Thinking and doing : the effects of dopamine and oxytocin genes and executive function on mothering behaviours

Thinking and doing: the effects of dopamine and

oxytocin genes and executive function on mothering behaviours

K. Tombeau Cost

, E. Unternaehrer

, A. Plamondon

, M. Steiner

,

M. Meaney

, L. Atkinson

,

J. L. Kennedy

, A. S. Fleming

and on behalf of the MAVAN Research Team

†Department of Psychology, University of Toronto Mississauga, Mississauga,^‡Fraser Mustard Institute for Human

Development, Toronto,^§Ludmer Centre for Neuroinformatics &

Mental Health,^¶Douglas Mental Health University, Institute of McGill University, Montreal,^∗∗Department of Educational Fundamentals and Practices, Laval University, Quebec,

††Department of Psychiatry and Behavioral Neurosciences, McMaster University, Hamilton,^‡‡Sackler Program for Epigenetics & Psychobiology at McGill University, Montreal, Canada,^§§Singapore Institute for Clinical Sciences, Singapore,

¶¶Department of Psychology, Ryerson University, and^∗∗∗Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada

*Corresponding author: K. Tombeau Cost, Department of Psy- chology, University of Toronto Mississauga, Mississauga, ON, Canada. E-mail: katherine.cost@utoronto.ca

Animal and human studies suggest that initial expres- sion of maternal behaviour depends on oxytocin and dopamine systems. However, the mechanism by which these systems affect parenting behaviours and the timing of these effects are not well understood. This article explores the role of mothers’ executive func- tion in mediating the relation between oxytocin and dopamine gene variants and maternal responsive- ness at 48 months post-partum. Participants (n=157) were mothers recruited in the Maternal Adversity, Vulnerability and Neurodevelopment Study, which assesses longitudinally two cohorts of mothers and children in Canada. We examined single nucleotide polymorphisms (SNPs) related to the dopamine and oxytocin systems (DRD1 rs686, DRD1 rs265976, OXTR rs237885 and OXTR rs2254298), assessed mothers’

decision-making at 48 months using the Cambridge Neurological Automated Testing Battery (CANTAB) and evaluated maternal responsiveness from videotaped interactions during the Etch-A-Sketch co-operation task. Mediation analyses showed thatOXTR rs2254298 A-carriers had an indirect effect on positive parent- ing which was mediated by mothers’ performance on decision-making task (estimate=0.115,P<0.005), while OXTR rs2254298 A-carriers had both direct and indirect

effects on physically controlling parenting, also medi- ated through enhanced performance on decision-making (estimate= −0.059,P<0.005). Dopamine SNPs were not associated with any measure of executive function or parenting (all P>0.05). While oxytocin has previ- ously been associated with only the early onset of maternal behaviour, we show that an OXTR polymor- phism is involved in maternal behaviour at 48 months post-partum through mothers’ executive function. This research highlights the importance of the oxytocin system to maternal parenting beyond infancy.

Keywords: CANTAB, decision-making, dopamine, Etch-A-S- ketch, executive functions, mediation, oxytocin, physically controlling parenting, positive parenting, SNP

Mothering is a dynamic and complex process, relying on myriad skills and multiple systems interacting as the mother develops and hones parenting skills to meet the chal- lenges inherent to child development. Much research has been conducted to understand the biological systems that underlie adaptive and sensitive parenting. While animal and human studies suggest that some components of maternal behaviour depend on the oxytocin and dopamine systems, the neurocognitive mechanisms linking these biological systems with maternal behaviour are not fully understood.

Oxytocin is a neuropeptide that was first characterized for its effects on parturition and milk ejection; more recently, it has been implicated in the expression of maternal behaviour in rats, sheep, rhesus macaques and humans (Feldman et al.2012). The oxytocin system, as quantified by salivary and plasma levels of the peptide and single nucleotide polymorphisms (SNPs) in both the genes for the ligand and for the receptor, has also been associated with social and emotional behaviours (Feldman et al. 2012). Several reports have focused on the effects of oxytocin ligand gene polymorphisms on maternal behaviour in the early life of the infant, finding associations between oxytocin SNPs and infant-directed speech and instrumental care (Jonas et al.

2013; Mileva-Seitzet al.2013). Oxytocin SNPs also associate with breastfeeding duration (Jonaset al.2013; Mileva-Seitz et al. 2013). Similarly, SNPs in the oxytocin receptor gene are also implicated in maternal behaviour, associating with affectionate maternal touch (Apter-Levyet al.2013), maternal depression (Apter-Levy et al.2013; Mendlewiczet al.2012) and quality of early maternal care (Feldman et al. 2013).

285 Erschienen in: Genes, brain, and behavior ; 16 (2017), 2. - S. 285-295

https://dx.doi.org/10.1111/gbb.12337

While the connection between early parental behaviours and oxytocin may be well established, the extent to which differences in oxytocin system function affects differences in later parenting has not been resolved.

The animal literature has broadly showed the role of oxytocin in the initiation of maternal behaviours. In rats (Fahrbachet al.1985; Pedersenet al.1982, 1985, 1994; van Leengoedet al.1987), sheep (as reviewed by Kendricket al.

1997) and mice (Richet al.2014), the initiation of maternal behaviour depends on the functional central oxytonergic system. Although many studies have shown that oxytocin is not necessary for the maintenance of maternal behaviour (Carter et al.1997; Fahrbachet al.1985; Numan & Corodi- mas 1985; Numan et al.2006), some work suggests that oxytocin is still important in regulating some aspects of later maternal behaviour. Pedersen and Boccia (2003) found that in rats, even after the establishment of maternal behaviour, infusions of an oxytocin antagonist reduced licking and grooming of pups, kyphotic nursing and high-arched nursing.

A study on mice also found that oxytocin was required for nursing, but that no differences between oxytocin knockout mice and wild type in any other maternal behaviour were apparent (Nishimoriet al.1996). While most animal literature agrees that oxytocin is involved in the initiation but not in the maintenance of maternal behaviours, there is abundant evidence from antagonist or genetic manipulation studies that dopamine is involved in both initiation and maintenance of maternal behaviours in model species (as reviewed by Bridges 2015). Despite these studies being conducted in animals, there are similarities in the physiological processes of the initiation and maintenance of maternal behaviour between model species and humans. In addition, model species and humans are both influenced by environmen- tal and experiential factors in the expression of maternal behaviours (as reviewed by Lonsteinet al.2015).

The dopaminergic system and polymorphisms in dopamine ligands and receptors have been implicated in impulsivity, addiction and gambling (Calduet al.2007; Chenet al.2011;

Comingset al.1997; Seegeret al.2004; Smithet al.2008;

Stice et al.2008; Stynet al.2009). Intriguingly, these psy- chopathologies involve reward circuits, which might have evolved to support social attachment and parent–child bond- ing (Insel 2003). Indeed, past research has already shown a link between dopamine polymorphisms and maternal behaviour in the first 6 months of life involving neurocognitive mechanisms (Mileva-Seitz et al. 2012). Specifically, mater- nal attention was associated with two dopamine receptor (DRD1) SNPs. Alleles in these two SNPs were associated with a decrease in the amount of time mothers spent attend- ing to their infants. The DRD1 might influence attention towards the infant by modulating infant salience. Moreover, genes of the dopaminergic system might also influence maternal sensitivity. The SNPs in both genes encoding for COMT and DRD4 are associated with the efficiency of dopamine transmission and with decreased maternal sensitivity in mothers with high levels of daily hassles (Lee et al. 2010; van IJzendoorn et al. 2008). Mothers with a 7-repeat DRD4 variant had increased sensitivity to fussy infants compared with mothers without the 7-repeat variant DRD4. Furthermore, differences in maternal vocalization and

speaking to the infant were associated with two dopamine receptor (DRD2) haplotypes (Kaitzet al.2010; Mileva-Seitz et al. 2012). The frequency of maternal verbal commands was also associated with dopamine transporter (DAT) poly- morphisms (Lee et al.2010). While previous research has focused on the effects of dopamine system polymorphisms on maternal behaviour in the early life of the infant, little is known about how dopamine polymorphisms might affect maternal cognition and behaviour as the child becomes more mobile, fluent and independent. As the child develops new skills and abilities, the mother must also develop her parenting skills and abilities in parallel with the child.

