Parental substance use history density and its influence on reward anticipation brain activation in late childhood and early adolescence

Abstract

Parental history of problematic substance use (PH) increases the risk for early adolescent substance use (SU), potentially due to premorbid differences in reward-processing brain regions (e.g., striatum). However, no studies have prospectively examined the separate contributions of parental history of alcohol (PHA) and drug (PHD) use or the impact of PH density (PH⁰, PH¹, PH²) on reward processing in preadolescents. This study analyzed data from 10,235 participants (ages 9–14) in the Adolescent Brain Cognitive Development Study^SM (ABCD). Reward processing was assessed using the Monetary Incentive Delay Task (MID) at baseline and two-year follow-up. Regions of interest included bilateral striatal activation elicited by neutral vs. anticipation of large rewards. Linear mixed-effect models evaluated PH, PHA, PHD, and PH density on ROI activation, controlling for relevant covariates. Results showed that youth with PHA¹ had greater nucleus accumbens activation during reward anticipation than those with no history (PHA⁰), but no significant differences were found between PHA² and PHA¹ or PHA² and PHA⁰. PHD and PH were not significantly associated with BOLD activation in striatal regions, nor were there changes over time. These findings highlight the need to consider both PH and environmental factors when assessing neurodevelopmental risk for early substance use.

Highlights

• •Parental substance use history impacts striatal activation during reward processing.
• •Parental substance use density and drug type affect youth reward-related activity.
• •Youths with one parent with problematic alcohol use show greater nucleus accumbens activation.
• •Youths with two parents with substance use issues report greater reward sensitivity.
• •Parental and environmental factors are linked with striatal BOLD response.

1. Introductions

Adolescence is a critical period characterized by neurodevelopmental changes and increased risk-taking behaviors, including substance use (SU) (D’Amico et al., 2005). In the U.S., 15.1 % of 8th graders reported alcohol use, while 10.9 % reported illicit drug use within the past year (Miech et al., 2024). Early SU initiation robustly predicts the development of substance use disorders (SUDs); however, a positive parental history of problematic SU (PH) further increases risk of SU initiation and escalation (Biederman et al., 2000, Lisdahl et al., 2021, Rusby et al., 2018). Specifically, adolescents with PH are at 3x greater risk for developing a SUD, suggesting that adolescents may be particularly sensitive to parental SU exposure (Biederman et al., 2000). This risk is likely heightened through a complex interplay between genetic risks affecting neurobiology (Hines et al., 2015), and exposure to environmental factors, including parental relationships (Smith et al., 2016), parenting behaviors (e.g., monitoring and warmth) (Arria et al., 2012, Donaldson et al., 2023, Rusby et al., 2018), prenatal substance use exposure (Minnes et al., 2017, Smith et al., 2016), and family SU modeling (Ennett et al., 2008).

One proposed neurobiological mechanism linking PH to adolescent SU suggests that adolescents with PH may exhibit differential development of reward processing prior to significant SU exposure, increasing engagement in risky, reward-driven behaviors like SU (Cope et al., 2019, Cservenka, 2016). Reward processing includes, but is not limited to, reward responsiveness and learning, anticipation of future rewards, and engaging in goal-directed behavior to obtain these rewards (Berridge and Robinson, 2003). Reward anticipation of future gains and losses is often studied using the Monetary Incentive Delay (MID) task (Knutson et al., 2000, Casey et al., 2018). During reward anticipation, the ventral striatum processes the expected values of rewards, while the dorsal striatum is involved in decision-making and selecting actions based on reward value (Balleine et al., 2007, Williams et al., 2021). Neuroimaging studies have identified key brain regions involved in reward anticipation, including the bilateral ventral striatum (e.g., nucleus accumbens (NAcc)) and dorsal striatum (e.g., caudate and putamen), which have been consistently observed in both children and adults (Knutson et al., 2000, Oldham et al., 2018, Bjork et al., 2004, Cao et al., 2019). In adults, meta-analyses show that individuals with substance use or gambling disorders exhibit reduced striatal activation during reward anticipation tasks compared to healthy controls (Luijten et al., 2017). While reward anticipation brain activation patterns are well-documented in adults, it remains unclear whether this pattern extends to adolescents, as ongoing neurodevelopment and the confounding effects of co-occurring substance use complicate interpretation.

During adolescence, the striatum undergoes significant developmental changes following an inverted U-shaped trajectory (Urošević et al., 2012). This pattern further varies by sex, with the caudate and putamen reaching peak volume earlier in females (around age 12) than in males (around age 14) (Raznahan et al., 2014). This developmental transition is also characterized by increases in striatal dopamine levels and receptor densities – dopamine type 1 (D1) and dopamine type 2 (D2) – from childhood into early adolescence (Ernst and Luciana, 2015). These receptors are of great interest in addiction neuroscience as they are closely associated with the rewarding effects of substances (Volkow and Morales, 2015). Individual variations in dopamine genes (e.g., DRD1, DRD2, DRD4, COMT) can further influence dopamine signaling, affecting risk for substance use (Blum et al., 2013). Additionally, pubertal development—specifically higher testosterone levels in preadolescents and adolescents—has been positively linked to increased reward-related BOLD activation in the striatum. This relationship has been demonstrated in both longitudinal (Braams et al., 2015; n = 299, ages 8–27) and cross-sectional studies (Forbes et al., 2010; n = 50, ages 10–16), suggesting that pubertal hormones shape reward sensitivity and susceptibility to risk-taking behaviors in youth.

Research has proposed two developmental theories regarding the adolescent pattern of reward anticipation. The hyposensitivity theory posits that healthy adolescents may exhibit diminished reward processing, resulting in low baseline activation in reward-related brain regions. As a result, adolescents may seek more intense or frequent rewards, including substances, to achieve similar levels of activation and reward experience compared to adults (Blum et al., 2000). In contrast, the prevailing hypersensitivity model argues that adolescents experience heightened reward sensitivity due to increased striatal activation, driving greater risk-taking and substance-seeking behaviors (Berridge and Robinson, 2003). Reward sensitivity can be operationalized as increased behavioral motivation to pursue rewards and heightened arousal in response to them (Galvan, 2013). Supporting this model, neuroimaging studies have shown greater BOLD activation in healthy adolescents compared to adults and children in the ventral striatum and mPFC (Galvan et al., 2006, Van Leijenhorst et al., 2010). This heightened sensitivity may be linked to increased dopamine release in the ventral striatum elicited by rewarding events (Jonasson et al., 2014, Koepp et al., 1998).

