Authors: Ginger McCorkle, Aline Andres, Clark R. Sims, Patrick H. Casey, Seth T. Sorensen, Trevor Durey, Jayne Bellando
Categories: Original Article, carey temperament scales, child development, child temperament, infant feeding, maternal affect
Source: Maternal & Child Nutrition
Doi: 10.1111/mcn.13728
Infant temperament is assumed to be primarily innate. However, newer research suggests that maternal affection impacts ratings of temperament and environmental factors, including feeding method, can also influence infant temperament. This study investigates child temperament and its relationships with maternal psychiatric symptoms, environmental variables and feeding method longitudinally in a cohort of children followed from 6 to 72 months. Differences in temperament by feeding group are also investigated. We hypothesized that maternal psychiatric symptoms, environmental stressors, and impaired family dynamics would have negative impact on child temperament, whereas breastfeeding would have a positive impact on child temperament.
Mothers' ratings of child's temperament, own psychiatric symptomatology, environmental stresses and family cohesion were obtained in 504 mother‐infant dyads via rating scales completed by mothers. Infants were breastfeed (BF), fed soy‐based infant formula (SF) or dairy‐based infant formula (MF). Linear mixed effect models investigated the relationship of variables on child's temperament while controlling for significant covariates and repeated measurements.
Mothers in this study did not endorse clinical‐level psychiatric symptomatology; however, when adjusted for significant covariates, higher psychiatric symptomatology significantly correlated with environmental stressors, impaired family dynamics and elevations in temperament ratings of infants' adaptability and mood. There were no lasting differences for temperament between feeding groups. However, some significant transient increases in rhythmicity and adaptability were found between SF and BF children.
Positive relationships between family environment stressors and maternal psychiatric ratings were found. Transient differences were found in child temperament based upon feeding method.
Keywords: carey temperament scales, child development, child temperament, infant feeding, maternal affect
Children's temperament has been of interest to researchers for several decades (Pluess & Belsky, 2010). Temperament typically is defined as the differences seen in early biological traits of infants that relate to reactivity and self‐regulation (Danzig et al., 2015). Seminal work in child temperament began in the 1950s with the New York Longitudinal Study by Thomas and Chess (1957). Thomas and Chess' study followed children over several decades and used direct observations along with parent interviews to assess behavioural and emotional traits. Their work serves as the theoretical basis for the Carey Temperament Scales (Britton, 2011), a well‐known measure to quantify temperament. Higher scores on each temperament dimension are indicative of more negative traits and temperament that is more difficult/intense. Temperament can be measured by trained clinicians who directly observe the child in standardized settings. More commonly, however, parents rate their child's temperament by the use of a rating scale such as the Carey Temperament Scales mentioned above (Carey & McDevitt, 1995). The questionnaires in the Carey series are all based upon the nine temperament dimensions that were postulated in the New York Longitudinal Study (Carey & McDevitt, 1995). The nine temperament dimensions include Rhythmicity (the predictable recurrence of the child's response to everyday events); Approach (the child's initial reaction to novel stimuli); Adaptability (the child's long‐term adjustment to new circumstances); Intensity (reflects the energy the child puts into a response or reacts to a situation); Mood (information about the child's dominant mood); Distractibility (how well a child is able to attend when external stimuli is present); Sensory threshold (how sensitive a child is to sensory input); Activity level (level of energy and movement); and Persistence (length of time engaged in activities in the face of obstacles) (Thomas & Chess, 1977).
Even in the early temperament literature, there was controversy about the validity of parent ratings of a child's temperament. When temperament is measured by parent report, there is the chance of a confound between parent ratings of the child's temperament and their own affective state. The research on the bidirectional relationship between mothers’ symptoms of depression or anxiety and their ratings of child temperament has been well documented (Ahmadzadeh et al., 2023; Britton, 2011; Cutrona & Troutman, 1986; Della Vedova, 2014; Hanington et al., 2010; Lin et al., 2017; Rode & Kiel, 2016; Vaughn et al., 1987; Whiffen, 1990). Understanding the bidirectional relationship is even more important for a parent with a mental health diagnosis. For example, maternal postpartum depression has been associated with ratings of more difficult infant temperament. In a study by Britton (2011), a depression and an anxiety measure were administered during the first postpartum month for mothers. Both of these measures correlated positively with temperament dimensions of infant activity, rhythmicity, approach, distractibility and overall temperament difficulties. Lin et al. (2017) found that higher Global Severity Index (GSI) scores on the Symptom Checklist‐90 (SCL‐90) among mothers were positively correlated with ratings of their child's temperament dimension scores of regularity, adaptability, intensity, mood and persistence.
