توجه به محرک های ترس در طول مواجهه: اثر حواس پرتی بر کاهش اضطراب، خودکارآمدی و کنترل درک شده

Abstract To replicate and extend the finding that distraction facilitates between session anxiety reduction (Oliver & Page (2003)), 27 spider phobics underwent three 10-min sessions of in vivo exposure followed by one 10-min exposure session at a 4-week follow-up, while having either stimulus-relevant focused conversation or stimulus-irrelevant distracting conversation with the experimenter. Physiological arousal and subjective anxiety were measured during exposure, and self-efficacy, perceived control and performance on a behavioural task were measured at pre-treatment, post session-3, and follow-up. Monitoring and blunting coping styles were also measured at pre-treatment to assess their impact on treatment outcome. Despite equal physiological activation between the groups, those who underwent distracted exposure showed greater reductions in subjective fear within and between sessions, and showed greater increases in self-efficacy ratings, internal perceived control and performance on a behavioural task. Coping style did not interact with the effect of distraction or focusing during exposure, however blunters had less subjective anxiety reduction overall, particularly when they underwent focused exposure. Results are discussed in terms of the emotional processing model and self-efficacy theory.

. Introduction Understanding the impact of attention to phobic stimuli during exposure is important for the treatment of phobic anxiety for several reasons. Firstly, during “natural” exposure an “attentional shift toward threat-stimuli is counteracted by a conscious attempt to cognitively avoid the stimulus” (Craske, Street, Jayaraman, & Barlow, 1991; p. 209), a tendency often seen in exposure treatment when patients distract attention away from the phobic object/situation as a means of coping with or reducing anxiety (Craske, Street, & Barlow, 1989). Secondly, some therapists instruct patients to use distraction when approaching feared situations (Crask et al., 1989) to increase the probability and duration of exposure (Beck, Emery, & Greengerg, 1985). Thirdly, models of exposure predict that the direction of attention influences the process of anxiety reduction. In particular, the emotional processing model (Foa & Kozak, 1986) argues that distraction interferes with sensory encoding of information presented during exposure, thereby inhibiting formation of a new memory of the feared event, resulting in less anxiety reduction. A number of studies have investigated the impact of distracting attention away from the phobic stimulus versus focusing attention towards the phobic stimulus, on anxiety reduction. However, given variation between studies in levels of participant’s stimulus bound anxiety, the types of anxiety-disordered participants treated, the duration and number of exposure sessions, and the indexes of therapeutic change, the findings are difficult to interpret. Perhaps the biggest obstacle to interpretation is the wide variation in the types of distraction used and consequently, the degree to which attention is distracted away from the phobic stimulus. Some research with animal-fearful participants has found that, compared to visual focusing, visual distraction during exposure slows the therapeutic response (e.g., Mohlman and Zinbarg, 2000 and Rodriguez and Craske, 1995). However, other studies have shown that when distraction was cognitive (e.g., verbal material presented auditorily) rather than visual, and when participants with a phobic level of fear underwent exposure, distraction did not have the same detrimental effects (e.g., Antony, McCabe, Leeuw, Sano and Swinson, 2001 and Craske, Street, Jayaraman and Barlow, 1991). The results of studies that included a follow-up period suggest that distracted exposure may enhance anxiety reduction during exposure, but that gains are not maintained on all indexes of change. For example, two studies using obsessive–compulsive participants, found within session reductions in subjective fear failed to generalise to between session reductions in those who underwent distraction (Grayson, Foa and Steketee, 1982 and Grayson, Foa and Steketee, 1986). Research using other participant populations has also found poorer longer-term treatment outcomes for those who underwent distraction during exposure (e.g., Haw & Dickerson, 1998: mildly spider-fearful participants with a same day follow-up; Kamphuis & Telch, 2000: moderate to severe claustrophobic participants with a 2-week follow-up; and Craske et al., 1989: agoraphobic participants with a 6-month follow-up). However, a recent