Previous studies on the neurocognitive processes associ- ated with maternal sensitivity and maternal cognition have shown a link between attention and spatial working memory tasks and maternal responsiveness to their 2–6-month-old infants (Atkinson et al. 2009; Gonzalez et al. 2012). Fur- thermore, teen mothers with poorer performance on a cognitive flexibility task also showed lower maternal sensitiv- ity (Chicoet al.2014). Adult mothers with greater cognitive flexibility, specifically the ability to switch attention and change strategies, displayed higher sensitivity towards their 2–6-month-old infants (Gonzalez et al. 2009). Additionally, we have found substantial associations between maternal sensitivity at 3–18 months and spatial working memory, cognitive flexibility and decision-making (A. Plamondonet al.

submitted). Executive functions may represent a possible mechanism through which the oxytocin and dopamine genes affect maternal behaviours. Quality parenting neces- sarily involves selective attention, sensitivity and synchrony, processes that are all controlled by the mother and require attention, cognitive flexibility and decision-making (Crandall et al. 2015; Deater-Deckard et al. 2010, 2012; Rutherford et al. 2015). Executive functions, as a global construct of top-down control of inhibition, working memory, planning and problem-solving, are involved in all aspects of daily living, whether achievement-oriented, health-related or in social interactions (as reviewed by Diamond 2013). Given the importance of executive functions in social contexts, the evidence from animal studies that the maintenance of maternal behaviours is less dependent on hormones, and in particular the demands associated with child development of self-regulation and autonomy for a parent (Bridgettet al.

2015; Changet al.2014; Cuevaset al.2014), we expect the relationship between maternal responsiveness and mater- nal executive function to become more salient during the offspring’s childhood as compared with infancy.

Both dopamine and oxytocin have been implicated in cog- nitive functions. Dopamine activity in the prefrontal cortex is critical for executive function, attention and cognitive flexibil- ity (Afonsoet al.2007; Arnsten 2006; Goldman-Rakic 1998).

Attention in humans is central to maternal sensitivity, and the logical question yet to be addressed is whether genetic pre- dictors of human attentional processes are also associated with differences in maternal behaviour and responsiveness.

Although not conducted with post-partum populations, poly- morphisms in several dopamine receptor genes have been associated with differences in cognitive tasks, working mem- ory and attentional function and/or in prefrontal activation (Bobb et al. 2005; Brookes et al. 2006; Caldu et al. 2007;

Lasky-Suet al.2007; Miseneret al.2004; Ribaseset al.2012;

Wilkosc et al.2010). While dopamine has been associated with executive functions, oxytocin has been related to mem- ory. Several studies have reported that oxytocin suppresses some types of memory in both rats (de Wied 1980; van Ree et al.1978) and humans (Heinrichs et al. 2004; Herzmann et al.2012; Wirth 2015). Furthermore, the likelihood that oxy- tocin, due to its evolutionarily ancient origins and its presence in non-social species, is involved in cognition apart from social interaction and/or cognition as a foundation of social interac- tion has been previously developed as a theory to explain the mnemonic effects of oxytocin (Wirth 2015). One study considered oxytocin system polymorphisms in the context of cognitive function and memory, Skuseet al.(2014) reported A-carriers ofOXTRrs237887 had impaired recognition mem- ory compared with non-A-carriers. In addition, dopamine and oxytocin are good candidate systems in the study of cognitive processes underlying maternal behaviours.

The current study

The current study aims to understand the relation- ship between dopamine and oxytocin SNPs, executive functions such as cognitive flexibility, working memory, decision-making, motor inhibition and maternal behaviour at 48 months post-partum. We first predicted that one of the mechanisms through which oxytocin and/or dopamine genes would affect maternal behaviour is through enhancing moth- ers’ executive function. We narrowed the genes of interest to the oxytocin receptor gene, known for its relation to plasma oxytocin levels, maternal behaviours and social cognition (Bakermans-Kranenburg & van IJzendoorn 2008; Feldman et al.2012; Jonaset al.2013; Luchtet al.2013; Masseyet al.

2015; Mileva-Seitz et al.2013; Park et al.2010; Slane et al.

2014; Wu et al. 2012) and the dopamine receptor 1 gene known for its relation to maternal orienting away, visuospa- tial working memory in human females, brain metabolism, executive functioning (Bobbet al.2005; Miseneret al.2004;

Ribaseset al.2012; Wilkoscet al.2010) and attention-deficit (hyperactivity) disorder (Bobbet al.2005; Mileva-Seitzet al.

2012; Miseneret al.2004; Ribaseset al.2012; Wilkoscet al.

2010). We then predicted that mothers carrying variants on the OXTR gene (rs2254298 and rs237885) that have been associated with increased empathy, positive mater- nal behaviours and increased plasma oxytocin, and DRD1 variants (DRD1rs686 andDRD1rs265976) that have been associated with less maternal orienting away, would show greater maternal flexibility and more attentiveness towards their children in a contingent and sensitive manner.

Materials and methods

Participants were part of the Maternal Adversity, Vulnerability and Neurodevelopment (MAVAN) Study, which longitudinally follows two cohorts of mothers and children in Hamilton, Ontario and Montreal, Quebec in Canada. The data in the primary analyses were restricted to the Hamilton sample, in which maternal cognitive functions were

assessed. Genetic data were available for 157 participants. A partial replication was performed with the Montreal sample.

Age of mother at delivery and SES

Demographic characteristics, such as age at delivery and socio-economic status (SES), were collected during the second trimester of pregnancy with the health and well-being of mothers and their newborns questionnaire, a composite questionnaire includ- ing short validated versions of a variety of questionnaires (O’Donnell et al.2014). The SES variable was calculated to include both income and education, with two groups represented, based on national census data (Statistics Canada 2008): 1=low SES, low education or low SES and high education or high SES, low education; 2=high SES, high education. Based on the national census data, low SES is defined as<$21 358 total family income after tax, while high SES is defined as>$21 359 total family income after tax. Low education is defined as less than high school graduation (Statistics Canada 2011;

Wendlandet al.2014).

Edinburgh Postnatal Depression Scale

The Edinburgh Postnatal Depression Scale (EPDS) is a questionnaire with 10 items to screen for post-partum depression. The EPDS has high sensitivity (86%) and high specificity (78%) (Coxet al. 1987) and is considered the gold standard to assess maternal symptoms of depression during pregnancy and the early post-partum. Participants completed the EPDS during the second trimester and at 3, 6, 12 and 18 months post-partum. Because of the stability of the scores across time (r=0.585–0.774,n=191,P<0.001), the five scores were used to compute a mean maternal depression score.

Genotyping

Participants’ DNA was collected by using buccal swabs (Epicentre, Madison, Wisconsin, U.S.A.). We selected SNPs based on the liter- ature on executive function and on parenting behaviours. The geno- types were determined using the Taqman assay on the ABI Prism 7000 (Applied Biosystems, Foster City, CA, USA). We tested the tagged SNPs in the dopamine receptor (DRD1 rs686 andDRD1 rs265976) and oxytocin receptor (OXTRrs237885;OXTRrs2254298, both in intron 3). Genotypes were tested for Hardy–Weinberg equi- librium (HWE; calculator for two alleles) and linkage disequilibrium (LD) within theOXTRandDRD1genes (HaploReg v2). A total of 10%

of samples were regenotyped. All data were checked for clustering, assay quality and blanks. Poorly performing SNPs/assays were dis- qualified. Our sample has over 80% power to detect anR² of as low as 0.048, 0.042, 0.037 and 0.041, respectively, for rs2254298, rs237885, rs686 and rs265976 [taking into account their correspond- ing minor allele frequencies (MAFs), sample sizes,𝛼of 0.05, additive genetic model and using a two-tailed analysis to detect effects in either direction] (Gauderman & Morrison 2006).

