OPEN High care and low overprotection

from both paternal and maternal

parents predict a secure
attachment style with a partner
among perinatal Japanese
women
Ekachaeryanti Zain1,2, Naoki Fukui1, Yuichiro Watanabe1, Koyo Hashijiri1, Takaharu Motegi1,
Maki Ogawa1, Jun Egawa1 & Toshiyuki Someya1
This study aimed to determine how paternal and maternal parenting before adolescence affects adult
attachment to a partner during the perinatal period, using three different models of attachment.
We used the Parental Bonding Instrument (PBI) and the Relationship Questionnaire (RQ) to examine
perceived parenting practices and adult attachment styles, respectively. The participants included
4586 Japanese women who were pregnant or who had given birth, up until one month after childbirth.
We performed structural equation modeling analysis between PBI and RQ scores with three different
category models, including the four-category model (secure, fearful, preoccupied, and dismissive
attachment) as Model 1, the two-category model (model of the self and others) as Model 2, and the
single-category model (total attachment style) as Model 3. Models 1 and 2 showed a good fit. Both
path models showed a significant association between adult attachment style and perceived paternal
and maternal parenting before adolescence, where high care and low overprotection from both
paternal and maternal parents predicted adult attachment. Our findings indicate that attachment
styles are best described using the four-category and two-category models, and suggest that both
paternal and maternal overprotection and care influence adult attachment with a partner during the
perinatal period.

The adult attachment style of a pregnant woman, specifically, attachment to a romantic partner, is an essential
factor in mentally preparing for the transformation in self-identity from a person who receives care to one who
provides care as a parent1. In addition, the quality of the maternal–fetal relationship, which relies on the care
provider, is associated with a pregnant woman’s attachment style to her partner 2. Furthermore, attachment can
influence how an expectant mother builds an emotional relationship, bonds, and attaches to the fetus 3. Thus,
the adult attachment style affects the mental well-being of mothers and fetuses during the perinatal period, and
continues to have an influence during the formation of antenatal attachment to the newborn. For instance, a
mother with insecure adult attachment to her partner may perceive her infant’s temperament as more difficult
compared with a mother with secure attachment. This could be caused by a less caring and less supportive
attitude toward children4. An insecure adult attachment style with respect to one’s partner has been related to
antenatal and postnatal depression5,6, maternal separation anxiety7, difficulties regulating emotions8, increased
vulnerability to stress9, greater parenting stress10,11, greater levels of suicidal ideation, and more frequent
suicide attempts in women during the perinatal period12. In contrast, individuals with a secure attachment style
appear to have better mental well-being and parenting motivation, and to provide better parental care13–15.
According to the former attachment theory developed by Bowlby 16,17, “the infant and young child should
experience a warm, intimate, and continuous relationship with his mother (or permanent mother substitute) in
which both find satisfaction and enjoyment” to develop an adult attachment style that is secure. Furthermore,
the relationship between a child and parent contributes to the formation of internal representations or work-
ing models of the self in relation to others, which determine one’s attachment style. The self-model represents

1
Department of Psychiatry, Niigata University Graduate School of Medical and Dental Sciences, 757
Asahimachidori-ichibancho, Chuo-ku, Niigata 951-8510, Japan. 2Department of Psychiatry, Faculty of Medicine,
Scientific Reports | (2023) 13:15684 | https://doi.org/10.1038/s41598-023-42674-1 1
 Mulawarman University, Samarinda, Indonesia. email: psy@med.niigata-u.ac.jp

Scientific Reports | (2023) 13:15684 | https://doi.org/10.1038/s41598-023-42674-1 2

