|Home | About | Journals | Submit | Contact Us | Français|
Previous research found that prenatal cocaine exposure (PCE) may increase children's vulnerability to behavior and cognition problems. Maturational changes in brain and social development make adolescence an ideal time to reexamine associations. The objective was to conduct a systematic review of published studies examining associations between PCE and adolescent development (behavior, cognition/school outcomes, physiologic responses, and brain morphology/functioning).
Articles were obtained from PubMed, PsycInfo, Web of Science, and CINAHL databases through July 2012 with search terms: prenatal drug, substance, or cocaine exposure; adolescence/adolescent; and in utero substance/drug exposure. Criteria for inclusion were nonexposed comparison group, human adolescents aged 11 to 19, peer-reviewed, English-language, and adolescent outcomes.
Twenty-seven studies representing 9 cohorts met the criteria. Four outcome categories were identified: behavior, cognition/school performance, brain structure/function, and physiologic responses. Eleven examined behavior; 7 found small but significant differences favoring nonexposed adolescents, with small effect sizes. Eight examined cognition/school performance; 6 reported significantly lower scores on language and memory tasks among adolescents with PCE, with varying effect sizes varied. Eight examined brain structure/function and reported morphologic differences with few functional differences. Three examined physiologic responses with discordant findings. Most studies controlled for other prenatal exposures, caregiving environment, and violence exposure; few examined mechanisms.
Consistent with findings among younger children, PCE increases the risk for small but significantly less favorable adolescent functioning. Although the clinical importance of differences is often unknown, the caregiving environment and violence exposure pose additional threats. Future research should investigate mechanisms linking PCE with adolescent functioning.
Prenatal drug exposure is a public health concern, with estimates of 4.4% of pregnant women aged 15 to 44 using illicit substances during pregnancy.1 Prenatal cocaine exposure (PCE) has received national attention since the presumed “crack baby epidemic” of the 1980s. PCE may act through both biological and environmental mechanisms. Not only does it cross the blood-brain barrier to the fetus and act as a stimulant by blocking the reuptake of dopamine, serotonin (5-HT), and norepinephrine,2 but the drug-using environment, often characterized by poverty, violence, a lifestyle of acquiring and using drugs, can undermine child development.3–5
Several reviews and meta-analyses have examined PCE through early childhood. Prenatal growth is often restricted in utero, resulting in shortened average gestation and lower birth weight.6,7 Compared with nonexposed children, children with PCE may demonstrate less optimal motor performance, reflexes, and orientation through 3 to 4 weeks postpartum8 and lower scores on measures of neurobehavioral functioning,9 attention and emotional expressivity,10 and speech and language development through ages 4 to 6 years.11 Effect sizes are often small, particularly after covariates, including prenatal alcohol/tobacco exposure and family/community factors, are introduced.11,12
A review on school-age children with PCE found negative associations with sustained attention and behavioral self-regulation but few associations with cognition, language, or growth.13 Another review through age 13 years found associations with behavior problems, attention, language, and cognition.14 Little is known about PCE and adolescent functioning.14 Adolescence is a transitional period characterized by increased social and cognitive demands, together with increasingly complex executive functioning abilities (eg, abstract reasoning).15 It is plausible that the small effects found in attention and emotion regulation among children with PCE early in life may increase during adolescence.9 Although many adolescents are capable of complex reasoning, their decisions are often influenced by psychosocial factors (eg, impulsivity), putting them at risk for adverse developmental outcomes.16 Adolescence is also a time of neural rewiring and potential recovery from earlier neural assaults (ie, PCE).17 This review was designed to examine how PCE relates to adolescent behavior, cognition/school performance, physiologic responses (cardiovascular or hypothalamic pituitary adrenal [HPA] axis), and brain morphology and function using neuroimaging.
Criteria used to select articles included (1) PCE as the primary independent variable; (2) nonexposed comparison group; (3) peer-reviewed, English-language; (4) adolescent (ages 11–19) outcomes; and (5) published in print by July 2012. Eight studies included both prenatal cocaine and opiate exposure,18–25 recognizing their frequent co-occurrence.26,27
Articles were obtained from PubMed, PsycInfo, Web of Science, and CINAHL databases with the following search terms: prenatal drug, substance, or cocaine exposure, adolescence/adolescent, and in utero substance, drug, or cocaine exposure. No additional articles were identified after searching references of selected articles.
