Background

From the fetal stage, the brain develops its structure and functions through interaction with the environment and sensory-motor experiences [1]. If "movement" is understood to be central to development [2], any motor deficits should affect the brain's perceptual, cognitive, and motivational functions [3, 4]. Thus the "motor" perspective is essential to exploring developmental stumbling blocks.

Studies dating back to the 1920s have described motor development as a series of universal milestones in infancy, including sitting, crawling, and walking [5, 6]. In the early 1980s, however, Esther Thelen demonstrated through treadmill stepping and other examples that the nervous system is only one component contributing to infant motor patterns [7, 8]. Within the framework of dynamic systems theory, she described how motor skills develop and are regulated. Dynamic systems theory focuses on children's own motivation in motor development and may help explain individual differences in the emergence of motor skills. In addition, Thelen also theorized about effective and appropriate early developmental interventions for each individual [4]. While research on individual differences in motor skill learning dates back more than 100 years, our understanding of how individual differences in aptitude affect motor development and its acquisition and control remains relatively poor [2, 9]. In recent years, though the relationship between individual differences in motor and executive functions has been actively investigated, little is known about children younger than 5 years of age [10,11, see 12 for a review].

The Executive Function (EF) consists of the capacity for goal-oriented regulation of oneʼs own thoughts, actions, and emotions and plays a significant role in early learning and transition to formal schooling. EF skills depend on neural networks, including the prefrontal cortex (PFC) and anterior cingulate cortex (ACC) [13, 14]. These networks are extensively interconnected with structures in the limbic system and brainstem [15], and these interrelationships between the PFC and ACC and other areas of the brain help establish EF as a control system that both influences and is influenced by cognitive, emotional, and behavioral functioning.

As neural substrates for EF have been proposed in early development [16, 17], these individual differences have been investigated and labeled Effortful Control within a temperamental framework [18], defined as the efficiency of executive attention. The idea that attention may be the first step toward demonstrating EF skills has been advocated [19, 20]. Referring to Baddeley’s model of working memory [21], the executive attention network in major attention theory is described as having an important association with EF [22]. The anterior cingulate cortex, the lateral ventral PFC and the basal ganglia have been found to involve the Executive attention network [23], which substantially overlaps with regions related to working memory [24,25,26].

Temperament is defined as a set of relatively consistent, biologically-based individual differences in reactivity and self-regulation [27]. Reactivity refers to an individual’s initial physiological and behavioral reactions to sensory stimuli of different qualities and intensities. These reactive tendencies, namely Surgency/Extraversion (i.e., positive emotionality) and Negative Affect are largely present at birth but become increasingly stable during childhood. In addition to reactivity, a behavioral system that enables voluntary control of attention and emotion is labeled Effortful Control, defined as the ability to inhibit a dominant response in order to perform a subdominant one. Though infants are highly limited in the temperamental factor of voluntary behavioral control (e.g., the ability to effortfully inhibit behavior upon command), this skill improves considerably in the third year of life, and such improvements continue into childhood [28, 29].

Motor control has been infrequently examined in temperament research from a developmental perspective e.g., [30,31,32]. However, the research trends described above have led to increased interest in identifying antecedents of EF in infant skills development. As the development of the regulatory function has been studied from early infancy within the framework of temperament, there is great merit in examining its relationship with motor development. That is, the orienting attention network exerts a great deal of control over other brain networks during infancy and early childhood in terms of temperamental regulation [33]. In addition, regarding reactivity, activity or perceptual sensitivity is known to affect the rate of motor development [34]. Previous studies have documented motor activity as a subcategory of behavior in order to describe temperament rather than as a primary outcome [35]. However, since studies show that change in one developmental domain influences changes in other domains [1, 36,37,38], we should reconsider the link between temperamental and motor development longitudinally.