series of studies has demonstrated the benefits of using distraction during exposure. Penfold and Page (1999) investigated the impact of stimulus-irrelevant neutral conver sation with mild to severe blood and injection-fearful participants in a single exposure session. Participants having a distracting conversation during exposure reported significantly lower subjective fear compared to those having a focusing conversation about the phobic stimuli, their thoughts, feelings and physiological responses. Aside from the 1982 study by Grayson et al. (replicated and extended by Grayson et al., 1986), Penfold and Page’s study is the first to be followed by a published replication and extension. Oliver and Page (2003) extended on the previous study by including three weekly exposure sessions and a 1-month follow-up, using the same manipulation of attention and participant group. The study found that from the end of exposure session-1 onwards, distracted participants had lower self-reported fear than focused participants, and the focusing group had the largest return of fear (ROF) at follow-up. The replicated finding that a clinically relevant form of attentional distraction facilitated treatment outcome, needs to be explained theoretically. The emotional processing model (Foa & Kozak, 1986) is the framework within which these findings are considered. 1.1. Emotional processing According to the emotional processing model (Foa & Kozak, 1986), attention plays a crucial role in anxiety reduction. In this model, the process of exposure begins with activation of the fear prototype in memory, followed by habituation of physiological responding. Since the stimulus is paired with decreased physiological responding, dissociation within the prototype between the response elements (programs for physiological mobilisation) and the stimulus elements (sensory information about the stimulus) follows. Lowered arousal results in corrective information, for example, decreased perceptions of harm and decreased negative valence, becoming incorporated into the meaning elements. Activation of the fear prototype during exposure is typically indexed by the presence of subjective fear and physiological arousal, and although “there is no means of directly measuring emotional processing” (Mohlman and Zinbarg, 2000; p. 127), modification of the network is indexed by within and between exposure session reductions on fear measures. Indicators of incomplete processing include a lack of between session habituation, and a ROF after a period where the fear has decreased for some time (Rachman, 1979)1. The emotional processing model identifies sensory encoding of the phobic information as a prerequisite for emotional processing. Therefore, “distraction strategies such as pretending to be somewhere else, distorting a fearful image, concentrating on non-fearful elements of a situation, and so on, diminish encoding of fear-relevant information and thus impede activation of fear” (Foa and Kozak, 1986; p. 30). Because the prototype is not completely activated during distraction, dissociation between the response and stimulus elements is impaired, and less corrective information is incorporated into the meaning elements, resulting in incomplete emotional processing, demonstrated by less subjective, behavioural, and physiological anxiety reduction. Alternatively, focusing on fear-relevant information during exposure should enhance encoding of fear-relevant information and facilitate activation of the fear structure, resulting in greater emotional processing, demonstrated by increased subjective, behavioural, and physiological anxiety reduction. The sequence of events underlying short-term “processing”, as described by the emotional processing model, is unable to account for how distracting conversation during exposure could produce greater anxiety reduction. However, the model describes the process of stable longer-term change in the fear structure which “constitute[s] additional information that accumulates to modify general beliefs and attitudes about ability to cope with feared situations” (Foa and Kozak, 1986; p. 29), and cites self-efficacy as one of these “global beliefs”. While remaining within the framework of the emotional processing model, an explanation for how distracting conversation impacts global beliefs, is found in self-efficacy theory. 