Sense of self in first-time pregnancy

The study examines how pregnancy, a transformative physiological and psychosocial event, relates to changes in the cognitive self-model. Grounded in predictive processing accounts, the authors conceptualize the self-model as a dynamic representation formed by integrating prior expectations and sensory inputs. They argue that pregnancy alters sources of self-related information (e.g., internal bodily signals, new social role of motherhood), potentially changing self-concept clarity, sense of agency, self-efficacy, body representation, and self-related learning. The research question asks whether, and in what way, these self-related constructs differ between first-time pregnant (primigravida) and never-pregnant (nulligravida) women, with the overarching hypothesis that pregnancy involves heightened self-learning and consequent shifts in multiple components of the self-model.

The paper reviews predictive processing frameworks describing self-model formation via minimizing prediction errors across cognitive hierarchies. It summarizes constructs relevant to self in pregnancy: (1) self-concept clarity—structure, consistency, stability of self-beliefs, with prior links to behavioral prediction and bodily illusions; (2) sense of agency—implicit (intentional binding) and explicit judgments, with past evidence that self-associated stimuli can enhance binding; (3) general self-efficacy—domain-general beliefs about one’s capabilities, previously studied in pregnancy largely within maternal or health-specific domains; (4) body representation—multisensory, action-oriented models of the body, with pregnancy-related changes including expansion of peripersonal space and complex body image experiences captured by subscales of body agency, estrangement, and visibility; and (5) self-prioritization in learning and perception—robust self-reference effects in memory and speeded processing for self-associated stimuli, with early postpartum studies suggesting shifts toward infant-bias. The authors note the scarcity of cognitive studies on maternal self-model transformation and call for examination across pregnancy stages.

Design: Preregistered, cross-sectional online study comparing primigravida and nulligravida women on questionnaires and two cognitive tasks, analyzed primarily with Bayesian statistics. Ethics and preregistration: Approved by Monash University Human Research Ethics Committee (Project 32109). Preregistered on OSF (https://osf.io/vua52). Data and preprocessing scripts available (https://osf.io/9mhbg/). Participants and recruitment: Data collected June 3–November 30, 2022. Recruited via Prolific (n=262 completed) and word of mouth (n=11). After quality control and extended data collection per preregistration, final analyzed sample: primigravida n=100; nulligravida n=102 (including 3 per group via word of mouth; 97 primigravida and 99 nulligravida via Prolific). Eligibility: assigned female at birth; fluent in English; normal/corrected vision; no previous pregnancy; no biological or non-biological children; no neurological damage/disorder; age ≥18; not menopausal or post-menopausal. Procedure and measures: Participants completed demographics, then questionnaires (random order), followed by two cognitive tasks hosted on Pavlovia (PsychoJS 2021.2.3) with enforced fullscreen. Questionnaires:

• Self-Concept Clarity: summed score; higher indicates greater clarity.
• Sense of Agency Scale: positive and negative subscales; higher positive scores = more frequent feelings of control; higher negative scores = more frequent feelings of being out of control.
• General Self-Efficacy: summed score; higher indicates stronger generalized efficacy beliefs.
• Body Experiences During Pregnancy (adapted for both groups): mean scores for three subscales: Body Agency (attractiveness, pride, competence), Body Estrangement (reduced control/ownership), Body Visibility (being stared at/touched/evaluated); higher scores indicate more of each experience. Cognitive tasks:
• Shape-label matching (self-bias in perceptual matching): Participants learned associations between geometric shapes and labels (self, relative, stranger). Outcomes: sensitivity (d′) for congruent matching; reaction time (RT) for congruent vs incongruent trials; self-bias indices computed.
• Intentional binding (implicit agency): Interval reproduction between action (key press) or visual onset and a subsequent shape; computed intentional binding as difference between visual and action intervals (greater values = more binding). Also included judgement of agency probes (Likert) on half of trials and identity memory probes on the remainder. Statistical analysis: Conducted in R 4.2.2 and JASP 0.16.4 with default Bayesian priors. Survey measures standardized as z-scores. Tests included Bayesian independent-samples t-tests for group differences, repeated-measures ANOVAs for within-subject factors (e.g., Identity, Congruency, Condition, Delay) and between-subject Group. Exploratory analyses examined trimester effects by coding nulligravida as 0 and primigravida by trimester (1–3). Bayes factors interpreted per Jeffreys’ scale; evidence for null (BF01) also reported. Correlation matrix reported in Supplement. Exploratory sensitivity analyses excluding participants with minimal agency judgements (n=10 per group) were conducted.

