RESEARCH REPOSITORY

This is the author’s final version of the work, as accepted for publication following peer review but without the publisher’s layout or pagination.

The definitive version is available at:

https://doi.org/10.1016/j.jad.2018.05.025

Galbally, M., Watson, S.J., Teti, D. and Lewis, A.J. (2018) Perinatal maternal depression, antidepressant use and infant sleep outcomes: Exploring cross-lagged associations in a

pregnancy cohort study. Journal of Affective Disorder

http://researchrepository.murdoch.edu.au/id/eprint/40996/

Copyright: © 2018 Elsevier B.V.

It is posted here for your personal use. No further distribution is permitted.

Accepted Manuscript

Perinatal maternal depression, antidepressant use and infant sleepoutcomes: exploring cross-lagged associations in a pregnancy cohortstudy

Megan Galbally , Stuart J. Watson , Doug Teti , Andrew J. Lewis

PII: S0165-0327(18)30396-3DOI: 10.1016/j.jad.2018.05.025Reference: JAD 9801

To appear in: Journal of Affective Disorders

Received date: 23 February 2018Revised date: 12 April 2018Accepted date: 18 May 2018

Please cite this article as: Megan Galbally , Stuart J. Watson , Doug Teti , Andrew J. Lewis , Perina-tal maternal depression, antidepressant use and infant sleep outcomes: exploring cross-lagged associ-ations in a pregnancy cohort study, Journal of Affective Disorders (2018), doi: 10.1016/j.jad.2018.05.025

This is a PDF file of an unedited manuscript that has been accepted for publication. As a serviceto our customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, andall legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

1

Highlights

No reciprocal associations between maternal depression and infant sleep problems.

Antidepressant exposure not associated with increased infant sleep problems.

Maternal depression not associated with increased infant sleep problems.

Reciprocal longitudinal effects between maternal cognitions and infant waking.

Early maternal cognitions and infant waking predicted later night settling behaviours.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

2

Perinatal maternal depression, antidepressant use and infant sleep outcomes: exploring

cross-lagged associations in a pregnancy cohort study.

Megan Galbally MBBS, PhD1,2,3

, Stuart J Watson PhD1,2

, Doug Teti PhD4, Andrew J. Lewis

PhD1

Author Affiliations:

1. School of Psychology and Exercise Science, Murdoch University, Australia

2. School of Medicine, University of Notre Dame, Australia

3. King Edward Memorial Hospital, Australia

4. Human Development and Family Studies, The Pennsylvania State University

Corresponding Author:

Professor Megan Galbally Murdoch University, 90- South Street, Murdoch 6150

Email: m.galbally@murdoch.edu.au Ph +61-8-9340 2844

Funding

This study is supported through the 2012 National Priority Funding Round of beyondblue in

a three-year research grant (ID 519240) and a 2015 National Health and Medical Research

Council (NHMRC) project grant for 5 years (APP1106823). Financial support has also been

obtained from the Academic Research Development Grants from Mercy Health and from the

Centre for Mental Health and Well-Being, Deakin University.

Declaration of Interest

MG has previously received honorarium for speaking from Lundbeck. The other authors

declare that they have no competing interests.

Abstract

Background: Both perinatal depression and infant sleep problems are common concerns in

many communities, with these problems often coinciding. Findings in this area conflict and

much of the research relies on poor measures of sleep and/or depression. Adding to this

complexity is the rise in antidepressant treatment for perinatal maternal depression and no

previous study has examined the relationship between such exposure and infant sleep.

Methods: This study draws on four waves of data (first and third trimesters, and six and 12

months postpartum) from 264 women in the Mercy Pregnancy and Emotional Wellbeing

Study, a prospective pregnancy cohort study of women recruited in first trimester in

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

3

Melbourne, Australia. Cross-lagged regression models were used to examine reciprocity of

longitudinal effects between depressive symptoms and sleep.

Results: Maternal antepartum depression and antidepressant use were not significant

predictors of infant sleep problems. Likewise, infant sleep problems were not significant

predictors of postpartum maternal depression. However, maternal cognitions about infant

sleep, characterised by maternal expectations to immediately attend to their crying child, did

demonstrate positive reciprocal effects with infant nocturnal waking between six and 12

months postpartum.

Limitations: Infant sleep outcomes were reported by the mother and the sample were

predominantly Anglophone, restricting generalizability of the models to other cultures.

Conclusions: Maternal depression and antidepressant use were not found to be significant

factors in infant sleep problems and, likewise, infant sleep problems were not associated with

maternal depression. However, postpartum maternal cognitions around six months

postpartum regarding limit-setting at night may predict increases in later nocturnal infant

signaling.

Keywords Depression, Antidepressants, Infant Sleep

Abbreviations

AD: Antidepressant; BISQ: Brief Infant Sleep Questionnaire; CFI: Comparative Fit Index;

EPDS: Edinburgh Postnatal Depression Scale; PIBBS: Parental Interactive Bedtime Behavior

Scale; RMSEA: Root Mean Square Error of Approximation; SCID-IV: Structured Clinical

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

4

Interview for the DSM-IV; SRMR: Standardized Root Mean Square; SSRI: Selective

Serotonin Reuptake Inhibitor

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

5

A range of factors including biological, socio-emotional, early parenting practices and

cultural practices, and parental expectations can potentially influence the development of

sleep patterns in infants. For many families there are challenges adjusting to infant sleep

patterns and unsettled infants can adversely affect both parental and infant wellbeing.

However, whether infant sleep patterns can influence the development of maternal depression

remains unclear. In several studies, maternal depression has been implicated as an associated

factor in infant sleep, but the true nature of the relationship between maternal mental health

and infant sleep outcomes is yet to be fully understood.

Studies relying only on maternal reports of depressive symptoms and of infant sleep

outcomes demonstrate a significant association between an infant’s unsettled sleep and

maternal depression. Indeed, unsettled infant sleep may predict maternal depression,

suggesting that interventions aimed at improving infant sleep outcomes may improve

maternal depression (Hiscock and Wake, 2001; Lam et al., 2003). However, research

examining this relationship using observational methods has demonstrated the relationship to

be more complex, with some maternal behaviors associated with depression affecting infant

sleep behaviors such as night waking (Teti and Crosby, 2012). Adding to this complexity,

there are key postpartum variables, such as breastfeeding, often unaccounted for in this area

of research that can influence maternal and infant sleep patterns, and may affect depression

symptoms (Dias and Figueiredo, 2015; Galbally et al., 2013).