While there is agreement between researchers that caregiver reporting can affect ratings of child temperament, there is evidence of some stability in mother's ratings of their child's temperament. The study by Ahmadzadeh et al. (2023) looked at the longitudinal relationship between mother's emotional state and the child's temperament traits and adjusted for genetic effects. Results showed that changes in mother's emotional symptoms showed small correlations with longitudinal changes in their ratings of the child's temperament. Another study by Asahara et al. (1992) showed that while child temperament ratings did tend to correspond to ease of handling the infant, the influence of child temperament and maternal perceptions of the rearing environment were not strong. There is also evidence that maternal ratings of temperament correlate with temperament ratings performed by trained observers (Wolk et al., 1992). This suggests that while some affective overlay is likely, there is validity in parental report of a child's temperament.
Recent research shows that temperament may also be affected by external variables such as the child's environments. A study by Abuhammad et al. (2020) looked at associations between infant temperament and environmental factors such as socioeconomic status (SES) and marital satisfaction. One interesting aspect of this study was that the study cohort was Jordanian, so data was not based on just a Western population. Results of this study showed three factors that significantly correlated with infant marital satisfaction, infant feeding method and assistance in caring for the infant. In terms of infant feeding method, they report that breastfeeding could result in higher scores on temperament domains due to fussiness. A study by Strickhouser and Sutin (2019) researched if SES in childhood was related to the development of a child's temperament. They hypothesized that children in more disadvantaged neighbourhoods would rate lower on sociability and persistence and higher on reactivity. Their hypotheses were supported and also showed that in this cohort of children studied from ages 4–15 years, the differences in temperaments did not decrease over time.
Family structure and stability are also shown to impact a child's temperament. Hong et al. (2015) explored the bi‐directionality of family cohesion and the temperament of the child. In their introduction, they cite that lack of family cohesion is related to negative parent–child interactions (Chen & Luster, 2002) while greater family cohesion is related to parent responsiveness and nurturing behaviour in parents (Behnke et al., 2008). In Hong's yearlong study of families, they found that family cohesion was positively related to mother's personality traits of Openness and negatively related to mother's scores on Neuroticism. Child temperament also contributed to parent behaviours during this study with children showing low effortful control (difficulty with ability to control behaviour) eliciting lower levels of parent‐positive support. The authors conclude that family cohesion changed over time as a function of both child and parent characteristics. Mothers with high positive support tended to increase this support over time and positive changes in mothers' negative control showed better effortful control in the child and more maternal agreeableness. Additionally, a meta‐analysis by Alving‐Jessep et al. (2022) reported emerging associations between child temperament and their gut microbiome. These studies point to external variables that may be associated with, and even influence, a child's temperament.
When looking at additional external variables, research shows that infant‐feeding mode may play a role in a child's temperament as well. The results, however, are not conclusive on which feeding mode is related to temperament domains and if these relationships are stable over time. Kielbratowska et al. (2015) found that mothers who breastfed (BF) rated their children higher in activity and reactions. Mothers who breastfed were more confident in their feeding and perceived their children to be less irritable during feeding. A study by Field et al. (2002) reports that mothers who BF rated their children more positively than mothers who formula‐fed (FF) their children. Abuhammed et al. (2020) found in a cohort of 278 women and their infants that in the first year of life, BF infants had higher temperament scores, suggesting more fussiness and being more ‘difficult to manage’ than FF infants. A study by Niegel et al. (2008) found that infants aged 6–18 months with more difficult temperament were significantly less likely to be exclusively BF for the recommended 6‐month period. At 12–14 months, there was no longer an association between infant temperament and BF. For the Niegel study, mother's demographic information was collected, but no measures of emotional state or family environmental factors were measured. Because of these inconsistencies, the interaction between early infant feeding and temperament is important to examine to determine if a clearer picture of this relationship can be found.
