Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Mar 11.
Published in final edited form as: Early Child Dev Care. 2018 Dec 5;190(12):1918–1930. doi: 10.1080/03004430.2018.1552947

The Effects of Early Postpartum Depression on Infant Temperament

Alyson F Shapiro a,*, Sandra N Jolley b, Ursula Hildebrandt c, Susan J Spieker d
PMCID: PMC7951142  NIHMSID: NIHMS1515462  PMID: 33716380

Abstract

Research linking postpartum depression (PPD) with negative child outcomes has predominantly examined PPD at six weeks postpartum or later, and has not controlled for depression during pregnancy. The present study examined associations between PPD at three weeks postpartum and temperament in 6-month-old infants in a sample of women who were not depressed during pregnancy. Depression was assessed at three weeks and six months postpartum using the Postpartum Depression Screening Scale (PDSS). Observed and maternal report of infant temperament was assessed when infants were 6-months-old. PPD symptoms significantly predicted observed temperament behaviour and marginally explained maternal report of infant temperament. Symptoms of PPD at three weeks postpartum were a stronger predictor than at six months. Findings suggest that early depressive symptoms may be particularly problematic, and have implications for early assessment and treatment of PPD even in women who were not depressed during pregnancy or are otherwise considered low risk.

Keywords: postpartum depression, maternal depression, infant temperament, infant emotion

Introduction

Postpartum depression (PPD) affects an estimated 12 to 40% of women, with higher prevalence rates being reported in larger sample studies (e.g. McCoy, Beal, Shipman, Payton, & Watson, 2006; Woody, Ferrari, Siskind, Whiteford, Harris, 2017). This is particularly problematic for child development because associations between negative child outcomes and PPD have been found in numerous studies across cultures (see Field, 2010; Kingston, Tough, & Whitefield, 2012). Research indicates that depressed mothers display less affectionate behaviour, are less responsive to infant cues, are more withdrawn, exhibit flatness of affect, and may exhibit intrusive or hostile behaviour toward their infants (e.g. Field, 1995; Field, 2010; Gauvain, 2001; Sohr-Preston & Scaramella, 2006). Infants of depressed mothers tend to have fussier, more avoidant temperaments, and make fewer positive facial expressions and vocalizations (see Field, 2010). The distressed, negative temperaments of infants of depressed mothers in turn are associated with the infants being more vulnerable to learning difficulties because the high levels of negative maternal affect increases infants’ arousal to levels that interfere with their early learning and cognitive development (Sohr-Preston & Scaramella, 2006).

The importance of examining early PPD.

The majority of research examining associations between PPD and child outcomes has measured PPD after two months post birth (see Field, 2010; Kingston, Tough, & Whitefield, 2012). Depression in the first few weeks postpartum also often goes undiagnosed because although the DSM-V criteria for having depression with perinatal onset indicates the beginning of symptoms either during pregnancy or within the first four weeks postpartum (American Psychiatric Association, 2013), the American Academy of Paediatrics recommends screening mothers for PPD at the first, second and fourth month baby visits (Earls, 2010). This is an important gap in the literature because: a) early parenting may have a unique long-term impact on infant development, and b) the earlier PPD is assessed, the less likely it is that infant temperament has had an impact on PPD rather than vice versa.

We were particularly interested in the potential influence of early PPD because the first year of life is a time of active brain growth and development (see Shonkoff & Phillips, 2000), and it is possible that early PPD may have an impact on temperament through influencing this early brain development. Seigel (2001) and Shore (2001) have drawn on research to create theory positing that early parent-infant interactions affect brain development, particularly right brain development and neural integration. There is also evidence that maternal PPD calibrates the hypothalamic-pituitary-adrenal (HPA) axis responses, with early PPD (3 months) having more of an impact than later PPD (Laurent, 2017). There is also substantial research indicating that temperament is associated with aspects of brain function and development, including: right frontal lobe asymmetry (Poole, Van Lieshout, & Schmidt, 2018), pre-frontal cortex functioning (Clauss et al., 2016), grey matter volume in various parts of the brain (e.g. Wei et al., 2018), and cortisol regulation (Poole, Jetha, & Schmidt, 2017).

The bulk of PPD research has also not assessed and accounted for prenatal depression (see Kingston et al., 2012), and the influence of PPD in the absence of depression during pregnancy is a severely understudied phenomenon. This is problematic because research has indicated that depression during pregnancy predicts later infant temperament, even after controlling for depressive symptoms post birth (e.g. Zhang et al., 2018). The current study aims to help fill these gaps in the literature by examining the impact of early PPD at three weeks post birth on later observed infant temperament in a sample of mothers who were not depressed during pregnancy. This is important for assessing whether there is an association between PPD and infant temperament in the absence of depression during pregnancy.

Temperament and PPD.

Temperament has been traditionally defined as stable, early-appearing individual differences in behavioural tendencies that have a constitutional basis (Goldsmith et al., 1987), and has historically been considered a set of genetic traits (Thomas & Chess, 1977). Temperament has also been referred to as individual differences across contexts in reaction to stimuli, including the expression of emotions, arousal, and self-regulation (Rothbart, Derryberry & Hershey, 2000; Kagen & Snidman, 1991). More recent research has re-conceptualized temperament as being the child’s fundamental disposition emerging early in life, which is influenced by the interactions of genetics and environmental influences (e.g. Shiner et al., 2012). The field of epigenetics has estimated the extent to which genetic and environmental factors contribute to behavioural variability. By decomposing the observed variance of a trait into genetic and environmental components, genetic differences account for approximately 20–60% of the variability of temperament, and environment is attributed to the remaining 40–80% of the variance (Saudino, 2005). Parents are the most predominant aspect of the infant’s environment, and thus parental mental health or illness has tremendous potential to impact the parent-infant relationship, infant development, and infant temperament.