1.2. Self-efficacy and perceived control Self-efficacy is defined as the beliefs and attitudes that people hold about their ability to cope, or perform various levels of a given behaviour. According to self-efficacy theory, “it is mainly perceived inefficacy in coping with potentially aversive events that makes them fearful” (Bandura, 1983; p. 465), and phobic behaviour is caused and maintained by beliefs about inability to cope. Self-efficacy is conceptualised as a “generative capability” that can be developed through behavioural accomplishment (Bandura, 1983; p. 467), and treatments that have emphasised behavioural mastery have indeed led to increased self-efficacy perceptions and superior treatment response (e.g., Williams, Dooseman, & Kleinfield, 1984). Following exposure, increased self-efficacy has been found to be predictive of several types of psychological change (e.g., Jones and Menzies, 2000 and Zoellner, Echiverri and Craske, 2000), with low self-efficacy related to relapse (Craske & Rachman, 1987). According to self-efficacy theory, it is both the quality and quantity of efficacy enhancing information provided during exposure, and how individuals weight and integrate this information, that mediates changes in anxiety. Consistent with self-efficacy theory, the rationale for the clinical use of distraction during exposure is conceptualised as helping the patient to cope, which increases the probability of successful “performance” in the presence of the phobic stimuli. From the perspective of self-efficacy theory, engaging in stimulus-irrelevant conversation that is personally meaningful in the presence of the feared stimuli, is powerful efficacy enhancing information. For example, the thought “I can have a normal conversation about myself in the presence of my worst fear” may be weighted more heavily in efficacy evaluations than simply feeling anxious and experiencing a slow decline in anxiety in the presence of the phobic stimulus. Both Penfold and Page (1999), and Oliver and Page (2003) describe their results as consistent with self-efficacy theory, however, thus far no studies manipulating attention towards the phobic stimulus have considered the therapeutic benefit of distraction in terms of its impact on self-efficacy. Closely related to self-efficacy, are beliefs about perceived control. According to self-efficacy theory “if people believe that they can exercise control over the occurrence of events that can be injurious, they do not fear them” (Bandura, 1983; p. 465). Several studies have demonstrated that non-specific aversive stimulation results in less anxiety and distress when participants have a sense of perceived control over the aversive stimulus (e.g., Glass and Singer, 1973 and Zvolensky, Eifert, Lejuez and McNeil, 1999). With regard to the effect of distraction on perceived control, Craske et al. (1989) found no differences between participants who underwent distraction and those who underwent focusing during exposure, on measures of predicted or actual control over responses to feared somatic sensations. However, Penfold and Page (1999) found that of those with low perceived control over anxiety at pre-treatment, those who underwent distracted exposure had greater anxiety reduction than those who underwent focused exposure or exposure alone, suggesting that distraction either increased perceived control or acted to compensate for low perceived control. Extending on this finding, Oliver and Page (2003) found all three groups had significant and equal increases in perceived control over anxiety from pre to post-treatment, but at follow-up, the distraction group had significantly higher perceived control compared with those in the focusing group, who demonstrated a significant reduction. Regardless of its causal relationship with anxiety reduction, perceived control appears to be modified following stimulus-irrelevant conversation during exposure. The current study replicated Penfold and Page (1999) and Oliver and Page (2003) while extending on these studies in several ways. These previous studies used mildly blood and injection-fearful participants, posing two potential problems for the generalisability of the study. Firstly, it is unclear whether the facilitative effect is generalisable to individuals with phobic levels of anxiety. Penfold and Page (1999) found that participants in the distraction condition with higher levels of stimulus bound anxiety, had more anxiety reduction, suggesting that the therapeutic effect of distraction may be even more pronounced in a phobic sample. Secondly, given the atypical nature of blood-injury phobia in terms of vasovagal syncope (Page, 1994), it remains to be determined whether the facilitative effect of distraction can be replicated with phobics who do not exhibit the fainting response. For these reasons, a spider phobic sample was used in the current study. Improved and additional indexes of change were also included in the current study. Penfold and