Furthermore, the infant sleep outcome can vary from night waking and signaling for

the parent, parental settling behaviors, difficulties going to sleep and sleep location. Defining

the sleep problem is crucial and varies in the research from parental perception of a problem,

numbers of wakes in the night, to the time taken for an infant to fall sleep. For instance, night

waking may be due to infant sleep problems, but equally, may be part of normal feeding

behavior (Galbally et al., 2013). Making sense of current research findings is important given

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

6

potential implications for clinical translation and recommendations, depending on whether

maternal factors or infant factors are stronger in this relationship and the nature of the infant

sleep concern. Teti and Crosby have conceptualized the association between infant sleep and

maternal depression as being either a maternal-driven model or an infant-driven model (Teti

and Crosby, 2012).

The maternal-driven model places the starting point at maternal depression, which

then influences maternal cognitive response and actions around infant sleep and ultimately

influencing infant sleep behaviors. This is supported by increased rates of depression in

mother-infant pairs presenting to early parenting sleep and settling services (Fisher et al.,

2004). It is also supported by prospective research demonstrating maternal depression in

pregnancy is associated with higher reported sleep difficulties in infants (Field et al., 2007;

Nevarez et al., 2010). Likewise, a systematic review of interventions at improving infant

sleep did not show a clear impact on infant night waking (signaling) or maternal depression

(Douglas and Hill, 2013). A recent meta-analysis found an effect of infant sleep interventions

on maternal depressive symptoms measured using the Edinburgh Postnatal Depression Scale

(EPDS). Although, the authors comment that the two highest quality studies in their review

did not report a significant association between infant sleep interventions and maternal mood

(Kempler et al., 2016).

An infant-driven model is conceptualized to start with an unsettled infant as a risk

factor for the development of maternal depressive symptoms. A study that prospectively

examined women who presented with an EPDS score less than 13 at one week postpartum

and elevated EPDS scores at four and eight weeks postpartum, also reported fatigue and an

unsettled infant with poor sleep (Dennis and Ross, 2005). A randomized control trial aimed at

reducing depression and infant sleep problems demonstrated a small but significant reduction

in depressive symptoms between four and six months postpartum for mothers in the

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

7

intervention arm. It should be noted, however, that for the intervention group, the average

EPDS was 6.3 (SD = 4.4) at four months and 5.1 (SD = 4) at 6 months, suggesting that this

sample was a low-risk sample for clinical depression (Hiscock et al., 2014). Few studies have

used diagnostic measures for depression, with the majority relying on EPDS; as such,

implications for depressive disorders cannot be drawn from this literature. Beyond the

maternal- and infant-driven models, a range of environmental factors are often not examined

in studies investigating infant sleep and maternal depression, but may equally influence both

maternal depression and infant sleep. These factors include in-utero exposures including

antidepressant medication and early parenting practices such as breastfeeding and infant sleep

location.

The use of antidepressant medication in pregnancy for women experiencing moderate

to severe depression is rising and hence understanding any influence on infant sleep

development is important. The most commonly prescribed antidepressant is a Selective

Serotonergic Reuptake Inhibitor (SSRI), which has been shown to alter maternal sleep

architecture, but as yet there is only limited exploration of the potential effect of fetal

exposure on the development of infant sleep architecture (Jindal et al., 2003). One study of

fetuses exposed to SSRIs throughout pregnancy reported disrupted development of quiet

sleep in late pregnancy, but follow-up into the postpartum was not undertaken (Mulder et al.,

2011). Studies that examine neonatal behavior following exposure to SSRIs identify sleep

changes as one of the most common symptoms (Galbally et al., 2017a; Zeskind and Stephens,

2004). However, again, follow-up beyond the neonatal period is limited. These studies also

rely on a script having been given or a hospital record yet many women discontinue treatment

with antidepressants in pregnancy (Ververs et al., 2006).

In this study, we built on existing research to examine the associations between

maternal depression and antidepressant use, and infant sleep. First, we assessed the role of

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

8

antidepressant exposure on infant sleep behaviors, including total nocturnal sleep time and

nocturnal waking. As per previous findings, we predicted that the sleep behaviors of infants

exposed to antidepressants would differ significantly compared to the sleep behaviors infants

not exposed to antidepressants (hypothesis 1). For infant sleep behaviors that did vary, we

predicted a dose effect of antidepressants, such that higher doses would be associated with

stronger effects (hypothesis 2).

Next, we assessed the role of clinically diagnosed maternal depressive disorder on

infant sleep location, perception of sleep problems, support seeking for infant sleep problems,

and use of parental bedtime settling strategies and behaviors. We predicted that mothers with

a depressive disorder would be more likely to report different sleeping locations, and infant

sleep perceptions, strategy use and behaviors compared to mothers not diagnosed with a

depressive disorder (hypothesis 3). We also examined the role of a maternal depressive

disorder on infant sleep behaviors at 6 and 12 months postpartum, testing the effect of

breastfeeding as a moderator. Specifically, we predicted a significant interaction between

maternal depressive disorder and breastfeeding on infant sleep behaviors (hypothesis 4).

Finally, we extended the models for maternal depressed mood and infant sleep

reported by Teti and Crosby (2012) by combining the mother- and infant-driven models,

testing simultaneously the longitudinal effects of maternal depressive disorder and symptoms

during pregnancy and the postpartum, and how these symptoms were associated with

nocturnal infant waking in the postpartum, controlling for breastfeeding. Specifically, we

predicted that the relations between postpartum maternal depressive symptoms and infant

waking would be bidirectional, after controlling for antepartum depressive disorder and

symptoms and breastfeeding in the postpartum (hypothesis 5). We also predict maternal

bedtime settling behavior and maternal cognitions about infant sleep in the postpartum will

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

9

be associated with maternal depressive symptoms and infant waking at 6 and 12 months

postpartum (hypothesis 6).