The current study explores the relationship between mother's ratings of emotional functioning and ratings of family environment and life stressors. This study further explores the differences in mothers' ratings of children's temperament in relation to maternal self‐ratings of emotional functioning. Data are also analysed to assess for differences in temperament based upon infant feeding mode. Based upon prior research, it was hypothesized that there would be a positive relationship between mother's psychiatric symptomatology and environmental stress. It was also hypothesized that a positive relationship would be found between mother's psychiatric symptomology (specifically anxiety and depression) and ratings of child's temperament. Given differences in the literature about BF infants and ratings of child's temperament, no a priori hypothesis is given for differences in temperament ratings for children who are BF.
Participants were 504 eligible mother–infant pairs enrolled in a longitudinal study between 2002 and 2011 when infants were less than 3 months of age. Participants who were 1–2 months postpartum were recruited from the Central Arkansas communities via advertisements such as flyers, print ads and radio ads. Table 1 reports on the cohort characteristics. Pregnancies were uncomplicated with no medical diagnoses (e.g., diabetes, pre‐eclampsia) or treatments known to affect foetal or infant growth and development (e.g., selective serotonin reuptake inhibitors, thyroid replacement). Mothers did not smoke, denied alcohol or substance use during pregnancy, and reported no use of soy products or other potential estrogenic compounds during pregnancy and/or lactation. Infants were term (>37 weeks), between 2.7 kg (6 lbs.) and 4.1 kg (9 lbs.) at birth, without known medical diagnoses and were not administered medications with recognized potential to affect human growth or development. Additional exclusion criteria included change of formula after age 2 months and before age 12 months; complementary foods before age 4 months (in compliance with AAP guidelines at the time of the study enrollment); and body weight at age 3 months of less than 5 kg (11 lbs). There were no inclusion or exclusion criteria relating to maternal emotional functioning, family environment, life stressors, or child temperament. Study visits included in this report were conducted at ages 6, 9, 12, 24, 36, 48, 60 and 72 months.
Parents selected their infants' diets prior to enrollment. Those deciding to formula feed chose between dairy‐based formulas (MF) (Similac® Advance® or Enfamil® Lipil®) or soy protein‐based formulas (SF) (Similac® Soy Isomil® or Enfamil® Prosobee®). Abbott Nutrition, Columbus, OH, manufactured Similac formulas and Mead Johnson, Evansville, IN, manufactured Enfamil formulas. All formulas contained supplemental docosahexaenoic acid (DHA) and arachidonic acid (ARA) and were provided to the study participants. Enrollment was performed in a diet‐type paced manner in an effort to maintain equal distribution across feeding groups throughout the study period. In this study, 174 children were BF, 168 were MF and 162 were SF. All formula‐fed infants remained on their selected formula from 2 to 12 months of age. Thus, formula‐fed participants did not change feeding group during the study period. For BF infants, exclusive breastfeeding (no infant formula intake) was encouraged until age 12 months. If this was not feasible, BF infants were supplemented with or weaned to MF between 6 and 12 months of age. Among the BF infants, 56% were exclusively breastfed until age 12 months, 23% were exclusively breastfed until age 9–12 months and 21% were exclusively breastfed until 6 months of age. If exclusive breastfeeding was not feasible until 6 months of age, the participants were not considered for the analyses. Complementary foods (e.g., juices, cereals and solid foods) could be introduced after 4 months for all diet groups.
Licensed master‐level psychologists, supervised by a licensed psychologist, administered all assessments and questionnaires during the participants' research study visit at the Arkansas Children's Nutrition Center. Examiners had specialized training in psychometric testing, were licensed by the state's Psychology board and were blinded to the study diet groups. The duration of typical study visits varied according to the participants' age (6, 9, 12, 24, 36, 48, 60 and 72 months) and ranged from approximately 1–4 h.