Further supporting the hypersensitivity model, Cope et al. (2019) found that in substance-naive children (M = 10.5 yrs.; SD = 1.2 yrs.), greater BOLD activation in the NAcc during reward anticipation on the MID task was linked to earlier SU initiation. Similarly, in an older adolescent sample (M = 15.3 yrs.; SD = 1.06 yrs), youth with greater BOLD activation in the caudate and putamen demonstrated greater odds of SU initiation at one-year follow-up during reward anticipation (Stice et al., 2013). Overall, these findings suggest that while adolescents, as a group, tend to exhibit a hyperactive reward response compared to children and adults, individual variations in reward sensitivity may contribute to differential risk for early substance use initiation. Specifically, adolescents with heightened reward sensitivity may be more prone to seeking out rewarding experiences, including substance use.

One key factor influencing individual differences in reward anticipation is a parental history of problematic substance use (PH), which may shape adolescents' reward-related neural responses and increase susceptibility for risk-taking behaviors. Several studies suggest that PH is associated with altered striatal responses during reward anticipation, particularly in key regions such as the NAcc and putamen. Cross-sectional studies have found that preadolescents (ages 9–10) and adolescents (ages 14–17) with PH display increased BOLD activation in the NAcc and putamen when anticipating rewards on the MID task (Kwarteng et al., 2021, Stice and Yokum, 2014). Supporting this, a meta-analysis by Tervo-Clemmens et al. (2020) found that adolescents at risk for problematic substance use –defined by either a family history (FH) of substance use disorder or neural activation patterns predictive of future use– consistently exhibited striatal hyperactivation, particularly in the putamen. However, this effect was largely driven by studies that included youth with externalizing psychopathology, suggesting that heightened striatal activation may reflect a general liability for behavioral dysregulation rather than a risk pathway specific to substance use.

Extending these findings into young adulthood, research has shown that individuals with a parental history of alcohol use (PHA) who reported early drunkenness (by age 15) exhibited heightened striatal dopamine release in response to monetary reward, compared to low-risk PHA individuals and controls. Dopamine release was negatively correlated with age of first drunkenness and positively associated with neural response to reward feedback; however, no group-level differences in BOLD activation were observed (Weiland et al., 2017). These findings suggest increased dopaminergic responsivity may serve as a neurobiological risk marker in PHA individuals. In contrast, other studies have found that adults with a PHA who did not develop substance use problems exhibited blunted NAcc activation during reward anticipation, a pattern which may reflect resilience rather than vulnerability (Andrews et al., 2011, Yau et al., 2012). Meanwhile, some studies report no significant differences in reward anticipation on the MID task between adolescents (Müller et al., 2015, Bjork et al., 2008) or adults (Musial et al., 2023) with a PH and those without (PH⁰). Taken together, these findings suggest that adolescence may be a critical period of heightened striatal sensitivity for individuals with PH.

Beyond binary classifications of PH, recent work has examined the impact of parental substance use density. One potential issue is the tendency to dichotomize PH as either present or absent; this simplified distinction may fail to capture the nuanced relationships between parental influence and adolescent SU or bi-directional patterns (i.e., different density patterns linked with hypo- or hyper-sensitivity). For example, Martz et al. (2022) conducted cross-sectional analyses to investigate the relationship between parental density of problematic alcohol use (PHA) and NAcc BOLD activation during reward anticipation using the MID task. Their study included two samples: baseline data from the Adolescent Brain Cognitive Development (ABCD) Study (ages 9 – 10) and the Michigan Longitudinal Study (MLS) (ages 7 – 15). Across both studies, youth with two parents with a history of problematic alcohol use (PHA²) demonstrated reduced right NAcc activation compared to those with one affected parent (PHA¹). However, in the MLS sample specifically, PHA² youth also showed significantly reduced right NAcc activation relative to those without a parental history of problematic alcohol use (PHA⁰). While the difference between PHA¹ and PHA⁰ did not reach statistical significance, there was a trend indicating that PHA¹ youth showed relatively greater BOLD activation across the right and left NAcc in both study samples. These findings suggest a potential unique vulnerability, with findings consistent with both hyperactivity and hypoactivity depending upon PH density, suggesting that the binary categorization of PH may obscure nuanced differences in neuronal mechanisms underlying reward processing.

Other potential factors may explain inconsistencies in the literature, such as sample size, demographics (i.e., age and sex), operationalization of PH, inclusion/exclusion criteria, and consideration of potential confounding factors (Bjork et al., 2008, Kwarteng et al., 2021, Martz et al., 2022, Müller et al., 2015, Stice and Yokum, 2014). Most research has focused exclusively on PHA (Martz et al., 2022, Müller et al., 2015, Bjork et al., 2008), making it unclear whether findings extend across other substances or if observed patterns are specific to PHA. In one study examining differences as a function of substance type, youth (ages 9–10) with PHA demonstrated increased activation in the right NAcc during reward anticipation compared to youth without PH of alcohol (PHA⁰), while youth with a PH of drug (PHD) exhibited increased left putamen activation during the same condition compared to youth without a PH of drug (PHD⁰; Kwarteng et al. 2021). Moreover, who is included in “parental” history varies; some studies focus exclusively on biological parents (Bjork et al., 2008, Kwarteng et al., 2021, Martz et al., 2022, Stice and Yokum, 2014), while others include extended family members (Müller et al., 2015, Musial et al., 2023, Andrews et al., 2011). Finally, factors – including externalizing and internalizing pathology, pubertal maturation, parental monitoring, parental warmth, and prenatal SU exposure – are independently associated with both striatal BOLD activation during reward anticipation (Forbes et al., 2010, Gatzke-Kopp et al., 2009, Tang et al., 2022, Tervo-Clemmens et al., 2020, Casement et al., 2014, Müller et al., 2013) and PH (Kelley et al., 2017, Arria et al., 2012, Barnow et al., 2002, Bountress and Chassin, 2015). Preclinical studies further highlight these relationships, showing that adolescent animals exposed to prenatal SU exhibit heightened reward-driven behavior, reduced ventral striatal D2 messenger RNA expression, and broader alterations in reward circuitry (DiNieri et al., 2011, Malanga and Kosofsky, 2003, Abu and Roy, 2021). Notably, prior studies generally have not accounted for these comorbid factors (Bjork et al., 2008, Kwarteng et al., 2021, Martz et al., 2022, Müller et al., 2015, Stice and Yokum, 2014). Thus, it is critical to account for the unique effects of these factors when examining reward development.