Executive function

Executive function at 48 months was assessed with the Cambridge Neuropsychological Automated Testing Battery (CANTAB; Cambridge Cognition, Ltd., Cambridge, UK). Four tests were administered on a touch-screen computer:

Motor inhibition was tested with the stop-signal task. Scores were computed as an average of the standardized scores on the proportion of time the subject was able to successfully stop and the stop-signal reaction time, which measured the average length of time the subject was able to inhibit the response in 50% of the trials. Higher scores indicate better motor inhibition.

Cognitive flexibility was tested with the intra/extra-dimensional task. Scores were obtained by reversing the standardized number of errors on the extra-dimensional shift condition. Higher scores indicate higher cognitive flexibility.

Working memory was tested with the spatial working memory task. Scores were computed as averages of the reversed standard- ized scores for the number of times a search strategy was used and the number of errors (visiting the same square twice within or

between searches). Higher scores indicate better usage of working memory strategies and better working memory.

Decision-making was tested with the information sampling task.

Boxes opened is a measure of information gathering prior to making a decision and is the averaged standardized score on the number of boxes opened in both a fixed and decreasing points conditions.

Higher scores on boxes opened indicate more information gathering and thereby better decision-making. Henceforth, decision-making ability is reflected in number of boxes opened.

Observed parenting behaviours

Behaviour was observed at 48 months post-partum while the mother–child dyad engaged with the Etch-A-Sketch task, which necessitates mother and child co-operation to reproduce a series of five images, increasing in difficulty. This task is useful to observe parent–child interaction in a challenging, co-operative and goal-directed activity. The mother was instructed to use only the left knob (to make horizontal lines) while the child was instructed to use only the right knob (to make vertical lines). The interaction was recorded and coded for the observed behaviours of the mother and the child on 16 scales. The scores were standardized and subjected to principle components factor analysis applying varimax rotations into four reliable factors: positive parenting, physically controlling behaviour, parental negativity and parental instruction, all with eigen- values>1. Parental instruction had low intraclass correlations for both intrarater and interrater reliability. Neither parental instruction nor parental negativity had good validity, largely because the items that comprised the factors were more reflective of mothers’ general mood and not to the interactions or parentingper se. In contrast, the factors, positive parenting and physically controlling parenting, had good face validity and good cohesion. To establish the latter, we computed a mean score on the standardized scales loading on the respective factor. Both factors had good cohesion as measured by Cronbach’s alpha: positive parenting (𝛼=0.757) explained 16.81% of the variance and physically controlling parenting (𝛼=0.531) explained 18.05% of the variance. The intraclass correlations for the measures included had high intrarater reliability (0.77–1.00) and higher interrater reliability (0.71–1.00).

Positive parenting

Positive parenting includes parent’s verbal personalized approval of the child, verbal approval and encouragement of the child’s actions, the extent to which the parent engages with the child to complete the task through looking, listening and reciprocity, the general atmosphere of the parent–child interaction, parental attune- ment to provide an atmosphere to guide the child to take on his/her own role in the task. Higher scores indicate more positive parenting.

Physically controlling

Physically controlling includes parent’s restraint of the child, such as physically restraining the child’s actions, taking over the Etch-A-Sketch or touching the right knob, as well as parent’s rough-handling such as smacking, forcefully pulling or pushing or roughly removing things from the child. Higher scores indicate more physical control.

Statistical analysis

In the first step, data were visually inspected for normality distribution and outliers. Then, we considered correlations between variables to construct a model. We then examined the four selected SNPs for OXTRandDRD1for each measure of executive function in a stepwise multiple regression model adjusting for basic demographics (SES and maternal age) and maternal depression. The SNPs that were significant after Bonferroni correction for multiple testing were investigated further. Each regression was confirmed with all outliers removed from analysis. Data preprocessing and multiple regression models were performed using the IBM SPSS Statistics software, version 22.

In the second step with R version 3.0.1, we ran a multiple imputation on a data set including all participants with available data for the significant SNPs using the predictive mean matching algorithm of the R-package ‘mice’ (10 imputations; van Buuren &

Groothuis-Oudshoorn 2011). We checked for outliers before and after data imputation and removed values with≥3 SDs from the mean of a respective variable. Mediation is a statistical model that pro- poses to explain the relationship between two variables (the inde- pendent variable and the dependent variable) through a third vari- able, which is explanatory or mechanistic, explaining how the inde- pendent and dependent variables are related (Hayes 2013; Hayes

& Preacher 2014). The mediation models on the imputed data set were conducted using the R-package ‘mediation’ (Tingleyet al.2014).

This R-package is designed for causal mediation analysis in empiri- cal research. Indirect effects through postulated mediators are esti- mated using a simulation-based approach to generate CIs. We used maternal behaviour measures on affirmative parenting and physically controlling behaviour (Etch-A-Sketch task) as outcome, SNPs that were significant in the stepwise regression as predictor and mea- sures of executive function (CANTAB) as mediator. The 95% CIs were estimated using 1000 bootstraps (non-parametric bootstrap- ping). We included basic demographics (SES and maternal age) and maternal symptoms of depression as covariates in all mediation analy- ses. Depression is known to impair both executive functions (Nakano et al.2008; Paelecke-Habermannet al.2005; Weiland-Fiedleret al.

2004) and maternal sensitivity (Gerdeset al. 2007; Lovejoyet al.

2000). Furthermore, depression has been associated with one of the SNPs of interest,OXTRrs2254298 (Apter-Levyet al.2013). The single path coefficients were obtained in a regression model using the linear regression (lm) function of the ‘stats’ package in R (R Development Core Team 2015). All results were pooled across the 10 imputations.

The third step consisted of a partial replication of the findings in the Hamilton cohort regarding the SNPs that predicted maternal behaviour measured on the Etch-A-Sketch in the second cohort of MAVAN participants from Montreal. We used lm to predict mater- nal behaviour on the Etch-A-Sketch by genotype, statistically con- trolling for SES, maternal age and maternal symptoms of depres- sion. Data imputation and lm were conducted using the R-package

‘mice’ (10 imputations; van Buuren & Groothuis-Oudshoorn 2011) and R version 3.2.0 ‘Full of Ingredients’ (R Development Core Team 2015). In all analyses, an 𝛼-value of <0.05 was considered significant.

Results

Genotypes

All SNPs were in HWE in our sample. All SNPs were in low LD

(r²<0.20) with the other SNPs on the same gene (HaploReg

v2). When groups were small (representing<5% of the sam- ple), minor homozygotes were combined with heterozygotes to create two groups for each genotype. Table 1 summarizes the HWE, MAF and genotype data.

Covariates

The mean age for all mothers at delivery was 30.50 years (5.13), with a minimum of 18 and a maximum of 43 years. As testing was conducted 4 years after delivery, mothers would be between 22 and 47 years of age at the time of cognitive testing and behavioural observations. For the combined measure of SES and education, 36 women reported low income and low or high education or high income and low education while 103 women reported high income and high education. The mean score on EPDS was 6.808 (4.339) with a range of 0–21.33.