The Carey Temperament Scales (Carey & McDevitt, 1995). The Carey Temperament Scales (Carey) were administered to mothers to assess the temperamental characteristics of their children. Derived from the work of Thomas and Chess, the Carey Temperament Scales have been widely used for decades by medical and psychological researchers interested in the study of temperament (Vaughn et al., 1987). There are three scales used in the Carey series depending on the child's age. The Revised Infant Temperament Questionnaire (RITQ) was given to mothers of infants at the 6‐ and 9‐month visits. At 12 and 24 months, mothers responded to the Toddler Temperament Scale (TTS), and the Behavioural Style Questionnaire (BSQ) was utilized at the 36‐, 48‐, 60‐ and 72‐month visits. Each Carey questionnaire is comprised of up to 100 items with each item rated on a 6‐point Likert scale of frequency ranging from ‘almost never’ to ‘almost always’. When calculated, the item scores provide a category score for each of the nine temperament areas, which are compared to the norms for the category. Higher scores on each temperament category are indicative of more negative traits and temperament that is more difficult. For this investigation, five temperament categories, rhythmicity, approach, adaptability, intensity and mood, were analysed. These categories were selected because they are generally the temperament domains from which ‘easy’ and ‘difficult’ classifications are derived (Kielbratowska et al., 2015; WestEd, 2014).
Test–retest reliability for the five scales of the Carey range from r = 0.68 for children 1–4 months to r = 0.88 for children 8–12 years of age. Alpha reliabilities for the Carey scales range from a lower alpha score of r = 0.57 for children aged 4–8 months to an alpha score of r = 0.82 for 8–12 year olds. Data in the manual reports that alpha reliability increases as the child ages and that internal reliability in infant measures can be low due to the instability of behaviour during infancy (Carey & McDevitt, 1995, p. 9). The authors of the Carey scales acknowledge the concerns in establishing validity due to ‘distortion by the rater’ which may ‘compromise the usefulness of the data obtained’. They reported, however, that questionnaire ratings have ‘at least a moderate level of validity’ (Behavioural‐Developmental Initiatives, 2007, p. 1).
Symptom Assessment‐45 Questionnaire (SA‐45) (Davison et al., 1997). The SA‐45, a brief multidimensional measure of psychiatric symptomatology, was administered to mothers at each study visit. Based on the original SCL‐90, the SA‐45 is comprised of 45 questions rated on a 5‐point Likert scale. This assessment yields T‐scores for nine symptom domains as well as a GSI, which provides an overall level of psychiatric symptomology. T‐scores of 65 or greater for individuals outside of a treatment setting suggest a potential problem area that could warrant further clinical investigation. In addition to the GSI, subscales investigated in this study Anxiety, which explores symptoms such as fearfulness, tension and restlessness; Depression, which assesses feelings of loneliness, hopelessness, worthlessness; Hostility, which examines experiences like uncontrollable temper outbursts, frequent arguments and destructive urges; and Interpersonal Sensitivity, which looks into the respondent's feelings of inferiority or self‐consciousness, in relation to others. Cronbach's α scores for the SA‐45 show internal consistency of 0.71 or greater for all domain scales. For adults, Cronbach's α correlations range from r = 0.74 (Anxiety and Psychoticism) to r = 0.87 (Depression). Test–retest reliability results for an adult nonpatient sample was r = 0.87 for the GSI (Strategic Advantage Inc, 2000).
Family Adaptability and Cohesion Evaluation Scale‐II (FACES‐II) (Olson et al., 1982). The FACES‐II, a self‐report questionnaire that measures an individual's perceptions of cohesion and adaptability within the family system, was administered to mothers. This 30‐item scale contains 16 cohesion items and 14 adaptability items. Cohesion is characterized as the degree of emotional bonding that family members have towards one another. Adaptability is described as the extent to which the family system is flexible and able to change in response to situational and developmental stress. High scores indicate greater cohesion and adaptability. Both Cohesion and Adaptability scores were obtained at the 12‐, 24‐, 36‐, 48‐, 60‐ and 72‐month visits.
