One of the leading frameworks for studying close relationships in adolescence and adulthood is attachment theory, a theory originally developed by Bowlby and Ainsworth to explain human infants’ emotional attachment to their caregivers (Ainsworth, Blehar, Waters, & Wall, 1978; Bowlby, 1982; see Fraley & Shaver, 2000, and Mikulincer & Shaver, 2003, for an overview of the theory as applied to adult relationships). According to Bowlby (1982), humans are born with an innate behavioral system, the attachment behavioral system, which ensures that people of all ages, but especially young children, will seek safety, protection, and support from selected other people (“attachment figures”) when threatened, injured, or ill. Ainsworth and her colleagues (1978) identified systematic patterns of attachment in infancy that seemed to result from certain kinds of parenting.
[...]
Here, we are particularly interested in the relative absence of certain positive behaviors and qualities in avoidant individuals’ close relationships. For example, avoidant individuals report experiencing fewer positive emotions (Searle & Meara, 1999) and rate everyday social interactions as boring and unengaging (Tidwell, Reis, & Shaver, 1996). They do not approach situations where self-disclosure is (for most people) appropriate, normative, and rewarding, nor do they approve of others’ self-disclosures (Anders & Tucker, 2000; Mikulincer & Nachshon, 1991). In stressful situations, avoidant individuals exhibit fewer caregiving behaviors and offer less support to their romantic partners (Feeney & Collins, 2001; Fraley & Shaver, 1998). During the Gulf War, avoidantly attached Israelis who lived in the most dangerous areas were more likely than non-avoidant individuals to use distancing, self-reliant coping strategies rather than seeking social support and comfort from close relationship partners (Mikulincer, Florian, & Weller, 1993).
Avoidant individuals do not show increases in creative thinking following a positive mood induction, as is common among secure individuals (i.e., Isen & Daubman, 1984; Mikulincer & Sheffi, 2000). Moreover, when primed (i.e., shown a stimulus so briefly that its appearance is not consciously perceived) with an attachment-related threat word (e.g., “separation”), avoidant individuals inhibited activation of their attachment figures’ names. For non-avoidant individuals, the pattern of results was exactly the opposite: Responses to the names of attachment figures were facilitated by an attachment-related threat word (Mikulincer, Gillath, & Shaver, 2002). These findings indicate that while non-avoidant individuals automatically activate mental representations of their attachment figures in times of threat, avoidant individuals, even on a pre-conscious level, steer clear of this kind of mental representation, at least when the issue of separation has been raised.
These findings illustrate two key features of the emotional and behavioral lives of avoidant individuals: 1) In their close relationships, they experience both less positivity and more negativity, and 2) they exhibit both fewer approach behaviors and more withdrawal behaviors. Valenced emotional dispositions and patterns of approach/withdrawal behavior are associated not only with patterns of adult attachment. In recent years, neuroscientists have begun to explore how the two cerebral hemispheres differentially process emotion-related information (see Davidson, Jackson, & Kalin, 2000, for a review). Researchers have not come to consensus on the best way to characterize these hemispheric asymmetries (Cacioppo & Gardner, 1999; Davidson & Irwin, 1999), but the gist of the findings is that emotions related to or caused by appetitive or approach motives are associated with greater activation in the left prefrontal cortex, whereas emotions related to or caused by avoidance or withdrawal motives are associated with greater activation in the right prefrontal cortex. We do not propose to resolve disagreements within that research area. Rather, we wish to use the general distinction, which is well documented, to investigate how individual differences in avoidant attachment relate to hemispheric differences in the processing of attachment-related information.
[...]
Given that negative emotions and withdrawal tendencies are associated with the right frontal cortex, it seems likely that avoidant attachment is a reflection of particular kinds of processing in that hemisphere of the brain. Avoidant individuals should be either more adept at processing negative attachment-related information or less adept at processing positive attachment-related information in the right hemisphere. The latter prediction is compatible with the extensive evidence reviewed above indicating that avoidant individuals fail to react positively to various kinds of positive affect inductions, fail to become engaged in social interactions, and fail to seek support from others when under stress.
Testing Asymmetry in the Brain
In order to explore attachment and hemispheric differences in emotion processing, we can take advantage of the fact that stimuli presented to one visual hemifield are first processed by the opposite hemisphere. For example, an image presented to the right side of a person’s visual field will first enter the primary visual area in the left hemisphere. By briefly presenting a stimulus on one side (e.g., the left side) of a fixation point, we can ensure that the stimulus is processed first by the contralateral, or opposite, hemisphere (in this case, the right hemisphere). Experimentally, this phenomenon can be used in a divided visual field task.
Divided visual field tasks are commonly used in the cognitive sciences to investigate how quickly and accurately different hemispheres process information. Typically, the task involves brief presentations of letter strings or images in either visual field. A participant makes a decision about that stimulus and responds using a keyboard or response box. This task is useful for studying how different kinds of emotional and attachment-related information are processed. Words that convey certain meanings and connotations can be presented to a particular visual field to be processed first by the contralateral hemisphere. In the study reported here, we wanted participants to be thinking about attachment-related issues so that their attachment systems would be engaged and attachment-related material in memory would be primed. We therefore had them write about attachment-related issues for 5 minutes before performing the divided visual field task.