Table 1: Genotype information and participant ethnicity within each analysis

Location Ethnicity

Gene SNP n

HWE P-value

MAF

(%) Genotype distributions

% Caucasian

% Mixed or non-Caucasian

% Missing

OXTR rs2254298 160 0.417 11.25 AA=1 AG=34 GG=125 Intronic 90.4 4.5 5.1

OXTR rs237885 187 0.611 49.20 AA=47 AC=90 CC=50 Intronic 93.6 5.3 1.1

DRD1 rs686 214 0.681 38.32 AA=80 GA=104 GG=30 3^′-Untranslated region 84.1 4.7 11.2

DRD1 rs265976 189 0.37 21.96 CC=113 GC=69 GG=7 Downstream region 84.4 4.6 11.0

Table 2: Correlation matrix with cognitive and behavioural phenotypes and maternal genotypes

Motor inhibition

Cognitive flexibility

Working memory

Decision- making

Positive parenting

Physically controlling parenting

DRD1 rs686

DRD1 rs265976

OXTR rs237885

OXTR rs2254298 Motor inhibition 1

Cognitive flexibility −0.014 1 Working memory −0.155 −0.186

(P=0.039) 1 Decision-making −0.008 −0.149 0.217

(P=0.017) 1 Positive parenting −0.127 −0.202 0.104 0.298

(P=0.011) 1 Physically

controlling parenting

−0.031 0.216 −0.121 −0.277

(P=0.011) −0.600 (P=0.000151)

1

DRD1rs686 −0.057 −0.101 0.131 0.079 −0.049 −0.014 1

DRD1rs265976 −0.013 −0.095 0.059 0.114 0.023 0.002 −0.168 1

OXTRrs237885 −0.028 −0.021 −0.049 0.202

(P=0.044)

−0.078 0.155 −0.031 −0.020 1

OXTRrs2254298 −0.026 −0.027 0.094 −0.291

(P=0.004) −0.006 −0.136 0.155 −0.113 0.011 1

Data preprocessing

One subject from Hamilton was excluded as an extreme value in the behavioural measures (>3 SDs from the mean).

In the stepwise multiple regression of the four genotypes on the four measures of executive function, only OXTR rs2254298 significantly predicted the decision-making variable (P=0.004) after the Bonferroni correction. Thus, subsequent analysis focused on OXTR rs2254298. The correlations between cognitive and behavioural phenotypes and maternal genotypes support the results of the stepwise regression (Table 2).

Mediation analyses

In the Hamilton cohort, the effects ofOXTRrs2254298 on positive parenting were fully mediated by decision-making (estimate=0.115, P<0.005). There were no significant direct effects of genotype on positive parenting (P=0.78).

In contrast, there was both a direct and an indirect effect of OXTR rs2254298 on mothers’ physically controlling behaviour, with an incomplete and inconsistent mediation by mothers’ decision-making (estimate= −0.059,P<0.005).

The direct effect of genotype on this physically controlling behaviour measure was positive, with A-carriers exerting

more physical control over their children than non-A-carriers (estimate=0.186,P=0.03). Mediation results are summa- rized in Table 3 and Fig. 1.

To exclude the possibility of a moderation of genotype on the association between executive function and maternal parenting, we ran moderation models with each measure of executive function as the independent variable, genotype as the moderator and each dimension of parenting behaviours as the outcome. None of these models were significant (data not shown).

Replication

We were able to partially replicate our findings in the Montreal cohort, who did not receive executive function testing, but for whom there was both genotyping and Etch-A-Sketch behavioural observations (n=152). We found a non-significant total effect of the genotype on physically controlling behaviour (P=0.15) with an estimated effect size of r=0.242. We used Fisher’s r-to-z transformation to test if the total effect sizes were comparable in both Hamil- ton and Montreal samples, which was indeed the case (P=0.204), providing a partial replication of our findings (Lowry 2015).

Table 3: Mediation of genotype by decision-making on positive parenting and physically controlling parenting by model, whereab=the mediated effect,c^′=the direct effect when the mediated effect has been partialled out andc=the total effect (direct effect+mediated effect)

Model Estimate 95% CI (lower) 95% CI (upper) P-value

OXTRrs2254298 Decision-making Positive parenting (Fig. 1a)

Average causal mediation effect (ab) 0.1145 0.032 0.221 0.01*

Average direct effect (c^′) 0.0468 −0.260 0.341 0.74

Total effect (c) 0.1613 −0.150 0.4523 0.33

Proportion mediated 0.7097 −6.732 6.090 0.33

OXTRrs2254298 Decision-making

Physically controlling parenting (Fig. 1b)

Average causal mediation effect (ab) −0.059 −0.122 −0.015 0.01*

Average direct effect (c^′) 0.186 0.025 0.349 0.02*

Total effect (c) 0.1272 −0.035 0.291 0.13

Proportion mediated −0.461 −4.626 2.453 0.14

*P<0.05.

(a) (b)

Decision making

Positive Parenting

Physically controlling OXTR

rs 2254298 GG = 1; A* = 2

OXTR rs 2254298 GG = 1; A* = 2 a = 0.500**

c’ = 0.047 c’ = 0.186*

c = 0.127 c = 0.161

b = 0.210 a = 0.500** b = -.044

Decision making

Figure 1: Mediation of genotype by decision-making on positive parenting and physically controlling parenting by model.

*P<0.05; **P<0.001.

Discussion

The OXTR rs2254298 was related to non-social decision-making, with A-carriers displaying more information gathering, measured by boxes opened, in a decision-making task than non-A-carriers. We did not find dopamine system SNPs to be related to either executive function or parenting behaviours in our sample. These null findings may be due to an underpowered analysis based on extremely small effect sizes. Conversely, it may be that other SNPs within the dopamine system are more strongly associated with execu- tive functions and/or parenting behaviours. Furthermore, the selected SNPs may not be functional variants. Also, our tests of executive functions may not accurately reflect aspects of executive functions in which dopamine is dominant. Finally, our study includes children, rather than infants, which may place emphasis on different brain network of the mothers.

In regard to ourOXTRfindings, the relationship between OXTR rs2254298 and positive parenting behaviours was mediated by non-social decision-making, in whom A-carriers who displayed more information gathering in a decision-making task display more positive parenting behaviours. We also found a direct effect ofOXTRrs2254298 on physically controlling behaviours, with A-carriers display- ing higher levels of physically controlling behaviours. There was also a mediated relationship betweenOXTRrs2254298

and physically controlling behaviours, withOXTRrs2254298 predicting more information gathering in a decision-making task, which in turn predicted lower physically controlling behaviours. The analyses showed an inconsistent mediation:

the OXTR rs2254298 A-carriers had increasing parental aggression towards the child as shown by increased phys- ically controlling behaviours and also increased executive functioning during decision-making, which suppressed parental aggression of physically controlling behaviours.

Thus, the mediation is inconsistent, in that the effects in theOXTRrs2254298 A-carriers were in opposite directions from the non-A-carriers: both increase physically controlling behaviours directly, but to also decrease physically controlling behaviours indirectly through increasing decision-making.

The inconsistent mediation hints at the pleiotropic effects of OXTR(Meyer-Lindenberg & Tost 2012). Our results can be considered in the context of three intersecting literatures:

the relationship of oxytocin to parenting behaviours, to social cognition and to aggression as well as decision-making in social interactions.

Oxytocin and parenting

Previous studies have found a relationship betweenOXTR rs2254298 non-A-carriers and parental behaviours, including reduced early parental care and engagement with their infants compared with rs2254298 A-carriers (Feldmanet al.

2013). Furthermore, non-A-carrier mothers were more likely

to be depressed than A-carrier mothers (67% compared with 34%; Apter-Levyet al.2013). In non-maternal participants, non-A-carriers also have an increased frequency of major depressive disorder (Mendlewicz et al.2012). This may be partially explained by an association between non-A-carriers and lower plasma oxytocin levels (Feldman et al. 2012).

Although it is important to note that plasma oxytocin levels are not central oxytocin levels, recent work shows that plasma and central oxytocin are strongly positively correlated (Carsonet al.2015). Moreover, rs2254298 genotype appears to interact with the experience of early adversity, with girls who are heterozygous (GA) having higher levels of social anxiety and depression as compared with homozygotes (Kumsta & Heinrichs 2013). Finally, non-A-carriers have increased unipolar depression and high levels of adult sep- aration anxiety (Costa et al. 2009). In addition, previous research consistently shows a link between A-carriers and increased parental care and involvement, possibly through differences in oxytocin levels and interaction with parental mood and early adversity. Our current findings add to this literature, showing a significant direct association between A-carriers and higher physically controlling parental behaviour towards the child and an indirect association between A-carriers and more positive parenting through higher scores on decision-making as an executive function.