1.1. Current study

The current study aimed to address these gaps by investigating whether the density (i.e., PH⁰, PH¹, or PH²) of PH was linked to neural activation patterns in the striatum (i.e., caudate, putamen, and NAcc) elicited by the reward anticipation contrast (anticipation of large reward > neutral) during the MID task, after accounting for additional environmental SU and psychopathology covariates. Additionally, we assessed whether neural activation patterns differed based on the type of parental SU history: overall SU (alcohol + drug; PH), alcohol-specific (PHA), or drug-specific (PHD). We also evaluated whether these patterns persisted as youth transitioned from late childhood to early adolescence (ages 9–14). To contextualize and aid in the interpretation of the directionality of these findings, we conducted further analyses to examine PH group differences in behavioral markers of reward sensitivity and motivation. The sample was drawn from a large, diverse, and generally substance-naïve (Lisdahl et al., 2021, Sullivan et al., 2022) cohort as they aged from late childhood into early adolescence.

Based on prior findings from Martz et al. (2022), we hypothesized that parental density (PH) would be associated with specific patterns of reward-related brain activation:

• 1.Youth with two affected parents (PH²/PHA²/PHD²) would demonstrate reduced anticipatory reward-related BOLD activation compared to those with one (PH¹/PHA¹/PHD¹) or no affected parents (PH⁰/PHA⁰/PHD⁰).
• 2.Youth with one affected parent (PH¹/PHA¹/PHD¹) would show greater BOLD activation than both those with no affected parents (PH⁰/PHA⁰/PHD⁰) and those with two affected parents (PH²/PHA²/PHD²).

Drawing from prior research (Kwarteng et al., 2021, Martz et al., 2022, Stice and Yokum, 2014), we expected these activation patterns to be evident in the bilateral nucleus accumbens for both PHA and PH analyses. Given the evidence that PHA and PHD exhibit distinct neural activation patterns, we anticipated that this BOLD response pattern would be found bilaterally in the putamen for PHD analyses. Based on the age range of the current sample, we hypothesized that the influence of PH on reward processing would remain consistent throughout the transition from late childhood to preadolescence. We also hypothesized that all youth with PH would demonstrate higher self-reported reward-seeking behaviors compared to youth without a parental history of substance use (PH⁰).

2. Methods

2.1. Participants and study overview

The current sample is drawn from the Adolescent Brain Cognitive Development (ABCD) Study®, a large-scale longitudinal study designed to examine brain development in children across the United States. The study recruited 11,876 children ages 9–10 from 2016 to 2018 (Garavan et al., 2018, Volkow et al., 2017) (data release 4.0; DOI:10.15154/1523041; https://nda.nih.gov/abcd). Caregivers and youth attended one to two sessions to complete a comprehensive battery of questionnaires, and youth completed magnetic resonance imaging (MRI) at baseline and two-year follow-up (2YR FU). The UCSD IRB approved all multi-site study procedures. Participants were excluded if they did not meet baseline ABCD study criteria (Garavan et al., 2018) or if they had missing PH data (n = 611) or were missing fMRI data at baseline or the 2-year follow-up (n = 970).

2.2. Measures

2.2.1. Parental history of alcohol & drug problems

The Family History Assessment was administered to caregivers at baseline to evaluate psychopathology and SU within the family (Rice et al., 1995); caregivers were asked whether either biological parent of their child has had problems related to alcohol and/or drugs (e.g., marital separation/divorce, work problems, arrests/DUIs, suspended or expelled from school two+ times; harmed health; in treatment; caused arguments or were intoxicated often). Youth were categorized as having 0–2 parents with problematic alcohol use (PHA⁰, PHA¹, PHA²), drug use (PHD⁰, PHD¹, PHD²), or any alcohol or drug use (PH⁰, PH¹, PH²).

2.2.2. Covariates

Sociodemographics. Demographic information, including sex assigned at birth, race, and ethnicity, was reported at baseline, while caregiver education attainment, household income, visit, and age were measured at baseline and 2YR FU, based on caregiver report (Barch et al., 2018).

Prenatal Drug Exposure. At baseline, caregivers completed the Developmental History Questionnaire to assess the youth’s biological mother’s SU (alcohol, cannabis, tobacco, cocaine, heroin, and oxycontin) before and after awareness of the pregnancy (Kessler et al., 2009, Merikangas et al., 2009). A dichotomous variable was created to indicate the presence or absence of any prenatal drug exposure.

Parental Warmth. The Child Report of Behavior Inventory [CRPI; (Schaefer, 1965)] was administered to youth at baseline to assess their perception of their caregiver’s warmth, acceptance, and responsiveness (e.g., “my caregiver makes me feel better after talking over my worries with him/her). A total score was computed separately for each caregiver by summing the relevant items, and these scores were then averaged to create an overall parental warmth score.

Parental Monitoring. At baseline and 2YR FU, the Parental Monitoring Scale (PMS) was given to youth to evaluate their perceptions of their primary caregiver’s awareness of their location and who they spend time with (Karoly et al., 2015); a total score across five items was calculated and used in the current analysis.