Each article was reviewed independently by 2 authors who documented (1) sample composition, (2) drug exposure (eg, definition, frequency), (3) design, (4) independent and outcome variables, (5) covariates, (6) results, and (7) adolescent outcome (eg, behavioral, cognitive/school, physiologic responses, and brain morphology and function using neuroimaging). A third author compiled the independent reviews and investigated discrepancies. Discrepancies were discussed among authors until agreement was reached.
Effect sizes were used to compare significant associations across studies. If effect sizes were not reported, we contacted the authors either to confirm our calculations based on existing data or to request data to calculate effect sizes. Because the analyses yielded multiple types of effect sizes (eg, η2, f2, d, etc.), we reported numeric and categorical effect sizes (eg, small [0.20–0.30], medium [0.50], large [0.80]) based on standard criteria (Tables 1, ,2,2, ,3,3, and and44).28–31
To examine the clinical importance of the findings for behavior and cognition/school performance, we compared the scores with standardized scores, when available. Scores within 0.5 SD of the standardized mean were within normal limits (WNL), scores 0.5 to 1.0 SD above the mean were high-WNL, scores 0.5 to 1.0 SD below the standardized mean were low-WNL, and lower scores were deficient. Data were extracted only for outcomes of interest and discussed only if scores were high-WNL, low-WNL or deficient.
We identified 27 articles that met selection criteria representing 9 unique cohorts (Tables 1, ,2,2, ,3,3, and and4).4). Covariates ranged from none to >20 and were organized into 7 categories: violence exposure, caregiving environment, caregiver characteristics, youth characteristics, birth factors, youth behaviors, and study factors (Tables 1, ,2,2, ,3,3, and and4).4). Covariate categories indicate that the study included at least one of the variables listed in that category or a variable similar to it (see footnotes to Tables 1, ,2,2, ,3,3, and and4).4). Prenatal exposure to tobacco, alcohol, and marijuana was cataloged (Table 5). All studies included a non-cocaine exposure (NCE) comparison group, matched on individual and environmental characteristics, with frequent prenatal exposure to alcohol, tobacco, and/or marijuana.
All articles represented low-income, predominantly African American samples. With the exception of 1 rural sample,32 all samples were urban. Most studies presented longitudinal data (n = 19); others analyzed the adolescent phase of longitudinal data sets (n = 8). Ages of participants ranged from 11 to 17 years. Several studies examined children aged <11, but only the adolescent results are reported here. All studies defined cocaine exposure through maternal and/or infant toxicology reports (urine, meconium) and/or self-reported prenatal drug use. Definitions differed across studies (Tables 1, ,2,2, ,3,3, and and4).4). The cohorts were relatively evenly distributed across adolescent outcomes (Fig 1).
Definitions varied across cohorts. Sixteen studies from 7 cohorts defined cocaine exposure dichotomously: “exposed” versus “nonexposed”18,23,24,32–43 or “heavy/persistent” versus “some/none.”21 Six studies from 1 cohort defined “heavy” as first trimester use ≥3 times per week.19,20,22,25,44,45 Four studies from 1 cohort defined “heavy” as at least 61 days of use and the top quartile of meconium metabolites.12,46–48 One cohort calculated a latent variable, based on maternal self-report in each trimester and quantitative and qualitative cocaine metabolite assays.49 Even though many studies identified high versus low PCE, many collapsed variables into any vs. none variables for analyses (see Tables 1, ,2,2, ,3,3, and and4).4). Twelve of the 16 studies that dichotomously defined PCE and 3 of the 11 that defined PCE dosage exposure reported significant findings.
The most widely used covariates were other prenatal exposures (alcohol, tobacco, and marijuana), the caregiving environment, and violence. Almost all studies considered ≥1 other prenatal drug exposure (N = 23), but measurements differed (Table 5).
The caregiving environment (examined in 20 studies) included caregiver stability, quality of the home environment, caregiver-child relationships, postnatal caregiver drug use, and caregiver depression/psychopathology (Table 6). Violence was examined in 16 studies and included witnessed/experienced community or domestic violence, sexual/physical abuse, community drug use, and violent friends (Table 6).
Eleven studies described behavioral problems. Nine examined externalizing/internalizing behaviors and/or drug use. Two examined behavior problems in the context of cognitive outcomes35 or physiologic responses34 and are included in both sections (Cognition/School and Physiologic Responses, Table 1).
