Abstract
The ways in which attachment and sensory processing may be integrated to inform occupational therapy practice have received some theoretical attention in the pediatric (Alers, 2005; Koomar, 2009) and adult (Champagne, 2011) mental health literature. In addition, these variables have each been empirically linked to a range of clinical phenomena, including mental illness (Brown, Cromwell, Filion, Dunn, & Tollefson, 2002; Mikulincer, Horesh, Eilati, & Kotler, 1999), perceptions of pain (Engel-Yeger & Dunn, 2011a; Meredith, Strong, & Feeney, 2006), coping (Meredith, Rappel, Strong, & Bailey, 2015; Schmidt, Nachtigall, Wuethrich-Martone, & Strauss, 2002), and psychological distress (Engel-Yeger & Dunn, 2011b; Jinyao et al., 2012). Nevertheless, specific associations between attachment and sensory processing have received little empirical consideration. The aim of this study was to replicate and extend previous findings regarding interrelationships among attachment and sensory variables to inform clinical approaches and guide future research.
Attachment Theory
Bowlby (1969) conceptualized attachment theory as a theoretical model of social and emotional development in infancy and childhood that has lifelong implications. Internalized early caregiving experiences have been proposed to influence individual perceptions of oneself as being worthy of care and of caregivers as being trustworthy to provide care (Bartholomew & Horowitz, 1991). When early childhood experiences are not adequate, negative perceptions of oneself or others are proposed to develop, reflecting attachment anxiety and attachment avoidance, respectively (Bartholomew & Horowitz, 1991). (A summary of terms used in this article is in Appendix 1, available online at http://otjournal.net; navigate to this article, and click on “Supplemental.”) On the basis of early experiences, individual internal working models are formed that are believed to influence perceptions throughout life (Bowlby, 1969).
Sensory Processing
People differ in the ways in which they process sensory information (Aron & Aron, 1997; Dunn, 2001), and Dunn’s (1997) Model of Sensory Processing offers a theoretical framework for these differences. According to this model, four distinct sensory processing categories are defined according to a person’s neurological threshold (the point at which stimulation is sufficient to elicit a response; the threshold can be high or low) and response strategy (active or passive; Dunn, 1997). People with a high neurological threshold are classified as having either a low registration pattern, reflecting passive responses to stimuli, or a sensation-seeking pattern, reflecting active responses (Dunn, 1997). People with a low neurological threshold are classified as having a sensation-avoiding pattern (they actively limit stimulation) or a sensory sensitivity pattern (they respond passively, experiencing discomfort and being distracted by stimulation; Dunn, 1997; see Appendix 1). As with attachment, Dunn (2001) proposed that sensory processing patterns remain largely consistent across the lifespan.
Associations Between Attachment and Sensory Patterns
A literature review yielded only three articles in which associations among attachment and sensory patterns were considered. A study by Jerome and Liss (2005) with a sample of 133 psychology students provided preliminary support for a relationship between attachment anxiety and avoidance and Dunn’s (1997) four sensory processing categories. Specifically, positive associations were demonstrated between sensory sensitivity and attachment anxiety, sensory avoidance and attachment avoidance, and low registration and both avoidant and anxious attachment patterns. The lack of association between sensation seeking and either attachment classification was speculated to indicate a relationship between sensation seeking and secure attachment.
Using different measures of attachment and sensory sensitivity, Meyer, Ajchenbrenner, and Bowles (2005) also found a link between sensory sensitivity and attachment in a nonclinical sample (N = 156) drawn from college and community populations. These authors found a correlation between the Highly Sensitive Persons Scale (HSPS; Aron & Aron, 1997) and the Upset and Misunderstood subscale of the Inventory of Parent and Peer Attachment (Armsden & Greenberg, 1987) indicating a link between attachment insecurity and sensory sensitivity.
More recently, Levit-Binnun, Szepsenwol, Stern-Ellran, and Engel-Yeger (2014) conducted a study with 194 first-year psychology students in which they considered the mediational role of attachment patterns in the relationship between Dunn’s (1997) sensory quadrants and anxiety. Although not the focus of the study, correlations between sensory responsiveness profiles and attachment were reported, and the following significant positive relationships were identified: sensory sensitivity and attachment anxiety, sensory avoidance and both attachment anxiety and avoidance, and low registration and anxious attachment. In addition, sensory seeking was negatively correlated with attachment avoidance, lending more confidence to associations between sensory seeking and secure attachment.
