrohizing entered as covariates, resectively, in searate analyses. Additional analyses were also erformed with anticiated and exerienced ain intensity during CPT erformance as the deendent variable, as this was considered a secondary outcome within the resent study. Continuous redictor variables (anxiety/ain catastrohizing/attention control) were centered rior to entering the analyses. Centering reduces the multicollinearity between redictors and any interaction terms among them and facilitates ost-hoc robing of significant interaction effects (Holmbeck, 00). Further, in case the analyses with total fixation duration revealed significant interactions with facial exressiveness, bias indices ( total fixation bias ) were calculated to further aid interretation of direction of effects. Searate total fixation bias scores were calculated for each level of facial ain exressiveness (HFE, MFE, LFE). In these analyses, ositive values of total fixation bias (i.e., mean total dwell time on ain faces subtracted from mean total dwell time on neutral faces) indicated that dwell time on ain faces was higher than dwell time on neutral faces, whereas a negative score indicated the reverse: higher dwell time on neutral faces. ANCOVA was erformed on these bias indices. In case of significant interactions, additional moderation analyses were erformed to interret the interaction effects (i.e., whether the association between the redictor variable and the outcome was significant only for high levels of the moderator variable, low levels of the moderator variable, or both). All moderation analyses followed the rocedure outlined by Holmbeck et al. (Holmbeck, 00), i.e., two new conditional continuous moderator variables were comuted by (1) subtracting 1 SD from the centered moderator variable (high attention control ) and () adding 1 SD to the centered moderator variable (low attention control). Next, two additional ANCOVAs were erformed - incororating each of these new conditional continuous moderator variables - to test the significance for high (+1 SD above 15
the mean) and low (-1 SD below the mean) values of the conditional centered moderator variable. To fully understand observed interactions, additional moderation analyses were run to examine the imact of attention control uon attention to ain as a function of low versus high levels of anxiety and catastrohizing, resectively, using a similar moderation rocedure as outlined above (Holmbeck, 00). To exlore any significant interaction effects further, we also conducted additional analyses examining effects at high and low levels of anxiety, using the same rocedure as above to create the high and low attention control. Whenever the shericity assumtion was violated (Mauchly s test of shericity was <.05), Greenhouse- Geisser corrections (with adjusted degrees of freedom, or NDf) were erformed. Bonferroni correction for multile comarisons was not aroriate as the current study did not meet any of the conditions for alying this adjustment (i.e., (a) a universal null hyothesis is of interest, (b) the same test is reeated in many subsamles, (c) searching for significant associations without a riori hyotheses for multile tests) (Perneger, 1998). Effect sizes were reorted using the Partial Eta Squared index ( ). Following Cohen (Cohen, 1988) and Oleknik and Algina (Olejnik and Algina, 000); small effect size = 0.01; medium effect size = 0.06; large effect size = 0.14. 3. RESULTS 3.1 Particiant characteristics Data from 17 articiants were discarded from analyses due to sub-otimal overall gaze track status (i.e., eye movements tracked less than 70% of total task viewing time). This is in line with revious eye-tracking studies (Vervoort et al., 013a). The final samle entered in all of the analyses reorted below therefore consisted of 37 children (14 boys) aged 8-17 years (M=1.1 years, SD=.45). Those articiants whose data were discarded did not differ from the articiants retained in the analysis on any of the indeendent or deendent 16
measures (all t (5) 1.58, ns). Mean scores, standard deviations, and correlations between measures are reorted in Table 1. The mean level of state anxiety (M=1.; SD=5.48), ain catastrohizing (M=15.8; SD=10.0), and attention control (M=33.3; SD=6.06) in the resent samle are comarable with those obtained in revious samles of school youths (Bakker et al., 1989; Muris et al., 008; Vervoort et al., 009). Particiants reorted moderate levels of anticiated (M=4.05; SD=.1) and exerienced mean ain intensity (M=5.00; SD=.05) during CPT erformance. Particiants CPT immersion time ranged from 18 to 40 seconds with a mean of 14.4 s (SD=99.1). Particiants ain tolerance in the current study is therefore in line with revious studies (Vervoort et al., 013b). About two third of the articiants (59.5 %) withdrew their hand from the cold water before the four minute ceiling was reached. Pearson correlation analyses indicated that self-reort measures and ain tolerance did not correlate significantly with articiant age (all r.05, ns). There were no significant differences between girls and boys for any of the measures (all t(35) 1.00, ns), excet for anticiated and exerienced ain intensity which was higher amongst girls (both t(35) -., <.05). State anxiety was significantly correlated with state ain catastrohizing (r=.41, <.05). State anxiety and state ain catastrohizing were not significantly correlated with attention control (r=-.1 and -.0, resectively, ns), and none of these measures were correlated with ain tolerance (all r.7, ns). However, both the articiant s anticiated and exerienced ain intensity during CPT were significantly correlated with state ain catastrohizing (both r.38, <.05). Exerienced ain intensity was also significantly negatively correlated with ain tolerance (r= -.34, <.05). 3. Picture ratings - INSERT TABLE 1 ABOUT HERE - 17