Reliability for the FACES‐II is strong with test–retest reliability scores of r = 0.84 for the full scale; r = 0.83 for Cohesion; and r = 0.80 for Adaptability. Time lapse between test administrations was 4–5 weeks. Cronbach's α analysis for the measure showed correlations of 0.87 for Cohesion, 0.78 for Adaptability and 0.90 for the total score. Discriminant and predictive validity were done using data previously collected by the authors and results ranged from 100% for childless couples to 82% for couples with young children (Olson et al., 1983).
Family inventory of life events and changes (FILE) (McCubbin & Patterson, 1983). The FILE is a self‐report instrument designed to record the events, changes and stressors experienced within a family unit in the past year. Higher scores on the FILE reflect more change and stress for the family. The 71 items are grouped into nine subscales (Intrafamily Strains; Martial Strains; Pregnancy and Childbearing Strains; Finance and Business Strains; Work–Family Transitions and Strains; Illness and Family ‘Care’ Strains; Losses; Transitions ‘In and Out’; and Family Legal Violations). Each item is assigned a standard weight indicating the relative magnitude of the particular event or strain. Scores for each subscale are computed by summing the items in the subscale, and then these overall nine subscale scores are summed to determine the Total Family Pile‐Up score, with higher scores indicating more stress and problematic events. Cut‐off scores are used to determine whether the family is experiencing a low, moderate or high stress level. The Total Pile‐Up scores were calculated at the 12‐, 24‐, 36‐, 48‐, 60‐ and 72‐month visits.
Test–retest reliability correlations for the nine subscales range from r = 0.64 (Finance and Business Strains) to r = 0.84 (Pregnancy and Childbearing Strains) with a Total Family Pile‐Up score correlation of r = 0.80. Internal reliabilities for these subscales are low and not reported. Authors state that the wide variability of family's life events make internal reliability parameters unstable and not appropriate to calculate. Validity reported by the authors reveal that family strain was significantly and inversely correlated with measures of changes in a child's health status over time (McCubbin & Patterson, 1983).
Information about mother's age, marital status, parental educational level, parental income, child sex and child race were obtained on self‐administered questionnaires upon enrollment. Mothers also provided information about child's birth weight, birth length, gestational age and developmental or mental health disorders.
Wechsler Abbreviated Scale of Intelligence (WASI) (Wechsler, 1999). The WASI was administered to mothers at the 3‐month study visit. The WASI, an abbreviated scale measuring an individual's intelligence quotient (IQ), assesses verbal (Verbal Comprehension Index), visual (Perceptual Reasoning Index), and general cognitive abilities (Full Scale IQ). Four subtests were used to measure cognitive skills (FSIQ‐4). The verbal subtests are Vocabulary and Similarities. Perceptual Reasoning subtests are Block Design and Matrix Reasoning. Subtest scores are converted to T‐scores, and Verbal Comprehension, Perceptual Reasoning, and Full Scale IQ scores are reported as standard scores. The Full Scale IQ score was used to adjust results to prevent potential confounding with mother's cognitive abilities.
Psychometrics for the WASI are strong. Internal consistency for adult norms shows correlations of r = 0.98 for the FSIQ‐4 for the adult sample. Test–retest reliability for the adult sample shows correlations of r = 0.92 for a mean retest interval of 31 days. Validity of the FSIQ ‐4 compared to the Wechsler Adult Intelligence Scale‐III reports was r = 0.92.