Method

This is a prospective cohort study of 282 pregnant women recruited before the 20th

week of pregnancy (Wave 1), and followed up during third trimester (Wave 2), birth of their

child (Wave 3), and six months (Wave 4), 12 months (Wave 5) and 36 months postpartum

(Wave 6). For this study, 264 women (Mage = 31.21, SDage = 4.69, range: 18 – 48 years old)

were included using data from Waves 1, 2, 4 and 5. Women are missing if they had

withdrawn prior to or after Wave 3 or if they were missing both postpartum waves. Study

participants comprised three groups: women taking antidepressant medication in pregnancy

(AD exposed; n = 47), non-medicated women who met diagnostic criteria for depression or

dysthymia at recruitment (Depressed; n = 29), and control women (n = 189). Further details

of the study are described in the published study protocol (Galbally et al., 2017b). Mercy

Health Human Research Ethics Committee approved this study and all participants provided

informed, written consent.

Measures

Mother reported whether they had ceased breastfeeding at 6 and 12 months

postpartum (0 = ceased breastfeeding, 1 = breastfeeding), and whether their infant slept in

the same room (0 = own room, 1 = parents’ room) and on the same surface (0 = own surface,

1 = parents’ surface).

A diagnostic measure was undertaken at recruitment: the Structured Clinical

Interview for DSM-IV (SCID-IV) Mood disorders schedule (First et al., 1997). In addition,

the Edinburgh Postnatal Depression Scale (EPDS) was administered at Waves 1, 2, 4 and 5

(Cox et al., 1987). The scale has been validated for use with Australian women during the

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

10

perinatal period, where sensitivity and specificity for cut-off scores of 12.5 and 13.5 were 100

and 95.7%, respectively (Boyce et al., 1993).

Antidepressant type, usage, dosage and timing during pregnancy was self-reported by

women, as well as obtained from hospital records at delivery. As the majority of participants

were on sertraline, all doses of antidepressants were converted to a sertraline-equivalent

dosage using a conversion chart (Procyshyn et al., 2015). Participant-reported antidepressant

dose during the first trimester is used for this analysis.

Maternal blood and umbilical cord blood were also collected within 30 minutes of

delivery and assayed for drug levels. Blood was centrifuged and the plasma stored at -80°C

within one hour of collection. All samples were analyzed with liquid chromatography with

mass spectrometry or tandem mass spectrometry, with methods described previously

(Andreassen et al., 2015; Reis et al., 2009). The assay levels for each antidepressant type

were standardized by relating the concentrations measured to the middle of the therapeutic

range for the same drug (Hiemke et al., 2011) so that concentration levels were comparable

across the variety of applied antidepressants.

The Brief Infant Sleep Questionnaire (BISQ), a 13-item parent report measure,

administered at both postpartum waves to assess duration and quality of infant sleep during

the previous week. Individual items measure the following: nocturnal and diurnal sleep

duration (hours), nocturnal infant waking (referred to as infant signaling) (frequency), and

duration of nocturnal wakefulness duration and nocturnal sleep-onset (minutes). Also

measured is the degree to which the mother believes the infant’s sleeping is a problem (A

very serious problem, A small problem, and Not a problem at all) (Sadeh, 2004).

The Parental Interactive Bedtime Behavior Scale (PIBBS) is a 17-item scale to assess

a range of parental behaviors used to settle infants to sleep and was administered at 12

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

11

months postpartum (Morrell and Cortina‐Borja, 2002). Domains of settling behavior include

physical comforting, encouraging autonomy, movement, passive physical comforting, and

social comforting. Each item is measured using a 5-point Likert scale (0 = Never, 4 = Very

Often). We used a total score, computed to represent the percentage (ranging 0 to 100%) of

time mothers report using these strategies to settle their child at bedtime. The subscales have

been demonstrated as reliable and valid in several samples (Morrell and Cortina‐Borja, 2002).

We used three items from the Setting Limits subscale of the Maternal Cognitions

about Infant Sleep Questionnaire to operationalize mothers’ cognitions about their child’s

sleep described as maternal sleep cognitions within this study (Morrell, 1999). Each item is

measured using a 6-point Likert scale (0 = Strongly agree, 5 = Strongly disagree). The items

were “It is alright to let my child cry at night”, “My child will feel abandoned if I don’t

respond immediately to his/her cries at night” and “I should respond straightaway when my

child wakes crying at night”. Prior to averaging, the second and third items were reverse-

coded, which resulted in higher scores indicating mothers’ who had trouble resisting their

children’s crying at night. In our data, these items demonstrated good internal consistency at

both six (α = .81) and 12 months (α = .76) postpartum.

Data Analyses

We used SPSS version 24 (IBM Corp., 2016) to conduct analyses for hypotheses 1 to

4. To address hypothesis 1, we presented infant sleep outcomes (BISQ) at 6 and 12 months

post-partum for women taking antidepressants and women not antidepressants, accompanied

by inferential tests (F and χ2

tests) to determine the significance of group differences. When

variance between groups differed significantly, we presented the result of the more robust

Welch F test. To address hypothesis 2, we presented Spearman’s correlations to test for

associations between antidepressant dose and sleep outcomes that differ between AD exposed

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

12

versus not exposed. These associations were only tested in the subset of women exposed to

antidepressants and provided maternal and cord blood following labor (n = 42).

To address hypothesis 3, we presented infant sleep location, and maternal perceptions

and behaviors relating to infant sleep for depressed and non-depressed women, comparing

differences inferentially using a series of bivariate logistic regressions. For hypothesis 4, we

conducted a series of factorial (2x2) analysis of variance tests, using the BISQ outcomes for

both postpartum waves, testing the interaction between depressive disorder groups and

breastfeeding.

To address hypothesis 5 that postnatal maternal depression and infant waking were

bidirectional associations even when controlling for antenatal depression and breastfeeding

and hypothesis 6 that postpartum depression and infant sleep will be associated with maternal

cognitions and behaviours we used Mplus version 8 (Muthén and Muthén, 1998-2012) to

analyse these models. The conceptual overall model, including nested models, to be tested are

presented in Figure 1. Figure 2 presents the best fitting model following model testing

analyses. We describe the steps to reach this below.

To undertake this analysis for hypothesis 5 and 6 we first conducted a series of nested

cross-lagged panel models, using repeat measures at first and third trimester, and six and 12

months postpartum. As data were multivariate non-normal, we used maximum likelihood

with robust standard errors (MLR) to estimate the models, with data missing at random

estimated using full-information maximum-likelihood. We then followed the cross-lagged

panel modelling process described by Geiser (2013). Progressively less constrained models

were compared using Satorra’s (Bryant and Satorra, 2012) scaled Chi-square difference tests

and the Akaike Information Criteria (AIC) value. The best fitting model was presented in

Figure 2 to address study hypotheses 5 and 6. We considered Comparative Fit Index (CFI) ≥

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

13

.95, Root Mean Square Error of Approximation (RMSEA) ≤ .05, and Standardized Root

Mean Square (SRMR) ≤ .08 to indicate a model that fit the data well Hu and Bentler (1999).