In order to examine whether the different icture categories of facial ain exressiveness matched differences in the articiant s ratings of ain intensity (therefore to confirm stimulus categories), articiants ain intensity ratings of the ictures were examined using a reeated measures ANOVA. Results revealed significant differences between ratings of the four exression levels (F(3,34)=309.4, є=.66; NDf(1.98;.44), <.0001). Differences were in the exected direction. Secifically, contrasts revealed that high exressive ain faces (M=68.4 SD=15.6) were rated significantly higher than moderate exressive ain faces (M=55.4; SD=14.5; F(1,36)=160.66, <.0001) which, in turn, were rated significantly higher than low exressive ain faces (M=44.4; SD=14.9; F(1,36)=8.6, <.0001) which, in turn, were rated significantly higher than neutral faces (M=9.0; SD=15.9; F(1,36)=43.4, <.0001). Adding either anxiety or ain catastrohizing, resectively, as a covariate to the reeated measures ANOVA revealed no main effects or interactions effects (all F(1,35)..40, ns). 3.3 The imact of anxiety and ain catastrohizing uon attention to ain, and the moderating role of attention control 3.3.1 Probability of first fixation We first investigated whether or not articiants first fixated on ain faces significantly more than neutral faces (i.e., whether or not there was an overall first fixation bias). As a rimary reeated measures ANOVA revealed that there were no significant differences between first fixation bias towards low (M=.53), medium (M=.53), and high exressive ain faces (M=.5) (F(,35)=.07, ns, = 0.00), we calculated a mean first fixation bias score across all three ain exressions. A one-samled t-test revealed a mean first fixation bias (M=.53) significantly greater than chance (.50) indicating that articiants first fixated significantly more on ain faces than neutral faces (t(36)=3.30, <.005). 18
To investigate the imact of anxiety and the moderating role of attention control uon the articiant s first fixation bias, a reeated measures ANCOVA was erformed with anxiety and attention control entered as covariates. Analysis indicated no main or interaction effects for anxiety, attention control, and/or facial exressiveness (all F 1.06, ns, all < 0.03). The analysis examining the imact of ain catastrohizing and the moderating role of attention control revealed no main or interaction effects for ain catastrohizing, attention control, and/or facial exressiveness (all F 1.43, ns, all 3.3. Total fixation duration < 0.001). We first investigated whether or not articiants dwelled longer on ain faces comared to neutral faces (i.e., whether or not there was an overall total fixation bias to ain). A reeated measures ANOVA with facial exressiveness as a within-subject factor (i.e., dwell time on neutral, low ain, moderate ain, and high ain faces) revealed a significant effect (F(3,34)=17.17; є=.61; NDf(1.83;0.74), <.0001, = 0.33). Contrasts revealed that dwell time on ain faces significantly increased with each level of ain exressiveness (i.e., M=TFD_lowain=1.8s; M=TFD_modain=1.34s, M=TFD_highain=1.39s) and was significantly longer comared to dwell time on neutral faces (MTFD_neutral=1.08s; all F(1,36) 17.51, <.0001). Adding anxiety and attention control as covariates to the reeated measures analysis indicated a significant anxiety x attention control x facial exressiveness interaction (F(3,30)=3.65, є=.6; NDf(1.86;18.6), <.05, = 0.10). This indicates that the interaction between anxiety and attention control is different deending on the tye and/or level of facial exressiveness (ie., neutral exression, low ain exression, moderate ain exression, and high ain exression). To interret this significant interaction, we first calculated total fixation bias scores for each level of ain facial exression (i.e., the degree to which articiants 19
dwelled on low, moderate, and high ain exressions in comarison to neutral faces). Using these total fixation bias indices, we erformed a reeated measures ANCOVA with total fixation bias indices for LFE, MFE, and HFE as deendent variables and attention control and anxiety as covariates. Findings indicated a significant anxiety x attention control interaction (F(,33) =4.84, <.05, = 0.0). No significant anxiety x attention control x facial exressiveness interaction (F(,3) =1.75, ns, = 0.11) was observed, indicating that total fixation bias indices were not different for low, moderate, or high ain facial exressiveness. Accordingly, two univariate ANCOVAs were conducted with the mean total fixation bias (i.e., dwell time averaged across low, moderate, and high exressive ain faces) to further interret differences for anxiety between articiants who reorted high levels of attention control and those who reorted low levels of attention control. As shown in Figure 