Pubertal Status. The Pubertal Development Scale (PDS) (Petersen et al., 1988) was used to assess perceived pubertal and menstrual status at baseline and 2YR FU (caregiver reports used due to age) (Huang et al., 2012). A summary variable for both male and female youths was calculated, categorizing youth into five distinct stages of puberty (from ‘pre-puberty’ to ‘post-puberty’).

Youth Externalizing & Internalizing Symptoms. Caregivers completed the Child Behavior Checklist (CBCL) at baseline and 2YR FU which captures youth psychopathology in the last six months (Achenbach, 2009, Barch et al., 2018); total age-corrected internalizing and externalizing T-scores were used here.

2.2.3. Reward processing measures

Behavioral Measurement of Reward Motivation. A modified 20-item PhenX BIS/BAS (Pagliaccio et al., 2016, Carver and White, 1994) assessed sensitivity to reward and punishment. This measure has been associated with reward-driven behaviors, including an increased risk of substance use in preadolescents and adolescents (Shao et al., 2025, Urošević et al., 2015). The BAS subscales measure different aspects of reward motivation: Drive reflects the motivation to pursue goals, Fun Seeking assesses the desire for new, rewarding experiences, and Reward Responsiveness captures positive reactions to receiving rewards. Total scores for each subscale were used in the analyses.

Task-Based fMRI Paradigm.Acquisition. Full T1 scan and fMRI acquisition and scanning parameters for the ABCD study are outlined elsewhere (Casey et al., 2018); scans underwent processing by the Data Analytics, Informatics, and Resource Core (DAIRC) of ABCD to ensure quality and consistency across sites (Hagler et al., 2019).

The Monetary Incentive Delay Task (MID). The MID was developed to measure domains of reward processing, including anticipation and receipt of rewards and losses (Knutson et al., 2000). Each trial starts with a 2000 ms incentive cue, with five possible trial types (Win $0.20, Win $5, Lose $0.20, Lose $5, or $0-no money at stake) (Casey et al., 2018). A jittered anticipation event is presented for 1500 – 4000 ms, and then, a variable target lasting 150 – 500 ms appears, prompting youth to respond. If youth pressed a response button when the target was on the screen, they could win money (win trial), avoid losing money (loss trial), or experience neither a win nor a loss (neutral trial). If they were too fast or too slow, they did not win money (win trial), lost money (loss trial), or neither won nor lost money (neutral trial). The target event is succeeded by a feedback message revealing the trial outcome. The MID task is individualized using the participant’s mean reaction time (RT) performed outside the scanner. This time is used to calibrate the starting response target duration so that each participant would reach a 60 % accuracy rate (Casey et al., 2018). The current study used anticipation of large reward > neutral BOLD contrast, representing reward anticipation. Data was collected from baseline and 2YR FU. Behavioral performance is included for descriptive purposes; measures include average task RT (ms) across trials within each condition (anticipation of large reward and neutral).

fMRI ROI Extraction. The Data Analysis, Informatics, and Resource Core (DAIRC) of ABCD preprocessed all scans to ensure quality and consistency across sites, applying corrections for head motion (Cox, 1996), B0 distortions (Andersson et al., 2003), non-linearity distortions (Jovicich et al., 2006), and between-scan motion for each participant. Task-related activation estimates were generated at the individual subject level using a general linear model (GLM) implemented in AFNI's 3dDeconvolve, and outputs included contrast beta weights (Cox, 1996). Striatal ROIs (putamen, caudate, and NAcc) were created from subcortical and cortical surface-based ROIs using FreeSurfer’s anatomically defined segmentations (Fischl et al., 2002) and average BOLD time beta coefficients and standard errors of main task contrast (anticipation of large reward > neutral contrast) were extracted. To control for head motion beyond ABCD DAIRC preprocessing, participants (n = 13) with fewer than 200 degrees of freedom across runs were excluded per the ABCD Study MRI Quality Control Guidelines(Task-Based Functional Magnetic Resonance Imaging, 2025, February 8). Additionally, participants who did not meet task-specific performance criteria (n = 47) (Hagler et al., 2019, Casey et al., 2018) or had individual BOLD responses exceeding ± 10 standard deviations from the sample mean (n = 10) were excluded.

2.3. Data analyses

Analyses were conducted in R [v4.2.2; (R Core Team.. (2022)]. Chi-square tests and ANOVAs were used to evaluate differences in sociodemographics and covariates across PH groups at baseline and the 2YR FU (Table 1). Linear mixed-effect (LME) models were used to examine the effect of PH group status on brain activation during the anticipation of large rewards > neutral outcomes (lme4 package) (Bates et al., 2015). The primary predictors included PH group (PH, PHA, PHD), time point (baseline and 2YR FU), and their interaction (PH x Time). Covariates included age, sex assigned at birth, caregiver education, household income, prenatal substance exposure, parental warmth, parental monitoring, pubertal status, and externalizing and internalizing symptoms. Race and ethnicity were not included as predictors but were used solely to describe the sample, as the focus was on measuring specific familial environmental factors rather than treating these as proxy variables (Cardenas-Iniguez and Gonzalez, 2024).

Table 1.

Demographic and study characteristics of the sample at baseline and year 2 follow-up (2YR FU) by parental history of overall problematic substance use (PH).

