The COVID-19 Pandemic and Early Child Cognitive Development: A Comparison of Development in Children Born During the Pandemic and Historical References

INTRODUCTION

The outbreak of the SARS-CoV-2 (COVID-19) pandemic in the USA in the latter part of 2019 and early months of 2020 has brought widespread change and disruption to our social, economic, and public health environments. While children have largely been spared the severe health and mortality complications of SARS-CoV-2 infection (1, 2), they have not been immune to the impact of the public health and social distancing policies that closed daycare centers, schools, and playgrounds (3, 4) and limited opportunities for play and interaction. While these policies helped to limit COVID-19 spread they also disrupted educational opportunities (5), limited interaction with other children and caregivers (6), and reduced opportunities for physical activity and play (7). More broadly, these policies have also shuttered businesses and led to large-scale employment shifts which may have affected the home, food, and financial security of children through parental or caregiver job loss or furloughs.

In light of these social and economic challenges, there has been concern amongst early child researchers regarding the impact of the pandemic on infant and early child development and mental health. While there are few past examples of non-conflict-related lock-downs with a similar geographic extent and prolonged timeline, past studies of acute and chronic environmental stress on child development provide important context. There is little doubt that chronic maternal stress, anxiety, and depression throughout pregnancy can impact fetal and post-natal development (8, 9). Maternal stress experience may be associated with increased fetal exposure to stress-related hormones (e.g., cortisol), which can result in differential brain structural and functional development and later cognitive impairments (10-13). For example, past analysis has revealed strong associations between maternal prenatal stress and anxiety related to maternal or paternal displacement and job loss and infant health (birth weight and gestation duration), mortality, temperament, and cognitive development (14). Severe acute stress, such as that associated with natural disasters (e.g., the 1998 snowstorm that affected central Canada, and the 2006 Hurricane Katrina in Louisiana) has been shown to yield long-term cognitive and intellectual consequences in children who were exposed in utero (15, 16). Of note, in utero exposure to maternal stress appears to more adversely affect the male fetus (17, 18), implying potential sexual dimorphism in stress sensitivity.

In addition to the potential impact on the developing fetus, children exposed to environmental stress and adversity may also be similarly affected. Prolonged stress, such as experienced during wartime or refugee replacement (e.g., the Bosnian war from 1992-1995) is associated with worsened long-term physical and mental health (19). Recent cross-sectional and longitudinal studies of children and adolescents during the early stages of the pandemic have revealed increased stress, anxiety, and depression (20) alongside reduced academic growth in math and language arts (4).

While pre and post-natal stress and adversity can clearly impact the developing child, these effects may be moderated by compounding social and structural factors such as socioeconomic status, access to healthcare, and other social determinants of health (21, 22). The COVID-19 pandemic has disproportionately affected Latino, Black, African American, and lower socioeconomic status (SES) individuals and families (23, 24). Of reported COVID-19 cases in pregnant women, the majority are Latino (42.5%) and Black and African American (26.5%) (25, 26). 46% of hospitalized children are Latino and 30% are Black or African American (27) and are more likely to develop severe illness (28). Likewise, Latino and Black/African American adults are three times more likely to be hospitalized with COVID-19, and 1.9-2.3 times more likely to suffer severe complications or death (29). Adults with incomes below $15,000 are at 2x greater risk of serious illness than those who make $50,000 or more (30) and they have been hardest hit by the closure of non-essential services, employee furloughs, and childcare and school disruptions.

Finally, changes in the children’s home environment may also have a developmental impact. For example, while associated with stress, reduced income from job loss or layoff may affect the ability to provide optimal nutrition and access to healthcare (31). Closed daycare centers, public parks, and gathering spaces can limit explorative play (32), child-child interaction, and environmental stimulation (22). And parents who may have taken on additional jobs may have experienced reduced time to read and interact with their children. All of these factors have previously been associated with variations in cognitive and behavioral development and outcomes (33-35).