1.1, findings indicated a cross-over interaction signifying that anxiety differentially imacted the articiant s dwell time on ain faces deending uon the articiant s level of attention control. Secifically, for articiants reorting high levels of attention control, higher anxiety was associated with increasing dwell time on ain faces (Low anxiety: M = 0.01 seconds, High anxiety: M = 0.3 seconds). Conversely, for articiants reorting low levels of attention control, higher anxiety was associated with decreasing dwell time on ain faces (Low anxiety: M = 0.49 seconds, High anxiety: M = 0.9 seconds). While these atterns did not significantly differ from 0 for articiants reorting low attention control (F(1,33)=1.69, ns, = 0.05) and only aroached significance for articiants reorting high attention control (F(1,33)=3.96, =.06, = 0.11), the significant cross-over interaction indicates that these atterns are significantly different from each other. Furthermore, as shown in Figure 1.1 (dotted lines), findings suggest that attention control has the most imact at lower levels of anxiety. Indeed, additional analyses indicated that for articiants with low levels of anxiety, 0
increasing attention control is associated with decreasing dwell time on ain faces (F(1,33)=9.8, <.005, = 0.). Attention control is not associated with dwelling on ain faces at high levels of anxiety (F(1,33)=.00, ns, = 0.00). Similar analysis but with ain catastrohizing entered as a covariate instead of anxiety indicated that the exected three-way interaction (i.e., catastrohizing x attention control x facial exressiveness) failed to reach significance (F(3,31)=1.93, є=.6; NDf(1.86;19.), ns, = 0.06). This suggests that the interaction between attention control and ain catastrohizing did not imact dwell time differently for neutral faces or ain faces (at any level of ain exressiveness). Analyses only revealed a significant attention control x facial exressiveness interaction (F(3,31)=4.69, є=.6; NDf(1.86;19.), <.05, = 0.1) and an attention control x ain catastrohizing interaction (F(1,33)=4.87, <.05, = 0.13). The former two-way interaction (i.e., attention control x facial exressiveness) is not exlored further since revious analyses indicated that this two-way interaction is subsumed by a threeway interaction between anxiety, attention control, and facial exressiveness. To interret the significant ain catastrohizing x attention control interaction, searate ANCOVAs were erformed for articiants reorting low and high attention control, resectively, and mean dwell time (averaged across the four exressiveness levels) entered as the deendent variable. As shown in Figure 1., findings indicated that ain catastrohizing was ositively associated with overall dwell time (on all faces) at high levels of attention control (Low catastrohizing: M = 1.18 seconds, High catastrohizing: M = 1.3 seconds) but negatively associated with overall dwell time (on all faces) at low levels of attention control (Low catastrohizing: M = 1.37 seconds, High catastrohizing: M = 1.9 seconds). While neither of these atterns reached significance (both F(1,33).3 ns, both < 0.07), the significant cross-over interaction suggests associations are significantly different from each 1
other. As shown in Figure 1. (dotted lines), findings suggest that attention control has the most imact at lower levels of ain catastrohizing. Indeed, additional analyses indicated that attention control is significantly negatively associated with overall dwell time (i.e., to all face tyes) at low levels of ain catastrohizing (F(1,33)=5.59, <.05, = 0.15). Attention control is not associated with overall dwell time at high levels of ain catastrohizing (F(1,33)=.38, ns, = 0.01). In sum, eye movement analyses indicated that articiants overall first fixated on ain faces more than neutral faces, and dwelled longer on ain faces than neutral faces. Whilst there was no imact of sychological variables (i.e., anxiety, ain catastrohizing, or attention control) on first fixations, sychological variables did modulate dwell time on ain versus neutral faces, articularly within the context of anxiety and attention control. Secifically, anxiety differentially imacted the articiant s dwell time on ain faces deending uon the articiant s level of attention control. Whilst the association between ain catastrohizing and attention control was redictive of dwell time on all faces (i.e., looking for longer at all faces, regardless of face tye or level of exressiveness), ain catastrohizing was not associated with biased attentional rocessing of ain, even when attention control was taken into account. -INSERT FIGURE 1 ABOUT HERE- 3.4 The imact of anxiety and ain catastrohizing uon ain outcomes, and the moderating role of attention control 3.4.1. Pain tolerance The analysis investigating the imact of anxiety and the moderating role of attention control uon the articiant s ain tolerance (i.e., the articiant s total immersion time
