Prefrontal Cortex Response to Threat: Race by Age Variation in 9-10 Year Old Children

Shervin Assari1,2*, Golnoush Akhlaghipour3, Mohammed Saqib4, Shanika Boyce5, Mohsen Bazargan1 1Department of Family Medicine, Charles R. Drew University of Medicine and Science, Los Angeles, CA, USA 2Department of Urban Public Health, Charles R. Drew University of Medicine and Science, Los Angeles, CA, USA 3Department of Neurology, UCLA, Los Angeles, CA, USA 4Department of Health Behavior and Health Education, University of Michigan, Ann Arbor, MI, USA 5Department of Pediatrics, Charles R. Drew University of Medicine and Science, Los Angeles, CA, USA

important, adolescence is a unique phase determined by a rapid change of brain maturation 9 . The National Institute of Health (NIH) notes about child brain development that: 1) the brain reaches its biggest size in early adolescence (11 and 14 years old for girls and boys respectively), 2) the prefrontal cortex undergoes maturation later than other parts of the brain, 3) a teenager's brain has considerable plasticity and resilience, meaning that it can successfully adapt and respond to its environment, 4) undergoes ongoing and rapid changes with 5) co-occurring physical, emotional, and social changes, and 6) exhibits the highest vulnerability to stress 9 .
Response to threat [10][11][12] , one of the major functions of the brain, has major implications for survival as well as social relations. Research has well shown that a large array of cortical and subcortical brain structures are involved in processing threat-and fear-related information [13][14][15][16] , all being essential for maintaining social behavior, group affiliation, and attachment 17 . To increase survival, the brain needs to strongly respond to threat, as it encounters and processes negative stimuli (e.g. threat or fear related), as well as responding to stimuli that are unpredictable, ambiguous, and potentially dangerous 18 . An increased response to threat is also a part of psychopathologies such as anxiety 19 , depression [20][21][22][23] , poor decision making [24][25][26] , aggression 27,28 , and risk behaviors 29 . Research has shown that age [30][31][32] , race [33][34][35][36][37][38] , and socioeconomic status (SES) [39][40][41] are among the main predictors of the brain response to threat.
While age may alter how the brain evaluates and responds to threat, it is not the only factor. Race, family socioeconomic status, are all proxies of exposure to stress, adversities, trauma, and all other salient determinants of children brain response to threat [39][40][41][42][43] . As social adversities, stress, and trauma impact brain response to threat 43 , there is a need to control if race and age still impact brain response to threat after socioeconomic status (as a proxy of stress and trauma) is controlled for.
Most of the existing research on the effect of age on brain response to threat is performed in White middleclass individuals. Research has, however, shown that what applied to the White middle-class may not necessarily apply to Blacks and other racial and ethnic groups. Any enhancement of these studies, however, require a large sample size. Such sample size requires large scale national studies that enhance statistical power for racial comparisons and also generate results that are generalizable to the US population. The Adolescents Brain Cognitive Development (ABCD), a national children's brain development study data with 10,000+ sample size [44][45][46][47][48] has generated an unprecedented opportunity to investigate how racial groups differ in the effects of risk and protective factors on brain development. As such, the ABCD study has also provided us a unique opportunity for comparison of Black and White children for age-related changes in those who live in low socioeconomic status families and neighborhoods. ABCD has a large sample size, national sampling, and has collected data on the brain imaging of 10,000+ American children [44][45][46][47][48] .

Aims
To understand the separate, additive, and multiplicative effects of the race (as a social factor) and age on the brain development of children, we borrowed data from a national sample of American children, with two aims: First to test the additive and multiplicative effects of age and race on children's brain response to threat. By testing multiplicative effects, we mean to explore the differential associations between age and brain response to threat across racial groups. Using the ABCD data, we expected an interaction between race and age, which was an indicator of a difference between White and Black children for the effects of age on brain function. It is not race or age, but their combined and interactive effects that shape brain response to threat (negative face). This expectation is in line with the other research on behaviors and brain development showing that protective factors may have diminished effects for Black than White children [49][50][51] .

Design and settings
This cross-sectional study was a secondary analysis of the existing data, borrowed from the ABCD study [44][45][46][47][48] . This analysis is limited to the wave 1 data of the ABCD study. As described elsewhere 44,52 , ABCD is one of the largest brain imaging studies of child brain development that has ever been conducted. Among the main advantages of the ABCD study are being a national sample, a large sample size, a large sample of Blacks, public availability of the data, robust measurement of the brain development, and considerable psychiatric and psychosocial variables [44][45][46][47][48] .