www.nature.com/scientificreports/

one’s self-worthiness or self-reliance, and the other-model represents the availability or the need for closeness
and presence of others in a relationship. This is known as the two-category model of adult attachment style. A
secure attachment style is characterized by both a positive self-model and a positive other-model, such that
one is self-assured while simultaneously valuing relationships with others18.
Bartholomew and Horowitz expanded the two-category attachment model proposed by Bowlby into a
four- category model, which includes secure, fearful, preoccupied, and dismissive attachment styles 19. They also
devel- oped the Relationship Questionnaire (RQ) to measure adult attachment styles to a partner 19. Individuals
with secure attachment styles are comfortable with autonomy and intimacy, while individuals with fearful
attachment styles are dependent and socially avoidant. Individuals with preoccupied attachment styles show
strong depend- ency on others and are preoccupied with intimacy. Meanwhile, individuals with dismissive
attachment styles are counter-dependent and avoidant of intimacy. The last three styles are considered to be
forms of insecure attachment.
Adult attachment has been associated with various factors, such as the quality of an individual’s childhood
caregiving environment (e.g., maternal depression, father absence), the perceived quality of the parenting they
received during childhood, their social competence, and the quality of their best friendships throughout their
lifetime20. In this study, we focused on the perceived quality of parenting before adolescence, as evaluated ret-
rospectively by expectant mothers. According to the Parental Bonding Instrument (PBI) 21, perceived parenting
practices can be categorized into two factors: care and overprotection. Parental care is defined as the provision
of affection, emotional warmth, and autonomy, whereas parental overprotection is characterized by emotional
coldness, control, and interference. A previous study suggested that a secure attachment style in women can
be predicted by low levels of maternal overprotection and high levels of maternal and paternal care 22. Another
prospective cohort study similarly reported that perceptions of optimal maternal parenting, i.e., high care and
low control from one’s mother, contributed to secure adult attachment in women 30 years after an initial
assess- ment of perceived parenting23. In addition, higher perceived levels of paternal care in women resulted
in a more positive self-image23. These findings highlight the importance of both paternal and maternal
parenting style, and suggest that parenting style is strongly associated with attachment formation throughout
the lifespan. However, no studies have examined whether paternal overprotection significantly predicts secure
attachment in female offspring. Furthermore, the way in which attachment style influences women in the
perinatal period, which includes the beginning of pregnancy and extends to the postpartum period, is
unknown. Perinatal women are considered to be a vulnerable population requiring specialized mental health
care because of the significant potential consequences of perinatal mental illness24. Therefore, understanding
the nature of adult attachment and the contribution of perceived parenting practices according to parental
gender could advance perinatal mental healthcare.
Previous studies have reported causal models with a good fit to the two-category14 and four-category22 mod-
els of adult attachment style in relation to perceived parenting style. However, no studies have examined this
topic using a single-category model of total attachment style (TAS) 25 with a path model analysis. In addition,
all previous studies used a model that combined the paternal and maternal PBI data in the latent variable of
the model. Our previous study indicated that parenting by one’s mother, as opposed to their father, might be
more important for maintaining satisfactory maternal-fetus bonding in women26. Among new mothers, the
effects of parenting by their mothers might differ from that of their fathers according to specific life events
and mental health status. To address this in the present study, we analyzed PBI data regarding the fathers and
mothers of new mothers using a path model in which the data from the two parents were not combined as the
latent variable.
The goal of this study was to determine how perceived pre-adolescence paternal and maternal parenting
affected adult attachment style using three attachment models. Prior to performing path analyses, we
formulated the following hypotheses regarding the expected relationships among the variables. First, we
hypothesized that the best-fitting path model would result in a reliable correlation between adult attachment
style and perceived paternal and maternal parenting. Second, we hypothesized that high levels of paternal care
and low levels of paternal overprotection would predict a secure attachment style, which is characterized by
high secure and low fearful, preoccupied, and dismissive attachment styles in Model 1, by a high positive self-
model and high positive other-model in Model 2, and by a high TAS in Model 3. Third, we hypothesized that
high levels of maternal care and low levels of maternal overprotection would similarly predict the
abovementioned relationships.

Methods
Ethics statement. This study involving human subjects followed the principles of the Declaration of Hel-
sinki. We obtained approval from the ethics committee of Niigata University (Approval Number: 2016–0019)
and the ethics committees of the participating obstetric institutions. We obtained written informed consent
from all participants.

Participants. This study was part of the Perinatal Mental Health Research Project 26–31, which included the
Department of Obstetrics and Gynecology at Niigata University Medical and Dental Hospital as well as 33
asso- ciated obstetric institutions in Niigata prefecture, Japan. Between March 2017 and March 2021, we
distributed a large-scale self-report questionnaire survey to pregnant Japanese women aged 18 years and older.
The exclusion criteria were serious physical complications, serious pregnancy complications, or severe
psychiatric disorders (e.g., severe schizophrenia or severe depression). Participants completed the RQ and PBI
questionnaires. Some participants were from the same dataset as our previous study, which found that
perceived negative parenting before adolescence had direct and indirect effects (via anxiety and depression)
on maternal-infant bonding in the perinatal period in 1301 pregnant Japanese women26.