Most of the research on associations between infant temperament and PPD has used maternal reports of infant temperament behaviour. For example, a cross-sectional survey by Britton (2011) found that anxiety and depression at one month postpartum was associated with mother-reported difficult infant temperament. Mothers with more symptoms of PPD between four and six months also reported sharper increases in indicators of difficult infant temperament in research by Hanington, Ramchandani, and Stein (2010). However, there is evidence that report of infant behaviour may not be completely accurate for depressed mothers. Specifically, landmark research by Field, Morrow and Adlestein (1993) compared mothers’ and observers’ ratings of infant negativity from the same video. They found that both observers and mothers reported that infants of depressed mothers were more negative than infants of non-depressed mothers, but depressed mothers reported significantly more infant negativity than observers. These results suggest both a bias toward negative perception of the infant on the part of depressed mothers, as well as a true increase in negativity on the part of infants of depressed mothers.

The limited research involving observations of infant temperament related behaviour has yielded mixed results. Forman and colleagues (2007) found that low risk mothers with PPD reported more infant negative affect at six months, but no associations were found between PPD and observed temperament. Observational research by Feldman and colleagues (2009) found associations between maternal depression and both negative emotionality and less mature regulation. Other research did not find an association between maternal depression and emotionality, but did find an association with infant fearfulness (Pauli-Pott, Mertesacker, & Beckman, 2004). Research by Parade, Armstrong, Dickstein and Seifer (2017) did not find associations between PPD and either observed temperament or change in temperament over time. However, they did find that maternal sensitivity moderated the impact of maternal PPD on change in temperament difficulty, such that increases in temperament difficulty were associated with PPD only when maternal sensitivity was low. These discrepancies in the research findings suggest the need for further research in this area.

Early PPD and later child outcomes.

The majority of PPD and postpartum distress (depression and anxiety) research begins examining maternal depressive symptoms after two months post birth and does not control for depression during pregnancy (see Kingston et al, 2012). One notable exception is a large multi-method study by Feldman and colleagues that started examining PPD at two days post birth. Findings from their publications regarding this study indicated that PPD was associated with decreased infant social engagement, more negative emotionality (Feldman et al., 2009), and compromised emotion regulation (Granat, Gadassi, Gilboa-Schectman, & Feldman, 2017). Beebe and colleagues (2008) measured PPD at six weeks postpartum, and found that PPD predicted lower self-contingency, and compromised interactive contingency when infants were 4-months-old. Little is known about how very early PPD affects later infant development overall, and temperament in particular.

PPD in low-risk samples.

Higher risk populations generally have a higher prevalence of PPD compared with lower risk groups (e.g. Mersky & Janczewski, 2018). However, research with low-risk community samples indicates that depressive symptoms can be high even in low risk samples shortly after the birth of a new baby. For example, Weisman and colleagues (2010) reported that 20.5% of mothers had high depressive symptoms two days after giving birth in their large low-risk sample, and Beebe and colleagues (2012) indicated that 25.8% of mothers had high depressive symptoms at six weeks post birth in their moderately sized sample. This research highlights the need to examine associations between early PPD and child outcomes in low-risk as well as high-risk samples.

The current study.

The goal of this study was to examine the influence of very early PPD at three weeks postpartum on both observed and maternal report of infant temperament behaviour in 6-month-old infants in an otherwise low-risk sample of women who were not depressed during pregnancy. Based on the findings of previous research, we expected that infants of depressed mothers would display sad or flat affect, be more withdrawn, be fussier, and express less positive affect. Because temperament reflects individual differences across contexts, we expected these associations to be evident both across temperament assessment tasks, and in both reported and observed temperament behaviour.

Methods

Participants

This research is part of a larger naturalistic longitudinal study (Jolley, Elmore, Barnard, & Carr, 2007). Participants were 16 mother-infant dyads who were followed from the eighth month of the mother’s pregnancy through the first six months of the infant’s life. Participants were recruited from obstetric, midwifery and paediatric practices, childbirth education classes, and university bulletin boards in the Seattle area. All research procedures received approval for research with human subjects by a university review board.

Eligibility criteria included: maternal age between 20–40 years, being in the third trimester of pregnancy, normal pregnancies, in good physical health, normal mood (no indications of depression or anxiety and depression during pregnancy), taking no medications, and being non-smokers. Screening for normal mood was done at enrolment using the Edinburgh Postpartum Depression Screening Scale (EPDS) (Cox, Holden, & Sagovsky, 1987).

The average maternal age was 31.25 years old. The ethnicity of the sample consisted of 87.5% of mothers who identified themselves as Caucasian, non-Hispanic and 12.5% mothers who identified themselves as Hispanic. The mean family income was $61,280. There was a broad range of education levels; 12.5% had only a high school degree, 18.8% had some college, and 68.8% obtained a college degree. Married women constituted 81.3% of the sample.

Measures

Depression.

The Edinburgh Postpartum Depression Screening Scale (EPDS) was used to screen potential participants for depression at enrolment when mothers were pregnant. The EPDS is a 10-item screening tool instrument that has been extensively used and validated for both postpartum (Cox et al., 1987), and pregnancy (Ryan, Mills & Misri, 2005). The Cronbach alpha for the current study was .67, reflecting acceptable internal consistency.

The Postpartum Depression Screening Scale (PDSS) (Beck & Gable, 2000) was used to measure PPD. The development of the PDSS was based on qualitative phenomenology research (Beck, 1992). The total score reflects overall postpartum depression symptoms, with a cut-off of 60 for major or minor depression, and scores of 80 or above indicating major postpartum depression. The 35 items utilize a 5-point Likert-type scale, from (1) “strongly disagree” to (5) “strongly agree”. Construct validity was confirmed with confirmatory factor analysis, goodness of fit of 0.87 in the validation study (Beck & Gable, 2000). Cronbach alphas for the current study were .88 at three weeks postpartum and .92 at six months postpartum, reflecting good internal consistency.

Maternal report of infant temperament.

The Infant Characteristics Questionnaire (ICQ) (Bates, Freeland, & Lounsbury, 1979) is a self-report measure of infant characteristics or temperament. Mothers filled out the ICQ questionnaire, rating their infant’s behaviour and reactivity on items with a 7-point scale. This assessment yields four dimensions: fussiness, inadaptability, unpredictability and dullness. Lower scores indicate easier temperament. The ICQ is a widely used measure with well-established validity, reliability, and factor structure (Bates, et al., 1979; Bates, Olson, Pettit, & Bayles, 1982). Cronbach alphas were: fussiness (α = .83), inadaptability (α = .83), unpredictability (α = .87), and dullness (α = .70).