A recent key issue in motor development is Developmental Coordination Disorder (DCD), which impairs the development of coordinated muscle functions associated with various motor activities such as dexterity, posture retention, and force coordination, and is reported to have a prevalence of 5 to 8% among school-age children [39]. Such children are at risk of lower self-esteem due to their poor ability to perform tasks [40, 41], which may lead to various maladjustments in school life and in friendships [42,43,44] and may cause them to become withdrawn [45, 46]. Additionally, DCD may convey early signs of developmental disorders, including autistic spectrum disorder (ASD), attention deficit hyperactivity disorder (ADHD), even high rates of comorbidity [47, 48]. Despite growing recognition of the clinical importance of DCD, the pathogenesis of DCD is still unknown, nor is anything known about its neurological basis, exacerbating factors, or protective factors [39]. Accordingly, early intervention in the motor development of children with DCD is desirable, and the importance of motivating target children is widely recognized [for a review, see 49].

Motor development is usually divided into “gross” and “fine” motor development. Gross motor development involves major movements of the whole body, primarily the trunk and legs, and culminates in independent walking, climbing, and running during toddlerhood. In contrast, fine motor development focuses on the use of the shoulders, arms, and hands, and is refined into small hand and arm movements such as grasping in late infancy and throwing in toddlers and preschoolers. It primarily relies on attentional rather than motor skill, and Effortful Control abilities may make a contribution to achievement [50]. Early motor delays in crawling or walking are often danger signals of the presence of other disruptions and may affect the development of skills in other domains. Yet only one longitudinal study has directly examined motivation in relation to motor development [30]. In that study, the developmental trajectory of motor motivation in infancy was investigated longitudinally by observing infants every 3 weeks from 7 to 12 months to describe the nature of motivation to move during locomotor transitions [35]. Results showed that motivation to move continued to increase as motor milestones were reached, and infants perceived to have stronger motivation to move showed earlier achievement of four milestones: independent sitting, pulling to stand, hands-and-knees crawling, and cruising.

With motivation defined as “the energization (instigation, activation) and direction (focus, aim) of behavior” [51], temperament-related dimensions of Surgency/Extraversion and Negative Affect are directly related to motivation [52]. That is, Positive Affect is linked to motivation in young children, including selection, engagement, and sustained interest in specific activities. In contrast, Negative Affect such as fearful inhibition may lead young children to avoid or withdraw from exciting or potentially punishing situations while providing them with the time necessary to solve a problem, meet a challenge, and plan the next step [53]. In both cases, Effortful Control constitutes a foundation for competent action as well as the ability to act—or withhold action—now in the interest of future achievements. While this may not constitute motivation, it offers a flexible means of achieving a motivationally appropriate end.

In response, as part of a national birth cohort under the Japan Environment and Children’s Study (JECS-A) and using items that are relatively easy for caregivers to respond to, we created a latent variable for the motor function at each age and examine its relationship with the three factors making up temperament: Surgency, Negative Affect, and Effortful Control. Since health check-ups for 3-year-old children are widespread in Japan and the results of this study could be used as part of such check-ups, we targeted children up to 3 years of age. Our hypotheses for early childhood were: 1) Temperament and motor functions development mutually influence each other at subsequent time points; and 2) From the beginning, the gross motor system may be strongly influenced by Surgency through motivation while the fine motor system may be tightly bound to the development of Effortful Control through the building of executive attention.

Method

Recruitment and procedure

The present study is an Adjunct Study of the Japan Environment and Children’s Study (JECS-A). The study protocol was approved by the Ethics Committee of the Nagoya City University Graduate School of Medical Sciences (Approval No. 60–00-0574). Participants were registered with the Aichi regional sub-cohort of JECS-A. A community-based recruitment strategy was adopted at 32 obstetric hospitals and clinics providing care for pregnant women in Ichinomiya and Nagoya. Women in the early stages of pregnancy who visited an obstetrics facility were invited to participate in JECS if they met the following criteria: (1) residence within the study area; (2) estimated delivery date after August 2011; and (3) the ability to read and write Japanese to complete the self-administered questionnaire. Of the 8,134 pregnant women deemed eligible during the recruitment period (January 2011–March 2014), 5,721 (70.3%) participants were enrolled as the baseline cohort of pregnant women [54].