• Self-concept clarity: Substantial evidence for no difference between groups; strong evidence against trimester effects.
• Sense of agency: Positive agency—evidence for no group or trimester differences. Negative agency—weak evidence for higher negative agency in pregnant group overall; decisive trimester effect with markedly higher negative agency in first trimester vs nulligravida and vs third trimester, and substantially higher than second trimester. Excluding participants with minimal agency judgements attenuated the group difference, suggesting alignment between explicit negative agency and task-based agency judgements in some participants.
• General self-efficacy: Evidence largely inconclusive overall; exploratory analyses indicated increased self-efficacy in second trimester vs nulligravida (BF10≈6.76), warranting further data.
• Body experiences (adapted scale): All three subscales higher in primigravida vs nulligravida—Body Agency (BF10≈6.8), Body Estrangement (BF10≈5056.8), Body Visibility (BF10≈57153.6). By trimester: Body Agency increased primarily in second trimester vs nulligravida; Body Estrangement and Body Visibility elevated across all trimesters with limited evidence for trimester-related change.
• Shape-label matching: Robust self-bias replicated in both groups (higher d′ and faster RT for self-associated shapes). Group effect: primigravida exhibited lower d′ accuracy than nulligravida across identities (no interaction), with deficits present in first and third trimesters; second trimester performance comparable to nulligravida. RT effects (Identity and Congruency, and their interaction) replicated canonical patterns; no group or trimester effects on RT; no between-group differences in self-bias indices for d′ or RT.
• Intentional binding: No classic binding effect detected (likely underpowered for this measure); no effects of Identity or Group on reproduction intervals. No group differences in binding or self-bias; participants nevertheless reported higher explicit agency in action vs visual conditions as expected. Memory probes showed a slight relative advantage for the ‘relative’ identity vs self; no group differences. Overall: Pregnancy is associated with higher body estrangement and visibility, elevated body agency (notably in second trimester), increased feelings of negative agency especially in the first trimester, and reduced accuracy for newly learned associations (first and third trimesters), while preserving typical self-prioritization effects and self-concept clarity.

Findings partially support the hypothesis that pregnancy alters the self-model. Despite no change in self-concept clarity, pregnant participants reported higher body estrangement and visibility and, overall, higher body agency, highlighting complex bodily and social experiences unique to pregnancy. The elevated negative agency, particularly in the first trimester, aligns with early pregnancy’s rapid physiological changes, perceived risks, secrecy, increased clinical monitoring, and reduced perceived control. Second trimester patterns (increased body agency and suggestive increase in self-efficacy) indicate a phase of consolidation and adaptation, whereas third trimester shows persistence of pregnancy-specific body experiences and a return of learning deficits. Task results show preserved self-prioritization across groups but reduced accuracy (d′) in pregnant participants, especially in first and third trimesters, suggesting trimester-related shifts in general learning processes rather than self-specific processing. The absence of an intentional binding effect, alongside intact explicit agency judgements, suggests methodological underpowering of the binding paradigm rather than a true lack of implicit agency differences. Together, the results indicate dynamic, trimester-dependent changes to components of the self-model during pregnancy: disruption and reduced perceived control early, consolidation mid-pregnancy, and a mixed profile late. These dynamics may inform understanding of perinatal cognitive and affective changes and their links to wellbeing.

The study contributes evidence that first-time pregnancy is associated with significant, nuanced changes in self-related constructs: increased body estrangement and visibility across pregnancy, heightened body agency (particularly second trimester), increased negative agency in early pregnancy, and reduced accuracy in learning newly formed associations (first and third trimesters), while maintaining typical self-prioritization and stable self-concept clarity. Exploratory trimester analyses suggest a dynamic trajectory—early disruption, mid-pregnancy consolidation, and late pregnancy reemergence of certain deficits. Future research should: (1) employ longitudinal designs to track within-person changes across trimesters and into the postpartum; (2) increase trials or use alternative implicit agency paradigms to robustly assess intentional binding; (3) compare pregnant cohorts with other relevant groups (e.g., eating disorders, schizophrenia, postpartum) to contextualize body experience constructs; (4) include diverse gender identities and compare primigravida with multigravida/multiparous cohorts; and (5) further investigate general self-efficacy trajectories and their clinical relevance to perinatal mental health.

• Cross-sectional design limits causal inference and within-person trajectory modeling; trimester effects are exploratory and require longitudinal confirmation.
• Intentional binding task likely underpowered (insufficient trials per condition) for detecting implicit agency differences, contributing to null binding effects.
• Adaptation of the Body Experiences During Pregnancy Scale for non-pregnant controls may limit direct comparability of certain constructs (e.g., estrangement, visibility) outside pregnancy.
• Sample consisted almost entirely of cisgender women; findings may not generalize to transgender men or non-binary gestational parents.
• Exclusion of multiparous cohorts prevents determination of whether effects are specific to first-time pregnancy or general to pregnancy.