Mean (SD) or %	ABCD Sample at Baseline			ABCD Sample at 2YR FU
	PH0 (n = 7523)	PH1 (n = 1455)	PH2 (n = 330)	PH0 (n = 5334)	PH1 (n = 1042)	PH2 (n = 221)
Age in years	9.93 (0.63)	9.92 (0.62)	9.96 (0.62)	11.94 (0.64)	11.94 (0.64)	11.99 (0.65)
Sex assigned at birth (female)a	48.29 %	50.93 %	48.79 %	45.97 %	48.75 %	49.77 %
Racea,*,^
White Black Asian American Multiracial	67.53 % 12.85 % 2.38 % 15.98 %	62.89 % 14.30 % 0.69 % 20.62 %	58.48 % 14.24 % 0.30 % 26.67 %	69.12 % 12.26 % 1.97 % 15.37 %	63.05 % 14.12 % 0.58 % 21.11 %	60.18 % 12.22 % 0.45 % 26.70 %
Ethnicity (Hispanic - Yes)a,*	19.83 %	23.92 %	16.97 %	18.80 %	21.02 %	17.19 %
Caregiver Education*,^
< HS Diploma HS Diploma/GED Some College Bachelor Postgraduate	4.03 % 7.59 % 21.48 % 27.34 % 39.49 %	5.91 % 11.62 % 37.94 % 22.89 % 21.65 %	5.15 % 13.64 % 48.79 % 16.97 % 14.85 %	3.75 % 7.52 % 20.85 % 27.69 % 40.01 %	5.57 % 11.32 % 36.56 % 23.90 % 21.98 %	4.52 % 14.48 % 49.77 % 15.38 % 15.38 %
Household Income*,^
< 50 K > = 50 K & < 100 K > =100 K	21.71 % 25.92 % 44.70 %	38.90 % 26.46 % 26.60 %	47.27 % 28.18 % 16.67 %	18.67 % 24.77 % 49.64 %	32.74 % 28.41 % 30.04 %	35.29 % 37.56 % 16.29 %
Pubertal Development Status*,^
Pre-Puberty Early-Puberty Mid-Puberty Late-Puberty Post-Puberty	51.35 % 22.96 % 21.02 % 1.30 % 0.03 %	45.22 % 22.75 %26.74 % 1.44 % 0.07 %	43.03 % 19.39 % 27.58 % 3.33 % 0.00 %	21.84 % 23.15 %31.33 % 17.81 % 0.69 %	17.27 % 21.79 %32.82 % 20.83 % 1.15 %	10.41 % 21.27 %36.20 % 23.53 % 0.00 %
Prenatal Substance Use – Yesa,*,^	28.80 %	43.02 %	77.58 %	29.06 %	43.38 %	76.02 %
Parental Warmtha,*,^	13.72 (1.43)	13.55 (1.53)	13.69 (1.49)	13.70 (1.44)	13.52 (1.53)	13.74(1.40)
Parental Monitoring*,^	22.05 (2.46)	21.78 (2.63)	21.65 (2.52)	22.58 (2.22)	22.17 (2.40)	21.62 (2.82)
CBCL Internalizing (t-scores)*,^	47.63 (10.24)	50.24 (10.85)	53.76 (11.58)	47.04 (10.18)	49.41 (10.98)	52.00 (12.03)
CBCL Externalizing (t-scores)*,^	44.53 (9.71)	47.64 (10.69)	52.51 (11.65)	43.62 (9.21)	46.36 (10.48)	50.47 (11.38)

Notes: PH = Parental History of Problematic Substance Use (drugs and alcohol); PH⁰ = Participants with no parental history of problematic substance use; PH¹ = Participants with a history of problematic substance use in one parent; PH² = Participants with a history of problematic substance use in both parents; 2 YR FU = Two-year follow-up timepoint; SD = Standard Deviation; HS = High School; CBCL = Child Behavior Checklist; RT = Reaction Time; ms = Millisecond; BAS = Behavioral Activation System. The table summarizes sociodemographic factors and youth and parent-level covariates for the current baseline and two-year follow-up samples, grouped by parental history (PH⁰, PH¹, PH²). It also indicates whether significant differences in these variables were observed across groups at baseline or the two-year follow-up. ^a Variables measured only at baseline. * Indicates significance between PH groups in the baseline data (p < 0.01). ^ Indicates significance between PH groups in the two-year follow-up data (p < 0.01).

PH group was coded as a categorical variable and dummy-coded in R, with PH⁰ as the reference category and separate binary indicators for PH¹ and PH². Outcome variables included BOLD response to reward anticipation in a priori ROIs at baseline and 2YR FU. The model included random intercepts for subject identification (accounting for repeated measures within individuals), family identification (accounting for clustering among related participants), and MRI manufacturer (controlling for scanner-related variability). Each term was modeled as a separate random term. To address convergence issues, the BOBYQA (Bound Optimisation BY Quadratic Approximation) optimizer was utilized (Powell, 2009). When convergence errors (i.e., singularity) occurred—indicating that a random effect contributed negligible variance and was redundant—it was removed to improve model convergence and stability (in all cases, this variable was subject identification likely due to limited repeated time points of observation). Post hoc tests were conducted to examine differences between each pair of parental history groups using a dummy coding approach. The Benjamini-Hochberg FDR correction was applied separately for each analysis type. Corrections were first applied across ROI analyses within each PH subgroup (PH, PHA, and PHD). For interaction models (PH × time, PHA × time, and PHD × time), corrections were applied across all ROI analyses within their respective subgroups (Benjamini and Hochberg, 1995).

To evaluate PH-behavior relationships and aid the interpretation of fMRI results, post-hoc LME models were run with PH group status as the primary predictor of RT on the MID task and BAS subscales. No covariates were included in these analyses. Furthermore, FDR correction was applied to all p-values examining the relationship between PH and the BAS subscales. Standardized beta coefficients (β) were computed using the MuMIn package in R and are presented for all models as an indication of effect size (Bartoń, 2010). All results were considered significant at the p < .05 threshold.

3. Results

3.1. Demographics & PH frequencies

The current sample included individual data from baseline (n = 9308; M age = 9.93 yrs., range = 9 – 11 yrs.) and 2YR FU (n = 6597; M age = 11.94 yrs., range = 11 – 14 yrs.). PH frequencies were as follows: alcohol: PHA⁰ = 7922; PHA¹ = 1177; PHA² = 183; drug: PHD⁰ = 8331; PHD¹ = 739; PHD² = 205; alcohol + drug: PH⁰ = 7523; PH¹ = 1455; PH² = 330. See Table 1 for detailed demographics by PH at baseline and 2YR FU.