Figure 1 depicts a flowchart of the inclusion process followed in this study. Of 5,721 caregivers initially contacted in 2011–2014, 3,426 agreed to participate. The adjunct questionnaire targeted 2,642 infants, for whom parents provided written informed consent. While three participants later withdrew consent, 2,639 received a questionnaire at 6 months, of which 2,448 were returned. 141 children with congenital diseases or disorders related to motor development (i.e., Apgar score < 7 1 min or 5 min after birth) or having been delivered preterm were excluded. Subsequently, 415 dropped out of the 24-month survey, and another 235 dropped out of the 42-month survey. Ultimately, responses from 1,657 caregivers, including three mothers of twins, were analyzed. Mothers in this dataset answered questionnaires specially constructed for this Adjunct Study from the onset of pregnancy until their child was 42 months old (full data). Participants’ characteristics were derived from the main JECS-A dataset. Participants were rewarded with a stored-value card for each participation. Starting in August 2013, self-administered questionnaires were sent out and completed by mothers. Responses were mailed and collected until April 2019.

Fig. 1
figure 1

Participant flow diagram

Full size image

Motor functions

Questionnaires about sensory motor functions were administered at 6, 24, and 42 months. Question items, including fine and gross motor activities at each age, were selected through discussions with pediatric neurologists specializing in developmental disorders while referring to previous research [55, 56] (see Supplementary Material 1).

At 6 months, regarding the frequency of their infant’s rolling over, caregivers were asked to choose from “never” to “frequently.” Caregivers also chose one from five pictures displaying the manner of rolling over. Optical righting reflexes were also assessed, with three choices offered along with illustrations. Caregivers were asked what happened to the baby’s head when the baby was held upright and tilted right or left. Reaching (Does your baby try to reach objects in front of him or her?), Posture when held (Do you sense the posture of your baby as tension in the body when he or she is being held?), and Suckling (Does your baby suckle well?) were assessed, with five choices per question; reaching frequency (1; Frequently — 5; Never); degree of tenseness (1; Very tense — 5; Unstable); and Suckling proficiency (1; Very well — 5; Not well at all).

At 24 months, questions probed the onset of independent stable sitting and walking. Skills in using cutlery and picking were assessed by caregivers being presented with a list of manners of doing the above. Caregivers were asked to circle the most applicable description of their child’s behavior. Regarding picking, the list was accompanied by pictures (e.g., pinching thin objects such as coins). Regarding oral dexterity, caregivers were shown six descriptions of a child eating (e.g., “My child has a strong tendency to swallow food without chewing”) and asked to circle all applicable options. As three out of six descriptions were related to a motor function problem, the number of circles was counted.

At 42 months, two-leg hopping and one-leg standing were assessed by caregivers being given a list of seven manners of doing the above and asked to circle the most applicable description of their child’s manner. Concerning fine motor movements, caregivers were given an illustration of a V-sign and asked to circle the most applicable from four options to describe their child’s behavior (e.g., She can make the V-sign for a brief moment, but her fingers untangle quickly). The skill of using cutlery was examined by asking caregivers whether their child used a spoon to put food in his or her mouth without spilling and shown a list of manners of doing so. If they circled the “My child can use a spoon” option, they were asked to choose one from four pictures displaying the manner of doing so.