Participant characteristics measured on a continuous scale were summarized as means and standard deviations and compared across feeding groups with a one‐way analysis of variance (ANOVA). Participant characteristics measured categorically were summarized as counts and percentages and compared across feeding groups using Pearson's χ ^2^ test. Scores for the SA‐45, Carey, FACES and FILE assessments were compared across feeding groups at each time point using ANOVAs. Linear mixed‐effect models were constructed to investigate the relationship of maternal affection and family environment throughout the study while controlling for significant covariates including infant sex, maternal IQ, paternal education and gestational age. Linear mixed effect models were also constructed to investigate the relationship of child temperament and maternal affect throughout the study while controlling for significant covariates including gestational age, infant sex, race, maternal IQ, maternal education and income, and paternal education and income. Corresponding estimated marginal means and their standard errors at each study visit were obtained. Sensitivity analyses were performed to identify whether differences were observed when considering only breastfed infants who were not exposed to formula during the 12 months period. Statistical tests were considered statistically significant at a prespecified 0.05 level. Analyses were performed in R software (version 3.5.1) using the tidyverse (version 1.3.0), furniture (version 1.9.7), lmerTest (version 3.1‐2) and emmeans (version 1.4.7) packages.
Institutional Review Board approval for the study was obtained and consent was obtained for all subjects.
Of the 600 participants initially enrolled in this study, 96 did not meet inclusion criteria, moved from the area, were lost to follow‐up, or chose to withdraw from the study before the 3‐month baseline visit. The final cohort analysis consisted of 504 subjects (174 BF, 168 MF and 162 SF mother–infant dyads) (Table 1). The final study visit conducted with 72‐month‐old children and their mothers included 370 (62%) participants.
This cohort was primarily Caucasian (88%, Table 1) and from married or cohabitating households (93%). No significant differences were found between feeding groups for paternal income, maternal age at delivery or child sex. Mothers of BF infants reported significantly higher educational achievement with 29.3% attaining a graduate degree or training compared to mothers of MF and SF infants (19.6% and 24.7%, respectively). However, less income data was reported from mothers who BF (49.4% not reported). Paternal education data showed more fathers of BF infants earning graduate degrees or training (17.8%) than fathers of MF infants (13.1%) or SF infants (6.8%). While all mothers’ mean cognitive scores were within the average range, as a group, mothers of BF infants had significantly higher Full Scale IQ scores (109.5 ± 10.1) than mothers of MF and SF infants (105.1 ± 9.2 and 103.0 ± 11.5, respectively, p < 0.001). Gestational age (39 weeks) was statistically different between BF infants compared to MF and SF infants, p < 0.001); however, actual differences were only within 3 days of delivery.
Differences between the subjects who completed the 72‐month study versus noncompleters' data showed completer mothers had two‐point higher IQ (statistically significant but not clinically significant) (p = 0.039); higher educational level (p = 0.034) and had more female children compared to mothers who did not complete the study (p = 0.011, Table 2).
For the total sample, mean T‐scores on the SA‐45 for ratings of global psychiatric difficulties (GSI) and the four domain scores assessed (Anxiety, Depression, Hostility and Interpersonal Sensitivity) were below the clinical threshold. No statistically significant differences were found in maternal ratings on any of these domains or the GSI when mothers were stratified based upon feeding group at any visit (Table 3).
Even with subclinical SA‐45 scores, results showed significant relationships between psychiatric difficulties and family environment. Ratings of environmental stressors (FILE) showed a significant positive correlation with the SA‐45 GSI (p < 0.001, Table 4). This significant correlation with environmental stressors was also present for all four SA‐45 domain scores after adjusting for significant covariates (p < 0.001 for Anxiety, Depression, Hostility and Interpersonal Hostility). These results suggest that households with more stressful events are associated with higher maternal psychiatric difficulties ratings.
Similar results were found for psychiatric difficulties and ratings of family Cohesion and Adaptability on the FACES‐II. The FACES‐II Cohesion score shows a significant negative correlation with mother's GSI (p < 0.001) and the four domain scores (p = 0.002 for Anxiety and p < 0.001 for Depression, Hostility and Interpersonal Sensitivity), indicating that lower ratings of family cohesion were associated with higher ratings of psychiatric difficulties.
The FACES‐II Adaptability score was also significantly negatively correlated with the mother's Depression domain score and the GSI score (p = 0.001 and 0.014, respectively, Table 4), demonstrating that lower adaptability of the family is associated with higher ratings of depression and global psychiatric symptoms.