Although findings of these studies support theoretical expectations, the participants’ attachment system was not activated before data gathering. Ravitz, Maunder, Hunter, Sthankiya, and Lancee (2010) noted that it is important to activate attachment phenomena so that they become manifest, thus increasing the validity of measurement. Activation requires evocation of fear, pain, or distress, which suggests that studies that have not provided this stimulus may misrepresent the strength of relationships among the variables.
Although no other studies have specifically investigated associations between attachment and sensory processing, evidence in the wider literature has supported these associations. For example, links have been demonstrated between sensory sensitivity and mental health diagnoses that involve strong attachment components, such as personality disorders and social phobia (Meyer et al., 2005). Moreover, traumatic childhood events and reduced ability to self-regulate (associated with insecure attachment) have been associated with susceptibility to sensory modulation difficulties (Alers, 2005), and early institutionalization has been associated with both attachment disorders (Zeanah, 2000) and sensory modulation difficulties (Wilbarger, Gunnar, Schneider, & Pollak, 2010).
Links With Psychological Distress
Attachment (Jinyao et al., 2012) and sensory processing (Engel-Yeger & Dunn, 2011b; Liss, Timmel, Baxley, & Killingsworth, 2005) have separately been linked with psychological distress (i.e., anxiety, stress, or depression). These associations have been investigated concurrently in one study (Levit-Binnun et al., 2014) in which attachment anxiety and avoidance mediated the relationship between anxiety symptoms and all four sensory profiles. Given these known relationships, level of distress needs to be controlled to ensure that associations between attachment and sensory processing are not better explained by shared variance with emotional distress.
Hypotheses
With attachment and sensory theories increasingly being integrated to inform clinical practice, the aim of this study was to gain insight into the empirical relationships between attachment and sensory patterns, with a view toward using this knowledge to guide interventions. To elaborate on existing evidence, the associations among adult attachment, sensory processing, and distress (anxiety, stress, depression) variables need to be examined concurrently while evoking the attachment system. On the basis of existing research, four hypotheses were proposed:
Positive associations will be found between sensory sensitivity and attachment anxiety when controlling for distress.
Positive associations will be found between sensory avoidance and attachment avoidance when controlling for distress.
Positive associations will be found between low registration and attachment anxiety when controlling for distress.
No association will be found between sensation seeking and attachment avoidance or anxiety.
Method
Participants
A convenience sample of 162 healthy adults was invited to participate in the study. Thirty-four people either did not respond to the invitation or declined to participate because of time constraints or conflicting commitments. Twelve other participants were excluded because of current pain complaints, resulting in a sample of 116 eligible participants. The demographic details are summarized in Table 1, and additional descriptive data are provided in Table 2. The majority of participants were Australian, single, and employed full-time and had completed Year 12 (senior) schooling.
Descriptive Details for Demographic Variables of the Study Sample (N = 116)
Note. TAFE = technical and further education.
Total >116 because some participants selected more than one option (e.g., studying full-time while working part-time).
Descriptive Data for Continuous Study Variables
Note. HSPS–SV = Highly Sensitive Persons Scale–Shortened Version.
Measures
Demographic Variables.
Participants were asked their age, gender, country of origin, relationship status, education level, employment status, and household income. In addition, on the basis of previous research (Andrews, Meredith, & Strong, 2011), whether the participant knew the researcher (participant type) was recorded for inclusion as a covariate. An additional variable (researcher) indicated which researcher had gathered the data.
Experiences in Close Relationships–Revised Questionnaire.
The Experiences in Close Relationships–Revised Questionnaire (ECR–R; Fraley, Waller, & Brennan, 2000) is a self-report measure of romantic attachment that consists of two 18-item scales: Attachment-Related Anxiety and Attachment-Related Avoidance. The 36 items are rated on a 7-point scale ranging from 1 (strongly disagree) to 7 (strongly agree). The ECR–R has demonstrated suitable validity and reliability (Sibley, Fischer, & Liu, 2005). In the current study, α coefficients for the two dimensions were .93 (anxiety) and .93 (avoidance). Before completing the ECR–R, participants were aware that they would participate in the cold pressor pain-inducement task, which was used to activate the attachment system.
Adolescent/Adult Sensory Profile.