The nested models presented conceptually in Figure 1 were tested and are described

below. The best fitting model from this process is presented in Figure 2. Model 1 is the

baseline model comprising paths a through e, inclusive; this model formed the basis of

comparison for subsequent less constrained models. In this model, all first- and second-order

autoregressive paths for depressive symptoms scores and first-order autoregressive paths for

cognitions about infant sleep and infant signaling were estimated (a paths). Pathways

between both antepartum depression and depressive symptoms at third trimester with six-

month postpartum infant signaling and cognitions about infant sleep were estimated (b paths).

Cross-sectional pathways between breastfeeding and infant signaling at six and 12 months

postpartum were also estimated (c paths). Finally, covariances were estimated between cross-

sectional measurements to account for contemporaneous associations.

<Figure 1>

In model 2, pathways were estimated originating from depressive symptoms at 6-

months to cognitions about infant sleep, bedtime settling behaviors and infant signaling all at

12 months (f paths). In model 3, pathways were estimated originating from infant signaling at

6-months to depressive symptoms, bedtime settling behaviors and cognitions about infant

sleep all at 12 months (g paths). In model 4, pathways were estimated originating from

cognitions about infant sleep at 6-months to depressive symptoms, bedtime settling behaviors

and infant signaling all at 12 months (h paths). In the final model (Model 5), all cross-lagged

pathways were included (paths f, g and h) to estimate the full cross-lagged model.

Results

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

14

As predicted in our first hypothesis, we expected antidepressant exposure to influence

sleep outcomes (Table 1). Infants differed only at six months postpartum in the average

number of hours they were sleeping at night, such that infants with AD exposure slept for a

significantly shorter period between 7pm and 7am. All other comparisons between the groups

were not significant (p > .01).

<Table 1>

None of the correlations for infant nocturnal sleep duration with sertraline-equivalent

dose, and maternal and cord plasma were significantly different from zero (Supplementary

Table 1). This suggests there is no evidence of a dosage effect of antidepressant exposure on

duration of sleep at night in infants at both 6 and 12 months post-partum.

Regarding maternal depression and infant sleep, depressed women were nearly three

times more likely to be sleeping on the same surface as their infant at six months, compared

to control women (Table 2). At 12 months postpartum, depressed women were twice as likely

to be sleeping in the same room as their child, compared to healthy women. The two groups

did not differ on any of the other measures, including bedtime settling behaviors.

<Table 2>

For the factorial ANOVAs, none of the main effects for the depression groups were

significant and none of the interactions between breastfeeding and depression for infant sleep

outcomes were significant. However, mothers who were still breastfeeding at six months

reported significantly more nocturnal infant signaling at six months (M = 2.28, SD = 1.70)

compared to those mothers who had stopped breastfeeding (M = 1.18, SD = 1.01), F(1, 213)

= 17.29, p < .001. This pattern was the same at 12 months, with mothers still breastfeeding

reporting significantly more infant signaling at night (M = 1.85, SD = 1.50) compared to

those mothers who had stopped breastfeeding (M = 1.12, SD = 1.02), F(1, 175) = 14.88, p <

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

15

.001. None of the other infant sleep outcomes (BISQ) differed between the breastfeeding

groups at either wave.

Table 3 displays zero-order bivariate correlation coefficients and descriptive statistics

for all variables included in the modelling. Table 4 displays the results of the nested cross-

lagged model comparison testing.

<Table 3>

<Table 4>

Model 1 was an adequate fit to the data (CFI = .94, RMSEA = .06 [.04, .08], SRMR =

.08). Adding the crossed effects of depressive symptoms at six months postpartum to infant

signaling, bedtime settling behavior and cognitions about infant sleep all at 12 months (i.e.,

Model 2), did not improve upon the fit of Model 1. This suggests that early postpartum

depressive symptoms are not predictive of later postpartum infant nocturnal signaling.

Adding the crossed effects of infant signaling at six months postpartum to depressive

symptoms, bedtime settling behavior and maternal limit setting cognitions about infant sleep

all at 12 months postpartum (i.e., Model 3), significantly improved the fit of Model 1. So too,

the addition of pathways between cognitions about infant sleep at six months and depressive

symptoms, bedtime settling behavior and infant signaling (i.e., Model 4), improved the fit of

Model 1. Given Model 2 did not improve on the fit of Model 1, the f paths in Figure 1

(online) were not included in the final cross-lagged model (Model 5). Model 5 was a better fit

compared to all other models (CFI = .98, RMSEA = .04 [.00, .06], SRMR = .05), suggesting

that over and above the effects of antepartum depression, depressive symptoms, breastfeeding

and autoregressive paths, early postpartum infant signaling and cognitions about infant sleep

are significantly associated with later postpartum infant signaling, cognitions about infant

sleep and bedtime settling behaviors.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

16

Figure 2 displays significant path coefficients for Model 5. Antepartum depressive

symptoms were only significantly associated with postpartum depressive symptoms via first-

and second-order autoregressive paths. Antepartum depressive disorder measured on the

SCID-IV did not predict any of the postpartum measures, including depressive symptoms. At

six months postpartum, breastfeeding was associated with significantly more infant signaling;

however, by 12 months, the cross-sectional association was not significant. There were no

significant bidirectional associations between depressive symptoms and infant signaling;

therefore, we reject hypothesis 5. Only one non-autoregressive path was a significant

predictor of depressive symptoms: higher maternal sleep cognitions regarding setting infant

sleep limits at six months were associated with higher depressive symptoms at 12 months.