Participants and Sampling
Participants of the ABCD study were selected across multiple cities across states. This ABCD sample was primarily recruited through the school systems with sampling (school selection) informed by race, ethnicity, age, sex, socioeconomic status, and urbanicity. For additional methodological details of ABCD sampling and recruitment, please see the paper published elsewhere 53 . Analytical sample for this study was limited to White or Black children who had data on n-Back task (n=5,066). negative relative to neutral face (Table 1). PFC response to threatening (negative) relative to neutral face was captured by functional MRI measures during the N-Back task. A detailed description of the ABCD study procedures and harmonization of the fMRI studies are available elsewhere 45 . N-Back task measures a wide range of brain functions such as emotion process, emotion regulation, impulsivity, working memory, encoding, retrieval, forgetting, and recognition. This task is ideal for measurement of dorsolateral prefrontal, parietal and premotor cortex, hippocampus, parahippocampus, and amygdala response.

Outcome
Outcomes in this study were brain response to threatening (negative) relative to neutral faces. These outcomes were selected because brain response to threatening (negative) faces is shown to be a leading brain function that is essential for socialization. These included the mean beta weight for n-Back run 1 threatening (negative) face versus neutral face contrast in APARC ROI left hemisphere-lateral orbito-frontal, left hemisphere -superior-frontal, right hemisphere -caudal middle frontal, and right hemisphere -superior frontal.

Independent Variables
Age. Parents reported the age of the children. Age was calculated as the difference from date of birth to the date of the enrollment to the study. Age was a continuous measure in years.

Confounders
Parental education (years): Parents were asked, "What is the highest grade or level of school you have completed or the highest degree you have received?". This variable tanged from 1 (no official education) to 21 (doctoral degree).

Family income:
Parents were asked "What is your total combined family income for the past 12 months? This should include income (before taxes and deductions) from all sources, wages, rent from properties, social security, disability and veteran's benefits, unemployment benefits, workman". Family income was a continuous measure ranging from 1 (Less than $5,000) to 10 ($200,000+), with a higher score indicating higher income. It is also parallel to the corpus callosum before it turns upwards to the superior margin of the brain hemisphere under the name pars marginalis (marginal part).  Financial difficulties: Financial difficulties were measured by the following seven items: "In the past 12 months, has there been a time when you and your immediate family experienced any of the following:" 1)"Needed food but could not afford to buy it or could not afford to go out to get it?", 2) "Were without telephone service because you could not afford it?" 3)" Did not pay the full amount of the rent or mortgage because you could not afford it?", 4) "Were evicted from your home for not paying the rent or mortgage?", 5)"Had services turned off by the gas or electric company, or the oil company would not deliver oil because payments were not made?", 6) "Had someone who needed to see a doctor or go to the hospital but did not go because you could not afford it?", and 7) "Had someone who needed a dentist but could not go because you could not afford it?" Responses to each item were either 0 or 1. We calculated a sum score with a potential range between 0 and 7, a higher score indicating lower family socioeconomic status. Our variable was a continuous measure. Financial difficulties are an accepted socioeconomic status indicator [54][55][56][57][58][59][60] .

Neighborhood Income (Socioeconomic Status):
Derived from ABCD residential history file, we used the median family income of the neighborhood as neighborhood socioeconomic status. This is in line with the Area Deprivation Index (ADI), based on the work done by the Health Resources & Services Administration (HRSA), Amy Kind, Ana Diez Roux, and others. We used the neighborhood income of the county-level / census block group/neighborhood. Extensive research suggests that ADI, and median family income, and neighborhood income are predictors of health. This variable was not calculated in terms of 1 USD but 100,000 USD [61][62][63][64] .
Ethnicity: Parents reported if they are of Hispanic ethnic background. This variable was coded as Hispanic = 1 and non-Hispanic = 0.
Sex: Sex was a dichotomous variable with males as 1 and females as 0.

Data Analysis
We used SPSS for data analysis. For descriptive purposes, we reported frequency (%) and mean (standard deviation [SD]) for categorical and continuous variables, respectively. To conduct our bivariate analyses, we used Chi-square and independent samples t-test to compare Blacks and White children. To perform our multivariable analyses, we performed two series of four linear regressions for each outcome. The independent variable was age. The outcome were one of the structures namely 1) left hemispherelateral orbito-frontal cortex, 2) left hemisphere -superiorfrontal cortex, 3) right hemisphere -caudal middle frontal cortex, and 4) right hemisphere -superior frontal cortex (Table 1). Our models first only controlled for sex. Then we ran models that also controlled for ethnicity and SES indicators namely parental education (years), family

Ethical Aspect
The ABCD study protocol received approval from the University of California, San Diego (UCSD) Institutional Review Board (cIRB). Some other universities that were a part of the ABCD study also obtained local IRB approval. All the participating children gave assent. Parents also signed informed consent. More detailed information on the ABCD study is available elsewhere 52 . Given that the data used in this study were fully de-identified, our investigation was non-human subject research. Thus, it found to be exempt from a full IRB review.