Taken together, these past findings strongly imply the likely impact of the ongoing COVID-19 pandemic and related environmental changes and stressors on child health and neurodevelopment. However, the specific impact on infant and pre-school cognitive development remains less clear, with few studies focused on this early age range from 0-3 years (36). However, emerging results are disturbing. For example, findings from a prenatal cohort in New York City recruiting during the earliest stages of the pandemic reveal reduced motor and cognitive development in infants at 6-months of age (37), and children 6 and 12-months of age born during the pandemic in Guangzhou, China, similarly show reduced communication skills compared to historical references (38).

From 2011, the RESONANCE longitudinal study of child health and neurodevelopment has been co-ordinated from the Warren Alpert Medical School at Brown University in Providence. This study includes approximately 1600 caregiver-child dyads who have been continuously enrolled between 0 and 5 years of age since 2010 and have been followed through infancy, childhood, and early adolescence. This cohort, therefore, offers a unique opportunity to examine the potential impact of the COVID-19 pandemic on child health trends in RI and to compare with other trends observed throughout the US.

Like many regions in the United States, RI implemented a series of school closures, public access, and travel restrictions, and limited social gatherings early on in the pandemic to limit virus spread. Elementary schools were closed to in-person instruction from March to September 2020 and many continued to operate fully online or with hybrid in-person learning until January 2021. Daycares were also closed in March 2020 but were allowed to reopen with limited capacity from June 2020 until June 2021. Broader state-wide travel restrictions and stay-at-home orders were enforced from March to May 2020, with many businesses operating with reduced on-site workforces and/or work-from-home options until mid-2021. Indoor and outdoor mask policies were also in place throughout 2020 and 2021 following CDC guidance. Despite being one of the smaller US states (population ∼1 million), RI has recorded ∼154,000 cases of COVID-19 illness and almost 3,000 deaths and mirrored national trends with respect to disproportionate infections and deaths in Hispanic, Latino, and African American communities (39, 40), and lower-income families (41). Thus pregnant individuals and families in RI experienced similar pandemic-related stresses as those elsewhere in the US.

Drawing from our RESONANCE dataset, we used linear mixed-effects models to examine trends in infant and early child neurodevelopment over the past decade, from 2011 to 2021. We find that even controlling for factors including age, biological sex, socioeconomic status, and maternal stress, measures of verbal, non-verbal, and overall cognitive functioning are significantly lower in 2021 compared to past years. These results provide additional supporting evidence that aspects associated with the COVID-19 pandemic, even in the absence of direct SARS-CoV-2 infection, have impacted infant and early child neurodevelopment.

METHODS

All data were collected in accordance with ethical approval and oversight by the Rhode Island Hospital institutional review board. Informed consent was obtained from all parents or legal guardians.

Between January 2011 and December 2021, a total of 1247 cognitive assessments have been performed on 700 healthy and neurotypically developing children from 3 months through 3 years of age. At the time of enrollment, all children met the following criteria for inclusion: Full-term gestation (38 weeks or greater); singleton pregnancy; no abnormalities on prenatal ultrasound; healthy birth weight (>2500g); no exposure to cigarette smoke, alcohol, marijuana, or illicit substances in utero; uncomplicated delivery (e.g., no history of birth asphyxia), no history of gestation diabetes in the mother; no major psychiatric illness in the mother including depression requiring medication in the year prior to pregnancy; no neurological trauma in the infant; and no diagnosed neurological disorder in the infant (e.g., epilepsy). Further, none of the children or pregnant mothers with newborns reported a positive antibody, antigen, or PCR-TR result for SARS-CoV-2 infection over the past year, or had a clinical diagnosis.

For this study, we examined the potential impact of pandemic-related environmental changes in the following groups: 1. 700 children between 3months and 3 years of age assessed between January 1, 2011, and November 30, 2021; and 2. In a comparison of 525 infants 0-16months of age born before the pandemic (n=388) and during the pandemic (n=137). A pictorial overview of all included child assessment timings is shown in Supplemental Figure 1, with demographic details provided in Table 1.

RESULTS

Comparing Pre-Pandemic and Current Mean Cognitive Measures

Mean ELC, VDQ, and NVDQ scores for each year from 2011 to 2021 are shown in Figure. 1 for (a) all children between 0-3 years of age; and (b) for the subcohort of infants 0-16months of age.

