Conference Agenda

A central question in attachment theory concerns how individual differences in attachment are organized and whether this variation is best conceptualized as dimensional or categorical. Prior research using the Strange Situation Procedure and the Adult Attachment Interview suggests that attachment variation can be captured by two dimensions–dismissing/avoidance and preoccupied/resistance–each of which is continuously distributed. However, it’s unclear whether this structure generalizes to attachment variation assessed via the Attachment Script Assessment (ASA). The ASA assesses attachment representations based on how repeated experiences of secure-base support (or lack thereof) are encoded as scripts in memory. The ASA is evaluated for secure base script knowledge (SBSK) (attachment problems/emotions acknowledged and effective support provided), deactivation (attachment problems/emotions dismissed), hyperactivation (attachment problems/emotions amplified and support is ineffective), and anomalous content (attachment problems evoke fear/disorientation and support exacerbates these states). The present study examined the latent structure of attachment variation across these four ASA scripts and evaluated whether this variation is best characterized as dimensional or categorical.

Participants were 986 adults drawn from multiple studies conducted in the Midwestern and Northeastern United States. Participants completed the ASA, a word-prompt procedure where participants generate 4 attachment-relevant narratives pertaining to parent-child and adult romantic relationships. These narratives were transcribed and continuously rated for SBSK, deactivation, hyperactivation, and anomalous content.

Exploratory factor analyses supported a two-factor solution (Table 1). The first factor included SBSK (reversed) and deactivation, reflecting the extent to which attachment-distress is minimized or ignored versus acknowledged and responded to. The second factor included hyperactivation and anomalous content, reflecting heightened negative affect, fear, and disorientation. Taxometric analyses indicated that the deactivating factor was dimensionally distributed, whereas evidence for the hyperactivating factor was ambiguous, though leaning dimensional. These findings extend dimensional models of attachment to the ASA and highlight potential heterogeneity in hyperactivating/disorganized attachment representations.

Although attachment representations have been linked to neural processing of infant faces, findings have not always replicated, draw on small samples, and lack control for depression (a correlated risk factor also implicated in neural responding to infants). Furthermore, prior work has focused on (in)secure attachment, leaving unknown whether specific insecure patterns are tied to neural signatures reflecting distinctive biases in infant cue processing. This study employed a novel Attachment Script Assessment (ASA) coding system to examine the significance of secure and insecure scripts for expectant women’s neural responding via ERPs (P1, N170, P3, LPP) to infant faces in an adequately-powered sample.

Expectant mothers (N=201; Mage=29 years) completed an infant emotion categorization task while electrophysiologically monitored, reported depressive symptoms (EPDS), and completed the ASA assessing secure (SBSK) and insecure (deactivation, hyperactivation, and anomalous content) attachment.

HLMs revealed that, controlling for depression, lower SBSK was associated with dampened neural responding reflective of diminished orientation toward (P100), attention to (P300), and sustained processing of (LPP) infant distress versus neutral expressions (b=0.23-0.54, p=.024-.040). Insecure and disorganized attachments were associated with different profiles of responding to infant signals (|b|=0.54-1.43, p=.009-0.035). Greater hyperactivation was associated with heightened saliency of infant cues as indicated by heightened P300 to infants. Greater deactivation was associated with dampened saliency of infant cues as indicated by dampened LPP to distress (v. neutral) faces and longer N170 latency to infants. Greater anomalous content was associated with dampened distinction of happy and neutral infant cues as indicated by comparable P300 amplitude and N170 latency to happy and neutral faces.

Findings demonstrate the validity of the novel ASA coding system and utility in distinguishing theory-consistent patterns of neural responding to infants. Findings indicate associations between attachment and neural responding to infants predate parenthood, suggesting pregnancy as an important window for early intervention.

Groh and Haydon (2024) proposed a set of scales designed to enrich our assessments of individuals’ attachment-related scripts during the Attachment Script Assessment. Groh and colleagues (2024) also provided initial evidence that these scales are uniquely associated with parenting behavior within a low-risk sample. The first goal of this study was to replicate these findings by evaluating whether adults’ deactivation, hyperactivation, and anomalous scripts were associated with their parenting behavior during interactions with their children. The second goal was to extend these findings by testing associations with parents’ self-reported mental health outcomes. The third goal was to examine whether these associations varied based on parents’ risk status.

Participants included 101 high-risk parents referred to Child Protective Services following allegations of child maltreatment and 82 low-risk parents recruited from the community. When children were approximately 8 years old, parents completed the Attachment Script Assessment, which were coded using the three scales developed by Groh and Haydon (2024). Parenting behavior was assessed during a non-distressing parent-child interaction, and parents reported their mental health problems using the Brief Symptom Inventory.

Results supported the utility of the attachment deactivation script scale. Among both groups of parents, higher deactivation ratings were significantly associated with lower parental sensitivity (r = -.42), lower positive regard (r = -.24), greater intrusiveness (r = .21), greater detachment (r = .26), and greater mental health symptoms (r = .17). In contrast, findings for the hyperactivation and anomalous script ratings were unexpected and/or subgroup-specific. Hyperactivation was not significantly associated with parenting behavior among low-risk parents and was counterintuitively associated with greater parental sensitivity (r = .26), greater positive regard (r = .27), and lower detachment (r = -.27) among high-risk parents. Anomalous content was significantly associated with greater overall mental health problems but only among low-risk parents (r = .23).

Attachment (in)security has critical implications for close relationship functioning across the lifespan (Bowlby, 1988). During adolescence, close friendships are central relationships for emotional support (Furman & Rose, 2015). The current study applied a coding scheme developed to assess insecurity in the adult Attachment Script Assessment (ASA) to the adolescent ASA. Attachment insecurity was examined in relation to adolescents’ disclosure and support behaviors with friends.

Participants were 162 adolescent best friends (81 dyads; 65% girls; M_age = 14.17 years; 91% White) who completed the adolescent ASA (Dykas et al., 2006). The transcribed interviews were coded for hyperactivation, deactivation, and anomalous content (ICCs = .82-.90). For each scale, separate averages were computed for mother and father stories.

Dyads completed the Problem Talk Task (Rose et al., 2014) in which they discussed problems for 16 minutes. The transcribed interactions were segmented into thought units (% agreement = 90.3%) and coded to identify problem disclosures (κs = 0.66-0.86). Responses to disclosures were coded as positive (e.g., validation) or negative (e.g., minimization) responses (κ = 0.78). The interactions were also coded for the extent that dyads engaged in co-rumination by assessing mutual encouragement, speculation, dwelling on negative affect, and rehashing (ICCs = 0.87-0.94). Participants reported on perceived support after the interaction.

Analyses indicated that for mother stories, hyperactivation was associated with conversational mechanisms that encourage problem discussion, including engaging in greater mutual encouragement (r=.20, p<.05), providing more positive support (r=.16, p<.05), and friends perceiving greater support (r=.20, p < .05). Deactivation was associated with conversational mechanisms that discourage problem discussion, including lower mutual encouragement (r=-.18, p<.05). For father stories, anomalous content was associated with friends perceiving lower support (r=-.20, p<.05) and deactivation was associated with adolescents perceiving lower support (r=-.17, p<.05).