Infant laboratory temperament behaviour.

The Pre-locomotor Laboratory Temperament Assessment Battery (Lab-TAB) (Goldsmith & Rothbart, 1999) was used to examine temperament behaviour in the context of standardized emotion-eliciting episodes. One episode appropriate for 6-month-olds was examined from each of four domains: basket of toys (activity level/ attention-eliciting), pop-up bear (joy-eliciting), attractive toy behind barrier (anger-eliciting), and masks (fear-eliciting). The episodes were administered in the same order for all participants and were videotaped in a laboratory setting. A modified Lab-TAB coding system was used to examine infant behaviour and expressions as outlined by Goldsmith and Rothbart (1999). All data were double coded to establish inter-rater reliability. Interclass correlations ranged from .66 to 1.0, with a mean of .89, reflecting good reliability. Infant affect was coded across episodes. The episodes and corresponding coding are described below.

The basket of toys episode was designed to elicit object-oriented interest and activity (3 minutes). Coding included the intensity of toy manipulation, and a composite variable for social play (bids to mother minus number of bouts of play). The pop-up bear episode (using a pop-up toy) was designed to elicit object-oriented enjoyment and surprise. Composite coding variables were created to reflect joy (sum of happiness and interest), and fearful responses (sum of fear, distress, and startle). The toy behind barrier episode was designed to elicit object-oriented frustration and anger associated with not being able to get to the attractive toy behind a clear barrier. Composite variables were created for sad or withdrawn affect (sadness plus negative vocalization minus interest), and infant approach to the barrier task (sum of struggling and bids to mother). The masks episode was designed to elicit mild fear. Masks were displayed by a hidden assistant. A composite variable was created to reflect fearful responses (sum of fear, stilling, distress, and escape attempts).

Study Procedures

The PDSS was administered at three weeks and six months postpartum. Referrals for treatment were made to mothers scoring above the cut-off for major depression in accordance with our ethical protocol. The Lab-TAB emotion-eliciting paradigms and the ICQ were administered when infants were six-months-old. During the Lab-TAB episodes, infants were seated in front of a table across from an experimenter, and mothers sat across the room. Mothers filled out questionnaires to reduce interference, while allowing infants visual access to their mothers. Tasks were aborted if the infant became very distressed.

Hypotheses

Based on previous research, we expected that infants of depressed mothers would be fussier, have less positive affect, show less intensity in play and be less interactive. We expected to see these temperament characteristics to be reflected to some extent across emotion-eliciting Lab TAB episodes. Infants of depressed mothers were also expected to be rated higher on the following ICQ scales: fussiness, inadaptability, and dullness. We did not have a specific hypothesis with regard to the unpredictability scale.

Data Analytic Plan

Goldsmith and Rothbart’s Lab-Tab scoring procedures (1999) were first used for data reduction. Means were computed across epochs within each episode. Descriptive statistics were calculated, and variables with little variance were removed from further analysis. Mean scores were transformed into Z-scores. Composite variables were then created from those z-scores. Composite variables were theoretically based, with correlations being conducted to verify that associations between variables were in the expected direction. In addition to the composite variables described above for each episode, composite variables were created to reflect affect infant across episodes. One variable reflected positive affect (sum of infant happiness and interest), and one variable reflected negative and withdrawn affect (sum of sadness, startle, distress and fear). Anger was not included in the negative affect composite because it has been associated with approach emotions in the literature (Harmon-Jones, Allen, 1998).

Descriptive statistics and t-tests were used to describe the prevalence of depressive symptoms over the time examined. Multiple linear regression models were used to examine the association between PPD symptoms and the infant temperament variables. The best fitting model for each temperament variable was found through the process of backwards elimination. The PDSS total scores at three weeks and six months post-birth were entered into the model together as independent variables. The regression approach using backwards elimination then removed the variables with the least change in R2 through a series of consecutive steps. Variables were not removed from the model if the change in R2 was large enough to reject the null hypothesis using a criterion of .1 (Norusis, 2000). This method was used because it was not assumed that early and concurrent maternal depressive symptoms would have an equivalent impact on infant temperament behaviour. The assumptions of regression (normality, homogeneity of variance, independence, having a linear relationship and lack of multi-collinearity) were tested and deemed acceptable, and Cook’s distance diagnostics were computed for each regression to identify outliers. In cases where the beta coefficients for PPD at three weeks and six months pospartum did not predict outcomes in the same direction, associations between predictor and outcome variables by change in PPD were examined by creating scatter plots where cases could be identified by the extent of decline in PPD symptoms (high, moderate, and low decline groups). One-tailed p-values were reported for tests of directional hypotheses and two-tailed p-values were reported for descriptive analyses.

Results

Postpartum Depression

Although none of the women were depressed during pregnancy, 62.5% had substantial symptoms of PPD (and 37.5% non-depressed) at three weeks post-partum, and 43.8% of the women continued to have substantial symptoms of depression at six months postpartum (56.2% non-depressed). Women with substantial PPD symptoms at both time points comprised 31.24% of the women, with depressive symptoms being significantly associated across time points (r(15) = .57, p = .02). Depressive symptoms reflected by the total PDSS scores declined significantly from three weeks (M = 74.12) to six months (M = 56.13) postpartum, t (15) = 3.37, p = .004. T-test results indicted that mothers who had extremely high PDSS scores (over 100) at three weeks postpartum had significantly steeper declines in depression (M = 42) compared to the rest of the sample (M = 10.5, t(14) = 3.14, p = .007).

Infant Laboratory Temperament Behaviour

Multiple linear regression revealed a significant positive association between symptoms of PPD, reflected by PDSS total scores, at three weeks post birth and expressed infant negative affect across LabTAB episodes when infants were 6-months-old, β = .57, F (1, 14) = 6.66, p = .01. Depressive symptoms at three weeks postpartum also significantly predicted less positive infant affect across episodes β = −.53, F (1, 14) = 4.43, p = .02. Concurrent PDSS scores did not contribute to the best fitting (explanatory) models in either of these cases.