• Download figure
• Open in new tab

Figure 1.

Visual comparison of yearly mean MSEL ELC, VDQ, and NVDQ composite scores for (top) all children 0 to 3 years of age in the cohort; and (bottom) infants 0 to 16 months of age born before or during the pandemic. In each panel, the black line represents the expected mean (100) and the grey region denotes the expected standard deviation (15). For 2020 we provide 2 measures, one for January 1 to April 1, 2020, and the second from April 2 to December 30, 2020. The box denotes measures obtained during the pandemic since April 1, 2020. In both child cohorts, we observe general decreases in cognitive performance in 2021 that, as shown in Table 2, are statistically significant.

In infants and toddlers (Fig. 1a) mean ELC values from 2011 to 2019 ranged from 95 to 107.3 with standard deviations ranging from 15.2 to 19.7 - in alignment with the expected mean of 100 and standard deviation of 15. Controlling for differences in age and maternal education via an ANCOVA (Table 2), we found inconsistent significant differences in mean ELC scores between 2011-2019 and the second half of 2020, but consistent and significant (p<0.05) reductions between 2011-2019 and 2021. Similar results were found for VDQ and NVDQ. Most differences between 2011-2019 and 2021 remained significant even after strict Bonferonni correction for the 30 tests (corrected p-value < 0.002). Effect sizes (Cohen’s f²) for significant differences are listed in the table and range from small (0.05) to large (0.4). The largest effect sizes were noted in the comparisons to the 2021 cognitive values.

View this table:

• View inline
• View popup
• Download powerpoint

Table 2.

Results of the yearly pairwise comparisons (ANCOVA) in overall, verbal, and non-verbal cognitive scores in all children 0 to 3 years of age. We note near consistent significant reductions in ELC, VDQ, and NVDQ measures in 2021 compared to the past decade (even correcting for multiple comparisons p < 0,002), with less consistent differences in 2020. Effect sizes for significant differences ranged from low (0.05) to high (0.3)

Focussing on infants born before and during the pandemic (Fig. 1b), we found similar reductions in performance in 2021 (mean ELC, VDQ, and NVDQ scores of 83.1+/-14.1, 75.5+/-24.1, and 91.1+/-24.1 compared to 2011-2019 (Table 3). However, not all comparisons remained significant after correction for the multiple comparisons. Similar effect sizes were noted as found in the larger cohort, ranging from 0.03 to 0.58.

View this table:

• View inline
• View popup
• Download powerpoint

Table 3.

Results of the yearly pairwise comparisons (ANCOVA) in overall, verbal, and non-verbal cognitive scores in infants 0 to 16-months of age. We note near consistent significant reductions in ELC, VDQ, and NVDQ measures in 2020 and 2021 compared to the past decade. Effect sizes for significant differences ranged from low (0.05) to high (0.3)

DISCUSSION

Children are inherently shaped by the environment in which they learn, grow, and play. Over the past 20 months, that environment has been fundamentally altered for many children. While the short and long-term neurodevelopmental impacts of the COVID-19 pandemic are not yet well understood in infants and young children (49), emerging reports are suggestive of reduced motor and communication skills (38). Results from this study add to this literature and are in agreement with these past studies. We find that verbal, non-verbal, and overall cognitive scores in children under 3 years of age are significantly lower than over the past decade with this difference amplified in infants <16-months of age and born since the beginning of the pandemic. As the pregnant individuals and children included in this study reported no symptoms of SARS-CoV-2 infection or had positive antibody or RT-PCR test results, this suggests that observed effects are related to environmental factors rather than due to direct effects of infection. However, we did not perform independent antibody testing to confirm past infection status.

Longitudinal modeling of children 0-3 years of age revealed positive associations between child cognitive performance and maternal education, gestation duration, and birth weight; and negative associations with male biological sex. In addition, we found a positive association between cognitive scores and the interaction of maternal education x PANDEMIC. Socioeconomic characteristics such as material education have well-known associations with child health and cognitive development and may be related to differences in the ability to provide adequate nutrition, supportive parenting, and stimulating and engaging environments. That increased maternal education was found to buffer the negative effects of the pandemic may be related to these families being able to provide continued stability (lower food and/or housing insecurity) and interactive stimulation (e.g., reading and play). In contrast, many lower-income families have dealt with job loss and financial insecurity that has necessitated taking on additional work and reducing the available family time and/or ability to provide quality daycare. Additional investigation of changes in food and housing insecurity, shifts in daycare, and employment status throughout the pandemic is needed to understand these individual factors but was unavailable here.