Oxytocin and social cognition

Oxytocin is most often studied through intranasal administra- tion and genetic polymorphisms in the context of social cogni- tion, and has been linked to an array of prosocial behaviours such as empathy, generosity and co-operation (for compre- hensive reviews, see Love 2014; Olff et al. 2013). In the parenting literature, higher oxytocin levels and SNPs that are thought to functionally enhance oxytocin are associ- ated with more approach behaviours and interaction with an infant (Apter-Levyet al.2013; Feldmanet al.2012, 2013;

Mileva-Seitzet al.2013; Wittfoth-Schardtet al.2012). While most research on social cognition and oxytocin uses an experimental approach with intranasal oxytocin administra- tion, some studies have considered genotypes. Specifically, on reading the mind in the eyes task, A-carriers did not recog- nize positive emotional pictures as often as G-carriers (Lucht et al.2013) and also had marginally lower verbal intelligence scores (Luchtet al.2009). Although rs2254298 was not asso- ciated with generosity, trustworthiness or trusting behaviours in a dictator game (Apicellaet al.2010), an earlier study found a marginal association with A-carrier males and increased generosity (Israelet al.2009). Within the context of social cognition, our results should be considered in the classic definition of maternal sensitivity, as a mother’s awareness of the child’s signals, an accurate interpretation of the sig- nals and a prompt and appropriate response to the signals.

While our measure of observed parenting is not sensitivity per se, it does invoke the components of maternal sensitivity.

In this study, A-carriers, who have previously been shown to have higher plasma oxytocin levels than non-A-carriers (Feld- manet al.2012), display higher levels of physically control- ling behaviours, which may relate to inaccurate interpreta- tion of the child’s signals. A study that measured oxytocin

plasma concentrations reported that low sensitivity moth- ers had higher baseline levels of oxytocin compared with high sensitivity mothers (Elmadihet al.2014). A study using intranasal oxytocin administration found that women given oxytocin had lower accuracy on identifying emotions in infant faces than women who had been given a placebo (Voorthuis et al.2014), which may be the mechanism by which higher oxytocin has been associated with lower sensitivity in our and others’ studies (Elmadih et al.2014). However, further research is required to understand the role of rs2254298 in social cognition and whether these effects are related to higher levels of plasma oxytocin.

Oxytocin and aggression

While oxytocin has for a long time been characterized as a

‘love hormone’, more recent research now characterizes oxy- tocin as a ‘double-edged sword’, acting to increase some affiliative behaviours and also some aggressive behaviours.

While the idea of divergent effects of oxytocin may seem irreconcilable, previous work has come to a similar conclusion with regard to attribution of emotional state in physiologi- cal arousal (Schachter & Singer 1962). Animal literature has consistently found an association between elevated oxytocin and maternal aggression (Bosch & Neumann 2012). How- ever, other factors may also be involved in whether oxytocin increases aggression. In a factor analysis of contextual effects interacting withOXTRgenetics, the factors of alcohol, provo- cation and ‘winning the task’ interacted withOXTRgenotype to increase aggression (LoParoet al.2016). Finally, a recent review sheds light on the importance of situational context as a factor that can moderate the effects of oxytocin (Olff et al.2013). In situational circumstances deemed ‘safe’ by the participant, oxytocin can facilitate prosocial behaviours, whereas in situations perceived as ‘unsafe’ by the partici- pant, oxytocin may increase the tendency towards antiso- cial behaviours (Olffet al.2013). Furthermore, both situational context and oxytocin can interact with personal psychological characteristics. It is possible that some participants perceived the videotaped mother–child interaction as ‘unsafe’, thereby increasing ‘antisocial’ behaviours, such as physically control- ling parenting. However, this effect of oxytocin on parental behaviour might have been tempered by higher executive functions, as shown on the CANTAB decision-making task.

In addition, our results corroborate the idea of oxytocin as a double-edged sword, with A-carriers displaying increased physically controlling behaviours, better performance on a decision-making task and indirectly more positive parenting than non-A-carriers.

Social interactions as decisions

While the idea of social interactions as a series of decisions may seem a bit calculated, the interpretation of the affective state of another person requires a decision, not only on what mental state to attribute to the other person, but also on how to respond to that perceived emotion in the particular present context (Lynn et al. 2014). In this characterization of social interactions as a series of decisions, there is a close overlap with Ainsworth et al.’s (1978) description of maternal sensitivity: to perceive infant cues, to correctly

interpret them and to respond contingently. Each aspect of maternal sensitivity therefore requires a decision to attend to the child instead of some other stimuli such as the television or a smartphone. For example, a decision as to how much information in terms of child-generated cues to gather and how to interpret that information, and finally a decision on if, when or how to act on that information. While most of the experimental paradigms researching oxytocin have thus far involved a social context, our work shows that even in a non-social, non-verbal task, one oxytocin receptor SNP is associated with differences in decision-making, which may be a major root of social interactions and an important fundamental component of mother–child interactions.

Limitations

As this study was conducted in a longitudinal manner, there are some limitations inherent to longitudinal data collection, such as sample biasing due to drop outs. Here, we attempt to control for drop out by using imputed data. We found no significant differences between included participants and excluded participants in SES, age, depression scores, positive parenting or physically controlling parenting (data not shown).

Also, while the SNPs selected for study have previously been associated with parenting behaviours and/or executive functions, there are no data to support a functional difference in genotypes for any of the SNPs within this study. Finally, we might have been underpowered in our analysis to detect very small but significant effect sizes, as these findings should be considered with caution. Also, our sample size is relatively modest for genetic studies, thus replication is important to substantiate these findings.

Conclusions

Previous research has shown that the oxytocin system is associated with the onset of maternal behaviour, maternal motivation and warmth, and is involved in bonding, close contact and nursing. We extend this literature by show- ing that the oxytocin system, through receptor polymor- phisms, is involved in maternal behaviour at 48 months post-partum through executive functions. However, the rela- tionships between oxytocin, executive functions and parent- ing behaviours are complex. The results suggest that an OXTRSNP related to increased oxytocin functioning is asso- ciated with improved decision-making, which in turn is associ- ated with increased positive parenting and decreased physi- cally controlling parenting. However, theOXTRSNP is also associated with increased physically controlling behaviour, when the variance associated with executive functions has been partialled out.

References

Afonso, V.M., Sison, M., Lovic, V. & Fleming, A.S. (2007) Medial pre- frontal cortex lesions in the female rat affect sexual and maternal behavior and their sequential organization.Behav Neurosci121, 515–526.

Ainsworth, M.D., Blehar, M., Water, E. & Wall, S. (1978)Patterns of Attachment: A Psychological Study of the Strange Situation.

Lawrence Erlbaum Associates, Inc., Hillsdale, NJ.

Apicella, C.L., Cesarini, D., Johannesson, M., Dawes, C.T., Licht- enstein, P., Wallace, B., Beauchamp, J. & Westberg, L. (2010) No association between oxytocin receptor (OXTR) gene polymor- phisms and experimentally elicited social preferences.PLoS One 5, e11153.

Apter-Levy, Y., Feldman, M., Vakart, A., Ebstein, R.P. & Feldman, R.

(2013) Impact of maternal depression across the first 6 years of life on the child’s mental health, social engagement, and empathy: the moderating role of oxytocin.Am J Psychiatry170, 1161–1168.

Arnsten, A.F. (2006) Fundamentals of attention-deficit/hyperactivity disorder: circuits and pathways.J Clin Psychiatry 67(Suppl. 8), 7–12.

Atkinson, L., Leung, E., Goldberg, S., Benoit, D., Poulton, L., Myhal, N., Blokland, K. & Kerr, S. (2009) Attachment and selective atten- tion: disorganization and emotional Stroop reaction time.Dev Psy- chopathol21, 99–126.

Bakermans-Kranenburg, M.J. & van IJzendoorn, M.H. (2008) Oxy- tocin receptor (OXTR) and serotonin transporter (5-HTT) genes associated with observed parenting.Soc Cogn Affect Neurosci3, 128–134.