The Adolescent/Adult Sensory Profile (AASP; Brown & Dunn, 2002) is a self-report measure of adult responses to everyday sensory experiences that is based on Dunn’s (1997) Model of Sensory Processing. Sixty items are scored on a 5-point Likert-type scale ranging from 1 (almost never) to 5 (almost always) according to the frequency with which a participant responds to specific sensory events in the manner described. These items are equally sorted among four quadrants (i.e., 15 per quadrant): low registration, sensation seeking, sensory sensitivity, and sensation avoiding. Each quadrant reflects a different sensory processing pattern. Raw scores for each quadrant are converted to reflect the extent to which a person identifies with each pattern, again rated on a 5-point scale ranging from 1 (much less than most people) to 5 (much more than most people). Results of a preliminary study have supported the measure’s validity and internal consistency (Brown, Tollefson, Dunn, Cromwell, & Filion, 2001). In the current study, α coefficients were .75 (low registration), .75 (sensation seeking), .72 (sensory sensitivity), and .78 (sensation avoiding).
Highly Sensitive Person Scale–Shortened Version.
The Highly Sensitive Person Scale–Shortened Version (HSPS–SV; Aron et al., 2010), a self-report measure of sensory processing sensitivity in adults, is a short, 11-item version of the original 27-item HSPS (Aron & Aron, 1997) used in the Meyer et al. (2005) study. These 11 items have been shown to be highly correlated with the overall scale, with internal consistency similar to that of the full version (Aron et al., 2010). Items are scored on an 8-point Likert-type scale ranging from 0 (not at all) to 7 (extremely). The α coefficient in the current study was .88.
Depression Anxiety Stress Scales 21.
The self-report Depression Anxiety Stress Scales 21 (DASS–21; Lovibond & Lovibond, 1995a, 1995b), an abbreviated 21-item version of the 42-item DASS, was used to measure distress. It comprises three 7-item scales: Depression, Anxiety, and Stress. Each item is scored on a 4-point scale ranging from 0 (did not apply to me at all) to 3 (applied to me very much or most of the time) on the basis of the previous week. Adequate internal consistency and validity have been reported across clinical and nonclinical samples (Henry & Crawford, 2005). Reliability scores in the current study were .82 (Depression), .59 (Anxiety), and .85 (Stress). Because of the relatively low α for anxiety, use of this variable was restricted to correlational analyses, and results pertaining to anxiety should be interpreted with caution.
Cold Pressor Pain-Inducement Task.
The cold pressor is a methodologically sound device that has frequently been used to induce temporary hand and forearm pain (e.g., Andrews et al., 2011; Meredith et al., 2006, 2015). The cold pressor consists of an insulated container filled with ice water maintained at a temperature of 0°–2°C. Participants are instructed to submerge their nondominant hand and forearm into the water for as long as possible. A 4-min time limit (of which participants are not aware) is imposed to minimize risk. At the conclusion of the cold pressor task, participants are debriefed. The entire procedure takes approximately 30 min. This task was included to activate the attachment system.
Procedure
The study protocol was approved by The University of Queensland’s institutional review board. Participants were sought through a word-of-mouth snowball sampling technique, beginning with friends and family of the research team. Testing was undertaken by two members of the research team (authors Bailey and Rappel). Involvement in the study was voluntary, and no incentives were offered. Potential participants were given an information sheet and provided written consent before completing the questionnaires. The cold pressor pain-inducement task was then conducted in a private, mutually convenient space following standard instructions (see Meredith et al., 2006). Although specific testing locations varied across researchers and participants, the cold pressor apparatus was the same for all participants.
Data Analysis
Statistical analyses were undertaken using IBM SPSS Statistics for Windows (Version 21; IBM Corporation, Armonk, NY). Initial data checking revealed six skewed variables: age, attachment anxiety, attachment avoidance, depression, anxiety, and stress. After screening of the demographic variables for multicollinearity, age and participant type were retained as control variables in further analyses. Attachment anxiety and attachment avoidance were transformed using the square root function for use in parametric tests. Untransformed variables were retained in nonparametric tests. As a result of high correlations among the DASS–21 subscales, only the Stress subscale was retained as a covariate in the regression analyses because it had high internal consistency and was most theoretically plausible.
A series of multiple linear regression analyses was then undertaken to examine the relationships between attachment and sensory variables when controlling for age, participant type, and stress. Assumptions for regression analyses were verified through formal (e.g., Shapiro–Wilk W test for normality of residuals) and informal (e.g., graphic) checks.