Study Overview and Hypotheses
Participants completed dispositional measures of attachment anxiety and attachment avoidance, which we expected to be related to hemispheric asymmetry, and were then primed to think about attachment issues in the following way. Half of them wrote about a secure attachment figure (someone they love very much, who is always there for them and is responsive to their needs) and half wrote about an attachment threat: abandonment (someone they love very much suddenly and without explanation leaving them for someone else). These priming manipulations were expected to make positive or negative attachment-related experiences temporarily more available in memory.
Our first hypothesis was that responses to emotionally valenced words would be lateralized in a pattern consistent with hemispheric asymmetry models; specifically, the right hemisphere would have a negative word advantage and the left hemisphere would have a positive word advantage. This difference was expected to appear in all experimental conditions. The second hypothesis was that the attachment prime would increase the availability of related verbal information, because thinking about a positive attachment situation (a loving, supportive attachment figure) would facilitate responses to related positive words, whereas thinking about a negative attachment situation (abandonment) would facilitate responses to related negative words. This effect was expected to occur independently of other experimental manipulations. The third hypothesis was that attachment-related avoidance would be associated with either a right hemisphere advantage for processing negative attachment-related words or a right hemisphere disadvantage for processing positive attachment-related words. If obtained, this finding would support the possibility that the psychological and behavioral correlates of avoidant attachment are related to differences in the way the two cerebral hemispheres process information.
[...]
Discussion
Our first hypothesis was supported. Consistent with affective-motivational models of hemispheric asymmetry, we found a left hemisphere advantage for categorization accuracy when processing positive emotion and attachment-related words, and a right hemisphere advantage for processing negative emotion and attachment-related words. These differences were larger in the right hemisphere. This is consistent with the fact that some behavioral studies of emotional lateralization find effects only for the right hemisphere (Atchley, Ilardi, & Enloe, 2003; Richards, French, & Dowd, 1995), while others find effects in both hemispheres (Bernat, Bunce, & Shevrin, 2001; Burton & Levy, 1991; Eviatar & Zaidel, 1991). This may reflect the fact that the right hemisphere is slower overall at language processing than the left hemisphere, and thus differences that exist equally in both hemispheres appear larger in the right hemisphere in tasks that involve language comprehension and decisions. It is also possible that the linguistic representation of emotion is not as strongly lateralized as other behavioral or response-related representations.
The second hypothesis, that writing about attachment issues would increase the availability of affectively related attachment information, was partially supported. When the participants were broken down into two groups—those who had responded with their left hand and those who had responded with their right hand—those who had responded with their right hand showed a pattern of results consistent with the hypotheses, whereas those who had responded with their left hand showed no differences. This may be due to a number of factors. First, the right hand is controlled by the left hemisphere, which is dominant for language production. Our prime was a language-intensive one (writing for approximately 5 minutes), so one would expect the left hemisphere to be more strongly primed than the right hemisphere, which has very limited language production capabilities.
Second, Koivisto (1998) has suggested that the right hemisphere is not sensitive to immediate priming effects, but shows priming effects later (i.e., post-lexical judgment). Our task demanded very rapid responses and thus may not be capable of detecting right hemispheric priming effects. Third, although semantic priming has been shown to occur in the right hemisphere (Chiarello et al., 1990; Chiarello & Richards, 1992; Long & Baynes, in press), these studies show that the right hemisphere is sensitive to priming for distantly related but not proximally related words (Chiarello & Richards, 1992; Chiarello et al., 2001). Since the words used in our study were closely related to positive and negative attachment issues, it is possible that the priming manipulation was a proximal one, and thus did not affect the right hemisphere (which controls left-handed responses).
The third hypothesis stated that dispositional attachment avoidance would be associated with either a right hemisphere advantage for processing negative attachment information or a right hemisphere disadvantage for processing positive words. The latter part of this hypothesis was supported: While there was no advantage for processing negative words in the right hemisphere, attachment avoidance was significantly associated with a disadvantage for processing positive words in the right hemisphere. Furthermore, attachment avoidance predicted lower scores on positive attachment words in the right hemisphere even after partialing out the effects of response hand, priming condition, negative affect, and positive affect.
We will consider two possible reasons for this effect. First, it is important to remember that although the visual information was presented to only one hemisphere, both hemispheres participated in the response decisions because information travels rapidly between the two hemispheres via the corpus callosum. Perhaps for individuals who score higher on attachment avoidance, positive attachment-related information does not transfer across the hemispheres as rapidly as other kinds of information, which reduces accuracy.
Second, individuals who score relatively high on measures of dispositional avoidance may differ from individuals who score low in the way they represent or make judgments about positive attachment information in the right hemisphere. Perhaps this kind of information is less well represented, or decision-making processes based on such information are more difficult when the information is presented to the right hemisphere. This interpretation is consistent with behavioral evidence that avoidance predicts less experience and expression of positive emotions (Searle & Meara, 1999; Tidwell et al., 1996). In the brain, this effect may originate from poorly represented positive attachment information networks in the right hemisphere, the hemisphere that is overall less efficient at processing positive emotional information.
[...]
In general, our findings suggest differences in the processing of attachment-related information in the brains of individuals who are high on attachment avoidance compared to those who are low. These differences may be important in understanding the nature of avoidance in close relationships.