Eight studies examined externalizing/internalizing behaviors,19–23,34,35,46 with mixed findings (Table 1). Two found no association between PCE and youth-reported behavioral problems,19,46 1 found associations between Stroop errors at 7.5 years and caregiver-reported aggression at 14 years, regardless of PCE status,35 and one found that the PCE group was more likely than the NCE group to respond to a laboratory social provocation paradigm with escape, rather than aggression.21 The remaining 4 found significant differences favoring the NCE group. Heavy PCE was associated with attention problems and externalizing/internalizing behaviors,20 PCE youth had higher teacher- and caregiver-reported externalizing problems,22 girls in the PCE group had more adolescent-reported anxiety than nonexposed girls,34 and anxiety/depression, but not aggression, in middle childhood mediated the association between PCE and adolescent peer victimization.23
Four studies examined adolescent drug use (Table 1).12,18,19,32 One found that PCE was associated with adolescent cocaine use (measured by hair, sweat, and urine).18 Another study found that heavy PCE was associated with youth-reported marijuana and alcohol use.12 The other 2, among younger adolescents, did not find associations between PCE and adolescent cocaine use (measured by hair)32 or any drug use (adolescent reported).19
One study did not mention other prenatal exposures,19 1 examined only alcohol and tobacco,23 and the others examined prenatal tobacco, alcohol, and marijuana exposure (Table 5).12,18,20–22,32,34,35,46 Ten studies examined the caregiving environment and violence exposure, primarily postnatal/current caregiver drug use (N = 7), quality of the home (N = 3), caregiver stability (N = 4), caregiver depression/psychopathology (N = 5), community violence (N = 8), and home violence (N = 7; Table 6). One study found the relationship between heavy PCE and behavioral problems was mediated by caregiver changes.20 Another found that PCE, together with current caregiver cocaine use, caregiver negativity, community violence exposure, and prenatal tobacco exposure, was associated with adolescent cocaine use18; another reported associations between heavy prenatal tobacco exposure (but not PCE) and violence, with delinquent behaviors.46 Five studies found that postnatal caregiver drug use/drugs in the home18,20–22,32 and caregiver psychiatric/depressive symptoms18–20,22,32 were associated with adolescent behavior problems. Another study found an association between community violence and peer victimization.23 Associations were direct and did not interact with PCE.
Seven of the 11 studies reporting behavior had significant findings; 3 reported small effects,18,20,21 3 reported small-moderate effects,12,23,34 and 1 reported small-large effects.22 Based on teacher-reported and caregiver-reported Child Behavior Checklist scores at ages 11 to 13, both exposure groups had high-WNL scores on externalizing, attention, and total problems.22 Means for other behavioral measures were WNL or not reported.
Eight studies addressed cognition/school performance (Table 2).24,35–37,43,45,48,49 Three examined language,36,43,49 4 examined neurocognitive and executive functioning (1 included here and in Neuroimaging),24,35–37 and 1 evaluated school outcomes.45
One study found an enduring relationship between PCE and lower language performance scores.49 Another found atypical language processing among the PCE group.43 The third found no relationship with receptive language.36
Five studies examined neurocognitive and executive functioning, with mixed results (Table 2).24,35–37 Two found no association between PCE and neurocognitive functioning (ie, cognitive control, working memory, reward processing, memory spatial cognition, visual cognition)37 or executive functioning.48 A third found no relations with working memory or inhibitory control, but the PCE group improved at a slower rate and had marginally lower scores than the NCE group on incidental memory tasks.36 Another study found that the PCE group demonstrated poorer performance on 2 memory tasks compared with the NCE group.24 A final study found that PCE group boys improved more slowly in inhibitory control than nonexposed boys and all girls.35
One study reported greater enrollment in an Individualized Education Plan (IEP) among the PCE group.45 Oppositional defiant disorder (ODD) moderated the association, with the PCE/ODD group more likely to have an IEP than youth with only 1 condition (ODD or PCE).45
Six of the 8 studies examined prenatal tobacco, alcohol, and marijuana exposure35–37,45,48,49; 1 included prenatal tobacco and alcohol only.24 Five measured exposures as any versus none (Table 5).24,35–37,45 One study reported negative associations with prenatal marijuana and alcohol exposure and positive associations with prenatal tobacco exposure and executive functioning.48 Another study reported that youth with prenatal marijuana exposure were less likely to have an IEP.45 Five studies examined the caregiving environment, including the quality of the home, caregiver stability, postnatal/current caregiver drug use, and caregiver depression/psychopathology (Table 6).24,36,37,45,49 Two studies found that caregiver depression was not associated with language development36 or neurocognitive function,37 and 1 found that environmental stimulation was associated with neurocognitive function.37 Three studies examined violence exposure,24,45,48 and 2 did not examine the caregiving environment or violence exposure.35,43 Associations were direct and did not interact with PCE.