<Figure 2>

Although there were no significant cross-lag paths between postpartum depressive

symptoms and infant signaling in the postpartum, there were significant cross-lag paths

between maternal cognitions about infant sleep and infant signaling. Higher maternal infant

sleep cognitions at six months predicted more infant signaling at 12 months, and more infant

signaling at six months predicted higher maternal sleep cognitions at 12 months. Comparison

of the standardized coefficients suggests that the earlier effect of sleep cognitions on later

signaling (β = .18) is slightly stronger than the earlier effect of infant signaling on later

maternal sleep cognitions (β = .16). More infant signaling and higher sleep cognitions at six

months were both significantly predictive of mothers spending more time engaged in bedtime

infant sleep behaviors at 12 months. As the cross-lagged paths for depressive symptoms were

not included in the final model and six month infant signaling did not directly predict

depressive symptoms, the two-wave mediation analyses were not undertaken. Therefore,

there is no support for hypothesis 6, which predicted mediation between infant signaling and

depressive symptoms through maternal sleep cognitions and bedtime settling behaviors.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

17

Discussion

Our findings do not support a relationship between either earlier maternal depression

predicting later infant sleep outcomes, or earlier increased infant nocturnal signaling

predicting later depression. Maternal cognitions regarding setting limits for infant sleep, such

as concern their baby will feel abandoned if not immediately responded to at night, did

predict depressive symptoms at 12 months of age and these maternal cognitions also

predicted increased nocturnal signaling at 12 months and increased active maternal bedtime

comforting behaviors. Despite the lack of support from the cross-lagged paths between

maternal depression and infant signaling, there was a significant positive contemporaneous

association between maternal depressive symptoms and infant signaling at six months.

Of note, there was no association with fetal antidepressant exposure and sleep

development in infants and this study utilized robust measures of exposure ranging from

prospective self-report and hospital records through to both maternal and cord blood levels to

increase certainty of exposure (Galbally et al., 2017a). Similarly, depression was not

associated with perceived infant sleep problems or settling behaviors using a diagnostic

measure of depression in pregnancy or self-report symptoms across the perinatal period.

However, depression was associated with an increase in likelihood that an infant would share

a bed surface six months and a room at 12 months postpartum. Breastfeeding was associated

with increased infant night signaling at six and 12 months and women who were on

antidepressant medication had infants who slept less overnight, and as previously, reported

also breastfed more frequently, which persisted to 12 months of age (Galbally et al., 2018, in-

press).

Infant sleep development begins in utero and continues across the neonatal, infancy

and early childhood periods. Sleep gradually decreases in quantity but increases in

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

18

consolidated periods following a rhythm across day and night, and sleeping through the night

is an important milestone in developmental maturity (Mirmiran et al., 2003). Research into

sleep development suggest a range of important biological factors, which include gestation at

birth, demand breastfeeding where maternal melatonin contained in milk will match the time

of day for the infant, light/dark exposure, and maternal sleep regulation (Mirmiran et al.,

2003). Beyond these biological factors is also the environmental context that infant sleep

development occurs within including parenting practices and expectations around infant sleep

that can vary across individuals, families and cultures. Therefore, understanding infant sleep

problems needs to be in the context of an understanding of the developmental and changing

nature of infant sleep over the first year of life.

Research into understanding the relationship between infant sleep and maternal

depression has been hampered by measurement and methodological concerns. These include

the definition of infant sleep problems as well as the measurement of maternal depression.

For instance, within a normal sleep cycle for infants, there is brief waking; however, reports

of infant signaling in the night suggests the infant wakes and is unable to re-settle and signals

to the parent, but this is often defined in studies without distinguishing if a baby is waking to

breastfeed or is truly unsettled overnight (Keener et al., 1988). Given the majority of

breastfed babies will continue to feed regularly overnight to at least 6 months of age and may

need to signal to ensure they are fed, the reason for waking is an important consideration

when defining a sleep problem (Galbally et al., 2013; Hörnell et al., 1999). Likewise, settling

to sleep at night is influenced by parenting sleep practices, sleep location and may reflect

aspects of the mother-infant relationship (Beijers et al., 2011; Sadeh et al., 2016; Teti et al.,

2010). Furthermore, an important aspect of understanding the quality of early mother-infant

relationship is the role of maternal depression given the well-characterized effect this may

have on the mother-infant relationship. Yet typically, when maternal depression is included in

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

19

studies of infant sleep, measurement has relied on screening self-report measures in low-risk

populations.

While poor infant sleep is associated with poorer child developmental outcomes,

research that has focused on breastfeeding has not found an association with adverse child

development outcomes (Quinn et al., 2001; Sadeh et al., 2015). It may be that signaling to a

caregiver overnight when it does not impact on sleep quantity may not be detrimental for

developmental outcomes, but represent a relational response to nighttime parenting practices.

While we found breastfeeding was associated with increased night signaling previously, we

have not found an increase in depression in breastfeeding women in this sample (Galbally et

al., 2018, in-press). Furthermore, a focus on interventions aimed at extinguishing infant

night-time signaling may have an unwanted effect of reducing the public health goal of

improving breastfeeding duration. Therefore, studies that examine depression and infant sleep

need to include feeding method and find methods that distinguish infant night waking for

feeding from true infant sleep problems.

Our findings were drawn from Anglophone populations and as such, it was not

possible to explore the cultural diversity of early parenting practices around infant sleep,

which may also be relevant for maternal depression. However, there is evidence that in

cultures where infant sleep practices are significantly different, for instance in Japan where

there are higher rates of bed sharing, both night waking and associated signaling through cry

is common and it is unclear if this may relate to maternal depression (Fukumizu et al., 2005;

Latz et al., 1999). Future research would benefit from replication in other cultural contexts

would be useful to understand if infant sleep practices influence this model. A further

limitation was the reliance on maternal report for infant sleep. This study would be enhanced

by the use of objective sleep measures for both mother and infant.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

20

The strengths of this study include the careful characterization of maternal depression

across pregnancy and the postpartum, and including data on treatment of depression with

antidepressant medication. Infant sleep outcomes and associated parenting behaviors around

infant sleep were also operationalized using several measures, including infant behavior,

parenting settling techniques and maternal cognitions about sleep. Breastfeeding, a common

confounding variable when considering infant sleep, was also controlled for in our models.

This allowed us to untangle the presence of additive effects of both maternal depressive

symptoms and cognitions on infant signaling, beyond normal waking for feeding.

Future research is required to understand how maternal mental health in pregnancy

may also influence these signals such as through maternal cortisol and melatonin in the

development of infant sleep. This could be through fetal programming mechanisms such as

explored in other areas of research on maternal depression and child outcomes (Gentile,

2017; Lewis et al., 2014). Equally though, studies need to examine mental health and key

postpartum variables such as the social, cultural and psychological factors including

breastfeeding, parenting strategies and sleep practices, which may also influence infant sleep.