3.2. MID behavioral performance

PH. Large reward RT was significantly faster for PH¹ compared to PH⁰ (t = -2.07, β = −1.24, SE = 0.60, p = 0.038) and PH² (t = 2.30, β = 1.37, SE = 0.59, p = 0.02); PH⁰ and PH² did not significantly differ. Neutral RT was significantly faster for PH¹ compared to PH² (t = 2.20, β = 1.52, SE = 0.69, p = 0.03) and PH⁰ (t = -1.97, β = −1.36, SE = 0.69, p = 0.048); there were no significant differences in neutral RT between PH⁰ and PH². PHA. Large reward RT was significantly faster for PHA¹ compared to PHA⁰ (t = -2.32, β = −1.39, SE = 0.60, p = 0.02) and PHA² (t = 2.29, β = 1.36, SE = 0.60, p = 0.02).Neutral RT was significantly faster for PHA¹ compared to PHA⁰ (t = -2.61, β = −1.81, SE = 0.69, p = 0.01) and PHA² (t = 2.29, β = 1.58, SE = 0.69, p = 0.02); there were no significant differences in large reward RT or neutral RT between PHA² and PHA⁰. PHD. Large reward RT and neutral RT did not significantly differ across PHD groups.

3.3. Primary analyses: PH & MID ROIs

PH. PH was initially associated with BOLD activation in the NAcc, with PH¹ youth exhibiting greater activation in the right (t = 2.17, β = 0.006, SE = 0.003, p = 0.03, p^FDR =0.41) and left (t = 1.99, β = 0.006, SE = 0.003, p = 0.046, p^FDR =0.41) NAcc compared to PH⁰ youth. However, these findings did not survive FDR correction. No significant differences were observed between PH¹ and PH² or PH⁰ and PH². No differences were seen in caudate and putamen across groups; no significant interactions between PH groups and time were observed (See Table 2 for main effects and Table 3 for interaction analyses.)

Table 2.

Main effect of parental history of substance use (PH) on brain activation: mixed-effects model results across ROIs.

Primary Predictor: Parental History of SU (PH)	Comparison	Standardized β Coefficients	SE	df	t	Uncorrected p Value	FDR p Value
Left Caudate	PH0 vs. PH1 PH0 vs. PH2 PH1 vs. PH2	0.003 0.002 0.0004	0.002 0.002 0.002	7079 7999 7930	1.31 0.83 0.18	0.19 0.41 0.86	0.92 0.92 0.99
Right Caudate	PH0 vs. PH1 PH0 vs. PH2 PH1 vs. PH2	0.002 0.002 0.001	0.002 0.002 0.002	7067 7992 7922	0.64 0.95 0.60	0.52 0.34 0.55	0.92 0.92 0.92
Left Putamen	PH0 vs. PH1 PH0 vs. PH2 PH1 vs. PH2	0.002 0.001 −0.0002	0.002 0.002 0.002	7204 8093 8025	1.01 0.38 −0.11	0.31 0.70 0.92	0.92 0.95 0.99
Right Putamen	PH0 vs. PH1 PH0 vs. PH2 PH1 vs. PH2	0.001 0.001 −0.00002	0.002 0.002 0.002	6989 7947 7878	0.70 0.33 −0.01	0.48 0.74 0.99	0.92 0.95 0.99
Left Nucleus Accumbens	PH0 vs. PH1 PH0 vs. PH2 PH1 vs. PH2	0.006 0.002 −0.001	0.003 0.003 0.003	6573 7628 7560	1.99 0.59 −1.99	0.046 0.56 0.71	0.41 0.92 0.95
Right Nucleus Accumbens	PH0 vs. PH1 PH0 vs. PH2 PH1 vs. PH2	0.006 0.003 −0.0001	0.003 0.003 0.002	6435 7555 7842	2.17 1.03 −0.02	0.03 0.30 0.98	0.41 0.92 0.99

Notes: PH, parental history of problematic substance use; SU, substance use; FDR, false discovery rate. ^ap < 0.05; p value remained significant after corrections.

Table 3.

Interaction effect of parental history of substance use (PH) and time on brain activation: mixed-effects model results across ROIs.

Primary Predictor: Parental History of SU (PH) x Time	Comparison	Standardized β Coefficients	SE	df	t	Uncorrected p Value	FDR p Value
Left Caudate	PH0 vs. PH1 PH0 vs. PH2 PH1 vs. PH2	−0.002 0.001 −0.001	0.002 0.002 0.002	11,330 11,580 11,550	−0.71 0.46 −0.75	0.48 0.64 0.45	0.97 0.97 0.97
Right Caudate	PH0 vs. PH1 PH0 vs. PH2 PH1 vs. PH2	−0.0001 0.001 0.0004	0.002 0.002 0.002	11,320 11,570 11,540	−0.03 0.20 0.20	0.97 0.84 0.84	0.97 0.97 0.97
Left Putamen	PH0 vs. PH1 PH0 vs. PH2 PH1 vs. PH2	0.001 0.001 0.001	0.002 0.002 0.002	11,370 11,610 11,580	0.31 0.43 0.27	0.76 0.66 0.79	0.97 0.97 0.97
Right Putamen	PH0 vs. PH1 PH0 vs. PH2 PH1 vs. PH2	−0.0001 0.001 0.001	0.002 0.002 0.002	11,320 11,560 11,530	−0.06 0.38 0.38	0.96 0.70 0.70	0.97 0.97 0.97
Left Nucleus Accumbens	PH0 vs. PH1 PH0 vs. PH2 PH1 vs. PH2	0.002 −0.001 −0.003	0.003 0.003 0.003	12,630 12,810 12,800	0.70 −0.64 −0.91	0.48 0.52 0.36	0.97 0.97 0.97
Right Nucleus Accumbens	PH0 vs. PH1 PH0 vs. PH2 PH1 vs. PH2	0.003 0.0003 −0.001	0.003 0.003 0.003	11,370 11,540 11,520	1.106 0.095 −0.41	0.27 0.92 0.68	0.97 0.97 0.97

Notes: PH, parental history of problematic substance use; SU, substance use; FDR, false discovery rate. ^ap < 0.05; p value remained significant after corrections.