Symptoms of maternal PPD were associated with infant behaviour and affect during some of the LabTAB episodes. Symptoms of PPD at three weeks postpartum were significantly and negatively associated with later joyful infant expressions during the pop-up toy joy-eliciting episode, β = −.31, F (1, 14) = 6.23, p = .01. Concurrent PDSS scores at six months post birth were significantly and positively associated with infant fearful affect during the pop-up toy episode, β = .33, F (1, 14) = 6.79, p = .01. Low PPD at three weeks postpartum predicted later infant approach behaviour during the anger-eliciting toys behind the barrier task, β = −.36, F (1, 14) = 7.90, p = .007. High PDSS scores at three weeks (β = .90, p = .001) in combination with low PDSS scores at six months postpartum (β = − .59, p = .01) were significantly associated with infant sad/withdrawn affect during the toy behind the barrier task, R2 = .53, F (2, 13) = 7.21, p = .004. Low PDSS scores at three weeks (β = −.71, p = .01) in combination with high PDSS scores at six months postpartum (β = .67, p = .014) were significantly associated with mean intensity of play during the basket of toys task, R2 = .40, F (2, 13) = 4.14, p = .02. In both cases where the combination of high PPD at three weeks and low PPD at six months predicted temperament outcomes, scatter plots indicated that the cases with the strongest associations between PPD and infant temperament also had the highest scores at three weeks and showed the most substantial decline in PPD. Maternal PDSS scores were not significantly associated with infant behaviour in the fear-eliciting masks episode or social play in the basket of toys episode. See Table 1 for all associations.

Table 1.

Multiple Regression Results of Associations between Maternal PPD Symptoms and Infant Temperament

Infant Temperament Variables at 6 months PPD 3 wks. PPD 6 mo. Model
β β R2 F
Observed affect across Lab-TAB Episodes
  Positive infant affect −.53* – .28 4.43*
  Negative infant affect .57** – .33 6.66**
Basket of toys (activity-eliciting)
  Intensity of play −.71** .67* .40 4.14*
  Social play .12 – .02 .16
Toy behind barrier (anger-eliciting)
  Approach behavior −.36** – .13 7.9**
  Sad/withdrawn affect .90*** −.59** .53 7.21**
Pup-up bear (joy-eliciting)
  Joyful affect −.31** – .10 6.23**
  Fearful response – .33** .11 6.79**
Masks (fear eliciting)
  Fearful response – .25 .06 .61
Maternal report of temperament (ICQ)
  Inadaptability .60* −.70* .35 3.53*
  Fussiness .67* −.54* .31 2.95*
  Dullness .32 – .10 1.48
  Unpredictable – −.15 .02 .30
Open in a new tab

Note.

*

p ≤ .05

**

p ≤ .01, p ≤ .005

Maternal Report of Infant Temperament (ICQ)

Higher PDSS scores reflecting depression at three weeks postpartum (β = .60, p = .03) in combination with lower concurrent PDSS score at six months postpartum (β = −.70, p = .02) were associated with higher maternal report of inadaptability on the ICQ when infants were 6-months-old, R2 = .35, F (2, 13) = 3.53, p = .03. There was also a significant association between PDSS scores and maternal report of infant fussiness, R2 = .31, F (2, 13) = 2.95, p = .05. Specifically, higher depressive symptoms at three weeks postpartum (β = .67, p = .018) in combination with lower depressive symptoms at six months postpartum were associated with reported infant fussiness (β = −.54, p = .04). Scatter plots indicated that the cases with the strongest associations between PPD and infant temperament also had the highest scores at three weeks and showed the most substantial decline in PPD. No significant associations were found between symptoms of PPD and maternal report of infant dullness or unpredictable temperament on the ICQ. See Table 1.

Discussion

Early PPD at three weeks postpartum predicted both observed challenging temperament behaviour in the Lab-TAB’s emotion-eliciting situation, and challenging aspects of mother reported infant temperament reflected in the ICQ scales when infants were six months old. This is particularly provocative given that only non-depressed pregnant women were enrolled in this study. Thus, these results reflect the impact of early PPD in the absence of depression during pregnancy, which is a phenomenon that rarely is explicitly studied through screenings during pregnancy and the transition to early postpartum. Indeed, to the best of our knowledge based on an extensive literature search using the PsycInfo database, this is the first study to demonstrate an association between PPD and infant temperament after controlling for depression during pregnancy. This is important in that it demonstrates an association between maternal PPD and infant temperament post birth, rather than reflecting an association that could largely reflect changes in the foetal development prior to birth.

The specific aspects of observed temperament behaviour in the Lab-TAB paradigm predicted by early PPD symptoms include more negative/withdrawn affect and less positive affect across emotion eliciting episodes. In the specific Lab-TAB episodes, early PPD symptoms at three weeks postpartum predicted: lack of joy in the pop-up toy joy-eliciting episode, sadness and a lack of action taken in the toys behind a barrier anger-eliciting episode, and lower intensity of play (reflecting more indifference) in the basket of toys episode. Concurrent PPD predicted infant fearful responses in the pop-up bear joy-eliciting episode. These results are all consistent with our hypotheses that symptoms of PPD would be related to more negative affect, and less positive affect and engagement. They are also consistent with past research indicating that infants of depressed mothers often exhibit depressed behaviour themselves (e.g. Field, 1998; 2010). It is of interest that symptoms of PPD were related to fearfulness in response to a pop-up bear joy-eliciting episode, but were not significantly related to infant fearful responses in the mask fear-eliciting episode where a fear response was expected. It may be that the children of the non-depressed mothers simply expressed similar levels of fear in the situation where fear was elicited because that may be an adaptive response in that situation. There were also significant associations between early PDSS scores and maternal report of infant fussiness and inadaptability. Early PPD did not significantly predict infant “dullness”. It may be that mothers have difficulty viewing their baby as both “fussy” and “dull” simultaneously, and “dullness” may be more difficult for depressed mothers in particular to pick up on and thus report.