The Japanese version of the Little Developmental Coordination Disorder Questionnaire (LDCDQ) [57] was administered at 42 months. This questionnaire was developed to screen for motor coordination difficulties in 3- and 4-year olds. The 15 items are divided into 3 subcategories: Control during movement, Fine motor, and General coordination, with five items each. The first category (Control during movement) contains items relating to motor control while either the child or an object is in motion such as striking a moving ball. For each item, caregivers are asked to compare their child’s performance with that of children of the same age and gender and to rate it on a five-point scale (1 = not at all relevant to my child — 5 = extremely relevant to my child), with Option 6 = never experienced added in our study. Three subcategory scores were calculated by adding ratings for all relevant items, with participants choosing Option 6 for any relevant item excluded. Higher scores indicate higher motor proficiency.

Temperament

In the present study of 6-, 24-, and 42-month-olds, temperament was assessed through the modified Japanese version of the parent-reported Infant Behavior Questionnaire-Revised (IBQ-R) [58], the Japanese short version of the Early Childhood Behavior Questionnaire (ECBQ) [59], and the Japanese short version of the Childhood Behavior Questionnaire (CBQ) [60]. In these questionnaires, to minimize parental biases associated with poor recall, aggregating information across situations prior to answering, or making comparative judgments, caregivers are given specified toddler or child reactions in concrete situations, such as “When being dressed or undressed during the last week, how often did the baby squirm or try to roll away?.” Then they were asked to rate the frequency of their child engaging in the given behaviors in on a seven-point scale (1 = never — 7 = always) over the past week for the IBQ-R and the past two weeks for part of the IBQ-R (two weeks for less frequent events during the first year) and the ECBQ. In the CBQ, caregivers were given statements such as “My child always seems in a hurry to get from one place to another” and asked to rate their child on a seven-point scale (1 = totally untrue of your child — 7 totally true of your child). Caregivers were also offered a “Not applicable” response option when the child had not been observed in the situation described.

Factor analyses of subscale scores on these questionnaires yielded three broad factors of temperament. As distinct from the two broad factors of Positive Emotionality/Surgency and Negative Affect, a third factor, Orienting/Effortful Control, were extracted for the Japanese short forms of the ECBQ and CBQ [59, 60]. Regarding the temperament questionnaire at 6 months, a modified shortened version of the IBQ-R was needed to match the number of questions allocated in consideration of the burden on JECS research collaborators. Referring to Putnam and Rothbart [61], we first excluded items if more than 20% of caregivers chose the Not Applicable option for the item. Next, for each subscale, we calculated item-total correlations for the dataset [58], which included 284 infants (129 girls) with an average age of 6.79 months (SD = 2.60, range = 3 to 12). The four items with the highest item-total correlations were then used to form tentative scales. Given our aim of a minimum alpha of 0.65 for data from each subscale, the scores from the tentative scales were found to have α < 0.65 for Cuddliness, Vocal reactivity, Low intensity pleasure, Smiling and laughter, Approach, Distress limitation, Fear, and Sadness. By adding a single item to each of these four tentative scales, we could raise internal consistency to acceptable levels for Vocal reactivity, Low intensity pleasure, Approach, and Distress limitation. Increasing two scales to six items improved internal consistency for Smiling and laughter and Fear. Based on the item-level principal axis factoring on each subscale of the standard form, items from the tentative scale were replaced by items not included in the tentative scale to ensure that all aspects of the multidimensional scale were represented. This procedure is what made us decide to retain the four-item scale, with alpha values of 0.60 and 0.61 for "Cuddliness" and "Sadness," respectively. In IBQ-R, shortening the scales had the most adverse effect on Cuddliness and Sadness, with a reliability under 0.70 [62]. Sadness was also reported to have poor consistency in the CBQ short form [61]. Due to the limitation in the number of questions in the 6-month-old JECS-A survey form, the tentative shortened version of the IBQ-R was used in the present study.