Descriptives
The current analysis was performed on 5,066 9-10 years old children who were either White (n = 3,782; %74.7) or Blacks (n = 1,284; %25.3). In Table 1 we have presented the descriptive demographic, socioeconomic status, and brain response to threat data. These data are presented for the pooled sample, as well as by race ( Table 1).
As Table 1 shows, White and Black children did not differ in age and sex but differed in family socioeconomic status and neighborhood socioeconomic status. Compared to White children, Black children had lower SES. Black and White children did not differ in brain response to threat (Table 1).  status indicators were not correlated with brain response to threatening versus neutral faces. These correlations were inconsistent between racial groups and were based on brain outcome. Table 3 reports the results of the two regression models in the pooled sample. Models 1-a, which only included the main effects, showed that age or race were not associated with activation of various brain regions in response to threatening than neutral faces. Model 2-a showed interactions between race and socioeconomic status on activation of the brain regions in response to threatening than neutral faces. Table 6 reports the results of race-specific models.

Model 3-a was performed in White children and Model 4-b
was performed in Black children. We found that age was not associated with brain response to threatening versus neutral faces in White or Black children. Table 5 reports the results of two regression models in the pooled sample. Models 1-b, which only included the main effects showed that race or age were not associated with activation of various brain regions in response to threatening than neutral faces. Model 2-b showed interactions between race and socioeconomic status on activation of the brain regions in response to threatening versus neutral faces. Table 6 reports the results of race-specific models. Model 3-b was performed in White children and Model 4-b was performed in Black children. We found that age was not associated with brain response to threatening versus neutral faces in White or Black children. However, the association was positive for Whites and threatening for Blacks.

Discussion
We found race by age effects in children brain response to threat. This is indicative of racial variation in age-related changes in brain response to threat. We, however, showed that these race by age effects remain significant when SES is controlled for. That is, race, per se, not as a proxy of SES, alters age-related brain development.
We found race per se not to be associated with brain response to threatening versus neutral faces in White but not Black children. Age was also not directly associated with activation of the brain in response to threatening  pattern for the outcomes, our four variables, response in the left hemisphere-lateral orbito-frontal, left hemisphere -superior-frontal, right hemisphere -caudal middle frontal, and right hemisphere -superior frontal ROIs also showed differences and nuances in their correlates.

Model 2-a All
Race was not found to change the brain response to threat. Race in our study was a proxy of adversities, stress, and poverty. Childhood poverty may be linked to reduced connectivity between the amygdala and hippocampus and some other regions, including the superior frontal cortex, lingual gyrus, posterior cingulate, and putamen [65][66][67][68] . The study showed that while childhood poverty predicts connectivity between 1) the left hippocampus and the right superior frontal cortex and 2) the right amygdala and the right lingual gyrus, this brain connectivity mediates the effect of childhood poverty on children depression 69 . Others have established hyperactivation of the reward network and hypoactivation of the executive network in low socioeconomic status individuals 70 . The effects of low socioeconomic status and stress on how the brain reacts to threat go beyond amygdala hyper-reactivity to threat and also impacts many other brain structures involved in memory and cognition 71 .
Research has shown that each additional year spent in poverty may be associated with a lower level of connectivity in neural networks involved in emotion regulation. However, these effects may be more pronounced for children who receive low levels of supportive parenting 72 . Blacks, compared to Whites, receive lower levels of protective parenting and live longer under poverty.
Our finding is indicative of double jeopardy for the brain development of Black children. The first risk is that race may be associated with poor brain development, as race is a proxy of exposure to poverty and racism. In addition, agerelated brain development may be delayed, again, probably as a result of racism and discrimination, for Black children. Studies have shown that racial discrimination results in an increase in amygdala connectivity with multiple brain regions. In the presence of racial discrimination, the amygdala shows more strong connection with the thalamus. Discrimination increased the connections between the amygdala and the putamen, caudate, anterior insula, anterior cingulate, and medial frontal gyrus 73 .
The more salient effect of age for White than Black children's brain development may be due to the lack of supportive elements and environment in Black lives. Other stressors such as racial and ethnic discrimination are unique in the life of racial minorities. Many studies have shown that social and physical environment is more enriched and resource filled for White than Black families. Libraries, schools, and houses where White families live have many books and educational resources. For Black families, however, such resources are scarce. As a result of growing in a rich or poor environment, age may not have the same effect on the brain development of White and Black children.
The results are related to a pattern called Minorities' Diminished Returns (MDRs) 50,51 . MDRs refer to protective and risk factors that differently influence outcomes for Black and White families. For example, socioeconomic status interact with race on anxiety 74 , depression 75 , poor health 76 , poor school performance 77,78 , as well as high-risk behaviors 49 such as aggression 49 and tobacco use 79,80 . Black-White differences in correlates of health and development seem to be rules rather than exceptions 76,[81][82][83][84] . For example, in the Fragile Families and Child Wellbeing Study (FFCWS), correlates of impulsivity, school performance, school bonding, attention-deficit/hyperactivity disorder (ADHD), obesity, aggression, depression, and self-rated health differed for White and Black children [84][85][86][87] . These differences between Black and White children 84-87 are due to differential effects of precursors on a wide range of developmental and health outcomes of Blacks and Whites 76,81,83,[88][89][90][91][92] . As a result of these MDRs, Black children show diminished returns on health outcomes regardless of their profile of exposure to risk and protective factors, while for Whites, risk and protective effects alter the outcomes 49,86,[93][94][95] . The same patterns of diminished returns are also shown for age 96 . Our study proposes that the same interaction may be seen for the effects of age on brain development.