during the cold ressor test) revealed a significant anxiety x attention control interaction (F(1,33)=6.49, <.05, = 0.16), indicating that the imact of anxiety uon tolerance is deendent uon the articiant s level of attention control. Searate ANCOVAs for articiants with high and low attention control indicated a cross-over interaction (see Figure 1.3). Similar to our findings on the imact of anxiety uon attentional maintenance to ain as a function of low vs. high attention control, findings indicated that the imact of anxiety uon ain tolerance shows oosite atterns deending uon low vs. high levels of attention control. Secifically, in line with exectations, findings indicated that higher levels of anxiety were associated with lower ain tolerance amongst articiants reorting low levels of attention control (Low anxiety: M = 189.91 seconds, High anxiety: M = 80.75 seconds) (F(1,33)=5.13, <.05, = 0.14). The oosite (but non-significant) attern was found for articiants reorting high levels of attention control (Low anxiety: M = 88.67 seconds, High anxiety: M = 160.01 seconds) (F(1,33)=.3, ns, = 0.07). Furthermore, additional analyses (see Figure 1.3; dotted lines) indicated that attention control exerts oosite effects deending on whether articiants are lowly or highly anxious. Secifically, for articiants with low levels of anxiety, increasing attention control is associated with decreasing ain tolerance (F(1,33)=4.0, =.05, = 0.11). The oosite (but non-significant) attern was found for articiants reorting high levels of anxiety (F(1,33)=3.16, ns, = 0.09). The analyses with ain catastrohizing and attention control revealed no significant main or interaction effects (all F.40, ns, all < 0.07). 3.4.. Anticiated and exerienced ain intensity The analyses with anxiety and attention control as indeendent variables and anticiated ain intensity as the deendent variable revealed a significant anxiety x attention 3
control interaction (F(1,33)=7.9, <.05, = 0.18). Searate ANCOVAs for articiants with high and low attention control indicated a crossover interaction (see Figure 1.4). Resembling findings on the imact of anxiety x attention control uon attentional maintenance to ain as well as ain tolerance, findings indicated that the imact of anxiety uon anticiated ain again shows oosite atterns deending uon low vs. high levels of attention control. Secifically, findings indicated that higher levels of anxiety were associated with higher anticiated ain amongst articiants reorting low levels of attention control (Low anxiety: M =.14, High anxiety: M = 5.3) (F(1,33)=9.63, <.005, = 0.3). Findings at high levels of attention control showed the oosite attern but failed to reach significance (Low anxiety: M = 4.57 seconds, High anxiety: M = 3.69 seconds) (F(1,33)=.38, ns, = 0.0). Furthermore, additional analyses (see Figure 1.4; dotted lines) indicated that attention control exerts oosite effects deending on whether articiants are lowly or highly anxious. Secifically, for articiants with low levels of anxiety, increasing attention control is associated with increasing anticiation of ain intensity (F(1,33)=5.11, <.05, = 0.13). Findings at high levels of anxiety showed the oosite attern but failed to reach significance (F(1,33)=3.00, ns, = 0.08). The analysis with exerienced ain intensity as deendent variable revealed no significant main or interaction effects (all F(1.33) <.59, ns, all < 0.0). The analyses with ain catastrohizing and attention control for both anticiated and exerienced ain intensity only revealed significant main effects for ain catastrohizing such that higher levels of ain catastrohizing were associated with higher anticiated and exerienced ain intensity (both F(1.33) 6.6, <.05, = 0.31 and 0.16 resectively). There were no other significant main or interaction effects (all F(1.33) 3.06, ns, all < 0.09). 4
4. DISCUSSION This study examined how anxiety-related variables and attention control interacted to redict children s attention to visual ain cues, and their tolerance of ain on the cold ressor test (CPT). Furthermore, we also exlored relationshis with anticiated and exerienced ain intensity. Childrens eye movements were tracked whilst viewing hotograhs of other children dislaying ainful facial exressions during the CPT, before comleting the CPT themselves. We hyothesized that, articularly when attention control is low, high anxiety and ain catastrohizing would contribute to (1) increased attention towards ain cues, and () decreased tolerance of cold ressor ain, which can be thought of as an avoidance of further ain (Boston and Share, 005; Vervoort et al., 013b). Results were in line with exectations regarding cold ressor ain outcomes, and artially in line with exectations regarding attention. Regarding attention; anxiety, ain catastrohizing, and attention control had no effect on where children first fixated their attention. That is, irresective of anxiety or attention control, children tended to first fixate on ain faces as oosed to neutral faces. Children also tended to dwell for longer on ain faces than neutral faces (and increasingly longer on highly exressive ain faces). However, dwell time on ain faces was deendent uon anxiety, moderated by attention control. Secifically, anxiety showed oosite relationshis with dwell time deending uon whether children reorted low or high levels of attention control, but with findings in the oosite direction than exected. Secifically, for children reorting low levels of attention control, higher anxiety was associated with decreasing dwell time on ain faces (although children still overall attended more towards ain than neutral faces, suggesting there was no avoidance of ain faces). Conversely, for children reorting high levels of attention control, higher anxiety was associated with increasing dwell time on ain 5