PHA. PHA was significantly linked with BOLD activation in the NAcc, with PHA¹ youth exhibiting greater activation in the r-NAcc (t = 2.95, β = 0.01, SE = 0.002, p = 0.003, p^FDR =0.03) and l-NAcc (t = 3.02, β = 0.009, SE = 0.003, p = 0.003, p^FDR =0.03) compared to PHA⁰ youth. No significant differences were observed between PHA¹ and PHA² or PHA⁰ and PHA² (See Fig. 1). BOLD activation in the caudate and putamen did not differ; no significant interactions between PHA groups and time were observed. (See Table 4 for main effects and Table 5 for interaction analyses.)

Fig. 1.

Nucleus accumbens activation during reward anticipation in the MID task across parental history of problematic alcohol use.Note: Violin plots showing the average beta coefficients for large reward > neutral contrast during the Monetary Incentive Delay Task (MID), stratified by Parental History of Problematic Alcohol Use (PHA⁰, PHA¹, PHA²). The data represents BOLD signaling in the left and right nucleus accumbens (NAcc), averaged across baseline and 2YR follow-up (2YR FU). Statistical significance: * p < 0.05.

Table 4.

Main effect of parental history of alcohol use (PHA) on brain activation: mixed-effects model results across ROIs.

Primary Predictor: Parental History of Alcohol Use (PHA)	Comparison	Standardized β Coefficients	SE	df	t	Uncorrected p Value	FDR p Value
Left Caudate	PHA0 vs. PHA1 PHA0 vs. PHA2 PHA1 vs. PHA2	0.003 0.0002 −0.001	0.002 0.002 0.002	6938 7965 7929	1.171 0.079 −0.38	0.24 0.94 0.70	0.93 0.97 0.97
Right Caudate	PHA0 vs. PHA1 PHA0 vs. PHA2 PHA1 vs. PHA2	0.002 −0.001 −0.001	0.002 0.002 0.002	6926 7951 7915	0.82 −0.26 −0.57	0.41 0.79 0.57	0.93 0.97 0.93
Left Putamen	PHA0 vs. PHA1 PHA0 vs. PHA2 PHA1 vs. PHA2	0.001 0.0003 −0.0002	0.002 0.002 0.002	7063 8073 8034	0.59 0.13 −0.11	0.56 0.90 0.92	0.93 0.97 0.97
Right Putamen	PHA0 vs. PHA1 PHA0 vs. PHA2 PHA1 vs. PHA2	0.002 −0.001 −0.001	0.002 0.002 0.002	6484 7910 7877	0.91 −0.33 −0.67	0.36 0.74 0.51	0.93 0.97 0.93
Left Nucleus Accumbens	PHA0 vs. PHA1 PHA0 vs. PHA2 PHA1 vs. PHA2	0.009 −0.002 −0.006	0.003 0.003 0.006	6444 7546 7529	3.02 −0.78 −1.91	0.003 0.43 0.06	0.03a 0.93 0.36
Right Nucleus Accumbens	PHA0 vs. PHA1 PHA0 vs. PHA2 PHA1 vs. PHA2	0.01 0.0001 −0.003	0.003 0.003 0.003	6308 7491 7457	2.95 0.041 −1.10	0.003 0.97 0.27	0.03a 0.97 0.93

Notes: PHA, parental history of problematic alcohol use; FDR, false discovery rate. ^ap < 0.05; p value remained significant after corrections.

Table 5.

Interaction effect of parental history of alcohol use (PHA) and time on brain activation: mixed-effects model results across ROIs.

Primary Predictor: Parental History of Alcohol Use (PHA) x Time	Comparison	Standardized β Coefficients	SE	df	t	Uncorrected p Value	FDR p Value
Left Caudate	PHA0 vs. PHA1 PHA0 vs. PHA2 PHA1 vs. PHA2	−0.003 −0.003 −0.001	0.002 0.002 0.002	11,230 11,480 11,450	−1.08 −1.06 −0.61	0.28 0.29 0.54	0.83 0.83 0.83
Right Caudate	PHA0 vs. PHA1 PHA0 vs. PHA2 PHA1 vs. PHA2	−0.001 −0.003 −0.002	0.002 0.002 0.002	11,210 11,470 11,440	−0.41 −1.46 −1.23	0.68 0.14 0.22	0.87 0.83 0.83
Left Putamen	PHA0 vs. PHA1 PHA0 vs. PHA2 PHA1 vs. PHA2	0.0003 −0.002 −0.002	0.002 0.002 0.002	11,260 11,510 11,490	0.13 −1.08 −1.08	0.89 0.28 0.28	0.94 0.83 0.83
Right Putamen	PHA0 vs. PHA1 PHA0 vs. PHA2 PHA1 vs. PHA2	0.0001 −0.002 −0.002	0.002 0.002 0.002	11,210 11,450 11,430	0.04 −0.77 −0.74	0.97 0.44 0.46	0.97 0.83 0.83
Left Nucleus Accumbens	PHA0 vs. PHA1 PHA0 vs. PHA2 PHA1 vs. PHA2	0.0005 −0.003 −0.003	0.003 0.003 0.003	11,920 12,150 12,130	0.16 −0.86 −0.88	0.87 0.39 0.38	0.94 0.83 0.83
Right Nucleus Accumbens	PHA0 vs. PHA1 PHA0 vs. PHA2 PHA1 vs. PHA2	0.001 −0.002 −0.002	0.003 0.003 0.003	11,530 11,700 11,680	0.30 −0.51 −0.59	0.76 0.61 0.55	0.91 0.84 0.83

Notes: PHA, parental history of problematic alcohol use; FDR, false discovery rate. ^Ap < 0.05; p value remained significant after corrections.

PHD. PHD was not significantly associated with BOLD activation in the bilateral caudate, nucleus accumbens, or putamen. No significant differences were found between PHD⁰ and PHD², PHD¹ and PHD², or PHD⁰ and PHD¹. Additionally, no significant interactions between PHD groups and time were observed. (See Table 6 for main effects and Table 7 for interaction analyses.)

Table 6.

Main effect of parental history of drug use (PHD) on brain activation: mixed-effects model results across ROIs.