Bobb, A.J., Addington, A.M., Sidransky, E., Gornick, M.C., Lerch, J.P., Greenstein, D.K., Clasen, L.S., Sharp, W.S., Inoff-Germain, G., Wavrant-De Vrieze, F., Arcos-Burgos, M., Straub, R.E., Hardy, J.A., Castellanos, F.X. & Rapoport, J.L. (2005) Support for association between ADHD and two candidate genes: NET1 and DRD1.Am J Med Genet B Neuropsychiatr Genet134B, 67–72.

Bosch, O.J. & Neumann, I.D. (2012) Both oxytocin and vasopressin are mediators of maternal care and aggression in rodents: from central release to sites of action.Horm Behav61, 293–303.

Bridges, R.S. (2015) Neuroendocrine regulation of maternal behavior.

Front Neuroendocrinol36, 178–196.

Bridgett, D.J., Burt, N.M., Edwards, E.S. & Deater-Deckard, K.

(2015) Intergenerational transmission of self-regulation: a multidis- ciplinary review and integrative conceptual framework.Psychol Bull 141, 602–654.

Brookes, K., Xu, X., Chen, W.et al.(2006) The analysis of 51 genes in DSM-IV combined type attention deficit hyperactivity disorder:

association signals in DRD4, DAT1 and 16 other genes.Mol Psy- chiatry11, 934–953.

van Buuren, S. & Groothuis-Oudshoorn, K. (2011) Mice: multivariate imputation by chained equations in R.J Stat Softw45, 1–67.

Caldu, X., Vendrell, P., Bartres-Faz, D., Clemente, I., Bargallo, N., Jurado, M.A., Serra-Grabulosa, J.M. & Junque, C. (2007) Impact of the COMT Val108/158 Met and DAT genotypes on prefrontal function in healthy subjects.Neuroimage37, 1437–1444.

Carson, D.S., Berquist, S.W., Trujillo, T.H., Garner, J.P., Hannah, S.L., Hyde, S.A., Sumiyoshi, R.D., Jackson, L.P., Moss, J.K., Strehlow, M.C., Cheshier, S.H., Partap, S., Hardan, A.Y. & Parker, K.J. (2015) Cerebrospinal fluid and plasma oxytocin concentrations are pos- itively correlated and negatively predict anxiety in children.Mol Psychiatry20, 1085–1090.

Carter, C.S., Lederhendler, I. & Kirkpatrick, B. (1997) The integrative neurobiology of affiliation. Introduction.Ann N Y Acad Sci 807, xiii–xviii.

Chang, H., Shaw, D.S., Dishion, T.J., Gardner, F. & Wilson, M.N. (2014) Direct and indirect effects of the family check-up on self-regulation from toddlerhood to early school-age.J Abnorm Child Psychol42, 1117–1128.

Chen, D., Liu, F., Shang, Q., Song, X., Miao, X. & Wang, Z. (2011) Association between polymorphisms of DRD2 and DRD4 and opioid dependence: evidence from the current studies.Am J Med Genet B Neuropsychiatr Genet156B, 661–670.

Chico, E., Gonzalez, A., Ali, N., Steiner, M. & Fleming, A.S. (2014) Executive function and mothering: challenges faced by teenage mothers.Dev Psychobiol56, 1027–1035.

Comings, D.E., Gade, R., Wu, S., Chiu, C., Dietz, G., Muhleman, D., Saucier, G., Ferry, L., Rosenthal, R.J., Lesieur, H.R., Rugle, L.J. &

MacMurray, P. (1997) Studies of the potential role of the dopamine D1 receptor gene in addictive behaviors.Mol Psychiatry2, 44–56.

Costa, B., Pini, S., Gabelloni, P., Abelli, M., Lari, L., Cardini, A., Muti, M., Gesi, C., Landi, S., Galderisi, S., Mucci, A., Lucacchini, A., Cas- sano, G.B. & Martini, C. (2009) Oxytocin receptor polymorphisms and adult attachment style in patients with depression.Psychoneu- roendocrinology34, 1506–1514.

Cox, J.L., Holden, J.M. & Sagovsky, R. (1987) Detection of postna- tal depression. Development of the 10-item Edinburgh Postnatal Depression Scale.Br J Psychiatry150, 782–786.

Crandall, A., Deater-Deckard, K. & Riley, A.W. (2015) Maternal emo- tion and cognitive control capacities and parenting: a conceptual framework.Dev Rev36, 105–126.

Cuevas, K., Deater-Deckard, K., Kim-Spoon, J., Watson, A.J., Morasch, K.C. & Bell, M.A. (2014) What’s mom got to do with it?

Contributions of maternal executive function and caregiving to the development of executive function across early childhood.Dev Sci 17, 224–238.

Deater-Deckard, K., Sewell, M.D., Petrill, S.A. & Thompson, L.A.

(2010) Maternal working memory and reactive negativity in parent- ing.Psychol Sci21, 75–79.

Deater-Deckard, K., Wang, Z., Chen, N. & Bell, M.A. (2012) Maternal executive function, harsh parenting, and child conduct problems.J Child Psychol Psychiatry53, 1084–1091.

Diamond, A. (2013) Executive functions. Annu Rev Psychol 64, 135–168.

Elmadih, A., Wan, M.W., Numan, M., Elliott, R., Downey, D. &

Abel, K.M. (2014) Does oxytocin modulate variation in maternal caregiving in healthy new mothers?Brain Res1580, 143–150.

Fahrbach, S.E., Morrell, J.I. & Pfaff, D.W. (1985) Possible role for endogenous oxytocin in estrogen-facilitated maternal behavior in rats.Neuroendocrinology40, 526–532.

Feldman, R., Zagoory-Sharon, O., Weisman, O., Schneiderman, I., Gordon, I., Maoz, R., Shalev, I. & Ebstein, R.P. (2012) Sensitive parenting is associated with plasma oxytocin and polymorphisms in the OXTR and CD38 genes.Biol Psychiatry72, 175–181.

Feldman, R., Gordon, I., Influs, M., Gutbir, T. & Ebstein, R.P. (2013) Parental oxytocin and early caregiving jointly shape children’s oxy- tocin response and social reciprocity.Neuropsychopharmacology 38, 1154–1162.

Gauderman, W.J. & Morrison, J.M. (2006) QUANTO 1.1: a com- puter program for power and sample size calculations for genetic-epidemiology studies. Retrieved July 29, 2016, from http://biostats.usc.edu/Quanto.html.

Gerdes, A.C., Hoza, B., Arnold, L.E., Pelham, W.E., Swanson, J.M., Wigal, T. & Jensen, P.S. (2007) Maternal depressive symptoma- tology and parenting behavior: exploration of possible mediators.

J Abnorm Child Psychol35, 705–714.

Goldman-Rakic, P.S. (1998) The cortical dopamine system: role in memory and cognition.Adv Pharmacol42, 707–711.

Gonzalez, A., Atkinson, L. & Fleming, A.S. (2009) Attachment and the comparative psychobiology of mothering. In Hann, M.D. & Gunnar, M.R. (eds), Handbook of Developmental Social Neuroscience.

Guilford Press, New York, NY, pp. 225–245.

Gonzalez, A., Jenkins, J.M., Steiner, M. & Fleming, A.S. (2012) Maternal early life experiences and parenting: the mediating role of cortisol and executive function.J Am Acad Child Adolesc Psychiatry 51, 673–682.

Hayes, A.F. (2013)Introduction to Mediation, Moderation, and Con- ditional Process Analysis: A Regression-Based Approach. Guilford Press, New York, NY.

Hayes, A.F. & Preacher, K.J. (2014) Statistical mediation analysis with a multicategorical independent variable.Br J Math Stat Psychol67, 451–470.

Heinrichs, M., Meinlschmidt, G., Wippich, W., Ehlert, U. & Hellham- mer, D.H. (2004) Selective amnesic effects of oxytocin on human memory.Physiol Behav83, 31–38.

