Attachment Across the Life Span

As 14-month Laura ventures across the playground, she looks over her shoulder at her father. It's a playful look, as if to say, "Isn't this great!" All of a sudden, a five-year-old whizzes past and almost knocks her down. Her eyes dart back to her father as if to say, "Is this safe? Am I going to be okay?" Her father's calm gaze reassures her that, indeed, everything is all right. The quality of Laura's exploration of her world depends on the relationship she has developed with her caregivers. If she knows them to be of comfort to her when she's distressed and they serve as a secure base from which to venture, then she can safely and securely explore and master her world. She has developed what Bowlby called a secure attachment. While this drama is quite obvious during the infancy and toddler phases of development, a similar set of dynamics appear to be played out during adolescence.

In Bowlby's theory, attachment refers to a biologically primed behavioral system that, under threatening conditions, enables infants to seek safety through proximity to their caregivers. Conceptually, he juxtaposed this motivational system to seek security and safety (to preserve the species) with the exploratory system whose function it is to motivate the child to master his environment. Through early interactions with others, children develop internal working models of themselves, others, and the interactions that are hypothesized to guide subsequent relationships and interactions.

In the conference plenary titled, "Attachment Across the Life Span," the presenters addressed the development of the internal working model and the reciprocal relationship between attachment and exploratory motivational systems, with Nancy Cohen, Ph.D. presenting her work with infants and their families, and Marlene Moretti, Ph.D. presenting her work with adolescents and their families.

Dr. Cohen described a psychodynamic parent-child psychotherapy called Watch-Wait-Wonder (WWW) and reported on the results of a study comparing it with a different form of parent-infant psychodynamic psychotherapy. The aim of WWW is to help infants and mothers achieve a more secure attachment relationship, improve the interactions between mother and child, and improve the child's cognitive development and affect regulation.

Both clinicians and researchers hypothesize that four aspects of parental behavior will help an infant in developing a secure attachment relationship: (1) perceiving the infant's emotional signals; (2) responding to them sensitively; (3) displaying affection; and (4) accepting the infant's behavior and feelings. The target of WWW is both the parent's own internal working model and the observable interactions with the child. What distinguishes WWW from other forms of parent-infant psychotherapy is that it is infant-led. In the first half of each session, the parent gets down on the floor with the child and observes her self-initiated activity, interacting only at the child's initiative. In the second half of the session, the parent discusses his observations and experiences of the infant-led play with the therapist and, together, they examine their meanings from the perspective of the parent and the themes and relational issues that the infant is trying to master. This enables the parent to examine his internal working models in relation to the infant and vice versa.

To evaluate the effectiveness of the WWW therapeutic model, mother-infant dyads coming for treatment were randomly assigned to either the WWW intervention or another form of psychodynamic parent-infant psychotherapy (PPT) that was therapist-led, rather than infant-led. Both therapies were successful in a number of aspects of treatment. Both resulted in a decrease in the presenting problem and parenting stress, and an increase in parenting confidence. In both therapies, mothers reported a strong and positive therapeutic alliance. There were also some specific differences in the outcomes of the two therapies. Mothers in the WWW group reported a larger decrease in depression at the end of treatment than did mothers in the PPT. The infants in the WWW treatment group shifted their attachment security either from insecure to secure (21%) or from disorganized to organized (15%) more than did the infants in the PPT treatment group (3% and 9%, respectively). Furthermore, the infants in the WWW group exhibited a greater capacity to regulate their emotions and a concomitant increase in cognitive development than did the infants in the PPT group. The gains made by these infants were maintained over a six-month follow-up period.

Although the mother-infant dyads in the WWW group exhibited the predicted changes in their attachment relationships, the hypothesized pathway that changes in attachment security was predicted to take place — an increase in mother's sensitivity and responsiveness — was not confirmed. Mothers in both groups displayed less intrusiveness in interactions with their children and no differences in measures of reciprocity, unresponsiveness, and conflict.

While the changes in attachment security were modest, it is important to remember that changing attachment security in a clinical sample has been notoriously difficult. Furthermore, the interventions were made when a majority of the children were between 18 and 24 months — well after the attachment relationship had been established. If interventions were done during the first year of life, it is highly possible that the gains made could have been even greater.

What is it about the WWW therapy model that might be effective in making an impact on both attachment security as well as infant cognitive development and affect regulation? Dr. Cohen hypothesizes that the mother down on the floor, following her child's lead, may influence both the emotional relationship with the child as well as the child's exploratory and mastery behaviors. the reciprocal relationship between these two aspects of development may have created a synergy between them that had a greater impact than either one alone.

Imagine infants whose families did not participate in a therapy like WWW as adolescents. Dr. Moretti presented her work with adolescents like these — with severe behavior problems — with the goal of understanding their attachment issues and examining the predictive importance of self representation and attachment security in emotion and behavior regulation.