Memory scores on the Children’s Memory Scales were near the 25th percentile for both exposure groups.24 Language scores were deficient to low-WNL for both groups on the Peabody Picture Vocabulary Test,36,37 Clinical Evaluation of Language Fundamentals,43,49 and Test for Reception of Grammar.37 Mental processing scores were low-WNL to WNL on the Delis-Kaplan Executive Function System.48
Many studies (n = 17) reported IQ scores. PCE group means were WNL in 1 study,42 low-WNL in 8,24,32,35,39–41,45,47 deficient in 4,36–38,49 and either WNL or low-WNL depending on high/low exposure in 2.12,48 NCE group means were WNL in 2 studies,32,42 low-WNL in 10,12,24,35,38–41,45,47,48 and deficient in 3.36,37,49 Two studies reported deficient IQ scores, without differentiating exposure groups.18,21
Eight studies investigated adolescent brain morphology and functioning using neuroimaging (Table 3).24,33,38–42,47 All had small sample sizes (N = 34–56), used few covariates other than other prenatal drug exposures, and had different neuroimaging methods (3 used similar methods).
One study found the PCE group had smaller left, right, and total caudate volumes than the NCE group.33 Another found lower cerebral blood flow intensities in the posterior and inferior brain regions in the PCE group. After adjusting for global cerebral blood flow, cerebral blood flow intensities in anterior and superior brain regions were higher in the PCE group with no differences in total gray or white matter volume or total cerebrospinal fluid volume.41 A third study found no group differences in subregions of the corpus callosum on measures of mean diffusivity or fractional anisotropy.42 One study found that differences in cortical gray matter, parenchymal volume, and head circumference were not significant after adjustment.47
Five neuroimaging studies examined how brain function relates to behavior (Table 3). One found no differences on working memory.38 Another found prefrontal cortex activity suppression in the PCE group, associated with negative emotion.40 A third study found associations between PCE and alterations in the excitatory/inhibitory balancing mechanism.39 In a fourth study, lower fractional anisotropy was not related to sensation seeking in youth with PCE.42 Finally, 1 study found that the PCE group had larger hippocampi than the NCE group, with associations between larger hippocampi and poorer performance on memory tasks.24
Of the 8 studies, 4 included prenatal tobacco, alcohol, and marijuana exposure;39–41,47 2 did not;33,38 1 included only tobacco exposure;42 and 1 included tobacco and alcohol exposure (Table 5).24 One study reported an inverse relation between number of prenatal exposures and head circumference.47 Two studies found that youth with PCE had more caregiver changes and violence exposure (Table 6).39,40 A third study reported that PCE group effects were often diminished after adjusting for the caregiving environment and violence exposure.24
Three studies examined physiologic response variability (Table 4).25,34,44 One found higher cortisol levels before and after stress exposure in the PCE group; PCE group boys had lower diastolic blood pressure than NCE boys, with no difference in girls, suggesting a lower level of cardiovascular arousal.34 There were no exposure or gender differences in heart rate or systolic blood pressure.34 Another study found the PCE group was more likely than the NCE group to have a blunted cortisol response to stress.44 The final study found no group differences in basal or morning cortisol, but the high PCE group demonstrated a blunted cortisol increase from evening to morning and greater likelihood of atypical changes compared with the low PCE and NCE groups.25
All studies examined prenatal tobacco, alcohol, and marijuana exposure and the caregiving environment (Table 5).25,34,44 One examined the quality of the home, postnatal/current caregiver drug use, caregiver stability, caregiver depression/psychopathology;44 1 included caregiver-child relationship quality;34 and the final study included caregiver stability, caregiver depression, postnatal/current caregiver drug use, domestic violence, child abuse, and community violence.25 One study found that caregiver depression was associated with an atypical diurnal cortisol pattern,25 and caregiver-child relationship quality did not moderate relationship between PCE and cortisol response.34 Another study found that the PCE group was more likely than the nonexposed group to demonstrate a blunted cortisol response that was amplified with exposure to domestic violence (Table 6).44
The 27 studies reviewed represent the initial wave of investigations of PCE among adolescents. Previous reviews reported persistent associations with PCE in infancy, childhood, and early adolescence, including small but significant differences in neurobehavioral functioning,9 attention and emotional expressivity,10 and speech and language development.11 This review does not support previous concerns that cognitive and behavior problems might be more prominent during adolescence in response to increased cognitive and social demands.9 Although 70% of the studies (19 of 27) reported significant findings favoring nonexposed youth, effects were small to medium, and scores were WNL, with uncertain clinical importance. Thus, the associations between PCE and multiple areas of development persist through adolescence but do not increase.