Herzmann, G., Young, B., Bird, C.W. & Curran, T. (2012) Oxytocin can impair memory for social and non-social visual objects: a within-subject investigation of oxytocin’s effects on human mem- ory.Brain Res1451, 65–73.

van IJzendoorn, M.H., Bakermans-Kranenburg, M.J. & Mesman, J.

(2008) Dopamine system genes associated with parenting in the context of daily hassles.Genes Brain Behav7, 403–410.

Insel, T.R. (2003) Is social attachment an addictive disorder?Physiol Behav79, 351–357.

Israel, S., Lerer, E., Shalev, I., Uzefovsky, F., Riebold, M., Laiba, E., Bachner-Melman, R., Maril, A., Bornstein, G., Knafo, A. &

Ebstein, R.P. (2009) The oxytocin receptor (OXTR) contributes to prosocial fund allocations in the dictator game and the social value orientations task.PLoS One4, e5535.

Jonas, W., Mileva-Seitz, V., Girard, A.W., Bisceglia, R., Kennedy, J.L., Sokolowski, M., Meaney, M.J., Fleming, A.S. & Steiner, M. (2013) Genetic variation in oxytocin rs2740210 and early adversity asso- ciated with postpartum depression and breastfeeding duration.

Genes Brain Behav12, 681–694.

Kaitz, M., Shalev, I., Sapir, N., Devor, N., Samet, Y., Mankuta, D.

& Ebstein, R.P. (2010) Mothers’ dopamine receptor polymorphism modulates the relation between infant fussiness and sensitive parenting.Dev Psychobiol52, 149–157.

Kendrick, K.M., Da Costa, A.P., Broad, K.D., Ohkura, S., Guevara, R., Levy, F. & Keverne, E.B. (1997) Neural control of maternal behaviour and olfactory recognition of offspring.Brain Res Bull44, 383–395.

Kumsta, R. & Heinrichs, M. (2013) Oxytocin, stress and social behav- ior: neurogenetics of the human oxytocin system.Curr Opin Neu- robiol23, 11–16.

Lasky-Su, J., Banaschewski, T., Buitelaar, J.et al.(2007) Partial replica- tion of a DRD4 association in ADHD individuals using a statistically derived quantitative trait for ADHD in a family-based association test.Biol Psychiatry62, 985–990.

Lee, S.S., Chronis-Tuscano, A., Keenan, K., Pelham, W.E., Loney, J., Van Hulle, C.A., Cook, E.H. & Lahey, B.B. (2010) Association of maternal dopamine transporter genotype with negative parenting:

evidence for gene x environment interaction with child disruptive behavior.Mol Psychiatry15, 548–558.

van Leengoed, E., Kerker, E. & Swanson, H.H. (1987) Inhibition of post-partum maternal behaviour in the rat by injecting an oxytocin antagonist into the cerebral ventricles.J Endocrinol112, 275–282.

Lonstein, J.S., Levy, F. & Fleming, A.S. (2015) Common and divergent psychobiological mechanisms underlying maternal behaviors in non-human and human mammals.Horm Behav73, 156–185.

LoParo, D., Johansson, A., Walum, H., Westberg, L., Santtila, P. &

Waldman, I. (2016) Rigorous tests of gene-environment interac- tions in a lab study of the oxytocin receptor gene (OXTR), alco- hol exposure, and aggression.Am J Med Genet B Neuropsychiatr Genet171, 589–602.

Love, T.M. (2014) Oxytocin, motivation and the role of dopamine.

Pharmacol Biochem Behav119, 49–60.

Lovejoy, M.C., Graczyk, P.A., O’Hare, E. & Neuman, G. (2000) Mater- nal depression and parenting behavior: a meta-analytic review.Clin Psychol Rev20, 561–592.

Lowry, R. (2015) Significance of the difference between two cor- relation coefficients. Retrieved January 14, 2016, from http://

vassarstats.net/rdiff.html.

Lucht, M.J., Barnow, S., Sonnenfeld, C., Rosenberger, A., Grabe, H.J., Schroeder, W., Volzke, H., Freyberger, H.J., Herrmann, F.H., Kroemer, H. & Rosskopf, D. (2009) Associations between the oxy- tocin receptor gene (OXTR) and affect, loneliness and intelligence in normal subjects.Prog Neuropsychopharmacol Biol Psychiatry33, 860–866.

Lucht, M.J., Barnow, S., Sonnenfeld, C., Ulrich, I., Grabe, H.J., Schroeder, W., Volzke, H., Freyberger, H.J., John, U., Her- rmann, F.H., Kroemer, H. & Rosskopf, D. (2013) Associations between the oxytocin receptor gene (OXTR) and “mind-reading”

in humans – an exploratory study.Nord J Psychiatry67, 15–21.

Lynn, S.K., Hoge, E.A., Fischer, L.E., Barrett, L.F. & Simon, N.M. (2014) Gender differences in oxytocin-associated disruption of decision bias during emotion perception.Psychiatry Res219, 198–203.

Massey, S.H., Estabrook, R., O’Brien, T.C., Pine, D.S., Burns, J.L., Jacob, S., Cook, E.H. & Wakschlag, L.S. (2015) Preliminary evidence for the interaction of the oxytocin receptor gene (oxtr) and face

processing in differentiating prenatal smoking patterns.Neurosci Lett584, 259–264.

Mendlewicz, J., Crisafulli, C., Calati, R., Kocabas, N.A., Massat, I., Linotte, S., Kasper, S., Fink, M., Sidoti, A., Scantamburlo, G., Ansseau, M., Antonijevic, I., Forray, C., Snyder, L., Bollen, J., Mont- gomery, S., Zohar, J., Souery, D. & Serretti, A. (2012) Influence of COX-2 and OXTR polymorphisms on treatment outcome in treat- ment resistant depression.Neurosci Lett516, 85–88.

Meyer-Lindenberg, A. & Tost, H. (2012) Neural mechanisms of social risk for psychiatric disorders.Nat Neurosci15, 663–668.

Mileva-Seitz, V., Fleming, A.S., Meaney, M.J., Mastroianni, A., Sin- nwell, J.P., Steiner, M., Atkinson, L., Levitan, R.D., Matthews, S.G., Kennedy, J.L. & Sokolowski, M.B. (2012) Dopamine receptors D1 and D2 are related to observed maternal behavior.Genes Brain Behav11, 684–694.

Mileva-Seitz, V., Steiner, M., Atkinson, L., Meaney, M.J., Levitan, R., Kennedy, J.L., Sokolowski, M.B. & Fleming, A.S. (2013) Interaction between oxytocin genotypes and early experience predicts quality of mothering and postpartum mood.PLoS One8, e61443.

Misener, V.L., Luca, P., Azeke, O., Crosbie, J., Waldman, I., Tannock, R., Roberts, W., Malone, M., Schachar, R., Ickowicz, A., Kennedy, J.L. & Barr, C.L. (2004) Linkage of the dopamine receptor D1 gene to attention-deficit/hyperactivity disorder.Mol Psychiatry 9, 500–509.

Nakano, Y., Baba, H., Maeshima, H., Kitajima, A., Sakai, Y., Baba, K., Suzuki, T., Mimura, M. & Arai, H. (2008) Executive dysfunction in medicated, remitted state of major depression.J Affect Disord111, 46–51.

Nishimori, K., Young, L.J., Guo, Q., Wang, Z., Insel, T.R. & Matzuk, M.M. (1996) Oxytocin is required for nursing but is not essential for parturition or reproductive behavior.Proc Natl Acad Sci U S A 93, 11699–11704.

Numan, M. & Corodimas, K.P. (1985) The effects of paraventricular hypothalamic lesions on maternal behavior in rats.Physiol Behav 35, 417–425.

Numan, M., Fleming, A.S. & Levy, F. (2006) Maternal behavior. In Neill, J.D. (ed),Knobil & Neill’s Physiology of Reproduction. Elsevier, San Diego, CA, pp. 1922–1993.