Scientific (2023) https://doi.org/10.1038/s41598-023- 3

www.nature.com/scientificreports/

Measures. The RQ measures adult attachment style in a relationship with a partner19. It measures four adult
attachment styles: secure, fearful, preoccupied, and dismissive. Each category is described using a paragraph
of statements and has a seven-point Likert-type scale ranging from 1 = “Does not apply to me at all” to 7 =
“Applies to me very much”. We instructed the participants to choose the number that best described their
relationship with their partner. Previous studies have reported the reliability 18 and validity19 of the RQ. We
used a Japanese version that was validated previously in Japanese adolescents 25. There are three different
model structures for categorizing and scoring adult attachment styles in the currently available literature. The
first is the four-category model (Model 1), which includes the original dimensions of the RQ: secure, fearful,
preoccupied, and dismissive attachment19. Each category in Model 1 is given a score according to the RQ’s
scoring method. The second is the two-category model (Model 2): self-model and other-model. Since we used
the RQ, we used the following formula to obtain a score for each category in Model 2: self-model = secure −
fearful − preoccupied + dismissive; other-model = secure − fearful + preoccupied − dismissive14,18. The
formula for scoring was in accordance with the category quadrants in Model 1, where both secure and
dismissive attachment styles had a positive quadrant in terms of self-reliance or autonomy, and both secure
and preoccupied attachment styles had a positive quad- rant in terms of the presence of others in a
relationship19. Thus, scores for RQ categories that were in a negative quadrant in Model 2 were included as a
deduction in the formula. The last is a single-category model (Model 3): total attachment style (TAS) score =
secure − fearful − preoccupied − dismissive25. This model is formulated from a single bipolar factor
representing the secure and insecure attachment styles, where all three insecure attach- ments are deduction
factors.
The PBI measures participants’ subjective experience of the parenting attitude they perceived their parents
as having, according to their memory, during their first 16 years21. We used two identical instruments,
adminis- tered independently, to assess perceived paternal and maternal parenting practices. The self-report
instrument consisted of 25 items, including 12 items representing parental care and 13 items representing
parental over- protection. We instructed the participants to rate each item, which was a statement about a
specific parenting behavior, according to how probable that had perceived that behavior to be. They used a
four-point Likert-type scale ranging from 0 = “Very unlikely” to 3 = “Very likely”. The PBI categorizes perceived
parenting practices into two subscales that we used in this study: care and overprotection. The care subscale
measures positive parenting, including items suggesting parental attitudes of care, affection, sensitivity,
cooperation, accessibility, and encour- agement of autonomy and independence. The overprotection subscale
measures negative parenting, including items suggesting parental attitudes of indifference, strictness,
punitiveness, rejection, interference, control, and constraint. We used the validated Japanese version of the
PBI32.
We administered the questionnaires to each participant on separate printed paper forms. We did not change
the order of the items, but presented them according to the previously validated version of each original ques-
tionnaire. We sequentially grouped the forms in the following order: paternal PBI, maternal PBI, and RQ, in
one set. Therefore, we expected that most participants responded to the questionnaires in the order in which
they were presented.

Statistical analyses. We performed structural equation modeling analysis to find correlations between the
RQ subscales and PBI subscales. We analyzed the paternal and maternal PBI forms interdependently. The PBI
subscales included care and overprotection. The RQ subscales were differentiated into three different models
of adult attachment styles, which varied according to the number of categories in the model. Model 1 used a
four- category model, which included secure, fearful, preoccupied, and dismissive attachment as the RQ
subscales. Model 2 used a two-category model, which included the self-model and other-model. Model 3 used
a single- category model, which included the TAS. In Models 1 and 2, we combined the RQ subscales in a
latent variable corresponding to attachment style. For each model, we drew paths from each of the PBI
subscales of care and overprotection to the latent variable of attachment style. Then, we performed covariance
structure analysis on each path model. After these analyses, we retained statistically significant paths (P <
0.05). We adopted the com- parative fit index (CFI) and the root mean square error of approximation
(RMSEA) as indices of the goodness of fit (CFI ≥ 0.95 and RMSEA ≤ 0.08)33 between the models and the data.
All statistical analyses were performed using SPSS version 25 (IBM Corp., Armonk, NY, USA) and Amos 25.0.0
(IBM Japan, Tokyo, Japan).

Results
We included all data from 4586 pregnant women who completed both the RQ and PBI questionnaires, with no
missing values (Table 1). We also included questionnaires from pregnant women who had given birth during
the course of questionnaire collection, such that they completed the questionnaire up to 1 month after
childbirth.
Model 1 and Model 2 met the criteria for a good fit to the data (CFI = 0.990 and 0.995, respectively;
RMSEA = 0.036 and 0.046, respectively), whereas Model 3 was not a good fit to the data (CFI = 1.000 and
RMSEA = 0.356; Table 2).
In Model 1, paternal care, paternal overprotection, maternal care, and maternal overprotection predicted
attachment style (r = 0.177, − 0.101, 0.131, and − 0.137, respectively; all P < 0.001; Fig. 1). Attachment style
included secure, fearful, preoccupied, and dismissive attachment styles (r = 0.439, − 0.748, − 0.602, and − 0.455,
respectively; all P < 0.001).
In Model 2, paternal care, paternal overprotection, maternal care, and maternal overprotection predicted
attachment style (r = 0.188, − 0.091, 0.116, and − 0.114, respectively; all P < 0.001; Fig. 2). Attachment style
included the self-model and the other-model (r = 0.695 and 0.553, respectively; all P < 0.001).
The correlation coefficients between the RQ subscales of Model 1 and Model 2 are given in

Scientific (2023) https://doi.org/10.1038/s41598-023- 4

www.nature.com/scientificreports/
Supplementary Tables 1a and 1b.

Scientific (2023) https://doi.org/10.1038/s41598-023- 5

www.nature.com/scientificreports/

Variable Value
Age (years) 31.94 ± 4.82
Parity (primipara/multipara) 2208/2378
Gestational age (T1/T2/T3) 2995/1238/353
PBI scores
Paternal care 26.32 ± 7.23
Paternal overprotection 7.37 ± 5.43
Maternal care 30.48 ± 6.02
Maternal overprotection 7.92 ± 6.39
RQ scores
Model 1
Secure 4.13 ± 1.62
Fearful 2.82 ± 1.71
Preoccupied 2.65 ± 1.72
Dismissive 2.01 ± 1.34
Model 2
Self-model 0.67 ± 3.42
Other-model 1.94 ± 3.06
Model 3
Total attachment style –3.36 ± 4.33

Table 1. Characteristics of participants (n = 4586). Data are expressed as mean ± standard deviation.
PBI, Parental Bonding Instrument; RQ, Relationship Questionnaire; T1, 12–15 weeks of pregnancy; T2,
30–34 weeks of pregnancy; T3, 4 weeks after childbirth.