There was a somewhat curious pattern of results in which a combination of high PPD symptoms at three weeks and low PPD symptoms at six months postpartum predicted more negative and less positive infant temperament in several instances. The beta coefficients in regression reflect the association between a predictor and outcome variable after accounting for other variables and holding them constant. Thus, they does not reflect a zero-order correlation, but rather what one variable predicts after taking others into account. Because higher early depressive symptoms were associated with a decline in PPD symptoms, this pattern suggests that it is the infants of the mothers with the most severe early PPD symptoms that predominantly exhibited the negative infant temperament related outcomes, even in the absence of continued depression. This interpretation is supported by scatter plots, which indicated that the cases with the strongest associations between early PPD and infant temperament also had the highest scores at three weeks and showed the most substantial decline in PPD.

The associations between early PPD and infant temperament even in the absence of later PPD are consistent with the findings of Forman and colleagues (2007) indicating associations between PPD and later child outcomes and mother-infant interactions even after mothers had recovered from depression. Tronick and Reck (2009) have studied the microanalysis of mother-infant interaction patterns in their laboratory, and found more mismatched communication in the mothers with depressive symptoms, which influences the infants’ affect and self-regulation. This impact is seen not only in severely and acutely depressed mothers, but also in mothers who have only high levels of depressive symptoms. The results from the current study predicting negative outcomes from early PPD in the absence of later PPD suggests that this influence of maternal depression on infant affect and responsiveness takes place very early in development, with infants learning very quickly that their bids for attention will not be responded to when their mothers are depressed.

This explanation about early PPD having an important impact on the development of infant temperament also makes sense given the pattern of PPD symptoms seen over time in the sample studied. Specifically, there was a significant decline in PPD symptoms overall from three weeks to six months postpartum, and an association between high PPD symptoms at three weeks postpartum and decline in symptoms over the time followed. This decline in PPD symptoms is likely due to the women with the most severe early depressive symptoms receiving treatment to alleviate their symptoms due to our human subjects ethical protocols. These women were referred to their healthcare providers for assessment and treatment in cases where women scored above the cut-off for depression. Follow-ups with the women who had the highest depression scores confirmed that treatment had indeed been received in the most severe cases. The overall decline in PPD symptoms would also make sense if some of the women initially experienced depressive symptoms associated with Baby Blues because it is considered more transient than major PPD. However, Baby Blues is typically considered to only occur for the first two weeks postpartum (e.g. Barbadoro, 2012), and the first assessment of PPD in this sample was at three weeks postpartum.

It is noteworthy that the majority of new mothers (62.5%) in this study had substantial symptoms of depression at three weeks post-partum despite not being depressed during pregnancy, and 43.8% had substantial symptoms of depression at both three weeks and six months postpartum. The high PPD rates seen in this sample may be partially due to the high sensitivity and specificity of the postpartum depression screening scale used (Beck & Gable, 2000). When the PDSS was compared with the EPDS and the Beck Depression Inventory-II, the PDSS had the highest combination of sensitivity and specificity at .94 and .98 respectively (Beck, & Gable, 2001). The PDSS is the only PPD screening tool that is written in the context of the postpartum period. The particularly high three weeks postpartum prevalence may also be partially reflective of the high rates of Baby Blues expected up to two weeks postpartum (Edhborg, Matthiesen, Lundh, & Widstrom, 2005).

PPD scores at three weeks postpartum were the strongest and most consistent predictors of infant temperament outcome variables. This is surprising given that the effects of concurrent depression are considered to be of primary importance in the field overall, and reviews and meta-analyses have found that concurrent self-report and observational measures of parenting have the strongest effects when compared to history of past depression (McLearn et al., 2006; Lovejoy, Graczyk, O’Hare, & Neuman, 2000). However, a few studies have found that the effects of early PPD symptoms in the first months of the child’s life may have a greater impact than later onset depression on parenting and child outcomes (Brennan, Hammen, Anderson, & Bor, 2000; Field, 1995; Laurent, 2017; Minkovitz et al., 2005). This is believed to be because of the importance of the formation of the early primary relationship between mother and infant, via the mother’s affect influencing her interaction with her infant, that impacts long-term interaction patterns, HPA axis regulation (Laurent, 2017), and may also influence brain development (Siegel, 2001). This rationale is supported by our findings of PPD as early as three weeks postpartum predicting infant temperament at six months of age even in the absence of later PPD symptoms. Because few studies measure depression as early as three weeks postpartum, it is possible that depression very early in the developmental process in the parent-child relationship and temperament formation may have a profound impact that isn’t revealed by research beginning with later measurements. This interpretation is consistent with results found by Britton (2011) identifying associations between PPD and reported temperament as early one month postpartum, suggesting that these associations can emerge very early (Britton, 2011). Future research is recommended to further examine these possible associations.

The results indicating the strong predictive ability of early PPD are also consistent with the research-based theoretical approach of Seigel and Shore (Shore, 2001; Siegel, 2001) positing that even micro aspects of early parent-infant interactions influence infant brain development as well as long-term interactive patterns. It is also consistent with research indicating that brain development, and the HPA axis in particular, are influenced by early maternal PPD (Lauraunt, 2017). The negative impact of early PPD on infant temperament through impacting brain development makes sense given the substantial research indicating associations between temperament and both brain morphology (e.g. Wei et al., 2018) and brain functioning (Clauss et al., 2016; Poole et al., 2018), and with HPA axis regulation in particular (Poole et al., 2017). Thus, parent-infant interactive patterns as early as three weeks post birth may lay the foundation for later development both in the infant and in the dynamic interactive social relationship between mother and child.

Limitations and Future Directions

A major limitation of the current investigation is the small sample size, which limits the generalizability of results as well as the statistical power for obtaining significant results. Given the sample size of 16 mother-infant dyads, we had sufficient power to only detect large effects. Because the sample size is small, we recommend further examining the associations between early PPD and later infant temperament with larger samples to confirm and expand upon the findings of the current research. It is noteworthy that the variance explained in the regression models in this study ranged from 10 to 53 percent. Because Cohen (1988) considered an R squared value above 13.8% to be a large effect size, we believe that our results are likely to reflect the robust nature of the association between PPD and infant temperament. Thus we believe further investigation of these associations is warranted.