Page (1999) found no group differences on a behavioural avoidance test, arguably due to its insensitivity (as seen in the ceiling effects), and Oliver and Page (2003) did not use a behavioural measure. The current study included a more sensitive behavioural avoidance test (with a larger number of steps than used by Penfold & Page), to explore whether undergoing exposure with distraction results in less behavioural avoidance. Unlike these previous studies, physiological indices of skin conductivity, heart rate, and blood pressure were also measured. Finally, a measure of coping style under threat was included. Informational preference under threat has been characterised as a stable behavioural style consisting of two dimensions: monitoring (the extent one monitors and seeks out threat information) and blunting (the extent one cognitively distracts and psychologically blunts threat information; Miller, 1987). Within the emotional processing framework, blunting is classified as a distraction strategy and should therefore have a detrimental impact on encoding of phobic information and activation of fear during exposure. Although coping style has been shown to impact anxiety reduction during exposure to phobic stimuli (e.g., Muris, Merckelbach and de Jong, 1995 and Steketee, Bransfield, Miller and Foa, 1989), the direction of its relationship with therapeutic benefit has been mixed. Miller, Brody and Summerton (1988) argued that high monitors/low blunters and low monitors/high blunters may cope better under stress if the conditions of the situation are consistent with their coping style preference. Further, Rodriguez and Craske (1993) suggested that monitors might experience more anxiety reduction during exposure if focused on the stimulus, while blunters might experience more anxiety reduction if distracted from the stimulus, and that this interaction may account for some of the inconsistencies in the distraction literature. Antony et al. (2001) conducted the only study to test this prediction, and found no interaction between coping style and attention on measures of heart rate, subjective fear or behavioural avoidance. However, in spite of evidence that they are independent dimensions (e.g., van Zuuren & Wolfs, 1991), the study collapsed monitoring and blunting. Therefore, the current study investigated how both high and low blunters and high and low monitors benefit from distraction versus focusing during exposure. Based on the findings of Penfold and Page (1999), and Oliver and Page (2003), the following predictions were made about a group of spider phobics undergoing distraction (stimulus-irrelevant conversation) during exposure: (a) that they would demonstrate greater within and between session reductions in subjective anxiety and greater between session reductions in behavioural avoidance, than those undergoing focusing (stimulus-relevant conversation during exposure)2,3; (b) that they would show greater reductions on a standardised self-report measure of spider fear than those undergoing focusing during exposure; (c) that they would maintain these gains at a 4-week follow-up and would not demonstrate a ROF; and (d) that they would show a larger increase in self-efficacy and perceived control than those undergoing focusing. It was also predicted that high blunters who underwent distraction during exposure and high monitors who underwent focusing during exposure would experience greater therapeutic benefit compared with others in their experimental condition.

3. Results Data from 27 participants were analysed. Missing physiological data (0.4%) were replaced by the mean of the closest three data points. Prior to analysis, variables within each condition with standardised scores exceeding 2.5 were assigned a raw score equivalent to the next most extreme score on that distribution (Tabachnick & Fidell, 1996). All t values reported were conducted in accordance with Bonferroni corrections. 3.1. Manipulation checks Of the 541 randomly selected 60-s excerpts from the exposure sessions, 100% were judged as belonging to the experimental condition from which they were drawn, supporting the assumption that the conversations were stimulus-irrelevant or stimulus-relevant. Also, compared to those who underwent focused exposure, participants who underwent distracted exposure reported spending less time thinking about the phobic stimulus during sessions 1, 2 and 3, (distraction: M=51.77%; focusing: M=96.11%; F(1,23)=27.20, η2=0.54, p<0.001), and during follow-up session-4 (distraction: M=26.04%; focusing: M=97.22%; t(21)=7.04, p<0.001), supporting the assumption that stimulus-irrelevant conversation directs cognitive attention