Other significant factors identified throughout our analysis were gestation length and infant birth weight. While this aligns with past studies of infant birth outcomes and subsequent development, most have focused on premature and low birth weight infants - infants excluded from our analysis. However, even amongst healthy birthweight infants, there is evidence that increased birth weight and later math and language processing (50). Similarly, even within children born full-term, later infant and children neurocognitive outcomes have been shown to improve for each week of gestation past 37 weeks, with a peak at 40-41 weeks (51).

In contrast, we found that males have overall lower neurodevelopment. The reason for this finding in our cohort is unclear, however, past analysis of male and female cognitive development using the MSEL has shown that in the context of children of mothers with depression, females performed better than males and improved more quickly between 6 and 18-months of age (52), overlapping with the age range investigated here.

One hypothesized mechanism by which the environment may affect neurocognitive development is via maternal pre-and postnatal stress. However, in contrast to other studies of mothers during the pandemic, such as the MOM-COPE study and a large online survey of pregnant individuals at the beginning of the pandemic (53, 36), we did not observe increased maternal stress in our study population, and it was not a significant predictive factor in our analysis. Two possible explanations for this are 1. Insensitivity of the PSS tool used to pandemic-specific stress; and 2. A potential selection bias in the families included in our study. Addressing these explanations, in turn, the PSS is a standardized tenitem questionnaire that asks about general life stressors and how stressful individuals find their lives. It, however, does not include specific questions related to health or wellbeing. In contrast, the MOM-COPE study utilized retrospective data collection using an ad-hoc developed questionnaire focused on worry and anxiety of COVID-19 infection, pregnancy risk, and their own and their infant’s health (36).

The online pregnancy survey study by Lebel and colleagues (53) also used a specially developed questionnaire to gauge maternal concern of the pandemic and its impact on their own and their infant’s health. Thus, it is possible that stress felt by parents and families specific to COVID-19 infection or complications during pregnancy was not identified on the PSS given its generic and broad nature that may have been revealed with more sensitive tools.

A second explanation as to why the stress was not found to be a significant factor is the potential selection bias of the families included in this study and, in particular, those who were assessed during the pandemic. Unlike many other academic research laboratories across the US, our facility has continued to operate throughout the pandemic with in-person study visits, with appropriate pre-screening and masking policies. As our center is located within a clinical setting, parents less concerned about the pandemic may have been more likely to participate than those with greater concerns. Thus, our observation that maternal stress (PSS) was not significantly increased may simply reflect the reality that we only tested less stressed and anxious families.

In the absence of stress-related changes, additional factors that may explain our findings are the use of face masks during testing. Although all study visits were performed in-person, the inability of infants to see full facial expressions may have eliminated non-verbal cues, muffled instructions, or otherwise impair the understanding of test questions and instructions. Without direct comparison of performance in the same children with and without face masks, it is difficult to rule in or out the potential influence of masks. However, some insight can be gleaned by examining the raw domain scores (Supplemental Figure 3). Over the first year of a child’s life, the MSEL focuses strongly on non-verbal domains such as 1. Gross motor control skills: e.g., supporting their head, pushing up on their forearms, rolling over, and pulling themselves up to standing; 2. Fine motor skills: e.g., holding rings and grasping and reaching for objects; and 3. Visual skills: e.g., fixation and face tracking. Tested verbal skills are also relatively basic, including orienting to sounds, responding to a voice, swallowing, making coos, chuckles, or laughing sounds, and babbling. For much of the assessment, the examiner is on the floor playing with and manipulating the child, and the assessment is not reliant on only verbal cues or instructions. In examining raw domain scores, we note that overall infants born since the pandemic are most affected in their basic motor and visual receptive skills rather than their expressive and receptive language abilities. Given the nature of these affected skills, it seems unlikely that mask-wearing by the experimenter could be the cause of the reduced scores in late 2020 and 2021. Of note, recent work using parent-reported development measures reported similar findings of reduced motor development in 6-month old infants born during the pandemic, further suggesting masks do not underly our observed findings.