Another limitation of the current research was the use of only four of the 17 possible pre-locomotor Lab-TAB episodes available for use in examining temperament with pre-locomotor children (Goldsmith & Rothbart, 1999). Research such as the current study examining temperament in young pre-locomotor infants is limited to emotion-eliciting episodes appropriate for young babies. However, the number of episodes used in the current research was also limited by the number of episodes we thought infants would be able to tolerate in a single visit at 6-months-of age. Future research could expand upon this measurement of early-observed temperament by bringing mother-infant dyads into their laboratory multiple times for a more in-depth assessment across a larger number of Lab-TAB episodes.

Finally, we recommend that future research examining associations between early PPD and infant temperament measure infant cortisol to enable the examination of physiological changes in the infant as a potential mediator between early PPD and later infant temperament. We have discussed altered brain development as a possible mechanism through which early PPD influences temperament, but the current study did not include the physiological assessments necessary to confirm this hypothesis.

Implications

The findings of the current research linking early PPD with later observed and reported difficult infant temperament highlights the importance of identifying and treating PPD as early as the first few weeks of the infant’s life because the early mother-infant interactions may be the foundation for further development. The sadness, lack of joy, lack of action, and lack of engagement in play associated with early PPD could also be signs of depression developing in the infant. This suggests that it is important to also assess the mental health and wellbeing of infants in cases where mothers are depressed during the postpartum period. It may also be beneficial for clinicians and intervention programs to target the mother-child dyad in addition to the maternal PPD both because past research has found that maternal sensitivity appears to ameliorate the impact of PPD on temperament (Parade et al., 2017), and because relationship-based approaches can help promote positive infant and child mental health (e.g. Lieberman & Van Horn, 2011). A dual focus on both treating maternal PPD and maternal-child interactions may be particularly important given findings by Foreman and colleagues (2017) indicating that challenging mother-infant interactions tend to persist even after treatment and recovery of PPD alone. It would be beneficial for early childhood education and health care facilities to recommend early PPD assessment, as well as for medical facilities to make this type of screening part of their regular care plan. The high predictive ability of PPD at three weeks postpartum in the current study suggests that even early transient depressive symptoms often categorized as the Baby Blues may have a negative impact on infant development. Furthermore, there are implications for assessing early PPD even in the absence of depression during pregnancy, because the links between early PPD and difficult infant temperament identified in the current research were with mothers who were not depressed during pregnancy.

Acknowledgements:

Funding Sources: NINR: Predoc NRSA (5 F31 NRO7528–03); Women’s Health Nursing Research Training Grant (T32 0739); University of Washington, Royalty Research Fund Grant 65–5266. We also want to express our appreciation to the study participants.

Biography

Alyson Shapiro, PhD is an assistant professor in the Department of Child and Family Development at San Diego State University. Her research focus is on the inter-related areas of family dynamics around the transition to parenthood period, and child development within the family context.

Sandra Jolley, PhD is a Paediatric Nurse Practitioner and retired faculty from the Department of Paediatrics at the University of Washington. Her paediatric focus is on primary paediatric care and infant mental health and wellbeing. Her research focuses on the transition of mothers from pregnancy to the postpartum period, the psychobiology of postpartum depression, and the impact of maternal mental health on infants and young children.

Ursula Hildebrandt received an M.S. in clinical social work and PhD in developmental psychology. She has been a post-doctoral fellowship at the Barnard Centre for Infant Mental Health and an assistant professor of psychology at Seattle Pacific University. She currently divides her time between conducting research, teaching, her private practice, and work at a non-profit organization serving young children in the foster care system.

Susan Spieker, PhD is the Kathryn Barnard Endowed Professor for Infant Mental Health in the Family and Child Nursing Department at the University of Washington. Her experience spans projects on infant childcare and attachment, adolescent childbearing and child and family development, and randomized trials on intervention in Early Head Start and child welfare.