Indicator Model 1 Model 2 Model 3

Factor Secure, fearful, preoccupied, and dismissive Self- and other-models Total attachment style
CFI 0.990 0.995 1.000
RMSEA 0.036 0.046 0.356

Table 2. Indicators for the three attachment models. CFI, comparative fit index; RMSEA, root mean square
error of approximation.

Figure 1. Structural equation modeling for Model 1. Abbreviations: OP, overprotection; PBI, parental bonding
instrument; RQ, relationship questionnaire. Secure, fearful, preoccupied, and dismissive attachment styles were
the observable variables. Attachment style was a latent variable composed of secure, fearful, preoccupied, and
dismissive attachment. Care and OP were observable variables.

Scientific (2023) https://doi.org/10.1038/s41598-023- 6

www.nature.com/scientificreports/

Figure 2. Structural equation modeling for Model 2. Abbreviations: OP, overprotection; PBI, parental bonding
instrument; RQ, relationship questionnaire. Self-model = secure − fearful − preoccupied + dismissive , and
other- model = secure – fearful + preoccupied − dismissive . Attachment style was a latent variable composed of
a self- model and an other-model. Care and OP were observable variables.

Discussion
The present study revealed that both Model 1, which was a four-category model, and Model 2, which was a
two-category model, had a path model with good fit. However, Model 3, which was a single-category model,
had a bad fit with the dataset. Since SEM is a powerful tool for examining the goodness of fit of a causal
model according to both scientific and statistical perspectives 34, the results of our analysis with Model 1 and
Model 2 strongly suggest that adult attachment style is related to perceived paternal and maternal parenting
before ado- lescence. It is essential to apply the correct model for assessing adult attachment. Model 3,
previously proposed by Matsuoka et al.25, has never been tested via SEM analysis. The present study revealed
that a single-category TAS model might not be appropriate for assessing adult attachment style. Furthermore,
it may not be possible to accurately assess attachment style using a model with a single category, as it may not
be possible to simplify the basic categorization enabled by a two-category or four-category model. Instead,
each category may describe a relationship, depending on specific life events and variations across
relationships. According to Bowlby17, attachment styles are formed by internal working models of
interpersonal interactions throughout life, as an individual forms self- and other-models that lead them to seek
or avoid various interactions in close relation- ships. The contributions of both fundamental dimensions
define the underlying measures of attachment style in social relationships18. We used a SEM analysis to test
the compatibility of Model 3 with the four-category model proposed by Bartholomew and Horowitz19. Our
data show that Model 1 and Model 2 can represent the constructs of the underlying theoretical framework of
adult attachment with good reliability and validity, while this was not the case for Model 3.
In this study, the good fit of the two models suggested that there exists a significant association between
adult
attachment style and perceived paternal and maternal parenting before adolescence. We analyzed parenting by
combining the paternal and maternal PBI forms but did not use latent variables of care and overprotection in
the models. We included all observable variables of care and overprotection from fathers and mothers,
respectively. Previous studies in the Japanese population that involved PBI variables and SEM analysis also
combined pater- nal and maternal parenting in their analyses and models 14,22,26. Among these, one study
combined paternal and maternal parenting in latent variables 14. However, analyzing the paternal and maternal
PBI without combining them in latent variables clarified that perceived parenting from a mother differs from
that of a father in terms of the influence on adult attachment style. The present study suggested that both
perceived paternal and maternal care are important during the perinatal period in terms of secure attachment to
a partner. In that regard, paternal parenting with high levels of care and low levels of overprotection predicted
a secure adult attachment style. This was the same for maternal parenting. This finding represents new
evidence for how perceived paternal care and overprotection might contribute equally to maternal parenting in
terms of forming adult attachment.
Previous studies using SEM analysis found that the latent variable of PBI overprotection was not associ-
ated with adult attachment style in 363 Japanese mothers 14, and that paternal overprotection without a latent
variable was not associated with adult attachment style in 2,709 Japanese female university students 22. In the
current study, which had a large sample size, we conducted SEM analysis without latent variables from the
PBI subscales of care and overprotection. Our data indicated that both paternal and maternal overprotection
and care influenced attachment with a partner in adulthood. However, Shiraishi et al. found that the latent
variable of PBI overprotection was not associated with adult attachment style in 363 Japanese mothers14. We
conducted direct measurements with observable variables to represent the collected scores of our PBI data.
This enabled us to test our hypotheses regarding the way in which adult attachment styles were independently