faces as oosed to neutral faces. While these atterns of anxiety and ain attending did not significantly differ from zero for articiants reorting low attention control, and only aroached significance for articiants reorting high attention control, the significant crossover interaction indicates that these atterns are significantly different from each other. Regarding cold ressor ain outcomes; attention control layed a similar moderating role. That is, at low levels of attention control, children with increasing anxiety anticiated more ain and were less tolerant for ain. The oosite attern (though not significant in itself) was observed for children with high levels of attention control. Thus, all three analyses with anxiety and ain attending, anticiation, and tolerance showed similar cross-over interactions, indicating a robust moderating role of attention control across self-reort and behavioural outcomes (i.e., eye movements and ain tolerance). Our findings therefore suggest that when a child is increasingly anxious in the face of ain, low attention control is imortant in understanding worse ain outcomes. Further, our findings relicate those within the anxiety literature, demonstrating that attention control moderates the association between anxiety and attending towards threatening stimuli (Bardeen and Orcutt, 011; Derryberry and Reed, 00; Helzer et al., 009; Schoorl et al., 014; Susa et al., 01). However, it is intriguing that our findings regarding ain attending were in the oosite direction than exected, with higher anxiety being associated with increasing dwell time on ain faces for children reorting high, not low attention control. One ossible but highly tentative exlanation for these findings is that increased attending to ain cues, facilitated by high attention control, can be conceived of as a strategy emloyed by increasingly anxious children that may have enabled them to seek further information about the ucoming otentially ainful exerience, thus allowing re-araisal. Accordingly, the observed attern may reflect effective use of emotion regulation, in which attending to information about ucoming ain allows the regulation of aversive emotions (such as anxiety) 6
by being able to focus on and re-araise the threatening nature of ain. Whilst this suggestion certainly requires indeendent exloration, there is indeed revious evidence that adults who are highly anxious benefit from attending to rather than avoiding information about ucoming ain, such that high anxious individuals show better emotion outcomes when effortfully attending to ain (e.g., (Vervoort et al., 014). If our findings are relicated in further studies, this may suggest that the function of attention control is different for children with differing levels of anxiety. Additional analyses further attest to this notion. Secifically, we found that whilst increasing attention control was associated with better ain-related outcomes in highly anxious children (i.e., anticiating less ain and tolerating ain for longer), it was associated with oorer outcomes in lowly anxious children (i.e., anticiating more ain and tolerating ain for less time). These findings may have imlications for interventions that attemt to change ain attending (e.g., Attention Bias Modification, see (Dehghani et al., 004)) or more general attention control caacities (e.g., Working Memory Training, see (Owens et al., 013)) for children with ersistent ain or ain-related anxiety. Interestingly, we found that catastrohizing about ain, unlike broader anxiety, did not interact with attention control to imact children s selective attending to ain. Catastrohizing did, however, interact with attention control to influence attending to all faces. Secifically, there was a cross-over interaction so that for children reorting high levels of attention control, higher catastrohizing was associated with increasing dwell time on all faces. Conversely, for children reorting low levels of attention control, higher catastrohizing was associated with decreasing dwell time on all faces. Nonetheless, catastrohizing did not imact selective attending to ain faces. This is somewhat in contrast with our revious findings (Heathcote et al., 015), wherein catastrohizing did interact with attention control to redict adolescents ain attending in a dot-robe task. This inconsistency may be due to the current study s methodological advances, for examle using eye-tracking and ersonally 7