It is only through this level of sophisticated research that we will understand the important

relationship between maternal depression and infant sleep and most important translate this

knowledge into evidence based guidance, recommendations and interventions for women and

children. Overall, our findings suggest a focus on maternal treatment is important for

understanding infant night waking and further research could examine whether effective

treatment of depression assists infant sleep behaviors and development.

Contributors

MG and AL proposed the original study. MG, AL and DT designed the overall study. SW,

MG and AL contributed to the planning and undertaking the statistical analysis and all

authors contributed to the interpretation of findings. MG and SW prepared the draft of the

paper and all authors critically reviewed and revised for content and gave approval to the

final to be published version of the manuscript.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

21

Role of the Funding Source

The organizations that funded the study did not contribute to any of the following elements of

the paper: study design, the collection, analysis, and interpretation of data, the writing of the

report, and the decision to submit the paper for publication.

Conflict of Interest

MG has previously received honorarium for speaking from Lundbeck. The other authors

declare that they have no competing interests.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

22

Acknowledgements

The authors would like to thank those who have both supported and given advice in the

development of MPEWS including: Marinus van IJzendoorn, Michael Permezel, Anne Buist,

Philip Boyce, Marie-Paule Austin, Peter Fonagy, Robert Emde, Marian Bakermans-

Kranenburg and Michelle Leech. The authors also thank the staff and students on the study

and research coordinators: Tina Vaiano and Nicole Brooks and volunteer Susan Pitchford for

their contribution to MPEWS. Thank you also to Carole Branch at Mercy Health Human

Research Ethics Committee. We are also sincerely grateful to the study participants who have

contributed a substantial amount of time to participating in this study.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

23

References

Andreassen, T.N., Falch, B.M., Spigset, O., 2015. A UPLC-MSMS method for the analysis of olanzapine in serum-with particular emphasis on drug stability testing. Journal of Chromatography B Analytical Technologies in the Biomedical and Life Sciences 1006, 112-120. Beijers, R., Jansen, J., Riksen-Walraven, M., de Weerth, C., 2011. Attachment and infant night waking: A longitudinal study from birth through the first year of life. Journal of Developmental & Behavioral Pediatrics 32, 635-643. Boyce, P., Stubbs, J., Todd, A., 1993. The Edinburgh Postnatal Depression Scale: Validation for an Australian Sample. Australasian Psychiatry 27, 472-476. Bryant, F.B., Satorra, A., 2012. Principles and Practice of Scaled Difference Chi-Square Testing. Structural Equation Modeling: A Multidisciplinary Journal 19, 372-398. Cox, J.L., Holden, J.M., Sagovsky, R., 1987. Detection of postnatal depression. Development of the 10-item Edinburgh Postnatal Depression Scale. The British journal of psychiatry 150, 782-786. Dennis, C.L., Ross, L., 2005. Relationships among infant sleep patterns, maternal fatigue, and development of depressive symptomatology. Birth 32, 187-193. Dias, C.C., Figueiredo, B., 2015. Breastfeeding and depression: a systematic review of the literature. Journal of Affective Disorders 171, 142-154. Douglas, P.S., Hill, P.S., 2013. Behavioral sleep interventions in the first six months of life do not improve outcomes for mothers or infants: a systematic review. Journal of Developmental & Behavioral Pediatrics 34, 497-507. Field, T., Diego, M., Hernandez-Reif, M., Figueiredo, B., Schanberg, S., Kuhn, C., 2007. Sleep disturbances in depressed pregnant women and their newborns. Infant Behavior and Development 30, 127-133. First, M., Spitzer, R., Gibbon, M., Williams, J., 1997. Stuructured Clinical Interview for DSM-IV Axis I Disorders (SCID-I), Clinical Version. Washington DC and London. American Psychiatric Press, Inc. Fisher, J., Feekery, C., Rowe, H., 2004. Treatment of maternal mood disorder and infant behaviour disturbance in an Australian private mothercraft unit: a follow-up study. Archives of Women's Mental Health 7, 89-93. Fukumizu, M., Kaga, M., Kohyama, J., Hayes, M.J., 2005. Sleep-related nighttime crying (yonaki) in Japan: a community-based study. Pediatrics 115, 217-224. Galbally, M., Lewis, A.J., McEgan, K., Scalzo, K., Islam, F.A., 2013. Breastfeeding and infant sleep patterns: an Australian population study. Journal of Paediatric Child Health 49, E147-152. Galbally, M., Spigset, O., Johnson, A.R., Kohan, R., Lappas, M., Lewis, A.J., 2017a. Neonatal adaptation following intrauterine antidepressant exposure: Assessment, drug assay levels and infant development outcomes. Pediatric Research. Galbally, M., van, I.M., Permezel, M., Saffery, R., Lappas, M., Ryan, J., van Rossum, E., Johnson, A.R., Teti, D., Lewis, A.J., 2017b. Mercy Pregnancy and Emotional Well-being Study (MPEWS): Understanding maternal mental health, fetal programming and child development. Study design and cohort profile. International Journal of Methods in Psychiatric Research. Galbally, M., Watson, S., Ball, H., Lewis, A.J., 2018, in-press. Breastfeeding, Antidepressants and Depression in the Mercy Pregnancy and Emotional Wellbeing Study. Journal of Human Lactation. Geiser, C., 2013. Data Analysis with Mplus. The Guildford Press, New York, NY. Gentile, S., 2017. Untreated depression during pregnancy: Short-and long-term effects in offspring. A systematic review. Neuroscience 342, 154-166. Hiemke, C., Baumann, P., Bergemann, N., Conca, A., Dietmaier, O., Egberts, K., Fric, M., Gerlach, M., Greiner, C., Gründer, G., Haen, E., Havemann-Reinecke, U., Jaquenoud Sirot, E., Kirchherr, H., Laux, G., Lutz, U.C., Messer, T., Müller, M.J., Pfuhlmann, B., Rambeck, B., Riederer, P., Schoppek, B., Stingl, J., Uhr, M., Ulrich, S., Waschgler, R., Zernig, G., 2011. AGNP Consensus Guidelines for Therapeutic Drug Monitoring in Psychiatry: Update 2011. Pharmacopsychiatry 21, 195-235.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