Scientific (2023) https://doi.org/10.1038/s41598-023- 7

www.nature.com/scientificreports/
associated with

Scientific (2023) https://doi.org/10.1038/s41598-023- 8

www.nature.com/scientificreports/

paternal care, paternal overprotection, maternal care, and maternal overprotection. The independent associa-
tions among these observable variables cannot be directly measured when latent variables are used 35. The use
of a latent variable generally corresponds with hypothetical constructs that have unobservable referents, but
their existence can be posited to explain the association among a specified class of observable variables 36. This
may have led to discrepancies between the results of our study and those by Shiraishi et al. Another study that,
like our study, used observable variables, reported that maternal but not paternal overprotection was associ-
ated with adult attachment style in 2709 Japanese female university students 22. However, considering that the
study populations varied across previous studies, different life phases or events might have influenced
perceived parenting throughout the lifespan.
The present study also suggested that during pregnancy, the contribution of perceived paternal care was
slightly greater than that for perceived maternal care in forming an adult attachment to a partner. This finding
was consistent with a previous study that reported that individuals who were able to form a secure romantic
relationship had a father who was caring, loving, humorous, and affectionate37. Notably, those study
participants tended to describe their opposite-sex parent more favorably than the same-sex parent in terms of
caregiver quali- ties. Another supporting study suggested that perceived paternal care contributed to a positive
self-model of adult attachment in women23. Our data indicate that in contrast to paternal parenting practices,
perceived low maternal overprotection may contribute slightly more than paternal overprotection to the
formation of adulthood attachment to one’s partner. As mentioned, a previous study also reported that
individuals who could form secure romantic relationships had a mother who was respectful, accepting, not
intrusive, and not demanding37, which represented the quality of low overprotection. Moreover, participants in
the previous study tended to overjudge the negative traits of their same-sex parents 37. Thus, the relationships
with each parent since childhood may separately contribute to the formation of attachment style to a partner in
adulthood under specific circumstances. Further studies are needed to characterize the individual effects of
paternal and maternal parenting.
In Model 1, the fearful attachment style had a higher load value than the other attachment styles (Fig. 1).
This result was in line with previous studies in Japanese populations22,25. However, this finding may be a
distinctive fea- ture of the Japanese population. A prevalent preoccupied attachment style was found mainly in
Asian regions38, and this attachment style was also prevalent in the current study, although less prevalent than
the fearful style. In contrast, a previous study of 5692 men and women of varying ethnicities, including 72%
non-Latino white individuals who provided attachment information using a self-reported questionnaire, found
that the attachment styles could be ordered from highest to lowest score as follows: secure, anxious, and
avoidant attachment12. The anxious dimension corresponds to low self-esteem, emotional closeness, and
dependency seeking, whereas the avoidant dimension corresponds to a degree of discomfort with intimacy.
According to the four-category model of adult attachment, the fearful style has dimensions of both anxiety
and avoidance19.
The cross-cultural aspects of attachment styles have been discussed in previous studies39. Indeed,
attachment styles might be distributed differently across countries because of varying cultural and social
systems. However, few studies have explored this38,39. According to the attachment theory developed by
Ainsworth40, individuals develop an attachment type according to their relationships with their mothers as
infants. In the ‘strange situation’ test of attachment style, most American infants (70% with standard
distribution) were found to have a secure attachment style. In contrast, studies using a similar method found
that German infants had a relatively high percentage of avoidant-type attachment41, while Japanese infants had
a relatively high percentage of resistant- type attachment42. The resistant type is defined as dependent behavior
with rejection of the attachment figure when they engage in interaction40. The avoidant-type of attachment has
been examined extensively over the last few years, and is defined as the fearful style in the four-category
model19. According to the two-category model, the fearful style arises from a combination of a negative view
of the self and a negative view of others19. However, the present study found high positive loading values for
the self-model and other-model, representing the secure attachment style based on Model 2. The emphasis in
Asian culture on maintaining a relatively selfless state might explain this, where developing a sense of
connectedness to the family and others is encouraged from childhood43. In contrast, developing a sense of self
might be more emphasized in European culture43. In addi- tion, demands with a high potential for emotional
conflict in romantic relationships may differ across cultures39. Japanese individuals might expect fewer
emotional conflicts, such that conflict defines emotional closeness in their relationships with others. Further,
the expectation of rejection from others may lead Japanese individuals to avoid intimacy23. In this regard,
although the fearful attachment style in the present study had a higher loading value in Model 1, there was
some dissonance associated with this attachment style in Model 2. This discrepancy may be explained by the
results of a study on romantic attachment, conducted in 62 cultural regions 38. The study reported a strongly
significant negative correlation between the secure and fearful subscales of the RQ in Western regions, while
this correlation was non-significant in some African and most Asian regions, including Japan38. Therefore, the
combination of a negative self-model and a negative other-model might not underlie a fearful attachment style
in Asian or specifically Japanese populations. Further studies are needed to construct a normative form of the
secure attachment style in an Asian population.
Our study has several limitations that merit discussion. First, we did not collect data on the participants’
family-structure history or related adverse life events such as a history of abuse or attachment figure separa-
tions from childhood to adulthood. Moreover, we did not consider data about the parental figures involved in
parenting, such as mother-father pairs, single mothers, single fathers, and substitute parental figures.
Therefore, we could not evaluate the differences in perceived parenting among groups that were stratified
using these data. Although we performed our data analysis using data for which both the paternal and maternal
PBI were available, the generalizability of various family-structure histories may influence our findings.
Second, we did not examine data on the current marital status of the participants. In addition, we did not

Scientific (2023) https://doi.org/10.1038/s41598-023- 9

www.nature.com/scientificreports/
collect data on the partner’s attach- ment style in the current romantic relationship. Since we focused on
finding the best model for use in future perinatal psychiatry studies, we only targeted women. Accordingly,
we evaluated both perceived maternal and