relevant ain cues. Imortantly, children s attending to ain facial exressions that inform about an ucoming ainful exerience may well differ from their attending to ain faces which are not relevant to or useful for their own exerience of ain (Van Damme et al., 010b). Discreancies may also be due to differences in samle ages and therefore cognitive caacities, with the current study including younger articiants. Indeed, it has recently been questioned whether younger children exerience truly catastrohizing thoughts, and it has instead been roosed that more general worry and anxiety is more relevant for younger children s ain outcomes (Eccleston et al., 01). This study has limitations. First, due to technical limitations of the eye-tracker, we excluded 17 articiants, resulting in a small final samle size. The small final samle also recluded analyses of age effects, which may lay a role in child rocessing of ain. Second, eye-tracking does not measure covert attentional cature receding saccades. Indeed, initial attention cature by threatening stimuli may occur covertly, rior to initial fixation (Armstrong and Olatunji, 01). Third, we recognize that this is a single exerimental observation and relication is needed, in articular by other laboratories. Desite these limitations, this study adds to the limited number of studies currently available on child attention and ain in three ways. 1) This is the first study to use eyetracking within the context of child ain, thereby overcoming issues of using manual reaction times to infer attention, and allowing us to discriminate in real time between initial attention cature and dwelling on ain cues in youth. ) This is one of few studies to use ecologically valid and ersonally salient stimuli, wherein hotograhs of ain facial exressions rovided children with information about an ucoming ersonal ain exerience (see also (Vervoort et al., 013b)). 3) This is the first study to examine the role of effortful attention control in understanding both ain attending and theoretically and clinically meaningful outcomes. Based on our findings, we believe three areas would be fruitful for further research. First, 8
whilst the Attention Control Scale (ACS) assesses an individual s general caacity to focus in the face of distraction and to shift attention, these caacities may differ from the ability to control attention in ain-secific contexts. In articular, anxious individuals may exhibit more difficulties controlling attention in ain than non-ain contexts. The fact that we found a significant moderation effect of general attention control attests to the imortance of this general caacity in influencing ain-secific constructs, however, examining attention control within the context of ain may further our understanding of this effect. Second, whilst the ACS has shown good utility in redicting attentional erformance in exerimental tasks (Derryberry and Reed, 00), it will be imortant for future studies to also include other measures of attention control such as flanker tasks (that are reliant on behavioural resonses and reaction times, e.g., (Eriksen and Eriksen, 1974; Fan et al., 00)) as well as antisaccade tasks (that are reliant on eye-movement atterns, e.g., (Rommelse et al., 008)). These measures may be articularly imortant in child and adolescent samles, as brain regions engaged in goal-directed attention are still changing during this eriod (Cohen Kadosh et al., 014; Monk et al., 003; Nelson et al., 00; Sear, 000), and so attention control may be more challenging to self-reort. Third, the extent to which the reorted attern of attending is clinically relevant will rely on a demonstration of these effects in clinical samles, and a demonstration that they are malleable under instruction, ractice, or theray. The current findings extend our understanding of the role of child and adolescent anxiety in attending to ain cues, and anticiation and tolerance of ain. Our findings suggest that attention control is an imortant moderating factor for multile outcomes relevant to these young eole s ain exeriences. 9
5. ACKNOWLEDGEMENTS The authors would like to thank Nele Decoene, Jela Van Bladel, Esther Van de Velde, Tineke Van Hove, and Laura Wyers for their assistance in recruitment and data collection. 30
6. AUTHOR CONTRIBUTIONS Lauren Heathcote: Study design / data analyses / interretation of results / writing of the manuscrit. Jennifer Lau: Provided thoughtful suggestions regarding theoretical concetualization of the manuscrit and interretation of study findings / editing of the manuscrit. Sven Mueller: Study design / rogramming tasks / editing of the manuscrit. Christoer Eccleston: Provided thoughtful suggestions regarding interretation of study findings / editing of the manuscrit. Elaine Fox: Interretations of study findings / editing of the manuscrit. Martijn Bosmans: Inerretation of results / editing of the manuscrit Tine Vervoort: Study design / data collection / data inut / data analyses / interretation of results / writing of the manuscrit. 31
7. REFERENCES Armstrong, T., Olatunji, B.O., 01. Eye tracking of attention in the affective disorders: A meta-analytic review and synthesis. Clin. Psychol. Rev. doi:10.1016/j.cr.01.09.004 Asmundson, G.J.G., Noel, M., Petter, M., Parkerson, H.A., 01. Pediatric fear-avoidance model of chronic ain: Foundation, alication and future directions. Pain Res. Manag. Bakker, F.C., Van Wieringen, P.C.W., Van der Ploeg, H.M., Sielberger, C.D., 1989. Handleiding bij de Zelfbeoordelingsvragenlijst voor Kinderen (ZBVK). Swets & Zeitlinger, Lisse, The Netherlands. Bardeen, J.R., Orcutt, H.K., 011. Attentional control as a moderator of the relationshi between osttraumatic stress symtoms and attentional threat bias. J. Anxiety Disord. 5, 1008 1018. doi:10.1016/j.janxdis.011.06.009 Beck, J.E., Liani, T. a., Baber, K.F., Dufton, L., Garber, J., Smith, C. a., Walker, L.S., 011. Attentional bias to ain and social threat in ediatric atients with functional abdominal ain and ain-free youth before and after erformance evaluation. Pain 15, 1061 1067. doi:10.1016/j.ain.011.01.09 Birnie, K. a, Caes, L., Wilson, A.C., Williams, S.E., Chambers, C.T., 014. A ractical guide and ersectives on the use of exerimental ain modalities with children and adolescents. Pain Manag. 