O’Donnell, K.A., Gaudreau, H., Colalillo, S., Steiner, M., Atkinson, L., Moss, E., Goldberg, S., Karama, S., Matthews, S.G., Lydon, J.E., Silveira, P.P., Wazana, A.D., Levitan, R.D., Sokolowski, M.B., Kennedy, J.L., Fleming, A. & Meaney, M.J. (2014) The maternal adversity, vulnerability and neurodevelopment project: theory and methodology.Can J Psychiatry59, 497–508.

Olff, M., Frijling, J.L., Kubzansky, L.D., Bradley, B., Ellenbogen, M.A., Cardoso, C., Bartz, J.A., Yee, J.R. & van Zuiden, M. (2013) The role of oxytocin in social bonding, stress regulation and mental health:

an update on the moderating effects of context and interindividual differences.Psychoneuroendocrinology38, 1883–1894.

Paelecke-Habermann, Y., Pohl, J. & Leplow, B. (2005) Attention and executive functions in remitted major depression patients.J Affect Disord89, 125–135.

Park, J., Willmott, M., Vetuz, G., Toye, C., Kirley, A., Hawi, Z., Brookes, K.J., Gill, M. & Kent, L. (2010) Evidence that genetic variation in the oxytocin receptor (OXTR) gene influences social cognition in ADHD.Prog Neuropsychopharmacol Biol Psychiatry34, 697–702.

Pedersen, C.A. & Boccia, M.L. (2003) Oxytocin antagonism alters rat dams’ oral grooming and upright posturing over pups.Physiol Behav80, 233–241.

Pedersen, C.A., Ascher, J.A., Monroe, Y.L. & Prange, A.J. Jr. (1982) Oxytocin induces maternal behavior in virgin female rats.Science 216, 648–650.

Pedersen, C.A., Caldwell, J.D., Johnson, M.F., Fort, S.A. & Prange, A.J. Jr. (1985) Oxytocin antiserum delays onset of ovarian steroid-induced maternal behavior.Neuropeptides6, 175–182.

Pedersen, C.A., Caldwell, J.D., Walker, C., Ayers, G. & Mason, G.A.

(1994) Oxytocin activates the postpartum onset of rat maternal behavior in the ventral tegmental and medial preoptic areas.Behav Neurosci108, 1163–1171.

R Development Core Team (2015)R: A language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna.

van Ree, J.M., Bohus, B., Versteeg, D.H.G. & de Wied, D. (1978) Neurohypophyseal principles and memory processes. Biochem Pharmacol27, 1793–1800.

Ribases, M., Ramos-Quiroga, J.A., Hervas, A., Sanchez-Mora, C., Bosch, R., Bielsa, A., Gastaminza, X., Lesch, K.P., Reif, A., Renner, T.J., Romanos, M., Warnke, A., Walitza, S., Freitag, C., Meyer, J., Palmason, H., Casas, M., Bayes, M. & Cormand, B. (2012) Candidate system analysis in ADHD: evaluation of nine genes involved in dopaminergic neurotransmission identifies association with DRD1.World J Biol Psychiatry13, 281–292.

Rich, M.E., deCardenas, E.J., Lee, H.J. & Caldwell, H.K. (2014) Impair- ments in the initiation of maternal behavior in oxytocin receptor knockout mice.PLoS One9, e98839.

Rutherford, H.J., Wallace, N.S., Laurent, H.K. & Mayes, L.C. (2015) Emotion regulation in parenthood.Dev Rev36, 1–14.

Schachter, S. & Singer, J.E. (1962) Cognitive, social, and physiological determinants of emotional state.Psychol Rev69, 379–399.

Seeger, G., Schloss, P., Schmidt, M.H., Ruter-Jungfleisch, A. & Henn, F.A. (2004) Gene-environment interaction in hyperkinetic conduct disorder (HD+CD) as indicated by season of birth variations in dopamine receptor (DRD4) gene polymorphism.Neurosci Lett366, 282–286.

Skuse, D.H., Lori, A., Cubells, J.F., Lee, I., Conneely, K.N., Puura, K., Lehtimaki, T., Binder, E.B. & Young, L.J. (2014) Common polymorphism in the oxytocin receptor gene (OXTR) is associated with human social recognition skills.Proc Natl Acad Sci U S A111, 1987–1992.

Slane, M.M., Lusk, L.G., Boomer, K.B., Hare, A.E., King, M.K. &

Evans, D.W. (2014) Social cognition, face processing, and oxytocin receptor single nucleotide polymorphisms in typically developing children.Dev Cogn Neurosci9, 160–171.

Smith, L., Watson, M., Gates, S., Ball, D. & Foxcroft, D. (2008) Meta-analysis of the association of the Taq1A polymorphism with the risk of alcohol dependency: a HuGE gene-disease association review.Am J Epidemiol167, 125–138.

Statistics Canada. (2008) Income and earnings. Retrieved March 9, 2015, from https://www12.statcan.gc.ca/census-recensement/

2006/rt-td/inc-rev-eng.cfm.

Statistics Canada. (2011) Education indicators in Canada. An inter- national perspective. Retrieved March 9, 2015, from http://

www.statcan.gc.ca/pub/81-604-x/81-604-x2011001-eng.htm.

Stice, E., Spoor, S., Bohon, C. & Small, D.M. (2008) Relation between obesity and blunted striatal response to food is moderated by TaqIA A1 allele.Science322, 449–452.

Styn, M.A., Nukui, T., Romkes, M., Perkins, K., Land, S.R. & Weiss- feld, J.L. (2009) The impact of genetic variation in DRD2 and SLC6A3 on smoking cessation in a cohort of participants 1 year after enrollment in a lung cancer screening study.Am J Med Genet B Neuropsychiatr Genet150B, 254–261.

Tingley, D., Yamamoto, T., Hirose, K., Keele, L. & Imai, K. (2014) mediation: R package for causal mediation analysis.J Stat Softw 59, 1–38.

Voorthuis, A., Riem, M.M., Van, I.M.H. & Bakermans-Kranenburg, M.J. (2014) Reading the mind in the infant eyes: paradoxical effects of oxytocin on neural activity and emotion recognition in watching pictures of infant faces.Brain Res1580, 151–159.

Weiland-Fiedler, P., Erickson, K., Waldeck, T., Luckenbaugh, D.A., Pike, D., Bonne, O., Charney, D.S. & Neumeister, A. (2004) Evidence for continuing neuropsychological impairments in depression.J Affect Disord82, 253–258.

Wendland, B.E., Atkinson, L., Steiner, M., Fleming, A.S., Pencharz, P., Moss, E., Gaudreau, H., Silveira, P.P., Arenovich, T., Matthews, S.G., Meaney, M.J. & Levitan, R.D. (2014) Low maternal sensitivity at 6 months of age predicts higher BMI in 48 month old girls but not boys.Appetite82, 97–102.

de Wied, D. (1980) Behavioural actions of neurohypophyseal pep- tides.Proc R Soc Lond210, 183–195.

Wilkosc, M., Hauser, J., Tomaszewska, M., Dmitrzak-Weglarz, M., Skibinska, M., Szczepankiewicz, A. & Borkowska, A. (2010) Influ- ence of dopaminergic and serotoninergic genes on working mem- ory in healthy subjects.Acta Neurobiol Exp (Wars)70, 86–94.

Wirth, M.M. (2015) Hormones, stress, and cognition: the effects of glucocorticoids and oxytocin on memory. Adapt Human Behav Physiol1, 177–201.

Wittfoth-Schardt, D., Grunding, J., Wittfoth, M., Lanfermann, H., Heinrichs, M., Domes, G., Buchheim, A., Gundel, H. & Waller, C. (2012) Oxytocin modulates neural reactivity to children’s faces as a function of social salience.Neuropsychopharmacology 37, 1799–1807.

Wu, N., Li, Z. & Su, Y. (2012) The association between oxytocin receptor gene polymorphism (OXTR) and trait empathy.J Affect Disord138, 468–472.

Acknowledgment

This research was funded by a grant from the Canadian Institutes of Health Research (FRN# 102658).