Scientific (2023) https://doi.org/10.1038/s41598-023- 1

www.nature.com/scientificreports/

paternal parenting styles using PBI only in women participants. We did not evaluate the parenting style of
their male partners, as this was beyond the scope of our study. Furthermore, we did not examine data
regarding the relationship circumstances in the current family, nor that associated with current stressful life
events. Therefore, we could not evaluate the effects of these potential contributing or confounding variables
on adult attachment styles. Finally, we did not balance the order of the paternal and maternal PBI. Therefore,
we cannot speculate regarding the influence of the order of presentation of the paternal and maternal PBI
questionnaires.
In conclusion, the present study showed that attachment styles are best described using four-category and
two-category models. These models suggested a significant association between adult attachment style and
per- ceived paternal and maternal parenting before adolescence, where high care and low overprotection from
both paternal and maternal parents predicted a secure attachment style. Our findings support the use of
attachment models when providing perinatal care, particularly in preparing parents for caregiver roles starting
from early pregnancy. However, further studies on the longitudinal implications of parenting practices in
terms of the development of attachment styles are needed to clarify the consistency of these effects throughout
the lifespan. It is also essential to determine secure attachment styles and positive parenting practices that are
normative in Asian populations.

Data availability
All relevant data are provided in the paper. We are not able to make the underlying data available to readers,
because we do not have permission from the participating institutions to do so.

Received: 11 January 2023; Accepted: 13 September 2023

References
1. George, C. & Solomon, J. Representational models of relationships: Links between caregiving and attachment. Infant Ment. Health
J. 17, 198–216. https://doi.org/10.1002/(SICI)1097-0355(199623)17:3%3c198::AID-IMHJ2%3e3.0.CO;2-L (1996).
2. Walsh, J., Hepper, E. G. & Marshall, B. J. Investigating attachment, caregiving, and mental health: a model of maternal-fetal
rela- tionships. BMC Pregnancy Childbirth. 14, 383. https://doi.org/10.1186/s12884-014-0383-1 (2014).
3. Røhder, K. et al. Maternal-fetal bonding among pregnant women at psychosocial risk: the roles of adult attachment style,
prenatal parental reflective functioning, and depressive symptoms. PLoS One. 15, e0239208.
https://doi.org/10.1371/journal.pone.02392 08 (2020).
4. Priel, B. & Besser, A. Adult attachment styles, early relationships, antenatal attachment, and perceptions of infant temperament:
a study of first-time mothers. Pers. Relatsh. 7, 291–310. https://doi.org/10.1111/j.1475-6811.2000.tb00018.x (2000).
5. Warfa, N., Harper, M., Nicolais, G. & Bhui, K. Adult attachment style as a risk factor for maternal postnatal depression: A
systematic review. BMC Psychol. 2, 1–11. https://doi.org/10.1186/s40359-014-0056-x (2014).
6. Bianciardi, E. et al. The anxious aspects of insecure attachment styles are associated with depression either in pregnancy or in
the postpartum period. Ann. Gen. Psychiatry. 19, 1–9. https://doi.org/10.1186/s12991-020-00301-7 (2020).
7. Lutz, W. J. & Hock, E. Maternal separation anxiety: Relations to adult attachment representations in mothers of infants. J.
Genet. Psychol. 156, 57–72. https://doi.org/10.1080/00221325.1995.9914806 (1995).
8. Ferraro, I. K. & Taylor, A. M. Adult attachment styles and emotional regulation: The role of interoceptive awareness and
alexithymia.
Pers. Individ. Differ. 173, 1–7. https://doi.org/10.1016/j.paid.2021.110641 (2021).
9. Maunder, R. G., Lancee, W. J., Nolan, R. P., Hunter, J. J. & Tannenbaum, D. W. The relationship of attachment insecurity to
subjec- tive stress and autonomic function during standardized acute stress in healthy adults. J. Psychosom. Res. 60, 283–290.
https://doi. org/10.1016/j.jpsychores.2005.08.013 (2006).
10. Nordahl, D. et al. Adult attachment style and maternal-infant bonding: the indirect path of parenting stress. BMC Psychol. 8, 1–
11. https://doi.org/10.1186/s40359-020-00424-2 (2020).
11. Rholes, W. S., Simpson, J. A. & Friedman, M. Avoidant attachment and the experience of parenting. Pers. Soc. Psychol. Bull. 32,
275–285. https://doi.org/10.1177/0146167205280910 (2006).
12. Palitsky, D., Mota, N., Afifi, T. O., Downs, A. C. & Sareen, J. The association between adult attachment style, mental disorders, and
suicidality: Findings from a population-based study. J. Nerv. Ment. Dis. 201, 579–586.
https://doi.org/10.1097/NMD.0b013e3182 9829ab (2013).
13. Mills-Koonce, W. R. et al. Adult attachment style and stress as risk factors for early maternal sensitivity and negativity. Infant
Ment. Health J. 32, 277–285. https://doi.org/10.1002/imhj.20296 (2011).
14. Shiraishi, N., Murakami, M., Goto, Y., & Kitamura, T. Intergenerational transmission of child-rearing styles mediated by adult
attachment: A structural equation model approach. Psychiatry Clin. Neurosci. Rep. 1, 22; https://doi.org/10.1002/pcn5.22
(2022).
15. Zaki, N., García, A. M. R. & Puga, J. L. Attachment style and prenatal expectations from a bayesian perspective. Psicothema
32, 138–144. https://doi.org/10.7334/psicothema2019.222 (2022).
16. Bowlby J. Attachment and loss. In: Vol. I: Attachment. (New York, 1969).
17. Bowlby J. Attachment and loss. In: Vol. II: Separation. (New York, 1973).
18. Griffin, D. W. & Bartholomew, K. Models of the self and other: Fundamental dimensions underlying measures of adult
attachment.
J. Pers. Soc. Psychol. 67, 430–445. https://doi.org/10.1037/0022-3514.67.3.430 (1994).
19. Bartholomew, K. & Horowitz, L. M. Attachment styles among young adults: A test of a four-category model. J. Pers. Soc. Psychol.
61, 226–244. https://doi.org/10.1037/0022-3514.61.2.226 (1991).
20. Fraley, R. C., Roisman, G. I., Booth-LaForce, C., Owen, M. T. & Holland, A. S. Interpersonal and genetic origins of adult
attachment styles: A longitudinal study from infancy to early adulthood. J. Pers. Soc. Psychol. 104, 817–838.
https://doi.org/10.1037/a0031435 (2013).
21. Parker, G., Tupling, H. & Brown, L. B. A parental bonding instrument. Br. J. Med. Psychol. 52, 1–10. https://doi.org/10.1111/j.
2044-8341.1979.tb02487.x (1979).
22. Liu, Q., Shono, M. & Kitamura, T. The effects of perceived parenting and family functioning on adult attachment: a sample of
Japanese university students. Open Fam. Stud. J. 1, 1–6. https://doi.org/10.2174/1874922400801010001 (2008).
23. Wilhelm, K., Gillis, I. & Parker, G. Parental bonding and adult attachment style: the relationship between four category models.
Int. J. Womens Health Wellness. 2, 1–7 https://doi.org/10.23937/2474-1353/1510016 (2016).
24. O’Hara, M. W. & Wisner, K. L. Perinatal mental illness: Definition, description and aetiology. Best Pract. Res. Clin. Obstet. Gynaecol.
28, 3–12. https://doi.org/10.1016/j.bpobgyn.2013.09.002 (2014).