4, 97 111. doi:10.17/mt.13.7 Boston, A., Share, L., 005. The role of threat-exectancy in acute ain: Effects on attentional bias, coing strategy effectiveness and resonse to ain. Pain 119, 168 175. doi:10.1016/j.ain.005.09.03 Boyer, M.C., Comas, B.E., Stanger, C., Colletti, R.B., Konik, B.S., Morrow, S.B., Thomsen, A.H., 006. Attentional biases to ain and social threat in children with recurrent abdominal ain. J. Pediatr. Psychol. 31, 09 0. doi:10.1093/jesy/jsj015 Chambers, C.T., McGrath, P.J., Gilbert, C.A., Craig, K.D., 1996. Child facial coding system - revised manual. IWK-Grace Health Centre, Dalhousie University & University of British Columbia, Halifax, NS, Canada. Cohen, J., 1988. Statistical ower analysis for the behavioral sciences, Statistical Power Analysis for the Behavioral Sciences. doi:10.134/1345678 Cohen Kadosh, K., Heathcote, L.C., Lau, J.Y.F., 014. Age-related changes in attentional control across adolescence: how does this imact emotion regulation caacities? Front. Psychol. 5, 111. doi:10.3389/fsyg.014.00111 Crombez, G., Bijttebier, P., Eccleston, C., Mascagni, T., Mertens, G., Goubert, L., Verstraeten, K., 003. The child version of the ain catastrohizing scale (PCS-C): A reliminary validation. Pain 104, 639 646. doi:10.1016/s0304-3959(03)0011-0 Crombez, G., Eccleston, C., Baeyens, F., Eelen, P., 1998. When somatic information threatens, catastrohic thinking enhances attentional interference. Pain 75, 187 198. doi:10.1016/s0304-3959(97)0019-4 Dehghani, M., Share, L., Nicholas, M.K., 004. Modification of attentional biases in chronic ain atients: a reliminary study. Eur J Pain 8, 585 594. doi:10.1016/j.ejain.004.0.003 Derryberry, D., Reed, M.A., 00. Anxiety-related attentional biases and their regulation by 3
attentional control. J. Abnorm. Psychol. 111, 5 36. doi:10.1037//001-843x.111..5 Eccleston, C., Crombez, G., 1999. Pain demands attention: a cognitive-affective model of the interrutive function of ain. Psychol. Bull. 15, 356 366. doi:10.1037/0033-909.15.3.356 Eccleston, C., Fisher, E.A., Vervoort, T., Crombez, G., 01. Worry and catastrohizing about ain in youth: A rearaisal. Pain 153, 1560 156. doi:10.1016/j.ain.01.0.039 Eriksen, B.A., Eriksen, C.W., 1974. Effects of noise letters uon the identification of a target letter in a nonsearch task. Percet. Psychohys. doi:10.3758/bf030367 Fan, J., McCandliss, B.D., Sommer, T., Raz, A., Posner, M.I., 00. Testing the efficiency and indeendence of attentional networks. J. Cogn. Neurosci. 14, 340 347. doi:10.116/0898990317361886 Heathcote, L.C., Vervoort, T., Eccleston, C., Fox, E., Jacobs, K., Van Ryckeghem, D.M.L., Lau, J.Y.F., 015. The relationshi between adolescents ain catastrohizing and attention bias to ain faces is moderated by attention control. Pain 156, 1334 1341. doi:10.1097/j.ain.0000000000000174 Helzer, E.G., Connor-Smith, J.K., Reed, M. a, 009. Traits, states, and attentional gates: temerament and threat relevance as redictors of attentional bias to social threat. Anxiety. Stress. Coing, 57 76. doi:10.1080/1061580080744 Holmbeck, G.N., 00. Post-hoc robing of significant moderational and mediational effects in studies of ediatric oulations. J. Pediatr. Psychol. 7, 87 96. doi:10.1093/jesy/7.1.87 King, S., Chambers, C.T., Huguet, A., MacNevin, R.C., McGrath, P.J., Parker, L., MacDonald, A.J., 011. The eidemiology of chronic ain in children and adolescents revisited: A systematic review. Pain 15, 79 738. doi:10.1016/j.ain.011.07.016 Komogortsev, O. V., Gobert, D. V., Jayarathna, S., Koh, D.H., Gowda, S.M., 010. Standardization of automated analyses of oculomotor fixation and saccadic behaviors. IEEE Trans. Biomed. Eng. 57, 635 645. doi:10.1109/tbme.010.05749 Meesters, C., Muris, P., Van Rooijen, B., 007. Relations of neuroticism and attentional control with symtoms of anxiety and aggression in non-clinical children. J. Psychoathol. Behav. Assess. 9, 149 158. doi:10.1007/s1086-006-9037-6 Monk, C.S., McClure, E.B., Nelson, E.E., Zarahn, E., Bilder, R.M., Leibenluft, E., Charney, D.S., Ernst, M., Pine, D.S., 003. Adolescent immaturity in attention-related brain engagement to emotional facial exressions. Neuroimage 0, 40 48. doi:10.1016/s1053-8119(03)00355-0 Muris, P., Van Der Pennen, E., Sigmond, R., Mayer, B., 008. Symtoms of anxiety, deression, and aggression in non-clinical children: Relationshis with self-reort and erformance-based measures of attention and effortful control. Child Psychiatry Hum. Dev. 39, 455 467. doi:10.1007/s10578-008-0101-1 Nelson, C.A., Bloom, F.E., Cameron, J.L., Amaral, D., Dahl, R.E., Pine, D., 00. An integrative, multidiscilinary aroach to the study of brain-behavior relations in the context of tyical and atyical develoment. Dev. Psychoathol. 14, 499 50. doi:10.1017/s095457940003061 33