Statistical analysis

Based on a previous study [63], we hypothesized a model in which we drew assumed paths between temperament and motor function variables. Statistical analyses were conducted using IBM SPSS Statistics 25.0 (SPSS Inc.). The compatibility of data with normal distribution was tested using the Kolmogorov–Smirnov test. To confirm normal distribution, Pearson's correlation coefficient or Spearman's correlation were used to determine associations between variables. Structural equation modeling was conducted using AMOS 25.0 (SPSS Inc.) to evaluate the possible relationship between motor development and child temperament. In our model, motor function and temperament (Surgency, Negative Affect, Effortful Control) at 6, 24, and 42 months interact at subsequent time points. Moreover, these sequential relationships influence LDCDQ scores for Control during movement, Fine motor activity, and General coordination at 42 months. To develop the model, we added correlations between “error” for the three observed temperament variables each at 24 and 42 months of age and for the three observed LDCDQ variables at 42 months. Furthermore, at 24 and 42 months, correlations between “error” of the latent motor function and “error” of the three observed temperament variables were also added, respectively. Model validity was evaluated using the goodness of fit index (GFI) and adjusted goodness of fit index (AGFI) as well as the following indices to assess model fit: the chi-square statistic (χ 2), the comparative fit index (CFI), and root mean square error of approximation (RMSEA). Missing data were handled by applying a Full Information Maximum Likelihood (FIML) estimation drawing on all available data to estimate model parameters without imputing missing values. The Akaike information criterion (AIC) was also used.

Results

Tables 1 and 2 summarize the characteristics of mothers and infants, with an attrition analysis also presented. Responders (n = 1,657) and non-responders (n = 650) were compared based on the following data from JECA registration: Child’s sex, Mother's highest level of education, Partner's highest level of education, Maternal age, Annual household income, Self-reported maternal smoking habit in the second or third trimester of pregnancy (non-smoker, ex-smoker, current smoker), and Existence of child siblings at around one month postpartum. Chi-square tests were conducted and showed that responders’ partners were more highly educated, and responders smoked less often during pregnancy. However, effect size was small (Cramer’s V ≤ 0.10).

Table 1 Comparison of general characteristics of responders and non-responders
Full size table
Table 2 Comparison of reported infant characteristics of responders and non-responders
Full size table

Table 2 shows reported infant characteristics and information on the temperamental and motor function variables at 6 months. Non-responders scored significantly higher than responders for Surgency, Negative Affect, and score of Rolling over though effect size (Cohen's d) was small (0.20 for Surgency, 0.07 for Negative Affect and 0.06 for Rolling over). Though not large, the difference in infants between responders and non-responders was significant. Table 3 presents the characteristics variables at 24 and 42 months. After testing for normal distribution, Pearson's correlation coefficient or Spearman's correlation were applied to determine associations between variables (See Table 4).

Table 3 Characteristics of variables at 24 months
Full size table
Table 4 Results of correlation analysis of variables used in Structural Equation Modeling
Full size table

Based on a previous study [63], we hypothesized a model in which we drew assumed paths between temperament and motor function variables. Figure 2 shows the final model, in which paths found to be statistically significant (p < 0.05) are presented. Fit statistics indicate that the model fits the data well: χ 2 (264) = 1013.463, p < 0.001, CFI = 0.901; GFI = 0.953; AGFI = 0.937; RMSEA = 0.041. We defined three LDCDQ scores as outcomes at 42 months. Manifest variables corresponding to the latent motor function variable loaded significantly (p < 0.01) for all three ages. For Stable sitting, Walking, and Oral dexterity at 24 months, lower values mean better performance, and these showed negative associations. At 42 months, the motor function showed positive associations in three categories: Control during movement (β = 0.47, p < 0.001), Fine motor movement (β = 0.35, p < 0.001), and General coordination (β = 0.28, p < 0.001).