Scientific (2023) https://doi.org/10.1038/s41598-023- 1

www.nature.com/scientificreports/
25. Matsuoka, N. et al. Adolescents’ attachment style and early experiences: A gender difference. Arch. Womens Ment. Health. 9, 23–29.
https://doi.org/10.1007/s00737-005-0105-9 (2006).

Scientific (2023) https://doi.org/10.1038/s41598-023- 1

www.nature.com/scientificreports/

26. Fukui, N. et al. Perceived parenting before adolescence and parity have direct and indirect effects via depression and anxiety on
maternal–infant bonding in the perinatal period. Psychiatry Clin. Neurosci. 75, 312–317. https://doi.org/10.1111/pcn.13289
(2021).
27. Fukui, N. et al. Exclusive breastfeeding is not associated with maternal-infant bonding in early postpartum, considering
depression, anxiety, and parity. Nutrients 13, 1184. https://doi.org/10.3390/nu13041184 (2021).
28. Ogawa, M. et al. Factor structure and measurement invariance of the hospital anxiety and depression scale across the
peripartum period among pregnant Japanese women. Neuropsychiatr. Dis. Treat. 17, 221–227.
https://doi.org/10.2147/NDT.S294918 (2021).
29. Motegi, T. et al. Depression, anxiety and primiparity are negatively associated with mother-infant bonding in Japanese mothers.
Neuropsychiatr. Dis. Treat. 16, 3117–3122. https://doi.org/10.2147/NDT.S287036 (2020).
30. Motegi, T. et al. Identifying the factor structure of the mother-to-infant bonding scale for post-partum women and examining its
consistency during pregnancy. Psychiatry Clin. Neurosci. 73, 661–662. https://doi.org/10.1111/pcn.12920 (2019).
31. Hashijiri, K. et al. Identification of bonding difficulties in the peripartum period using the mother-to-infant bonding scale-
Japanese version and its tentative cutoff points. Neuropsychiatr. Dis. Treat. 17, 3407–3413. https://doi.org/10.2147/NDT.S336819
(2021).
32. Kitamura, T. & Suzuki, T. A validation study of the parental bonding instrument in a Japanese population. Jpn. J. Psychiatry.
Neurol.
47, 29–36. https://doi.org/10.1111/j.1440-1819.1993.tb02026.x (1993).
33. Schermelleh-Engel, K., Moosbrugger, H. & Müller, H. Evaluating the fit of structural equation models: Tests of significance
and descriptive goodness-of-fit measures. Methods Psychol. Res. 8, 23–74 (2003).
34. Kano, Y. Does structural equation modeling outperform traditional factor analysis, analysis of variance and path analysis?.
Behav- iormetrics. 29, 138–159. https://doi.org/10.2333/jbhmk.29.138 (2002).
35. Kline, B. R. Principles and Practice of Structural Equation Modeling, 4th Edition (ed. Little, T. D.). 12–13 (New York, 2016).
36. Nathalie, L. & Kathleen, L. S. What makes a hypothetical construct “hypothetical”? Tracing the origins and uses of the
‘hypothetical construct’ concept in psychological science. New Ideas Psychol. 31, 22–31.
https://doi.org/10.1016/j.newideapsych.2011.02.005 (2013).
37. Hazan, C. & Shaver, P. Romantic love conceptualized as an attachment process. J. Pers. Soc. Psychol. 52, 511–524. https://doi.org/
10.1037//0022-3514.52.3.511 (1987).
38. Schmitt, D. P. et al. Patterns and universals of adult romantic attachment across 62 cultural regions: are models of self and of
other pancultural constructs?. J. Cross. Cult. Psychol. 35, 367–402. https://doi.org/10.1177/0022022104266105 (2004).
39. Grossmann, K. E. & Grossmann, K. The wider concept of attachment in cross-cultural research. Hum. Dev. 33, 31–47.
https://doi. org/10.1159/000276501 (1990).
40. Ainsworth, M. D. & Bell, S. M. Attachment, exploration, and separation: illustrated by the behavior of one-year-olds in a
strange situation. Child Dev. 41, 49–67. https://doi.org/10.2307/1127388 (1970).
41. Grossmann, K. E. & Grossmann, K. Preliminary observations on Japanese infants’ behavior in Ainsworth’s strange situation.
Res. Clin. Center Child Dev. Ann. Rep. 11, 1–12 (1989).
42. Takahashi, K. Examining the strange-situation procedure with Japanese mothers and 12-month-old infants. Dev. Psychol. 22,
265–270. https://doi.org/10.1037/0012-1649.22.2.265 (1986).
43. Wang, Q. & Leichtman, M. D. Same beginnings, different stories: a comparison of American and Chinese children’s narratives.
Child Dev. 71, 1329–1346. https://doi.org/10.1111/1467-8624.00231 (2000).