Olejnik, S., Algina, J., 000. Measures of Effect Size for Comarative Studies: Alications, Interretations, and Limitations. Contem. Educ. Psychol. 5, 41 86. doi:10.1006/ces.000.1040 Owens, M., Koster, E.H.W., Derakshan, N., 013. Imroving attention control in dyshoria through cognitive training: Transfer effects on working memory caacity and filtering efficiency. Psychohysiology 50, 97 307. doi:10.1111/sy.1010 Perneger, T. V, 1998. What s wrong with Bonferroni adjustments. BMJ 316, 136 138. doi:10.1136/bmj.316.7139.136 Perquin, C.W., Hazebroek-Kamschreur, A. a J.M., Hunfeld, J. a M., Bohnen, A.M., Van Suijlekom-Smit, L.W. a, Passchier, J., Van Der Wouden, J.C., 000. Pain in children and adolescents: A common exerience. Pain 87, 51 58. doi:10.1016/s0304-3959(00)0069-4 Roelofs, J., Peters, M.L., van der Zijden, M., Thielen, F.G.J.M., Vlaeyen, J.W.S., 003. Selective attention and avoidance of ain-related stimuli: a dot-robe evaluation in a ain-free oulation. J. Pain 4, 3 38. doi:htt://dx.doi.org/10.1016/s156-5900(03)00634-5 Rommelse, N.N.J., Van der Stigchel, S., Sergeant, J.A., 008. A review on eye movement studies in childhood and adolescent sychiatry. Brain Cogn. 68, 391 414. doi:10.1016/j.bandc.008.08.05 Salvucci, D.D., Salvucci, D.D., Goldberg, J.H., Goldberg, J.H., 000. Identifying Fixations and Saccades in Eye-Tracking Protocols. Proc. Eye Track. Res. Al. Sym. 71 78. doi:10.1145/355017.35508 Schoorl, M., Putman, P., Van Der Werff, S., Van Der Does, a. J.W., 014. Attentional bias and attentional control in Posttraumatic Stress Disorder. J. Anxiety Disord. 8, 03 10. doi:10.1016/j.janxdis.013.10.001 Sear, L.P., 000. The adolescent brain and age-related behavioral manifestations. Neurosci. Biobehav. Rev. doi:10.1016/s0149-7634(00)00014- Sielberger, C.D., Edwards, C.D., Lushene, R.E., Montuori, J., Platzek, D., 1973. The State- Trait Anxiety Inventory for Children (reliminary manual). Consulting Psychologists Press, Palo Alto, CA. Susa, G., Pitică, I., Benga, O., Miclea, M., 01. The self regulatory effect of attentional control in modulating the relationshi between attentional biases toward threat and anxiety symtoms in children. Cogn. Emot. 6, 1069 1083. doi:10.1080/0699931.011.638910 Van Damme, S., Crombez, G., Eccleston, C., 004. Disengagement from ain: the role of catastrohic thinking about ain. Pain 107, 70 76. doi:htt://dx.doi.org/10.1016/j.ain.003.09.03 Van Damme, S., Legrain, V., Vogt, J., Crombez, G., 010a. Keeing ain in mind: A motivational account of attention to ain. Neurosci. Biobehav. Rev. 34, 04 13. doi:10.1016/j.neubiorev.009.01.005 Van Damme, S., Legrain, V., Vogt, J., Crombez, G., 010b. Keeing ain in mind: A motivational account of attention to ain. Neurosci. Biobehav. Rev. doi:10.1016/j.neubiorev.009.01.005 34
Van Ryckeghem, D.M.L., Crombez, G., Van Hulle, L., Van Damme, S., 01. Attentional bias towards ain-related information diminishes the efficacy of distraction. Pain 153, 345 351. doi:10.1016/j.ain.01.07.03 Verhoeven, K., Crombez, G., Eccleston, C., Van Ryckeghem, D.M.L., Morley, S., Van Damme, S., 010. The role of motivation in distracting attention away from ain: An exerimental study. Pain 149, 9 34. doi:htt://dx.doi.org/10.1016/j.ain.010.01.019 Vervoort, T., Caes, L., Crombez, G., Koster, E., Van Damme, S., Dewitte, M., Goubert, L., 011. Parental catastrohizing about children s ain and selective attention to varying levels of facial exression of ain in children: A dot-robe study. Pain 15, 1751 1757. doi:10.1016/j.ain.011.03.015 Vervoort, T., Caes, L., Trost, Z., Notebaert, L., Goubert, L., 01. Parental attention to their child s ain is modulated by threat-value of ain. Heal. Psychol. doi:10.1037/a0099 Vervoort, T., Goubert, L., Crombez, G., 009. The relationshi between high catastrohizing children s facial dislay of ain and arental judgment of their child's ain. Pain 14, 14 148. doi:10.1016/j.ain.008.1.08 Vervoort, T., Trost, Z., Prkachin, K.M., Mueller, S.C., 013a. Attentional rocessing of other s facial dislay of ain: An eye tracking study. Pain 154, 836 844. doi:10.1016/j.ain.013.0.017 Vervoort, T., Trost, Z., Sütterlin, S., Caes, L., Moors, A., 014. Emotion regulatory function of arent attention to child ain and associated imlications for arental ain control behaviour. Pain 155, 1453 1463. doi:10.1016/j.ain.014.04.015 Vervoort, T., Trost, Z., Van Ryckeghem, D.M.L., 013b. Children s selective attention to ain and avoidance behaviour: The role of child and arental catastrohizing about ain. Pain 154, 1979 1988. doi:10.1016/j.ain.013.05.05 Vlaeyen, J.W.S., Linton, S.J., 000. Fear-avoidance and its consequences in chronic musculoskeletal ain: A state of the art. Pain. doi:10.1016/s0304-3959(99)004-0 Von Baeyer, C.L., Piira, T., Chambers, C.T., Traanotto, M., Zeltzer, L.K., 005. Guidelines for the cold ressor task as an exerimental ain stimulus for use with children. J. Pain. doi:10.1016/j.jain.005.01.349 35
8. FIGURE LEGENDS Figure 1: (1) Mean total fixation bias to ain cues (averaged across low, moderate, and high facial ain exressiveness) as a function of low (1 SD below the mean) and high (1 SD above the mean) levels of children s anxiety and attention control () Mean total fixation duration (averaged across all faces) as a function of low (1 SD below the mean) and high (1 SD above the mean) levels of children s anxiety and attention control. (3) Pain tolerance (in seconds) as a function of low (1 SD below the mean) and high (1 SD above the mean) levels of children s anxiety and attention control. (4) Anticiated ain intensity as a function of low (1 SD below the mean) and high (1 SD above the mean) levels of children s anxiety and attention control. All: **.005; *.05; 36
37