away from phobic stimulus related thoughts, relative to stimulus-relevant conversation. There were no group differences on ratings of the first three exposure sessions on; percentage of time participants were judged to hold their gaze towards the spider (distraction: M=96.36%; focusing: M=96.62%; F(1,25)=0.90, p=ns), self-reported estimates of the percentage of time they spent looking at the spider (distraction: M=97.46%; focusing: M=94.64%; F(1,23)=1.12, p=ns), or self-reported estimates of the percentage of time the spider was moving (distraction: M=39.26%; focusing: M=48.19%; F(1,23)=2.52, p=ns). There were also no group differences at follow-up for the three measures, respectively: (distraction: M=97.93%; focusing: M=97.00%; t(21)=0.96, p=ns); (distraction; M=98.5%; focusing: M=97.78%; t(21)=0.652, p=ns); and (distraction: M=45.71%; focusing: M=41.67%; t(21)=0.37, p=ns). These findings support the assumption that distracting or focusing conversations do not interfere with the degree of visual attention towards the phobic stimulus. Further, the absence of group differences in perceived movement ratings, supports the assumption that exposure characteristics that influence anxiety reduction (e.g., the intensity of exposure) were equal between conditions. There were no group differences in self-report ratings of “similarity to typically fearful spiders” for the phobic stimuli used in sessions 1 and 3, (distraction: M=6.78; focusing: M=5.44; t(16)=1.13, p=ns), session-2 (distraction: M=5.44; focusing: M=5.11; t(16)=0.33, p=ns), session-4 (distraction: M=5.86; focusing: M=4.89; t(21)=1.16, p=ns), or the BAT (distraction: M=6.89; focusing: M=7.22; t(16)=0.33, p=ns), removing this variable as a potential explanation for group differences in anxiety reduction. Finally, all groups experienced physiological activation from baseline to the first measurement taken in session-1, as revealed by main effects of time for systolic blood pressure (F(1,25)=43.84, η2=0.64, p<0.001), diastolic blood pressure (F(1,25)=46.03, η2=0.65, p<0.001), heart rate (F(1,25)=12.77, η2=0.34, p=0.001), and SCL (F(1,23)=39.38, η2=63, p<0.001). Those undergoing distracted exposure experienced greater activation in SCL from baseline to the first measurement than those undergoing focused exposure (F(1,23)=5.50, η2=0.19, p=0.028), but there were no group differences in the alter measures of physiological activation. There were also no differences in the activation of subjective fear in the initial 2 minutes of exposure between those undergoing distracted versus those undergoing focused exposure (1st SUD rating: t(25)=0.233, p=ns; 2nd SUD rating: t(25)=0.98, p=ns). 3.2. Dependent variables pre-treatment There were no group differences on any dependent variable at pre-treatment (see Table 1 for physiological data, Table 2 for self-report psychometric data, and Fig. 4 for BAT data). There were no group differences in monitoring scores (distraction: M=14.00; focusing: M=13.38; t(25)=0.79, p=ns) or blunting scores (distraction: M=5.71; focusing: M=3.92; t(25)=1.88, p=ns). Table 1. Mean physiological reactivity for distraction and focusing conditions at pre-treatment and during exposure sessions 1–4 Measure Condition Time Pre-treatment Session-1 Session-2 Session-3 Follow-up session-4 Heart rate Distraction 78.57 (3.00) 83.84 (2.78) 79.63 (2.95) 80.50 (2.93) 88.97 (2.54) Focusing 76.23 (4.03)a 80.82 (2.88) 76.99 (3.06) 76.08 (3.04) 78.73 (3.17) Skin conductivity level Distraction 2.32 (0.45) 3.03 (0.39) 2.92 (0.38) 3.19 (0.44) 2.63 (0.36) Focusing 2.53 (0.46)b 2.95 (0.41) 3.44 (0.40) 3.48 (0.46) 2.97 (0.46) Systolic blood pressure Distraction 107.00 (3.01) 120.83 (2.46) 118.07 (2.35) 114.17 (2.15) 117.26 (2.60) Focusing 109.38 (2.71)c 115.97 (2.74) 112.80 (2.43) 109.26 (2.23) 110.49 (3.24) Diastolic blood pressure Distraction 69.71 (1.45) 79.96 (1.93) 78.49 (2.09) 76.19 (1.68) 77.76 (1.54) Focusing 64.92 (2.16)d 74.40 (2.00) 72.77 (2.17) 70.39 (1.74) 69.51 (1.92) Note: Means for each session represent estimated marginal means calculated from the five measures taken in each session. Values in parenthesis represent the standard error of the mean. a At pre-treatment there were no differences between conditions; t(25)=0.47, p=ns. b At pre-treatment there were no differences between conditions; t(25)=−0.34, p=ns. c At pre-treatment there were no differences between conditions; t(25)=−0.59, p=ns. d At pre-treatment there were no differences between conditions; t(25)=0.79, p=ns. Table options Table 2. Mean self-report psychometric scores for distraction and focusing conditions at pre-treatment, post session-3, and post follow-up Measure Condition Time