Some cautionary notes
Here we list six points that need caution. First, the use of age as a primary independent variable using crosssectional data from a longitudinal cohort study may be inappropriate. There is some age variability, with a range of 8-11 years old, but this cohort was recruited intentionally to be roughly the same age. Future data releases will allow for within-person evaluations of longitudinal change in neural activation that will be better suited for answering the questions posed by this paper. Second, this study is simplistic as it used age as a primary independent variable using cross-sectional data from a longitudinal cohort study. Although there is some age variability, with a range of 8-11 years old, the ABCD cohort was recruited intentionally to be with low age range at baseline. Future data releases will allow for within-person evaluations of longitudinal change in neural activation that will be better suited for answering questions posed in this paper. Third, we should emphasize that we see race as a social factor (a proxy of poverty and socioeconomic status) on how the brain is affected by low or high socioeconomic status (parental education). Across various brain mechanisms, we focused on the amygdala, which is highly involved in emotional regulation, emotional expression, aggression, and impulsivity. An alteration of the amygdala response is expected to be involved in a wide range of emotional, cognitive, and behavioral outcomes. To be more specific, we tested if race, as a proxy of racism, adversities, and stress alters age-related brain development of children, with a specific focus on brain response to threatening (negative) relative to neutral face, which is shown to be one of the brain functions profoundly affected by race, socioeconomic status, stress, and adversity 40,41,[97][98][99] . Fourth, this study was exploratory. The ABCD data provides an opportunity to answer important questions with high statistical power. Studies utilizing the ABCD data have the potential to have a major contribution to the field. This paper is the first step in this regard. Fifth, no matter how responsible we attempt to be in our interpretations, studies arguing for racial differences in biological functioning are controversial. They have the potential to be abused. They are at an increased risk of it being used to harm or stigmatize already marginalized groups. Sixth, in almost all epidemiological studies, race has an imbalanced distribution, and this has implications for the power of the study. Thus, the results of racestratified models should be interpreted with caution. However, most of our analyses were in the pooled sample, which is less affected by the distribution of race.

Conclusions
In summary, in a large national sample of American children, age and race do not have independent (separate) but inter-dependent (multiplicative) effects on PFC response to threatening (negative) relative to neutral faces. This suggests that age-related brain development of children might be under the influence of race (as a social rather than biological construct). More research is needed on how social factors such as race, racism, social environment, adversities, and stress influence children brain development as a function of age.
ABCD Funding: Data used in the preparation of this article were obtained from the ABCD Study, held in the National Institutes of Mental Health Data Archive. The ABCD Study is supported by awards U01DA041022, U01DA041028, U01DA041048, U01DA041089, U01DA041106, U01DA041117, U01DA041120, U01DA041134, U01DA041148, U01DA041156, U01DA041174, U24DA041123, and U24DA041147 from the NIH and other federal partners. A full list of supporters is available at https://abcdstudy.org/federal-partners.html.