One major component of the internal working model is the person's view of self. To tap into adolescents' self-representations, they were asked to list the characteristics they believed described themselves as they are, the attributes they wish to be or feel they should be, the attributes that describe how their parents view them to be, and those they believe their parents desire of them. A number of interesting findings emerged from these data:

• In mid-adolescence (ages 14-15), adolescents and adults are farthest apart, sharing only 25% of the standards for identity development.
  
  
• In contrast, in late adolescence and early adulthood (ages 18-24), there is considerably more congruence (40%) between adolescent and adult standards of identity.
  
  
• High-risk conduct disordered adolescents' own and inferred of parents' representations of self are profoundly more negative than adolescents in non-clinical samples. Girls' self-representations are more negative than boys'.
  
  
• In adolescents with conduct disorder, there is less differentiation between own and inferred parental views of self than in non-clinical samples.
  
  
• The negativity of self representations is a robust predictor of relational aggression (i.e., acts harmful to relationships) in girls and overt aggression in both girls and boys.
  
  
• Girls who perceived themselves as different from what their mothers wished them to be and experienced low support from their mothers for the development of autonomy had low self esteem. Conversely, adolescent girls who also perceived themselves to be different from who they thought their mothers wished them to be but also received high support for their autonomy had high self esteem.

To examine the potential attachment dynamics in conduct disordered adolescents, they were classified into one of four attachment categories using Bartholomew's family attachment interview. The four categories were based on two underlying dimensions of the internal working model: models of self and models of other. Those in the secure category had a positive view of self and other. Self is viewed as competent and others are viewed as helpful. Those in the preoccupied category had a negative view of self and positive view of others. Self is viewed as unable to consistently evoke positive responses from others. Those in the fearful category had a negative view of self and negative view of others. They desire connection but fear rejection and are mistrustful. Those in the dismissing category view others as incompetent. They denigrate close relationships and maintain a stance of vulnerability and independence.

In two clinical samples of conduct disordered adolescents, less than 10% were classified as securely attached. Almost half were classified as fearful, and about a quarter each were classified as preoccupied or dismissing. A gender effect was found such that, across the two studies, boys were equally divided between the dismissing and fearful categories and a few in the preoccupied category. Girls were equally divided between the fearful and preoccupied categories with only a few in the dismissing group.

While the prevailing assumption about aggressive youth is that they would typically be described similarly to those in the dismissing category, the results were surprisingly different. It was the classification as preoccupied that predicted aggressive and delinquent behavior. Dismissing attachment predicted low anxiety, delinquency, and low aggression. Finally, while a negative self-representation as well as preoccupied attachment were both related to aggressive behavior, each was also a unique and significant predictor of aggressive behavior.

Dr. Moretti and her colleagues have concluded that aggressive and delinquent behaviors often (though not always) reflect difficulties in the process of identity individuation within an insecure parent-child relationship. Further, one hypothesized function of aggression is to engage caregivers who are perceived as reluctant, unavailable, or inconsistently engaged. As a result, these adolescents appear to be overly dependent on their parents because of the deficits in their feelings of self-worth and in effective emotion and behavior regulation.

These rich findings have numerous implications for interventions. First, adolescents do not need to detach from their parents to develop autonomy. In fact, the ability to develop one's identity may require healthy conflict and difference in the context of secure relationships. Second, typical programs for adolescents rely on strengthening parental authority and containment of adolescent behavior. A different model, currently being implemented by Dr. Moretti and colleagues at the Maples Adolescent Center, is described as an "attachment-based multisystemic approach," with the underlying guiding principle "moving from control to connection." A panoply of specific interventions is used in the program, but a hallmark of the approach is the assignment of consistent caregivers/staff to each adolescent in the residential program. It is hoped that the development of a close, empathic, and responsive relationship with these staff members will enable the adolescents to construct new internal working models of self and other.

The conceptual, empirical and clinical richness of both Dr. Cohen's and Dr. Moretti's work stimulated a lively and appreciative exchange of ideas. Three additional ideas emerged from this discussion. First, when attachment is discussed, the focus is typically on the mother and child relationship. Very few researchers have examined the father-child attachment relationship and the implications for the family-level unit of analysis. Second, attachment appears to be a critical aspect of healthy development, but certainly not the only one. While it may be necessary, it is not sufficient to explain the course of a child's early development. Finally, these concepts and empirical findings are inextricably bound to our culture and we must not assume that they apply to other parents, infants, and adolescents. As always, answering the important questions raised in the presenters' work has raised a wonderfully diverse array of additional questions to answer and issues to tackle.

Martha Edwards, Ph.D., is an AFTA board member and on the faculty and Director of the Early Prevention and Enrichment Project at the Ackerman Institute for the Family. At New York University Medical School's Child Study Center, she is also a member of the Unique Minds team and co-principal investigator on a project to develop family-focused treatment for children with bipolar disorder and their families.