Pre-treatment Post session-3 Post follow-up FSQ Distraction 67.21 (3.12) 50.14 (4.16) 42.29 (5.40) Focusing 61.54 (3.24)a 61.77 (4.31) 57.67 (6.74) SSES Distraction 2.55 (0.23) 5.23 (0.32) 5.65 (0.37) Focusing 2.36 (0.24)b 3.49 (0.33) 4.31 (0.48) ACQ Full scale Distraction 88.00 (4.28) 90.29 (3.91) 91.00 (4.57) Focusing 83.00 (4.45)c 81.31 (4.05) 92.11 (5.69) Internal subscale Distraction 37.57 (2.35) 40.86 (2.04) 40.64 (2.13) Focusing 36.39 (2.43) 35.62 (2.11) 41.00 (2.65) External subscale Distraction 50.43 (2.23) 49.43 (2.20) 50.36 (2.62) Focusing 46.62 (2.32) 45.69 (2.28) 51.11 (3.26) Note: Values in parenthesis represent the standard error of the mean. a At pre-treatment there were no differences between conditions; t(25)=1.26, p=ns. b At pre-treatment there were no differences between conditions; t(25)=−0.59, p=ns. c At pre-treatment there were no differences between conditions; t (25)=0.81, p=ns. Table options Mean number of steps achieved on the BATs (+SE) for those who underwent ... Fig. 4. Mean number of steps achieved on the BATs (+SE) for those who underwent distracted exposure and focused exposure, from pre-treatment (BAT-1) to follow-up post exposure session-4 (BAT-6). Figure options 3.3. Analysis of dependent variables Two separate univariate repeated measures ANOVAs were conducted on each dependent variable, one assessing change from sessions 1 to 3 and the other from session-3 to follow-up session-4. As four participants from the focusing condition did not attend the follow-up, conducting a separate ANOVA ensured results from sessions 1 to 3 were not biased by their exclusion. Also, sensitivity of detecting an increase in fear indexes from session-3 to follow-up would be reduced if data were collapsed across four sessions. 3.4. Within and between exposure session analysis 3.4.1. SUDS As seen in Fig. 2, SUD ratings for participants in both conditions decreased within and between each of the first three exposure sessions (linear trends: F(1,25)=62.13, η2=0.71, p<0.001; F(1,25)=25.34, η2=0.50, p<0.001). As predicted, those undergoing distracted exposure reported lower SUD ratings overall (F(1,25)=6.41, η2=0.20, p=0.018), and their ratings decreased more rapidly within sessions (linear trend: F(1,25)=10.11, η2=0.29, p=0.004). It can also be seen in Fig. 2 that SUD ratings for participants in both conditions decreased within and between sessions 3 and 4 (linear trends: F(1,20)=48.39, η2=0.71, p<0.001; F(1,20)=22.96, η2=0.53, p<0.001). Again, those undergoing distracted exposure reported lower SUD ratings overall (F(1,20)=5.06, η2=0.20, p=0.036), and their ratings decreased more rapidly within sessions (linear trend: F(1,20)=5.59, η2=0.22, p=0.028). Simple effects analysis revealed those undergoing distracted exposure had lower SUD ratings during session-2 (F(1,25)=9.97, η2=0.29, p=0.004), session-3 (F(1,25)=5.18, η2=0.17, p=0.032), and session-4 (F(1,20)=4.85, η2=0.20, p=0.04). Mean SUD ratings (+SE) during exposure sessions for those undergoing distracted ... Fig. 2. Mean SUD ratings (+SE) during exposure sessions for those undergoing distracted exposure and focused exposure. Figure options To determine if distraction during exposure resulted in greater SUD ratings reductions for those with higher levels of anxiety (as found by Penfold & Page, 1999), participants were divided into high and low anxiety (based on the number of steps they achieved in the pre-treatment BAT) using a median split procedure. As seen in Fig. 3, both high and low initial anxiety participants undergoing distracted exposure demonstrated significant reductions in SUD ratings between sessions, whereas only low initial anxiety participants undergoing focused exposure demonstrated reductions (linear trend: F(1,23)=7.03, η2=0.23, p=0.014). High anxiety participants undergoing focused exposure, experienced no reduction in subjective anxiety. Mean SUD ratings (+SE) during exposure sessions for participants with low and ... Fig. 3. Mean SUD ratings (+SE) during exposure sessions for participants with low and high anxiety undergoing distracted exposure, and low and high anxiety undergoing focused exposure. Figure options 3.4.2. BAT As seen in Fig. 4, there was an overall increase in the number of steps achieved on the BAT over time (linear trend: F(1,25)=72.71, η2=0.74, p<0.001). As predicted, those who underwent distracted exposure completed significantly more steps overall (F(1,25)=5.48, η2=0.18, p=0.028), qualified by a more rapid increase in steps achieved (linear trend: F(1,25)=17.59, η2=0.41, p<0.001). It can also been seen in Fig. 4, that from BAT-4 (post session-3) to BAT-6 (follow-up post session-4), there was an overall increase in the number of steps achieved (linear trend: F(1,21)=5.60, η2=0.21, p=0.028). Again, participants who underwent distracted exposure completed significantly more steps overall (F(1,21)=13.48, η2=0.39, p=0.001). By BAT-3 (post session-2), participants in the distraction condition had completed a greater number of steps than those in the focusing condition (t(25)=2.44, p=0.044), and continued to do so at BAT-4 (t(25)=4.47, p=0.001), BAT-5 (t(25)=3.53, p=0.011), and BAT-6 (t(25)=3.58, p=0.011). 