24

Hiscock, H., Cook, F., Bayer, J., Le, H.N., Mensah, F., Cann, W., Symon, B., St James-Roberts, I., 2014. Preventing early infant sleep and crying problems and postnatal depression: a randomized trial. Pediatrics 133, e346-354. Hiscock, H., Wake, M., 2001. Infant sleep problems and postnatal depression: a community-based study. Pediatrics 107, 1317-1322. Hörnell, A., Aarts, C., Kylberg, E., Hofvander, Y., Gebre‐Medhin, M., 1999. Breastfeeding patterns in exclusively breastfed infants: a longitudinal prospective study in Uppsala, Sweden. Acta paediatrica 88, 203-211. Hu, L.t., Bentler, P.M., 1999. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal 6, 1-55. IBM Corp., 2016. IBM SPSS Statistics for Windows, 24 ed. IBM Cop., Armonk, NY. Jindal, R.D., Friedman, E.S., Berman, S.R., Fasiczka, A.L., Howland, R.H., Thase, M.E., 2003. Effects of sertraline on sleep architecture in patients with depression. Journal of clinical psychopharmacology 23, 540-548. Keener, M.A., Zeanah, C.H., Anders, T.F., 1988. Infant temperament, sleep organization, and nighttime parental interventions. Pediatrics 81, 762-771. Kempler, L., Sharpe, L., Miller, C.B., Bartlett, D.J., 2016. Do psychosocial sleep interventions improve infant sleep or maternal mood in the postnatal period? A systematic review and meta-analysis of randomised controlled trials. Sleep medicine reviews 29, 15-22. Lam, P., Hiscock, H., Wake, M., 2003. Outcomes of infant sleep problems: a longitudinal study of sleep, behavior, and maternal well-being. Pediatrics 111, e203-207. Latz, S., Wolf, A.W., Lozoff, B., 1999. Cosleeping in context: sleep practices and problems in young children in Japan and the United States. Archives of pediatrics & adolescent medicine 153, 339-346. Lewis, A.J., Galbally, M., Gannon, T., Symeonides, C., 2014. Early life programming as a target for prevention of child and adolescent mental disorders. BMC medicine 12, 33. Mirmiran, M., Maas, Y.G., Ariagno, R.L., 2003. Development of fetal and neonatal sleep and circadian rhythms. Sleep Medicine Reviews 7, 321-334. Morrell, J., Cortina‐Borja, M., 2002. The developmental change in strategies parents employ to settle young children to sleep, and their relationship to infant sleeping problems, as assessed by a new questionnaire: The Parental Interactive Bedtime Behaviour Scale. Infant and Child Development 11, 17-41. Morrell, J.M.B., 1999. The Role of Maternal Cognitions in Infant Sleep Problems as Assessed by a New Instrument, the Maternal Cognitions about Infant Sleep Questionnaire. The Journal of Child Psychology and Psychiatry and Allied Disciplines 40, 247-258. Mulder, E.J., Ververs, F.F., De Heus, R., Visser, G.H., 2011. Selective serotonin reuptake inhibitors affect neurobehavioral development in the human fetus. Neuropsychopharmacology 36, 1961. Muthén, L.K., Muthén, B.O., 1998-2012. Mplus User's Guide, 7th ed. Muthén & Muthén, Los Angeles, CA. Nevarez, M.D., Rifas-Shiman, S.L., Kleinman, K.P., Gillman, M.W., Taveras, E.M., 2010. Associations of early life risk factors with infant sleep duration. Academic pediatrics 10, 187-193. Procyshyn, R.M., Bezchlibnyk-Butler, K.Z., Jeffries, J.J., 2015. Clinical handbook of psychotropic drugs. Hogrefe Publishing. Quinn, P.J., O'Callaghan, M., Williams, G.M., Najman, J.M., Andersen, M., Bor, W., 2001. The effect of breastfeeding on child development at 5 years: a cohort study. Journal of paediatrics and child health 37, 465-469. Reis, M., Aamo, T., Spigset, O., Ahlner, J., 2009. Serum concentrations of antidepressant drugs in a naturalistic setting: compilation based on a large therapeutic drug monitoring database. Therapeutic Drug Monitoring 31, 42-56. Sadeh, A., 2004. A brief screening questionnaire for infant sleep problems: validation and findings for an Internet sample. Pediatrics 113, e570-577.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

25

Sadeh, A., De Marcas, G., Guri, Y., Berger, A., Tikotzky, L., Bar-Haim, Y., 2015. Infant sleep predicts attention regulation and behavior problems at 3–4 years of age. Developmental Neuropsychology 40, 122-137. Sadeh, A., Juda‐Hanael, M., Livne‐Karp, E., Kahn, M., Tikotzky, L., Anders, T.F., Calkins, S., Sivan, Y., 2016. Low parental tolerance for infant crying: an underlying factor in infant sleep problems? Journal of Sleep Research 25, 501-507. Teti, D., Crosby, B., 2012. Maternal Depressive Symptoms, Dysfunctional Cognitions, and Infant Night Waking: The Role of Maternal Nighttime Behavior. Child Development 83, 939-953. Teti, D., Kim, B.R., Mayer, G., Countermine, M., 2010. Maternal emotional availability at bedtime predicts infant sleep quality. Journal of Family Psychology 24, 307-315. Ververs, T., Kaasenbrood, H., Visser, G., Schobben, F., de Jong-van den Berg, L., Egberts, T., 2006. Prevalence and patterns of antidepressant drug use during pregnancy. European journal of clinical pharmacology 62, 863-870. Zeskind, P.S., Stephens, L.E., 2004. Maternal selective serotonin reuptake inhibitor use during pregnancy and newborn neurobehavior. Pediatrics 113, 368-375.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

26

Figure 1. Full cross-lagged panel model.

Figure 2. Model 5 cross-lag panel model results (N = 254). Greyed-out, dashed lines denote

non-significant paths. Percentages represent explained variance (i.e., R2). *p < .05 **p < .01

***p < .001.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

27

Table 1. Mean (SD) differences for infant sleep outcomes (BISQ) at 6 and 12 months postpartum

between mothers not exposed to antidepressants and mothers exposed to antidepressants (N = 264).