Fig. 2
figure 2

Path-diagram with standardized coefficients for the structural equation model showing associations between temperamental factors and latent motor functions at 6–42 months. The error term is omitted from the figure

Full size image

With regard to temperament, all paths in line with development for Surgency, Negative Affect, and Orienting/Effortful Control were significant. Orienting/Regulation at 6 months was positively associated with Surgency at 24 months (β = 0.07, p < 0.01), while Negative Affect at 6 months was negatively associated with Effortful Control at 24 months (β = -0.07, p < 0.01). Subsequently, Surgency and Negative Affect at 24 months showed positive paths toward Effortful Control at 42 months (respectively, β = 0.10, p < 0.001; β = 0.10, p < 0.001). On the other hand, Effortful Control at 24 months was negatively associated with Surgency and Negative Affect at 42 months (respectively, β = -0.24, p < 0.001; β = -0.12, p < 0.001). Concerning the relationship between temperament and the motor function, at 6 months, Surgency and Orienting/Regulation were positively correlated with the motor function (respectively, r = 0.57, p < 0.001; r = 0.40, p < 0.001). Moreover, the motor function at 6 and 24 months was positively associated with Effortful control at 24 months (β = 0.13, p < 0.01) and 42 months (β = 0.11, p < 0.01), respectively. That is, good motor function leads to good Effectful control from infancy onward. Moreover, the motor function at 24 months was also positively associated with Surgency at 42 months (β = 0.28, p < 0.001). On the other hand, there was a negative path from Orienting/Regulation at 6 months to the motor function at 24 months (β = -0.09, p < 0.05) and a positive path from Effortful Control at 24 months to the motor function at 42 months (β = 0.23, p < 0.001). Then, as the outcome at 42 months, Surgency was positively associated with two subcategory scores for motor coordination (Control during movement and General coordination) (respectively, β = 0.18, p < 0.001; β = 0.06, p < 0.001) except Fine motor movement, while Effortful control was positively associated with all three categories (Control during movement, Fine motor movement, and General coordination) (respectively, β = 0.14, p < 0.001; β = 0.30, p < 0.001; β = 0.37, p < 0.001). Conversely, Negative Affect had a negative impact on Fine motor movement (β = -0.08, p < 0.001) and General coordination (β = -0.08, p < 0.001).

The minimum sample size required for our final model was 71 based on a preliminary calculation (Alpha = 0.05, degrees of freedom in the SEM model = 267, desired power = 0.80, null root mean square error of approximation [RMSEA] = 0.05, alternative RMSEA = 0.08). We therefore judged that our sample size was satisfactory.

Discussion

The relationship between temperament and motor control was analyzed longitudinally from 6 to 42 months as part of JECS-A. As our respondents did not differ significantly in demographic data from the non-respondents who dropped out, we believe that our results can be generalized. That is, in infancy, Surgency showed a strong relationship with the motor function, while in toddlers, Effortful Control emerges to control both the motor function and temperamental reactivity in general.

High surgency in infancy can promote motor function through high activity or curiosity. Infants are fond of high intensity stimuli and highly responsive to them, eliciting encouragement from their surroundings. Later on, high motor function has positive effects on Surgency at 42 m. Surgency was also reported to be the most significant predictor of physical activity among children [64]. At 42 m, this facilitated two LDCDQ subscores, except for Fine motor movement. High surgency can be associated with roughness and clumsiness, not with improving fine motor movement.

On the other hand, Negative Affect has a negative impact on Fine motor movement and General coordination at 42 months. As these movements require motivation as well as persistence, if negative emotions are strong, children may be led to rejection or to giving up too quickly and thus not improving. Moreover, missed opportunities due to anxiety, easy anger or discomfort when things go wrong, or difficulty in being calmed may interfere with improving motor coordination. On the other hand, Control during movement is not affected by negative emotions because the questions in the subscale that ask whether or not children have difficulty with basic gross motor skills such as "running," "kicking," "throwing," or "flying." In addition, the finding that Negative Affect at 6 months has a negative impact on Effortful Control at 24 months might indicate that strong Negative affect interferes with interactions with caregivers, which are thought to promote the development of orienting attention in infancy [33].