3.4.3. Physiological data Means and standard errors of physiological measurements during sessions 1, 2, 3 and 4 are presented in Table 1. 3.4.3.1. Heart rate There was an overall decrease in heart rate across sessions 1, 2 and 3 (linear trend: F(1,25)=18.01, η2=0.42, p<0.001), but there were no differences between those undergoing distracted and focused exposure. There was an overall increase in heart rate from session-3 to follow-up session-4 (F(1,21)=10.54, η2=0.33, p=0.004), with those undergoing distracted exposure demonstrating higher heart rates (F(1,21)=4.44, η2=0.18, p=0.047), specifically during session-4 (F(1,21)=6.36, η2=0.23, p=0.02). 3.4.3.2. Skin conductivity level There was an overall increase in SCL across sessions 1, 2 and 3 (linear trend: F(1,23)=5.63, η2=0.20, p=0.026). However, SCL reduced within sessions (linear trend: F(1,423)=9.30, η2=0.29, p=0.006), with those undergoing distracted exposure demonstrating a more rapid reduction (linear trend: F(1,23)=4.76, η2=0.17, p=0.040). There were no significant changes in SCL from session-3 to follow-up session-4. 3.4.3.3. Blood pressure Systolic blood pressure: There were overall decreases within and between sessions 1, 2 and 3 (linear trends: F(1,25)=23.27, η2=0.48, p<0.001; F(1,25)=20.24, η2=0.45, p<0.001), with no differences between groups. From session-3 to follow-up session-4, no further reductions between sessions occurred, however those undergoing focused exposure experienced greater reduction within sessions (linear trend: F(1,21)=4.71, η2=0.18, p=0.042), specifically within session-3 (F(4,100)=3.43, η2=0.12, p=0.011). Diastolic blood pressure: There were overall decreases both within and between sessions 1, 2 and 3 (linear trends: F(1,25)=6.39, η2=0.20, p=0.018; F(1,25)=22.82, η2=0.48, p<0.001). Further, those undergoing distracted exposure had higher diastolic blood pressure overall (F(1,25)=5.05, η2=0.17, p=0.034). From session-3 to follow-up session-4, no further between or within session reductions occurred, with those undergoing distracted exposure continuing to experience higher diastolic blood pressure (F(1,21)=11.16, η2=0.35, p=0.003). Those undergoing distracted exposure had higher diastolic blood pressure during session-3 (F(1,25)=5.76, η2=0.19, p=0.024) and session-4 (F(1,21)=11.22, η2=0.35, p=0.003). 3.5. Pre-treatment vs. post session-3 vs. follow-up post session-4 Descriptive statistics for the psychometric measures are presented in Table 2. 3.5.1. FSQ As shown in Table 2 there was an overall decrease in scores on the FSQ from pre-treatment to post session-3 (F(1,25)=14.65, η2=0.37, p=0.001). As predicted, those who underwent distracted exposure demonstrated decreased FSQ scores, while those who underwent focused exposure demonstrated no decrease (F(1,25)=15.46, η2=0.38, p=0.001). From post session-3 to follow-up post session-4 there was an overall decrease in FSQ scores (F(1,21)=6.63, η2=0.24, p=0.018), with those who underwent distraction showing a larger reduction than those who underwent focusing (F(1,21)=3.34, η2=0.14, p=0.082). 3.5.2. SSES As shown in Table 2, self-efficacy scores increased from pre-treatment to post session-3 (F(1,25)=70.30, η2=0.74, p<0.001). As predicted, those who underwent distracted exposure had higher scores than those who underwent focused exposure (F(1,25)=8.54, η2=0.26, p=0.007), qualified by a more rapid increase in scores over time (F(1,25)=11.53, η2=0.32, p=0.002). From post session-3 to follow-up post session-4 there was also an overall increase in scores (F(1,20)=7.79, η2=0.28, p=0.011), with the scores of those who underwent distracted exposure remaining higher (F(1,20)=7.39, η2=0.27, p=0.013). Those who underwent distracted exposure had significantly higher self-efficacy scores at post session-3 (t(25)=3.76, p=0.02) and at follow-up post session-4 (t(20)=2.21, p=0.039). As shown in Fig. 5, experimental self-efficacy scores increased from pre-treatment to post session-3 (F(1,25)=121.75, η2=0.83, p<0.001), with those who underwent distracted exposure demonstrating higher scores overall (F(1,25)=5.54, η2=0.18, p=0.027), qualified by a more rapid increase (F(1,25)=11.45, η2=0.31, p=0.002). There