Although some studies attempted to quantify PCE, most dichotomized it into any versus no exposure. Definitions of heavy exposure varied across studies, potentially contributing to discrepant findings; however, significant findings appeared to be evenly distributed between dichotomous and dosage definitions. Conclusions regarding the long-term effects of PCE would be strengthened by differential dose relationships.
Covariates varied in number and type, potentially contributing to discrepant findings. Most studies measured prenatal marijuana, tobacco, and/or alcohol exposure, often using direct methods (urine and/or meconium, plus caregiver report). Although measurement strategies differed, prenatal exposures were often associated with negative adolescent outcomes. As with cocaine, clarity in examining dose relationships may increase validity. Eight studies included both cocaine and heroin exposure. However, analyses were rarely adjusted, thus interfering with attribution.
The environmental aspects of PCE have been well recognized.13 Many studies included measures of caregiver stability, postnatal/current caregiver drug use, home quality, and caregiver depression/psychopathology, along with violence exposure. With the exception of the neuroimaging studies, many used the same or comparable measures, often finding that caregiving environment and/or violence were directly related to adolescent functioning, regardless of PCE status.
Covariates were frequently included in analyses in an attempt to isolate the effects of PCE, and analyses often examined the direct effects of PCE. Recent developmental theories recommend focusing on individual differences in response to environmental stressors and opportunities.50 Thus, models that examine the mechanisms linking PCE and the caregiving environment may yield a more comprehensive understanding of PCE and adolescent functioning.
Behavioral problems were the largest area of research with 11 studies. Although almost two-thirds (7/11) found associations between PCE and behavior problems, across multiple sources (adolescent-, caregiver-, teacher-report, and laboratory paradigms [Table 1]) effect sizes were small with differences WNL, raising concerns about their clinical importance. Two studies that did not find a link with PCE examined illegal behaviors (eg, property damage, violence), based on youth report.19,46 Two studies conducted with youth aged 14 to 16 found a link with adolescent use of psychoactive substances.12,18 Findings suggest the need for additional investigation between PCE and nonillegal behaviors and among older adolescents on their vulnerability to substance use.
Although 74% (6/8) of cognition/school studies had significant findings, results lacked specificity, because most areas (language, memory, neurocognitive functioning, executive control, and inhibitory control) were investigated by only 1 or 2 studies. The most consistent findings were among language and memory tasks.24,35,36,43,49 Cognition/school performance warrants additional investigation with models that incorporate indirect effects and environmental conditions, as well as attention to the clinical importance of the findings. Several studies reported means with either the PCE group or both groups low-WNL. Almost all reported low-WNL or deficient IQ scores for both exposure groups, suggesting the negative effects of poverty.
Most of the 8 studies examining brain morphology using neuroimaging found significant differences; however, morphologic differences rarely translated to behavioral differences. Thus, the significance of the morphologic differences is unknown.
Effect sizes tended to be medium to large, and neuroimaging studies had small sample sizes, often with limited covariate control, little standardization in methods, and infrequent replication. Neuroimaging is an emerging area in the investigation of PCE and next steps include (1) using larger sample sizes, (2) other prenatal drug exposures, (3) direct and indirect relations with environmental conditions, (4) relations with behavioral functioning, and (5) replication.
The relationship between PCE and the neuroendocrine system is an emerging topic of theoretical interest.25,34,44,51 Although the 3 studies that examined associations between PCE and the HPA system through the cortisol response had inconsistent findings, the HPA axis was altered, suggesting additional research studies to understand how PCE relates to cortisol reactivity and developmental variability.