In contrast to reactive aspects, the temperamental regulatory function may reflect a longitudinal mediation model. That is, motor development is mediated by temperamental self-regulation, and vice versa. Although the regulatory function at 6 months negatively affects the motor function at 24 months, as shown in Fig. 2, we found positive correlations between these variables (Table 4). Thus this sign reversal could be caused by strong correlation between the motor function at 6 months and 24 months [65]. In addition, Reactivity, Surgency, and Negative Affect at 24 months positively affect Effortful Control at 42 months. This may be consistent with the finding that infants with higher levels of Orienting/Regulation also show evidence of stronger levels of emotionality in a task undertaken while in distress along with evidence of efforts to self-regulate emotional reactivity [66]. On the other hand, from 24 to 42 months, Effortful Control works in the direction of suppressing reactivity, i.e., Surgency and Negative Affect. The widely-used behavior tasks assessing Effortful Control require suppressing dominant emotional or motor responses and performing subdominant behaviors [29] such as slowing down fine skills (e.g., drawing) or gross skills (e.g., walking along a 6-foot line slowly). In general, observed paths in Effortful Control for emotional reactivity and movement from 24 to 42 months were consistent with previous studies [18, 28, 33, 66].

As the LDCDQ includes items requiring attention and self-control, our results showing a positive relationship between Effortful Control and LDCDQ scores are in line with expectations. The combination of motor impairment and executive function deficits is common in children with DCD, especially those with persistent DCD over a 2-year period from Times 1 to Time 2, the 2-year follow-up [67]. As higher-order control functions develop separately from the motor function and may work as a compensatory strategy for reduced efficiency [68], careful assessments for DCD and the executive function could prove useful.

Our study has several limitations. First, we used original question items to investigate the development of the motor function in which movement is not directly observed by an evaluator with specialized knowledge. To help caregivers without such knowledge evaluate this function, we asked questions, whenever possible using illustrations along with explanations. Although multiple-choice questions are easy to answer, some actions observed by caregivers may not be available as options, or caregivers may be unable to choose between options. Further refinement of each item is needed in the future in order to make this questionnaire more effective in identifying the need for intervention at a young age. Second, temperament was also assessed through a parent-reported questionnaire. In particular, the questionnaire for 6-month-olds was tentative because the number of items was limited in light of the overall volume of the questionnaire and of the burden it placed on participants. Though the temperament questionnaire at 6 months did not show a high reliability coefficient, we accepted it because we did not wish to unnecessarily narrow the conceptual range by making it a short scale in the present study. While the results observed for temperament at 6 months were generally consistent with previous studies, we should be aware of the limitations of the tentative version of the 6-month temperament scale. As Putnam et al. argue [61], the standard version should be used for temperament studies whenever possible. Third, approximately 40% of children born in the study area have taken part in the study to date, reflecting a representative population in the regional subcohort of the JECS-A study [54]. However, in our study, a person was classified as “lost to follow-up” if any one of the three follow-up questionnaires was not returned within one year after the last one was sent out. Our attrition analysis showed that there was little difference between the maintenance and dropout groups. However, it is possible that the infants in the dropout group were slightly reactive.

As noted above, as Effortful Control has a generally positive effect on motor development, we should encourage it from the early stages. On the other hand, Surgency seems to have both positive and negative effects (positive on the gross motor system and negative on the fine motor system). Thus it will be necessary to adjust Surgency appropriately, making it neither too strong nor too weak. In addition, since Negative Affect may have increased negative effects on coordination movements at 42 months, it will be desirable to avoid giving the child excessive negative experiences by refraining from scolding or criticizing the child's behavior. On the other hand, high Effortful Control scores among preschool children have been reported to be associated with lower physical activity and greater sedentary time [69]. Understanding how the development of Effortful control relates to motor development and even adaptive functioning may lead to effective early intervention [70]. At the same time, it will be desirable to develop screening items that can be easily answered by parents with no expertise in infant motor development.