References

  1. American Psychiatric Association (2013) Diagnostic and statistical manual of mental disorders: DSM-5, 5th ed. Arlington, VA: American Psychiatric Publishing. 10.1176/appi.books.9780890425596, ISBN 978-0-89042-555-8. [DOI] [Google Scholar]
  2. Barbadoro P, Cotichelli G, Chiatti C, Simonetti ML, Marigliano A, Di Stanislao F, & Prospero E (2012). Socio-economic determinants and self-reported depressive symptoms during postpartum period. Women & Health, 52(4), 352–368. 10.1080/03630242.2012.674090 [DOI] [PubMed] [Google Scholar]
  3. Bates JE, Freeland CA, & Lounsbury ML (1979). Measurement of infant difficultness. Child Development, 50(3), 794–803. [PubMed] [Google Scholar]
  4. Bates JE, Olson SL, Pettit GS, & Bayles K (1982). Dimensions of individuality in the mother–infant relationship at six months of age. Child Development, 53(2), 446–461. 10.2307/1128988 [DOI] [PubMed] [Google Scholar]
  5. Beck CT (1992). The lived experience of postpartum depression: A phenomenological study. Nursing Research, 41(3), 166–170. 10.1097/00006199-199205000-00008 [DOI] [PubMed] [Google Scholar]
  6. Beck CT, & Gable RK (2000). Postpartum Depression Screen Scale: Development and psychometric testing. Nursing Research, 49(5), 272–282. 10.1097/00006199-200009000-00006 [DOI] [PubMed] [Google Scholar]
  7. Beck CT & Gable RK (2001). Comparative Analysis of the Performance of the Postpartum Depression Screening Scale with Two other Depression Instruments. Nursing Research, 50, 242–250. 10.1097/00006199-200107000-00008 [DOI] [PubMed] [Google Scholar]
  8. Beebe B, Badalamenti A, Jaffe J, Feldstein S, Marquette L, Helbraun E, et al. (2008). Distressed mothers and their infants use a less efficient timing mechanism in creating expectancies of each other’s looking patterns. Journal of Psycholinguistic Research, 37(6), 453–454. 10.1007/s10936-008-9081-3 [DOI] [PubMed] [Google Scholar]
  9. Brennan PA, Hammen C, Andersen MJ, Bor W, Najman JM, & Williams GM (2000). Chronicity, severity, and timing of maternal depressive symptoms: Relationships with child outcomes at age 5. Developmental Psychology, 36(6), 759–766. 10.1037/0012-1649.36.6.759 [DOI] [PubMed] [Google Scholar]
  10. Britton JR (2011). Infant temperament and maternal anxiety and depressed mood in the early postpartum period. Women & Health, 51(1), 55–71. 10.1080/03630242.2011.540741 [DOI] [PubMed] [Google Scholar]
  11. Clauss JA, Benningfield MM, Rao U, & Blackford JU (2016). Altered prefrontal cortex function marks heightened anxiety risk in children. Journal of the American Academy of Child & Adolescent Psychiatry, 55(9), 809–816. 10.1016/j.jaac.2016.05.024 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cohen J (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates. [Google Scholar]
  13. Cox JL, Holden JM, & Sagovsky R (1987). Detection of postnatal depression: Development of the 10-item Edinburgh postnatal depression scale. The British Journal of Psychiatry, 150, 782–786. 10.1192/bjp.150.6.782 [DOI] [PubMed] [Google Scholar]
  14. Edhborg M, Matthiesen A, Lundh W, & Widström A (2005). Some early indicators for depressive symptoms and bonding 2 months postpartum--A study of new mothers and fathers. Archives of Women’s Mental Health, 8(4), 221–231. 10.1007/s00737-005-0097-5 [DOI] [PubMed] [Google Scholar]
  15. Feldman R, Granat A, Pariente C, Kanety H, Kuint J, & Gilboa-Schechtman E (2009). Maternal depression and anxiety across the postpartum year and infant social engagement, fear regulation, and stress reactivity. Journal of the American Academy of Child & Adolescent Psychiatry, 48, 919–927. 10.1097/CHI.0b013e3181b21651 [DOI] [PubMed] [Google Scholar]
  16. Field T (1995). Infants of depressed mothers. Infant Behavior & Development, 18, 1–13. [DOI] [PubMed] [Google Scholar]
  17. Field T (1998). Maternal depression effects on infants and early interventions. Preventive Medicine: An International Journal Devoted to Practice and Theory, 27, 200–203. 10.1006/pmed.1998.0293 [DOI] [PubMed] [Google Scholar]
  18. Field T (2010). Postpartum depression effects on early interactions, parenting, and safely practices: A review. Infant Behavior and Development, 33(1), 1–6, 10.1016/j.infbeh.2009.10.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Field T, Morrow C, & Adlestein D (1993). Depressed Mothers Perceptions of Infant Behavior. Infant Behavior & Development, 16), 99–108. 10.1016/0163-6383(93)80031-3 [DOI]
  20. Forman DR, O’Hara MW, Stuart S, Gorman LL, Larsen KE, & Coy KC (2007). Effective treatment for postpartum depression is not sufficient to improve the developing mother-child relationship. Development and Psychopathology, 19(2), 585–602. 10.1017/S0954579407070289 [DOI] [PubMed] [Google Scholar]
  21. Gauvain M (2001). The Social Context of Cognitive Development Guilford Press, New York, NY. [Google Scholar]
  22. Goldsmith HH, Buss AH, Plomin R, Rothbart MK, Thomas A, Chess S, et al. (1987). What is temperament? Four approaches. Child Development, 58(2), 505–529. 10.1111/1467-8624.ep7253747. [DOI] [PubMed] [Google Scholar]
  23. Goldsmith HH, & Rothbart MK (1999). Laboratory Temperament Assessment Battery-Prelocomotor University of Wisconsin, Madison. [Google Scholar]
  24. Granat A, Gadassi R, Gilboa-Schectman E, & Feldman R (2017). Maternal depression and anxiety, social synchrony, and infant regulation of negative and positive emotions. Emotion, 17(1), 11–27. 10.1037/emo0000204 [DOI] [PubMed] [Google Scholar]
  25. Harmon-Jones E, & Allen JB (1998). Anger and frontal brain activity: EEG asymmetry consistent with approach motivation despite negative affective valence. Journal of Personality and Social Psychology, 74, 1310–1316. 10.1037/0022-3514.74.5.1310. [DOI] [PubMed] [Google Scholar]
  26. Jolley SN, Elmore S, Barnard KE, & Carr DR (2007). Dysregulation of the hypothalamic-pituitary-adrenal axis in postpartum depression. Biological Research For Nursing, 8(3), 210–222. 10.1177/1099800406294598 [DOI] [PubMed] [Google Scholar]
  27. Kingston D, Tough S, & Whitfiend H (2012). Prenatal and postpartum maternal psychological distress and infant development: A systematic review. Child Psychiatry and Human Development, 43(5), 683–714. 10.1007/s10578-12-0291-4 [DOI] [PubMed] [Google Scholar]
  28. Lovejoy MC, Graczyk PA, O’Hare E, & Neuman G (2000). Maternal depression and parenting behavior: A meta-analytic review. Clinical Psychology Review, 20(5), 561–592. 10.1016/S0272-7358(98)00100-7 [DOI] [PubMed] [Google Scholar]