Several methodologic considerations limit interpretations. First, the 27 studies relied on 9 cohorts, potentially limiting generalizability. However, the studies were distributed evenly across the 4 outcomes of interest, indicating that data from multiple cohorts informed the conclusions. Second, authors reported statistical significance but little information on clinical importance. Including the percentage of children with clinically relevant findings would increase our understanding of the implications of PCE. Third, although there are examples of research in the 4 areas we identified (behavior, cognitive performance, neuroimaging, and physiologic response), there has been limited standardization regarding methods or measures, making it difficult to interpret discordant finding. Fourth, there is a risk of publication bias against null findings and selective reporting. Finally, because PCE includes both substance exposure and environmental exposure to a drug-using lifestyle, comparison groups are difficult to define. Although analytical models may adjust for environmental differences, investigations of mechanisms and synergistic effects are often overlooked. For example, responsive caregiving and low stress may alter brain structure52 and functioning,53 thereby potentially protecting children from the negative consequences of PCE.
Compared with previous reviews, PCE continues to be associated with elevated behavioral problems, particularly externalizing problems.13,14 A promising area of behavioral research is adolescent drug use, particularly as youth age and are exposed to drug-using opportunities. Although the language vulnerability reported in previous reviews has continued into adolescence,11,14 effect sizes are small, raising concerns about their clinical importance. Executive functioning and neurocognition, identified in previous work as vulnerable areas for children with PCE,9,14 had few associations in adolescence; however, findings were mixed.
On the basis of this review, we identified 6 areas for future research. First, in addition to examining the direct effects of PCE and adjusting for potential confounders, investigators could use recent developmental theories50 to examine the indirect effects of PCE on adolescent functioning. Second, most PCE/NCE differences are small and of unknown clinical importance. Understanding the clinical implications of differences (eg, the need for special educational services) and their response to intervention would be helpful. Third, although parent-child relationships related to PCE were an important area of research during early childhood, there has been little research on parent-child relationships during school-age and adolescence, including their potential ability to mitigate the negative effects of PCE on child and adolescent functioning.14 Fourth, only 3 studies considered adolescent internalizing problems. Because internalizing problems can be a precursor to subsequent mental health problems, additional attention is warranted.20 Fifth, little is known about the effects of the timing and dose of PCE. Although several investigators attempted to differentiate between heavy and light prenatal exposure, definitions differed, and there were few findings. Future research is needed to identify precise measures of the timing and dose of PCE. Finally, longitudinal investigations of patterns of strengths and weaknesses among children, adolescents, and adults with PCE would aid in our understanding of resilience and developmental consistency.
In conclusion, the field of PCE has advanced significantly since the misleading “crack baby” scare of the 1980s. Investigators have recognized (1) that the physiologic disruptions associated with PCE do not necessarily lead to behavioral or cognitive deficits, (2) that the impact of PCE should be considered in the context of other prenatal and environmental exposures, and (3) that understanding life span development may depend on the relationships that physiologic responses and brain morphology/functioning play in response to stressful events, such as PCE, along with environmental and caregiving threats and opportunities.
Children in longitudinal research cohorts have reached adolescence, enabling us to examine long-term associations with PCE. As in earlier reviews, findings are often mixed, with evidence of small, but significant associations with development. However, there has not been a strong emergence of clinically relevant findings during adolescence, perhaps because development is influenced by multiple biological and environmental factors. What is evident are the pervasive effects of poverty on both PCE and nonexposed youth.5 Many adolescents, regardless of their prenatal exposures, obtained scores on standardized measures at the low end of the normal range, possibly increasing their risk for future challenges across the life span. A better understanding of the mechanisms linking early stressors, including both poverty and PCE, with subsequent functioning will benefit the field of PCE and enhance our understanding of the determinants of adolescent development.
We thank Samantha P. Bento for assistance with editing and reviewing the manuscript and the anonymous reviewers for their comprehensive comments and excellent suggestions.
Dr Buckingham-Howes was responsible for idea formation and conception, literature searches, reading articles, abstracting information from articles, drawing parallels and finding differences between conclusions in studies, and the majority of manuscript writing; Dr Berger was responsible for idea formation and conception, literature searches, reading articles, abstracting information from articles, and some manuscript writing; Ms Scaletti was responsible for reading articles, abstracting information from articles, reconciling disagreements, data analysis, some manuscript writing, and providing valuable feedback during the manuscript writing process; and Dr Black was responsible for obtaining funding, idea formation and conception, some manuscript writing, and providing valuable guidance and feedback during the manuscript writing process.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: This study was supported by the National Institute on Drug Abuse (R01-DA07432 and R01-DA021059). Funded by the National Institutes of Health (NIH).