Relationships between ratings of maternal psychiatric difficulties and child temperament, adjusting for covariates, are provided in Table 5. Mother's GSI showed a positive association with the child's temperament ratings of Rhythmicity (p = 0.012), Adaptability (p < 0.001) and Mood (p < 0.001). The SA‐45 domain of mother's Hostility showed a positive association with ratings of child's Rhythmicity (p = 0.024), as well. No significant associations were seen between GSI or SA‐45 domain scores on other child temperament ratings.
Differences in child temperament by feeding group during the first 6 years of life adjusted for covariates are shown in Figure 1. Significant transient differences in Rhythmicity and Adaptability were seen for SF and BF children. For Rhythmicity, BF infants had higher scores at 6 months (p = 0.01) compared to MF infants. SF infants had higher Rhythmicity at 4 years compared to MF infants (p < 0.05). SF infants also had higher Rhythmicity values at 5 and 6 years compared to BF infants (p < 0.05 and 0.01, respectively). For Adaptability, BF infants had significantly higher scores at 6 months (p = 0.05) compared to MF infants. There were no lasting differences in temperament scores throughout the first 6 years of life based upon infant feeding group. Among the BF group, we further examined whether the length of BF affected child temperament over time and found no significant association with any child temperament trait studied (data not shown). Sensitivity analyses were conducted on children who were breastfed until age 12 months without switching to formula (56%) for each outcome, which found similar results.
Figure 1 Temperament category scores by feeding group (a) rhythmicity, (b) approach, (c) adaptability, (d) hostility, and (e) mood.
This longitudinal analysis of mother–infant pairs explores the potential association between maternal psychiatric symptomatology, environment and child temperament from ages 6–72 months. The study population was comprised of mothers with subthreshold self‐ratings of psychiatric difficulties. However, ratings showed more psychiatric difficulties correlated with environmental stressors, impaired family dynamics and elevations in some temperament ratings when adjusted for significant covariates. Infants stratified based upon feeding groups (breastfed; soy‐fed; milk‐fed) showed transient difficulties in Rhythmicity and Adaptability between soy and breastfed children, but these differences did not remain.
Mean scores for the measure used to rate psychiatric issues (SA‐45) were below the clinical threshold for all subjects. However, higher maternal ratings on the SA‐45 Global Stress Index and the four domain scales chosen for this study were all significantly related to environmental stressors (FILE) and lack of family Cohesion and family Adaptability (FACES‐II). While causality is not implied, these results show that there is a relationship between family environment and mothers’ ratings of psychiatric difficulties even in mothers whose self‐ratings are subthreshold for psychiatric symptoms. Previous studies of maternal mood have similarly suggested that factors such as family stress, negative life events, poor resiliency and perceived lack of social support are associated with mother's self‐report of psychiatric difficulties (Beck, 2001; Britton, 2008). This study is consistent with these findings even in a group where significant levels of psychiatric difficulties are not reported.
This study also finds that there was a positive relationship between two of the mothers' psychiatric difficulties ratings (GSI and Hostility) and the ratings of their children's Rhythmicity, Mood and Adaptability temperament scores. In this study, no ratings of difficulties in child temperament were significantly correlated with mother's ratings of her own Depression or Anxiety. This was a surprising outcome because most studies report that mothers with elevated Anxiety and Depression have been associated with more difficult child temperament (Ahmadzadeh et al., 2023; Hanington et al., 2010; Vaughn et al., 1987). It is also interesting that ratings of the child's Approach or Intensity were not significantly correlated with mother's psychiatric ratings on any of the scales on the SA‐45. One reason for this difference may be due to specific enrollment criteria in the study, which excluded mothers with diagnosed mental health disorders. Another contributing factor may be more recent research findings that suggest that elevations are not always found in just a few variables for mother or infants. Ahmadzadeh et al. (2023) state that mothers who self‐report higher levels of emotional symptoms in themselves are more likely to report higher levels of emotionality in their infants, but it may not correlate with the range of temperament traits typically assumed.
This study found intermittent differences for the SF and BF infants' temperament ratings in the areas of child's Mood and Rhythmicity, but these changes were not sustained over time. Literature in temperament related to feeding method is not expansive, and it is also inconsistent. One study, Abuhammad et al. (2020) found BF babies were rated fussier than FF babies. Data from the Norwegian Mother, Father and Child Cohort Study (Niegel et al., 2008) found that infants with difficult temperaments were less likely to be breastfed, but by 14 months, these differences were no longer observed. This study's conclusion concurs with the Neigel data that any differences over time do not persist.
This study is strengthened by a large sample of mother–child pairs who were investigated longitudinally from the children's infancy to early childhood. It is also strengthened by adherence to a strict infant diet in the first year of life that was provided to them. Subjects were well characterized, and appropriate covariates were controlled for in data analysis.
Some limitations of this study are important to consider. The participants in this study are a nonclinical sample of mothers and infants. Most of the mothers are Caucasian and have higher education and average cognitive scores. This does skew the data and may limit generalizability.
The study did not randomize infants to feeding group for ethical reasons. Because parents self‐selected their babies’ diets, motivation and personality traits may be associated with certain affective states by feeding group (Kielbratowska et al., 2015). There is the lack of racial diversity in the present cohort, which may limit the generalizability of our results to other racial groups. Utilizing parent‐report measures to assess child temperament is often considered a study limitation as a potentially biased means of data collection (Britton, 2011; Lin et al., 2017; Vaughn et al., 1987) although it has been reported that ratings of temperament made by mothers correlate significantly with assessments of temperament made in a laboratory environment by trained examiners (Britton, 2011). Finally, the investigation of an association between child temperament and duration of breastfeeding may have been compromised by lower power since the analysis included just one‐third of our sample at 72 months.
In conclusion, although mothers in this study were not endorsing clinical‐level psychiatric symptomatology, higher psychiatric symptomatology significantly correlated with environmental stressors, impaired family dynamics and elevations in temperament ratings of infants' Rhythmicity, Adaptability and Mood when adjusted for significant covariates. There were no lasting differences for temperament between feeding groups. However, some transient increases in Rhythmicity and Adaptability were found between SF and BF children. These results suggest that researchers and clinicians should be aware of family environmental stress and mother's affective state even if these variables are not at a clinically elevated level. Mothers' ratings of own affect and family stress may be related to their ratings of child's temperament, but this longitudinal study suggests that these ratings are transient and are not associated with feeding methods for the mother and child dyad. This information may be helpful for mothers as they are beginning their journey as new mothers and trying to navigate new experiences with infant feeding. Future research should explore family environmental stress when exploring child temperament and consider feeding practices at young ages.
Ginger McCorkle performed research. Aline Andres, Jayne Bellando and Patrick H. Casey designed research study. Clark R. Sims, Aline Andres, Jayne Bellando and Ginger McCorkle analysed and interpreted the data. Ginger McCorkle, Jayne Bellando, Clark R. Sims, Aline Andres and Trevor Durey wrote the paper. Aline Andres, Ginger McCorkle, Jayne Bellando, Patrick H. Casey, Trevor Durey and Seth T. Sorensen provided expert input on text and editing submission.
The authors declare no conflict of interest.
The authors are grateful to the children and families who participated in this study and to the clinical core team for their assistance in data collection. All phases of this study were supported by USDA‐ARS Project 6026‐51000‐010‐05S.
McCorkle, G. , Andres, A. , Sims, C. R. , Casey, P. H. , Sorensen, S. T. , Durey, T. , & Bellando, J. (2025). Exploring the relationship between child temperament, maternal psychiatric symptoms, family environment and infant feeding. Maternal & Child Nutrition, 21, e13728. 10.1111/mcn.13728
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. Individual participant data will not be made available. Study of Early Infant Diet on Growth and Development (Beginnings), NCT00616395, https://clinicaltrials.gov/ct2/show/NCT00616395.
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. Individual participant data will not be made available. Study of Early Infant Diet on Growth and Development (Beginnings), NCT00616395, https://clinicaltrials.gov/ct2/show/NCT00616395.