  29. Laurent H (2017). Early calibration of the HPA axis by maternal psychopathology. Psychoneuroendocrinology, 78, 177–184. 10.1016/j.psyneuen.2017.01.034 [DOI] [PubMed] [Google Scholar]
  30. Lieberman AF, & Van Horn P (2011). Psychotherapy with infants and young children: Repairing the effects of stress and trauma on early attachment Guilford Press. ISBN 9781609182403 [Google Scholar]
  31. McCoy SJ, Beal JM, Shipman SB, Payton ME, & Watson G (2006). Risk factors for postpartum depression: A retrospective investigation at 4-weeks postnatal and a review of the literature. Journal of the American Osteopathic Association, 106, 193–198.
 [PubMed] [Google Scholar]
  32. McLearn KT, Minkovitz CS, Strobino D, Marks E, & Hou W (2006). The timing of maternal depressive symptoms and mothers’ parenting practices with young children: Implications for pediatric practice. Pediatrics, 118(1), e174–e182. 10.1542/peds.2005-1551 [DOI] [PubMed] [Google Scholar]
  33. Mersky JP, & Janczewsky CE (2018). Adverse childhood experiences and postpartum depression in home visiting programs: Prevalence, association, and mediating mechanisms. Maternal and Child Health Journal, 22, 1051–1058. [DOI] [PubMed] [Google Scholar]
  34. Minkovitz CS, Strobino D, Scharfstein D, Hou W, T M, Mistry KB, et al. (2005). Maternal depressive symptoms and children’s receipt of health care in the first 3 years of life. Pediatrics, 115(2), 306–314. 10.1542/peds.2004-0341 [DOI] [PubMed] [Google Scholar]
  35. Norusis JM (2000). SPSS 11.1: Guide to Data Analysis Upper Saddle River, NJ: Prentice Hall. [Google Scholar]
  36. Parade SH, Armstrong LM, Dickstein S, & Seifer R (2017). Family context moderates the association of maternal depression and stability of infant temperament. Child Development, 88, 1–18. 10.1111/cdev.12895. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Parade S & Leerkes E (2008). The reliability and validity of the Infant Behavior Questionnaire-Revised. Infant Behavior and Development, 31(4), 637–46. 10.1016/j.infbeh.2008.07.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Pauli-Pott U, Mertesacker B, & Beckmann D (2004). Predicting the development of infant emotionality from maternal characteristics. Development and Psychopathology, 16, 19–42. 10.1017/S0954579404044396 [DOI] [PubMed] [Google Scholar]
  39. Poole KL, Jetha MK, & Schmidt LA (2017). Linking child temperament, physiology, and adult personality: Relations among retrospective behavioural inhibition, salivary cortisol, and shyness. Personality and Individual Differences, 113, 68–73. 10.1016/j.paid.2017.03.013 [DOI] [Google Scholar]
  40. Poole KL, Santesso DL, Van Lieshout RJ, & Schmidt LA (2018). Trajectories of frontal brain activity and socio-emotional development in children. Developmental Psychobiology, 60(4), 353–363. doi: 0.1002/dev.21620 [DOI] [PubMed] [Google Scholar]
  41. Ryan D, Mills L, & Misri N, (2005). Depression during pregnancy. Canadian Family Physician, 51, 1087–1093 [PMC free article] [PubMed] [Google Scholar]
  42. Rothbart MK, Derryberry D, & Hershey K (2000). Stability of temperament in childhood: Laboratory infant assessment to parent report at seven years. In Molfese VJ & Molfese DL (Eds.), Temperament and personality development across the life span, (pp. 85–119). Hillsdale, NJ: Erlbaum. [Google Scholar]
  43. Saudino KJ (2005). Behavioral genetics and child temperament. Developmental and Behavioral Pediatrics, 26(3), 214–223. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Shiner RL, Buss KA, McClowry SG, Putnam SP, Saudino KJ, & Zentner M (2012). What is temperament now? Assessing progress in temperament research on the twenty-fifth anniversary of Goldsmith et al. (1987). Child Development Perspectives, 6, 436–444. 10.1111/j.1750-8606.2012.00254.x [DOI] [Google Scholar]
  45. Shonkoff JP, & Phillips DA (2000). From neurons to neighbourhood: The science of early childhood development (Shonkoff JP & Phillips DA, Eds.). Washington, DC: National Academy Press. ISBN: 0-309-06836-3 [PubMed] [Google Scholar]
  46. Shore AN (2001). Effects of a secure attachment relationship on right brain development, affect regulation, and infant mental health. Infant Mental Health Journal, 22(1–2), 7–66. 10.1002/1097-0355(200101/04)22:1 [DOI] [Google Scholar]
  47. Siegel DJ (2001). Toward an interpersonal neurobiology of the developing mind: Attachment relationships, “mindsight,” and neural integration. Infant Mental Health Journal, 22(1–2), 67–94. doi:3.0.CO;2-G [Google Scholar]
  48. Sohr-Preston SL, & Scaramella LV (2006). Implications of timing of maternal depressive symptoms for early cognitive and language development. Clinical Child and Family Psychology Review, 9(1), 65–83. [DOI] [PubMed] [Google Scholar]
  49. Stone SD, & Menken AD (2008). Perspectives on perinatal mood disorders: A comprehensive treatment guide New York: Springer. [Google Scholar]
  50. Thomas A, & Chess S (1977). Temperament and Development Oxford: Brunner/Matzel. [Google Scholar]
  51. Tronick EZ, & Reck C (2009). Infants of depressed mothers. Harvard Review of Psychiatry, 17, 147–56. 10.1080/10673220902899714 [DOI] [PubMed] [Google Scholar]
  52. Wei L, Guo N, Baeken C, Bi M, Wang X, Qiu J, & Wu G-R (2018). Grey matter volumes in the executive attention system predict individual differences in effortful control in young adults. Brain Topography 10.1007/s10548-018-0676-1 [DOI] [PubMed]
  53. Weisman O, Granat A, Gilboa-Schectman E, Singer M, Gordon I, Azulay H, Kuint J, & Feldman R (2010). The experience of labor, paternal perceptions of the infant, and the mother’s postpartum mood in a low-risk community cohort. Archives of Women’s Mental Health, 13(6), 505–513. 10.1007/s00737-010-0169-z. [DOI] [PubMed] [Google Scholar]
  54. Woody CA, Ferrari AJ, Siskind DJ, Whiteford MG, Harris MG (2017). A systematic review and meta-regression of the prevalence and incidence of perinatal depression. Journal of Affective Disorders, 219, 86–92. 10.1016/j.jad.2017.05.003 [DOI] [PubMed] [Google Scholar]
  55. Zhang W, Finik J, Dana K, Glover V, Ham J, & Nomura Y (2018). Prenatal depression and infant temperament: The moderating role of placental gene expression. Infancy, 23(2), 211–231. 10.1111/infa.12215 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES