Prenatal cocaine exposure places infants at risk for developmental delays, particularly language delays (Chapman, 2000; Lester, LaGasse, & Seifer, 1998), which increase the likelihood that they will require special education services as children and show poor developmental outcomes (Delgado, Vagi, & Scott, 2006). This study used second-order latent growth curve modeling to identify the trajectory of language development of infants from 12 to 36 months of age who had completed a three-year early intervention program. This study also investigated the predictive ability of six factors related to the child, family, and intervention. Child factors included sex, birthweight, and cognitive ability at 12 months, a family factor included mother's education level, and intervention factors included intervention type and age at enrollment. Results revealed that language growth across the two-year period was quadratic, with growth declining initially, then accelerating after 24 months, compared to the nationally normed sample upon which standardized language scores were based. Being a girl, being enrolled in the program soon after birth, and having higher cognitive ability at 12 months predicted higher initial language ability. Intervention group predicted growth in language, such that both intervention groups predicted less or no initial decline compared to the primary care non-intervention group, and the center-based intervention group showed faster acceleration after 24 months compared to the home-based intervention group. Implications of these results on the research literature and practical applications were discussed.

The positive link between children's joint attention skills and subsequent language outcomes is well established. There is also abundant evidence that responsive caregiver behaviors lead to optimal language outcomes. Though directive behaviors are generally considered detrimental to children's growth, specific types of directive behaviors which extend or build upon a child's behavior are thought to promote children's learning. No study has examined how caregiver behaviors interact with children's joint attention skills to affect subsequent language. Therefore, the objective of the present study was to examine how three different caregiver behaviors (Responsiveness, Supportiveness, and Intrusiveness) measured at 18 months moderated the relation between joint attention (also measured at 18 months) and language in two separate samples (a 24-month and a 36-month outcome sample). Intrusiveness was a significant moderator of the relation between RJA and 24-month language. RJA was a significant predictor of 24-month receptive and expressive language only in children whose caregivers had a low to moderate level of intrusiveness. Understanding the child and caregiver factors that promote or hinder children's language outcomes in children at risk for delay may help inform and target interventions that will help improve children's school readiness outcomes.

Prenatal cocaine exposure has been associated with behavior problems at school age. However, the correspondence between use of cocaine and alcohol during pregnancy is often high, making appropriate allocation of variance and control for other exposures and their interactions difficult. Additionally, gender-specific effects are not typically reported. The purpose of the current study was to determine the degree to which gender-specific effects of prenatal cocaine exposure on teacher-reported child externalizing behavior problems were evident when evaluated in relation to prenatal alcohol exposure. Subjects were singleton infants of mothers who were prospectively evaluated during pregnancy. At age seven, 499 children (214 exposed prenatally to cocaine) were evaluated in our laboratory and teacher reports were solicited. Analyses stratified by gender and prenatal alcohol exposure status, and controlled for significant pre- and postnatal confounders, revealed that among boys with prenatal alcohol exposure, those with persistent cocaine exposure throughout pregnancy had significantly higher levels of Delinquent Behavior compared to boys with no cocaine exposure. Boys with any prenatal cocaine exposure were twice as likely as unexposed boys to have clinically significant Externalizing Behavior scores. However, no association was found between prenatal cocaine exposure and scores on Externalizing Behavior and specific syndromes for boys with no prenatal alcohol exposure. Among girls with no prenatal alcohol exposure, those with persistent cocaine exposure had significantly higher levels of Externalizing Behaviors and Aggressive Behaviors compared to girls with no prenatal cocaine exposure after control for confounding, and were almost five times as likely to have clinically significant Externalizing Behavior scores. However, for girls with prenatal alcohol exposure, no association between prenatal cocaine exposure and scores on Externalizing Behavior and specific syndromes was found after control for confounding. The current findings support gender- and alcohol-moderated effects of prenatal cocaine exposure on school-age teacher-reported child behavior problems. These findings are similar to what we have reported for independent parent-reported behavioral evaluation.

The current study examined the development of responding to joint attention (RJA), a prelinguistic skill, in a sample of children prenatally cocaine exposed. The sample used was part of a larger population of children randomly assigned to three levels of intervention. The growth of RJA in the current sample was best characterized by two linear growth groups determined by a semi-parametric growth modeling program. Each trajectory group was differentially associated with three language outcomes. Gender, treatment group, and birthweight were three risk factors that influenced the likelihood of belonging to either growth cluster. RJA?s predictive significance in terms of concurrent and subsequent language was also established, accounting for the variance associated with contemporaneous measures of cognition. The findings (regarding the relationship between RJA and language) were consistent with previous research examining joint attention behaviors in other types of samples. Additionally, this study contributed uniquely to the body of research on joint attention by exploring the growth of RJA, a precursor of language, in a sample of children at risk for language impairment.

Mutual regulation of attention was investigated in a group of prenatally cocaine-exposed and non-exposed mother-infant dyads during a 5-minute videotaped free play session. Mutual regulation was measured using a state-based coding scheme designed to categorize dyadic interactions into three mutually exclusive and exhaustive states: maternal bid, mutual engagement, and non-involved. Results revealed no significant differences between cocaine-exposed and non-exposed dyads in overall amount of mutual engagement displayed. Cocaine-exposed dyads exhibited significantly longer mutual engagement episodes. Mothers in the two groups did not differ in the number or quality of bids for mutual engagement, and infants in both groups were equally responsive to maternal bids. No ecological variables were found to predict mutual engagement.

Background: Pregnancy substance use is linked to low birth weight. However, less is known about relative contributions of various substances and whether effects are due to decreased gestational duration, restriction of fetal growth, or both. The study goal was to use causal modeling to evaluate the individual impact of alcohol, tobacco, cocaine, and marijuana on gestational duration and fetal growth. Methods: Participants were 3164 urban black women recruited at entry to prenatal care and followed to delivery, with all gestational dating ultrasound supported. Pregnancy substance use was assessed via self-report (alcohol, tobacco, cocaine, and marijuana). Results: Alcohol, cigarette, and cocaine use were all individually and negatively related to gestational age at delivery. However, only alcohol, cigarette, and marijuana use predicted fetal growth, with effects for alcohol and cigarette greater and more discrepant for older women. Overall, heavy cigarette smoking had the greatest individual impact on birth weight (up to 431 g). Heavy levels of use of all 4 substances by older women decreased birth weight by 26% (806 g). Conclusions: For perhaps the first time, reduced birth weight is apportioned both by type of substance and mechanism of effect. The use of alcohol and/or cigarettes was clearly more harmful to fetal growth than cocaine use. Findings demonstrate the need for continued emphasis on intervention efforts to address legal and illicit pregnancy substance use.

Although the hazards of prenatal exposure to addictive substances have been documented for decades, it continues to be a prevalent social and health concern today. Alcohol and cocaine are two commonly abused substances during pregnancy, often leading to behavioral and cognitive disorders in exposed children. At present, the relationship between teratogenic effects of prenatal alcohol exposure (PAE) and prenatal cocaine exposure (PCE) on the brain and observed behavioral outcomes is still unclear. A primary reason for this incomplete understanding is the lack of information regarding neuronal functioning in these populations. Functional MRI, which measures real-time brain activation in response to certain stimuli, can be utilized to bridge the gap between known structural damage and observed behavioral outcomes. This thesis aims to examine structural and functional alterations in PAE and PCE populations as compared to unexposed, socio-economic status-matched populations. As the PAE population is highly affected by structural dysmorphology, the applicability of a newly developed diffeomorphic image registration method to this population is examined. Additionally, task-positive and task-negative functional connectivity and activity are investigated in the PAE population, and related to underlying structural alterations. Neural correlates of global arousal and emotional regulation are investigated in the PCE population, as these behavioral outcomes are most notable. Similarly, functional connectivity and activation in task-positive and task-negative networks, as well as correlated structural measures, are examined in the PCE population. The diffeomorphic image registration algorithm was found to improve both structural and functional image registration for the PAE population. In the examination of specific deficits in arithmetic processing, poorer performance in the PAE group was attributed to a multi-level effect produced by altered structural and functional connectivity and functional activity in calculation and default mode networks. Baseline arousal levels were found to be higher in adolescents with PCE as compared to healthy controls (by altered default mode network functioning); emotional regulation also appeared to be affected in the PCE group by a prefrontal-amygdala structural and functional disconnect. The findings of this thesis give insights into the relationship between task-positive and task-negative duality and cognitive impairment, and contribute to a more comprehensive understanding of the spectrum of clinical disorders caused by prenatal exposure to alcohol and cocaine.