Not Exposed

to AD

(n = 217)

Exposed to

AD

(n = 47) F p

BISQ at 6 Months Postpartum

Nocturnal sleep duration (hours) 10.08 (1.34) 9.52 (1.64) 5.43 .021

Daytime sleep duration (hours) 3.60 (1.51) 3.82 (1.36) 0.72 .397

Nocturnal infant signalling (frequency) 1.98 (1.44) 2.44 (2.17) 1.67† .203

Duration of night wakefulness (minutes) 38.04 (43.77) 52.61 (62.83) 3.11 .079

Nocturnal sleep onset (minutes) 28.14 (29.49) 31.10 (30.93) 0.34 .560

BISQ at 12 Months Postpartum

Nocturnal sleep duration (hours) 10.50 (1.24) 10.16 (1.74) 2.05 .154

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

28

Daytime sleep duration (hours) 2.51 (1.23) 2.28 (.85) 1.14 .286

Nocturnal infant signalling (frequency) 1.46 (1.45) 1.50 (.96) 0.05† .822

Duration of night wakefulness (minutes) 33.16 (78.36) 31.78 (35.93) 0.01 .922

Nocturnal sleep onset (minutes) 21.39 (16.52) 25.62 (25.49) 0.86† .359

† Welch F Test due to significant heterogeneity of group variances.

Table 2. Infant sleep outcomes (BISQ; n, %) and bedtime settling behaviours (PIBBS; M,

SD) between depressed and non-depressed mothers (diagnosed using the SCID-IV; N = 264).

Not Depressed

(n = 206)

Depressed

(n = 58)

n (%) n (%) OR [95% CIs]

Infant Sleep Arrangements

Sharing surface at 6 months 20 (10.8) 11 (21.6) 2.83 [1.01, 5.14]

Sharing surface at 12 months 44 (23.7) 12 (23.5) 1.72 [.77, 3.82]

Sleeping in parents' room at 6

months

64 (34.4) 23 (45.1) 1.57 [.84, 2.94]

Sleeping in parents' room at 12

months

43 (25.9) 17 (42.5) 2.11 [1.03, 4.33]

Perceived Infant Sleep Problem at 6

months

86 (45.7) 22 (42.3) .87 [.47, 1.62]

Perceived Infant Sleep Problem at 12

months

88 (50.0) 20 (47.6) .91 [.46, 1.78]

Sought Help for Infant Sleep Problem at

12 months (%)

54 (30.7) 13 (31.0) 1.01 [.49, 2.10]

Use Bedtime Sleep Strategy at 12 72 (41.1) 18 (43.9) 1.12 [.56, 2.22]

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

29

months (%)

M (SD) M (SD) F (p-value)

PIBBS Total at 12 Months 30.05 (8.87) 28.57 (7.50) 0.349

Active Physical Comforting 20.86 (14.34) 20.94 (13.50) 0.974

Encourage Autonomy 18.91 (14.11) 17.31 (12.88) 0.517

Settle by Movement 11.56 (17.10) 8.13 (14.86) 0.243

Passive Physical Comforting 14.97 (17.53) 9.94 (15.22) 0.099

Social Comforting 22.11 (16.28) 23.28 (16.81) 0.684

Note. BISQ, Brief Infant Sleep Questionnaire; PIBBS, Parent Interactive Bedtime Behaviours Scale;

SCID-IV; Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental

Disorders, Fourth Edition.

Table 3. Descriptive statistics and zero-order bivariate correlations between all variables

included in the cross-lagged models (N = 264)

1 2 3 4 5 6 7 8 9 10 11

1

2

1. SCID-IV

Depressiona

(

W

1) -

2. Depressive

Symptoms

(

W

1)

.39

**

* -

3. Depressive

Symptoms

(

W

2)

.34

**

*

.67

**

* -

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

30

4. Depressive

Symptoms

(

W

4)

.28

**

*

.50

**

*

.56

**

* -

5. Depressive

Symptoms

(

W

5)

.30

**

*

.53

**

*

.59

**

*

.68

**

* -

6. Cognitive

Limit-setting

(

W

4)

-

.09

-

.08

-

.07 .01

-

.17

* -

7. Cognitive

Limit-setting

(

W

5)

-

.11

-

.12

-

.10

-

.05

-

.17

*

.68

*** -

8. Bedtime

Settling

Behaviours

(

W

5)

-

.04 .06 .03 .11

.15

*

-

.23

**

-

.19

** -

9. Nocturnal

Infant

Signalling

(

W

4) .05 .00 .04

.17

* .13

-

.19

**

-

.29

*** .22** -

10. Nocturnal

Infant

Signalling

(

W

5)

-

.07 .03

-

.04

.15

*

.17

*

-

.30

***

-

.32

***

.31**

*

.60

**

* -

11.

Breastfeedinga

(

W

4)

-

.05

-

.11

-

.14

*

-

.03

-

.19

** -.02 -.08 .19*

.24

**

*

.1

3 -

12. ( - .03 0.0 .02 - -.06 - .21** .27 .2 .50 -

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

31

Breastfeedinga W

5)

.10 1 .04 .16

*

**

*

5*

*

**

*

Mean

.22

6.5

5

6.3

5

5.9

5

6.4

6

3.0

0

3.3

9 29.91

2.0

4

1.

41 .76

.

3

7

Standard

Deviation

.41

4.7

4

4.5

3

4.5

5

4.8

9

1.2

4

1.2

4 8.65

1.6

0

1.

29 .43

.

4

8

Range

0-1

0-

27

0-

25

0-

24

0-

25 1-6 1-6

13.33-

65.83

0-

10

0-

7 0-1

0

-

1

bCorrelations with SCID-IV Depression and Breastfeeding Cessation are point-biserial

coefficients.

*p < .05 **p < .01 ***p < .001.

Table 4. Comparison of nested cross-lagged models

Model χ2 d.f.

Satorra-

Bentler

Scaled

Correction

Factor

Model

Comparison

Satorra-

Bentler

Scaled

Δχ2 Δd.f. AIC

M1: No Cross-lagged

Effects

75.07 37 1.03

8340.84

M2:Depressive

symptoms effects

72.12 34 1.02 M1 vs. M2 2.58 3 8343.19

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIP

T

32

M3: Infant Signalling

Effects

57.94 34 1.03 M1 vs. M3 16.63** 3 8329.52

M4: Infant Sleep

Cognition Effects

54.85 34 1.03 M1 vs. M4 19.63*** 3 8325.83

M5: Full Cross-lagged

Model

41.39 31 1.03 M1 vs. M5 32.74*** 3 8318.19

M2 vs. M5 33.60*** 3

M3 vs. M5 16.08** 3

M4 vs. M5 13.08** 3

**p < .01 ***p < .001