Results
Correlations Among Variables
An examination of the correlations between attachment, sensory processing, distress, and the demographic variables was conducted using Spearman rank-order correlation analysis. Although attachment anxiety was correlated with all distress variables, it was linked only with sensory sensitivity and the HSPS–SV (Table 3). Attachment avoidance was correlated only with sensory sensitivity.
Intercorrelations Between Variables Using Spearman’s ρ
Note. N = 116. Distress is represented by the depression, anxiety, and stress variables. HSPS–SV = Highly Sensitive Persons Scale–Shortened Version.
p < .05. **p < .01. ***p < .001.
Associations Between Attachment and Sensory Processing
Four multiple linear regression analyses were conducted, one for each pair of attachment (anxious–avoidant) and sensory (sensory sensitivity–HSPS–SV) variables. The sensory sensitivity variables were entered as independent variables, along with age, participant type, and stress (Table 4). Analyses were not conducted for the remaining three sensory processing variables (low registration, sensation seeking, and sensory avoidance) because they were not correlated with either attachment variable.
Results of Regression Analyses for Each Attachment Dimension
Note. HSPS–SV = Highly Sensitive Persons Scale–Shortened Version.
p < .05. **p < .01.
In the first step of the regression models analyzing attachment anxiety and sensory variables, attachment anxiety was significantly associated with both AASP sensory sensitivity, F(1, 112) = 5.35, p = .02, and the HSPS–SV, F(1, 112) = 6.00, p = .02. When stress and control variables were included in these analyses, stress was the only significant contributor to the model. In the regression models analyzing attachment avoidance and sensory variables, attachment avoidance was significantly associated with only sensory sensitivity, F(1, 112) = 7.75, p = .006. In analyses including stress and control variables, sensory sensitivity was retained as the only significant contributor to this model.
Discussion
This study offers valuable insights into the relationship between adult attachment and sensory processing variables in healthy adults. It is the first investigation of these variables that sought to activate the attachment system before testing and controlled for participant stress and relationship with the researcher. It is also the first to incorporate two different measures of sensory sensitivity.
Consistent with previous findings and with Hypothesis 1, attachment anxiety was related to sensory sensitivity as measured by both the Sensory Profile (Jerome & Liss, 2005; Levit-Binnun et al., 2014) and the HSPS–SV (Meyer et al., 2005). These associations were not retained when controlling for stress, however, which suggests that the association between attachment anxiety and sensory sensitivity was largely accounted for by stress. This result highlights the importance of controlling for distress in future studies in this field.
Contrary to expectations, attachment avoidance was unrelated to sensory avoidance (Hypothesis 2) and was linked instead with sensory sensitivity. This relationship was retained when controlling for stress. Sensory sensitivity is similar to sensory avoidance in that they both represent low thresholds for sensory stimuli. According to Dunn (1997), people with a low sensory threshold respond either passively (sensory sensitivity) or actively (sensory avoidance). Thus, although participants who reported high levels of attachment avoidance in all studies reported lowered sensory thresholds, their preferred coping approach differed.
In contrast to the findings of both Jerome and Liss (2005) and Levit-Binnun et al. (2014), attachment anxiety was found to be unrelated to low registration. This lack of a relationship, although it did not support Hypothesis 3, was in line with Jerome and Liss’s theoretical expectations. This inconsistency warrants further empirical attention.
Of the sensory processing variables, sensation seeking was the only one that was unrelated to any of the distress or attachment variables, providing support for Hypothesis 4. This finding was consistent with our theoretical expectations because sensation seeking indicates both a high sensory threshold and an active coping approach. Indeed, Engel-Yeger and Dunn (2011b) found that sensation seeking was correlated with positive affect. Although an association between sensation seeking and secure attachment was suggested, further validation using an attachment measure that specifically includes a secure variable is required.
The inclusion of distress variables (anxiety, depression, and stress) in this research afforded further insight into the relationships between sensory processing and attachment. As expected, attachment anxiety was significantly associated with all three distress variables. In contrast, although attachment avoidance was highly correlated with attachment anxiety, it was unrelated to any of the distress variables. Kobak, Cole, Ferenz-Gillies, Fleming, and Gamble (1993) suggested that people with an avoidant attachment pattern adopt a deactivating coping strategy that serves to deny attachment needs and to deny or minimize the emotions and cognitions associated with these needs. Thus, the lack of a significant association between attachment avoidance and distress suggests that the deactivating coping strategies may have been effective in this sample of healthy adults.
