The mediating role of maternal responsiveness in some longer term effects of postnatal depression on infant development

https://doi.org/10.1016/j.infbeh.2004.03.003Get rights and content

Abstract

We hypothesized a mediating role for low maternal responsiveness in certain child developmental deficits found previously to be associated with postnatal depression. Forty depressed inpatients and 48 control (non-depressed) mothers and their infants were followed until 42 months postpartum. Mother–infant interactions were impaired at 6 months postpartum in the depressed group. Subsequent cognitive deficits were found in children's Full Scale IQ on the Wechsler Preschool Primary Scale of Intelligence (Revised) at 42 months of age. Statistical modeling of the mediation hypothesis confirmed that these effects were overwhelmingly explained by lowered maternal responsiveness at 6 months. Similarly, increased temperamental difficulties were apparent in children of depressed mothers but were not correlated with maternal responsiveness. Male infants of depressed mothers were disproportionately vulnerable (compared to females) to impaired cognitive abilities associated with maternal depression. The data concur with previous work and provide empirical support for the theoretical prediction that early disturbances of the mother–infant interaction will mediate some developmental deficits in the children of depressed mothers.

Introduction

Postnatal depression (PND) affects around 10% of all mothers. During PND, women suffer a cluster of depressive symptoms over extended periods (weeks, months or sometimes years: O’Hara & Zekoski, 1988).

The detrimental effects of PND on child development have been detailed in several small longitudinal studies (see Cooper & Murray, 1997; Milgrom, Martin, & Negri, 1999; Murray & Cooper, 1997). Meta-analyses of the available published data have found small but significant effects of PND on both emotional and cognitive child outcomes at 1 year of age (e.g. Beck, 1998). For example, Lyons-Ruth, Zoll, Connell, and Grunebaum (1986) reported lower scores on standardized scales of mental and motor development in infants of depressed mothers. Other researchers have recorded behavioural and social adjustment difficulties up to the fifth year of life including children's behaviour towards their mothers, the presence of behavioural disturbance at home and the social patterning of play at school (Murray et al., 1999). The effects of PND may even persist into later childhood, such that 11-year-old schoolchildren whose mothers were depressed 3 months postpartum have been found to have lowered IQs, more attentional problems and greater difficulty with mathematical reasoning (Hay et al., 2001).

How might these effects be mediated? One strong possibility is via impairment of the early mother–infant interaction (Murray, Cooper, Wilson, & Romaniuk, 2003). Depressed mothers speak less, are less responsive, present with a ‘flat’ affect and express more negative emotions (Frankel & Harman, 1996; Righetti-Veltema, Conne-Perreard, Bousquet, & Manzano, 2002). Exposure to non-optimal maternal interactive behaviour may interfere not only with information processing abilities but also with the capacity to develop behavioural self-control (Tronick & Weinberg, 1997).

Recently, Cicchetti, Rogosch, and Toth (2000) reported the efficacy of a preventative psychotherapeutic intervention aimed specifically at improving the early mother–child relationship in terms of positive interaction and maternal affect. By improving these aspects of the interaction, the negative impact of PND on infant IQ was abolished. This result is consistent with the idea that impaired mother–infant interaction mediates PND's effects on cognitive development. However, a non-causal correlation between impaired mother–infant interaction and impaired child development could arise if both are (independently) correlated with PND.

Furthermore, research has consistently found a strong association between maternal depression and infant temperament (e.g. McMahon, Barnett, Kowalenko, Tennant, & Don, 2001). It has been reported that aspects of neonatal temperament can influence maternal mood and maternal interaction style (e.g. Murray, Stanley, Hooper, King, & Fiori-Cowley, 1996; see meta-analysis by Beck, 1996). Conversely, others have found that maternal depression, child-rearing practices and early life events can themselves affect aspects of child temperament (e.g. de Vries & Sameroff, 1984; Sugawara, Kitamura, Toda, & Shima, 1999).