Acknowledgements
We are grateful to Dr. Takuro Sugai, Dr. Ryusuke Tsuboya, Dr. Koichi Takakuwa, Dr. Masayuki Yamaguchi,
Dr. Kazufumi Haino, Ms. Rie Araki, and Ms. Setsuko Mitome for their collaboration during the early stages
of this work. We greatly appreciate all participants and staff at the following participating obstetric
institutions: Kameda Daiichi Hospital, Niitsu Obstetrics and Gynecology Clinic, Tomita Obstetrics and
Gynecology Clinic, Honda Ladies Clinic, Agano City Hospital, Chihara Clinic, Sekizuka Clinic, Hirohashi
Obstetrics and Gyne- cology Clinic, Saiseikai Sanjo Hospital, Saiseikai Niigata Hospital, Itoigawa Sogo
Hospital, Niigata Tokamachi Hospital, Takeyama Hospital, Watanabe Kinen Clinic, Kashiwazaki General
Hospital and Medical Center, Kido Hospital, Tachikawa Medical Center, Saintpoulia Women’s Clinic,
Nagaoka Chuo General Hospital, Arakawa Ladies Clinic, Ueda Ladies Clinic, Uonuma Kikan Hospital,
Murakami General Hospital, Sado General Hospital, Tokunaga Ladies Clinic, Nagaoka Red Cross Hospital,
Niigata Shibata Hospital, Ladies Clinic Ishiguro, Angel Mother Clinic, Joetsu General Hospital, Niigata City
General Hospital, Ojiya General Hospital, Watanabe Clinic, and Niigata University Medical and Dental
Hospital. This work was supported by JSPS KAKENHI (grant num- ber JP19K08040 to NF), a grant from the
Niigata Medical Association (to TM), and a grant from the Medical Association of Niigata City (to MO). We
thank Sydney Koke, MFA, from Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript.

Author contributions
E.Z., N.F., Y.W., J.E., and T.S. conceived the original study concept and developed the study design. N.F. and T.M.
collected the data. N.F., K.H., T.M., and M.O. entered the data into a database. N.F. performed the statistical
analyses. E.Z., N.F., and Y.W. wrote the first draft of the manuscript. All authors contributed to the revision
of the manuscript and read and approved the final manuscript.

Funding
This work was supported by JSPS KAKENHI (grant number JP19K08040 to NF), a grant from the Niigata Medical
Association (to TM), and a grant from the Medical Association of Niigata City (to MO).

Competing interests
The authors declare no competing interests.

Additional information
Supplementary Information The online version contains supplementary material available at https://doi.org/
10.1038/s41598-023-42674-1.
Correspondence and requests for materials should be addressed to T.S.
Reprints and permissions information is available at www.nature.com/reprints.
Scientific (2023) https://doi.org/10.1038/s41598-023- 1
www.nature.com/scientificreports/

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons licence, and indicate if changes were made. The images or other third party material in
this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is
not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly
from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

© The Author(s) 2023

Scientific (2023) https://doi.org/10.1038/s41598-023- 1