was a continued increase from post session-3 to follow-up post session-4 (F(1,21)=9.67, η2=0.32, p=0.005), with those who underwent distracted exposure demonstrating higher scores (F(1,21)=6.00, η2=0.22, p=0.023). Likewise, general self-efficacy scores showed an increase from pre-treatment to post session-3 (F(1,25)=20.12, η2=0.45, p<0.001), with those who underwent distracted exposure demonstrating higher scores (F(1,25)=4.94, η2=0.17, p=0.036), qualified by a more rapid increase over time (F(1,25)=7.47, η2=0.23, p=0.011). There was no further change in scores from post session-3 to follow-up post session-4. Hypothetical–experimental self-efficacy scores increased from pre-treatment to post session-3 (F(1,25)=10.16, η2=0.29, p=0.004), and from post session-3 to follow-up (F(1,21)=4.71, η2=0.18, p=0.042). There were no differences between those who underwent distracted and focused exposure. Mean SSES item scores (+SE) for experimental, general and ... Fig. 5. Mean SSES item scores (+SE) for experimental, general and hypothetical–experimental self-efficacy for those who underwent distracted exposure and focused exposure. Figure options 3.5.3. ACQ As seen in Table 2, there were no significant changes from pre-treatment to post session-3 for full-scale scores, and subsequently no differences between conditions. External subscale scores showed the same pattern. However, on the internal subscale, those who underwent distracted exposure had a greater increase from pre-treatment to post session-3, than those who underwent focused exposure (F(1,25)=6.22, η2=0.20, p=0.020). From post session-3 to follow-up post session-4, there was an overall increase in full-scale ACQ scores (F(1,21)=16.02, η2=0.43, p=0.001), with those who underwent focused exposure demonstrating a more rapid increase (F(1,21)=12.29, η2=0.37, p=0.002). However, there was no overall significant difference between the conditions. Subscale scores showed the same pattern, with overall increases (external: F(1,21)=11.89, η2=0.36, p=0.002; internal: F(1,21)=8.08, η2=0.28, p=0.01) particularly for those who underwent focused exposure (external: F(1,21)=6.44, η2=0.24, p=0.019; internal: F(1,21)=9.71, η2=0.32, p=0.005). Overall, as predicted, those who underwent distracted exposure had a greater increase in internal control from pre-treatment to post session-3. However, contrary to predictions, those who underwent focused exposure had a greater increase in full-scale and subscale scores from post session-3 to follow-up post session-4. 3.6. Interaction of coping style and SUD reductions Participants were divided into low and high monitors and low and high blunters using a median split procedure. There were no differences in SUD ratings between high and low monitors across sessions 1–4. However, as seen in Fig. 6, low blunters reported lower SUD ratings than high blunters from session-1 to session-3, (F(1,23)=4.64, η2=0.17, p=0.042), with low blunter’s SUD ratings decreasing more rapidly (linear trend: F(1,23)=5.32, η2=0.19, p=0.030). The same profile was evident from session-3 to follow-up session-4, with low blunters reporting lower SUD scores (F(1,18)=9.42, η2=0.34, p=0.007). Overall, high blunters reported higher SUD ratings during session-2 (F(1,23)=6.54, η2=0.22, p=0.018), session-3 (F(1,23)=5.45, η2=0.19, p=0.029), and session-4 (F(1,23)=8.31, η2=0.32, p=0.01). Furthermore, high blunters undergoing distracted exposure reported lower SUD ratings than high blunters who were undergoing focused exposure from session 1 to 3 (F(1,23)=4.01, η2=0.15, p=0.057) and from session 3 to 4 (F(1,18)=7.41, η2=0.29, p=0.014). Mean SUD ratings (+SE) during exposure sessions for low blunters (n=5) and high ... Fig. 6. Mean SUD ratings (+SE) during exposure sessions for low blunters (n=5) and high blunters (n=9) undergoing distracted exposure, and low blunters (n=10) and high blunters (n=3) undergoing focused exposure. Figure options 3.7. Predictors of behavioural avoidance With the number of steps completed on BAT-4 (post session-3) as the criterion variable, SUD ratings and self-efficacy ratings made at post session-3 were entered into a multiple regression analysis using a stepwise method. A significant model was found (adjusted R2=0.47; F(1,25)=24.33, p<0.001), with self-efficacy at post session-3 as a significant predictor of the number of steps completed on BAT-4 (β=0.70, p<0.001), indicating that achieving a greater number of steps on the BAT was predicted by higher self-efficacy. Self-efficacy at follow-up post session-4 was also a predictor of the number of steps achieved on BAT-6 at follow-up post session-4 (adjusted R2=0.59; F(1,21)=30.81, p<0.001; β=0.78, p<0.001).