In this paper we present an empirical test of the hypothesis that maternal responsiveness in early infancy is the primary mediator between maternal depression and specific infant developmental outcomes, namely cognitive functioning and temperament at 42 months of age. This explicit mediational model is summarized by the path diagram in Fig. 1, where a and b represent the path coefficients of interest. For simplicity, all other direct, additive and multiplicative effects of depression on IQ are here subsumed in the path coefficient labeled c.

Baron and Kenny (1986) have detailed the methodological requirements for identifying mediator variables in psychological data. The logic of the method relies on estimating the relationships among a series of regression coefficients. First, the potential mediator must be shown to depend on the underlying causal factor of interest (in this case, PND). Next, the causal factor must successfully predict the outcome variable (child developmental measures). Last, the variance in the outcome variable ascribable to the causal factor must be significantly reduced by inclusion of the mediating variable in the statistical model. This approach has been used successfully to test models of father–child attachment (Lundy, 2002) and has been systematized in a formal statistical test by Sobel (1982).

Section snippets

Participants

Forty, depressed, recently delivered mothers were recruited while being treated as inpatients at mother–baby psychiatric units in Victoria, Australia. A control group of 48 well mothers was recruited from community health centres in metropolitan Melbourne. Recruits came from a broad socio-geographic range. All participants were volunteers competent to give signed consent, which was obtained at the time of enrolment.

Infants were aged on average 15.8 weeks old (S.D. = 7.1) at the time of

Participants

Demographic data from the two groups displayed very close similarity. The mean age of mothers was depressed = 30.4 years (range 19–39); non-depressed = 31.1 years (range 20–38). Depressed and non-depressed mothers had virtually identical years of schooling (12.3 and 12.7, respectively). Infant sex ratios in both groups were very close to 1:1. Forty percent of depressed mothers and 35% of non-depressed mothers had other children. No significant differences existed between the two groups on any

Discussion

As outlined in Section 1, there is strong evidence associating maternal mood disorder with problems in mother–infant interactions such as affective discordance, abrupt breaks in engagement and poor sensitivity to infants.

This study focused on testing an explicit mediation hypothesis linking the role of PND-impaired mother–infant interaction to negative child outcomes. Our results concur with this model and suggest that decreased levels of maternal responsiveness to and general stimulation of

Acknowledgement

This research was supported by a grant from the Victorian Health Promotion Foundation.

References (32)

  • P.J. Cooper et al.

    The impact of psychological treatments of postpartum depression on maternal mood and infant development

  • M.W. de Vries et al.

    Culture and temperament: Influences on infant development in three East African societies

    American Journal of Orthopsychiatry

    (1984)
  • W. Fullard et al.

    Assessing temperament in one to three year old children

    Journal of Pediatric Psychology

    (1984)
  • Hamilton, M. (1976). Hamilton Depression Scale. In W. Guy (Ed.), ECDEU assessment manual for psychopharmacology...
  • P.L. Harrison et al.

    AGS early screening profiles

    Journal of Psychoeducational Assessment

    (1990)
  • D.F. Hay et al.

    Intellectual problems shown by 11-year-old children whose mothers had postnatal depression

    Journal of Child Psychology and Psychiatry

    (2001)
  • Cited by (163)

    • Cognitive behavioral therapy for perinatal depression—Tailored face-to-face treatment and translation for online delivery of MumMoodBooster

      2023, Handbook of Lifespan Cognitive Behavioral Therapy: Childhood, Adolescence, Pregnancy, Adulthood, and Aging
    • Postpartum depression and ADHD in the offspring: Systematic review and meta-analysis

      2022, Journal of Affective Disorders
      Citation Excerpt :

      A systematic review in the first 3 months postpartum indicated a point prevalence of depression of 13 % (Howard et al., 2014). The research underlines that women with PPD, when interacting with their children, show less sensitivity and responsiveness (Milgrom et al., 2004) which indicates higher proportions of negative emotional expression (Murray et al., 1993). The health, social, emotional, cognitive and physical development of children are affected by the long-term PPD of mothers (Field, 2010).

    View all citing articles on Scopus
    View full text