Secondary to masks, however, may be the unfamiliarity of infants and toddlers to individuals outside of their immediate family. Infants may have been distracted in the unfamiliar laboratory setting or suffered “performance anxiety” in front of the unknown research staff and testers testing strangers even with their parents nearby. Our research staff and neuropsychologists are highly trained and well versed in interacting with children across this age group and working with the child and their parents to obtain the most representative assessment.

Our findings raise important questions. First, do they generalize across the US or are they specific to the RI population and the cohort investigated? Despite the relatively restricted sample size, particularly within the infant cohort, we note the high effect sizes in our findings suggest our findings should be reproducible. Indeed, our results align with other emerging reports from other cohorts that used different assessment tools. It is unclear how different state-level responses to the pandemic will affect child outcomes, for example, Arkansas, Iowa, Nebraska, North Dakota, South Dakota, Utah, and Wyoming that did not issue stay-at-home orders. Second, are observed differences in infant neurocognitive performance since the pandemic temporary, and will they normalize as children grow and return to prepandemic levels of play, interaction, schooling, etc. Unfortunately, the appearance of new virus variants and continued high infection and hospitalization rates have prolonged the much-anticipated return to “normal”. Our data, currently, is not sufficient to determine if observed reductions are transient or longer-lasting. However, the longitudinal nature of our study will allow us to address this in time as individuals return. It is clear, however, that young infants and children are being affected now. Programs such as unemployment insurance, Supplemental Nutrition Assistance Program (SNAP), the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), and housing assistance, may help minimize the impact of the pandemic on the most sensitive of children. In addition, further research directly exploring aspects of parent-child attachment, interaction, nutrition, food security, and environmental stimulation is needed to understand the primary causative factors underlying the results presented here.

CONCLUSION

The COVID-19 pandemic has fundamentally altered the child health landscape, with families living in strikingly different economic, psychosocial, and educational environments than what was present just 18 months ago. Against this environmental backdrop, unanswered questions remain regarding the impact of the work-from-home, shelter-in-place, and other public health policies that have limited social interaction and typical childhood experiences on early child neurodevelopment. In this work, we provide early evidence suggestive of significant reductions in attained cognitive function and performance in children born over the past 18 months during the pandemic. While maternal education appears to mitigate against the negative consequences of the pandemic, the primary factors underlying this observation remain unclear. Understanding these factors is critical to helping ensure affected children rebound as the pandemic winds down and they re-enter daycares and schools; as well as implementing additional public health and educational policies that address the most affected of children, particularly those in lower-income families.

DATA SHARING

All data acquired and presented here is available upon request to the authors.

ROLE OF THE FUNDING SOURCE

Funding for this study was provided by the National Institutes of Health (SCD). Neither funder played any role in the acquisition, analysis, or interpretation of the data, or was involved in the drafting or approval of this manuscript.

CONTRIBUTOR ROLES

All listed authors were involved in the study design, data acquisition, and analysis, drafting and revising this manuscript, and providing final and accountable approval for its contents. SCD and DV verify the underlying data.

FINANCIAL DISCLOSURES

The authors report no significant financial conflicts of interest with respect to the subject matter of this manuscript.

Supplemental Table Legends

Supplemental Figure Legends

• Download figure
• Open in new tab

Supplemental Figure 1.

Visual overview of all child study visits used in each set of analysis. (a) All children from 0 to 3 years of age; (b) Infants born prior to (blue) or during (red) the pandemic.

• Download figure
• Open in new tab

Supplemental Figure 2.

Plots of the mean gestation period, birth weight, maternal education, and maternal perceived stress for our child sample that were measured each year.

• Download figure
• Open in new tab

Supplemental Figure 3.

Raw MSEL domain scores in 0 to 16-month-old infants born before the pandemic (blue) and during the pandemic (red).