Also noteworthy was the lack of significant associations between sensory avoidance and both depression and stress. Although inconsistent with Engel-Yeger and Dunn (2011a), who found a link between sensory avoidance and negative affect, this finding is theoretically plausible: Although people who are sensory avoidant have a low sensory threshold, they adopt active coping strategies that appear to provide some protection from distress. Despite this explanation, a significant result was obtained between sensory avoidance and anxiety. Although this result was tentative, given the low level of internal consistency obtained for anxiety, it does support the link between sensory avoidance and trait anxiety reported by Engel-Yeger and Dunn (2011b).
Consistent with expectations, low registration and sensory sensitivity (measured using both the Sensory Profile and the HSPS–SV) were all positively linked with distress. Given that these two sensory variables indicate a passive coping approach, these findings are consistent with the coping literature, which has suggested that passive coping is less adaptive than active coping (Snow-Turek, Norris, & Tan, 1996).
Limitations and Future Research Directions
The results of this study should be interpreted cautiously. First, the cross-sectional nature of the study did not permit causal conclusions, so the direction of the associations remains unclear. Both insecure attachment (Jinyao et al., 2012) and sensory processing patterns (Engel-Yeger & Dunn, 2011b) have previously been conceptualized as vulnerability factors for the development of anxious symptomatology. In contrast, Liss et al. (2005) suggested that anxiety may be a risk factor for developing extreme sensory processing patterns, and anxiety may also increase the salience of insecure attachment (Ravitz et al., 2010). The complex interrelationships among these variables warrant further longitudinal investigation to inform clinical approaches.
Second, the use of a convenience sampling method in the current study may have introduced sampling bias. To minimize this bias, two researchers gathered data, both remained blind to participants’ attachment and sensory processing classifications, and whether participants were known to the researcher was retained as a control variable. Third, the need to use transformed dependent variables complicates interpretation of results.
Reliance on self-report measures is also a limitation of this study because these measures are known to be influenced by error and method variance factors including social desirability, memory, and mood. Consequently, future studies should include objective measures and measures of social desirability.
Replication of this study is recommended with a larger, more representative sample and a longitudinal study design. Using an attachment measure that conceptualizes attachment as a four-category (secure, fearful, preoccupied, dismissing) or three-dimensional (secure, anxious, avoidant) construct will also improve clarity regarding the associations between sensory variables and attachment security.
Implications for Occupational Therapy Practice
Although our conclusions are tentative, the results of this study lend support for the association between insecure attachment and sensory sensitivity and highlight the role of distress in this relationship. These associations have some implications for occupational therapy practice:
A person who is insecurely attached, sensory sensitive, and distressed may feel more compromised in therapy than others who do not have these characteristics.
Awareness of these interrelationships alerts practitioners to possible relationship-based or sensory informed therapeutic approaches that may be acceptable to this potentially vulnerable group.
It is possible that working with one domain (sensory/relationship) may improve outcomes in all three interrelated domains (sensory, attachment, and distress). For example, for someone who has an avoidant attachment pattern, commencing treatment with sensory-informed approaches may prove less threatening than providing emotional support.
Further research is needed to better understand the complex ways in which attachment insecurity and sensory sensitivities coexist to inform early intervention and prevention practices for a range of clinical populations across the lifespan.
Conclusion
The findings of this study offer support for associations between insecure attachment and high scores on measures of sensory sensitivity in healthy adults. The findings further suggest that, for insecurely attached people and people with passive coping strategies, treatment providing education on these factors and instruction in active sensory-related coping approaches may minimize distress. Additional research is needed to validate and extend these findings to offer insight into how best to integrate these concepts to achieve better informed and more client-centered clinical practice.
Supplemental Material
Supplementary material for Adult Attachment, Sensory Processing, and Distress in Healthy Adults
Supplementary material, sj-pdf-1-aot-10.5014_ajot.2016.017376.pdf for Adult Attachment, Sensory Processing, and Distress in Healthy Adults by Pamela J. Meredith, Kirsty J. Bailey, Jenny Strong and Georgia Rappel in The American Journal of Occupational Therapy
Footnotes
Acknowledgments
The authors acknowledge the statistical advice of Asad Khan. This research was approved by The University of Queensland’s Behavioral and Social Science Ethical Review Committee (No. 2012000013).
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