Primary Predictor: Parental History of Drug Use (PHD)	Comparison	Standardized β Coefficients	SE	df	t	Uncorrected p Value	FDR p Value
Left Caudate	PHD0 vs. PHD1 PHD0 vs. PHD2 PHD1 vs. PHD2	0.004 0.002 −0.0001	0.002 0.002 0.002	7410 8224 8120	1.64 0.82 −0.02	0.10 0.41 0.98	0.83 0.83 0.99
Right Caudate	PHD0 vs. PHD1 PHD0 vs. PHD2 PHD1 vs. PHD2	0.002 0.003 0.002	0.002 0.002 0.002	7401 8222 7401	0.675 1.06 0.66	0.50 0.29 0.51	0.83 0.83 0.83
Left Putamen	PHD0 vs. PHD1 PHD0 vs. PHD2 PHD1 vs. PHD2	0.003 0.0004 −0.001	0.002 0.002 0.002	7522 8296 8197	1.59 0.22 −0.55	0.11 0.83 0.58	0.83 0.93 0.87
Right Putamen	PHD0 vs. PHD1 PHD0 vs. PHD2 PHD1 vs. PHD2	0.002 −0.001 −0.007	0.002 0.002 0.002	7353 8203 8098	1.17 −0.29 −0.82	0.24 0.77 0.41	0.83 0.93 0.83
Left Nucleus Accumbens	PHD0 vs. PHD1 PHD0 vs. PHD2 PHD1 vs. PHD2	0.004 0.001 −0.001	0.003 0.003 0.003	6978 8082 7940	1.36 0.43 −0.25	0.17 0.67 0.80	0.83 0.93 0.93
Right Nucleus Accumbens	PHD0 vs. PHD1 PHD0 vs. PHD2 PHD1 vs. PHD2	0.004 0.002 - 0.00003	0.003 0.003 0.003	6878 7931 7798	1.31 0.66 −0.01	0.19 0.51 0.99	0.83 0.83 0.99

Notes: PHD, parental history of problematic drug use; FDR, false discovery rate. ^ap < .05; p value remained significant after corrections.

Table 7.

Interaction effect of parental history of drug use (PHD) and time on brain activation: mixed-effects model results across ROIs.

Primary Predictor: Parental History of Drug (PHD)x Time	Comparison	Standardized β Coefficients	SE	df	t	Uncorrected p Value	FDR p Value
Left Caudate	PHD0 vs. PHD1 PHD0 vs. PHD2 PHD1 vs. PHD2	0.0001 0.001 0.001	0.002 0.002 0.002	11,290 11,630 11,570	0.17 0.45 0.33	0.87 0.65 0.74	0.96 0.96 0.96
Right Caudate	PHD0 vs. PHD1 PHD0 vs. PHD2 PHD1 vs. PHD2	0.002 0.0002 0.001	0.002 0.002 0.002	11,300 11,650 11,590	0.65 0.08 0.23	0.52 0.94 0.82	0.96 0.96 0.96
Left Putamen	PHD0 vs. PHD1 PHD0 vs. PHD2 PHD1 vs. PHD2	−0.001 0.001 0.001	0.002 0.002 0.002	11,330 11,680 11,620	−0.05 0.54 0.51	0.96 0.59 0.61	0.96 0.96 0.96
Right Putamen	PHD0 vs. PHD1 PHD0 vs. PHD2 PHD1 vs. PHD2	−0.0004 0.001 0.001	0.002 0.002 0.002	11,280 11,640 11,580	−0.21 0.57 0.61	0.83 0.57 0.54	0.96 0.96 0.96
Left Nucleus Accumbens	PHD0 vs. PHD1 PHD0 vs. PHD2 PHD1 vs. PHD2	−0.002 −0.001 - 0.0002	0.003 0.003 0.003	11,850 12,140 12,090	−0.77 −0.46 −0.07	0.44 0.64 0.94	0.96 0.96 0.96
Right Nucleus Accumbens	PHD0 vs. PHD1 PHD0 vs. PHD2 PHD1 vs. PHD2	−0.0004 0.0004 0.001	0.003 0.003 0.003	11,680 11,960 11,910	−0.12 0.14 0.18	0.90 0.89 0.86	0.96 0.96 0.96

Notes: PHD, parental history of problematic drug use; FDR, false discovery rate. ^ap < .05; p value remained significant after corrections.

3.4. Primary analyses: covariates & MID ROIs

Although not the primary focus, some covariates were consistent independent predictors of striatal BOLD response. This section summarizes key findings from the PH analyses; findings across PHA and PHD analyses were broadly consistent with these findings.

Prenatal Substance Use Exposure. Individuals exposed prenatally showed reduced BOLD activation in the right putamen compared to those without exposure (all ps < 0.05). No differences were observed in other ROIs. Parental Monitoring. Parental monitoring was positively associated with BOLD responses in the bilateral putamen and caudate, as well as the right NAcc (all ps < 0.05). Internalizing Symptoms. Higher internalizing symptoms were significantly associated with reduced BOLD responses in the bilateral NAcc (all ps < 0.05); no significant associations were found in other ROIs.

3.5. Post-hoc BAS behavioral analyses

BAS Reward Responsiveness. PHA. PHA² scored significantly higher than PHA¹ on the Reward Responsiveness BAS subscales; however, this effect did not remain significant after FDR correction (t = -2.38, β = −0.43, SE = 0.18, p = 0.04, p^FDR =0.12). PHA⁰ did not significantly differ from PHA¹ or PHA². Reward Responsiveness did not significantly differ across PH or PHD groups. BAS Drive. No significant differences were found between PH groups (PH0 vs. PH1, PH0 vs. PH2, PH1 vs. PH2) on the BAS Drive subscale, and this pattern remained consistent across PHA and PHD. BAS Fun Seeking. PH. PH² scored significantly higher on the Fun Seeking BAS subscale compared to PH⁰ (t = -4.88, β = −0.54, SE = 0.11, p = 0.0003, p^FDR =0.003) and PH¹ (t = -3.65, β = −0.44, SE = 0.12, p = 0.01, p^FDR =0.06); however, the PH² vs. PH¹ comparison did not remain significant after FDR correction. PH⁰ was not significantly different from PH¹. PHA. PHA² scored significantly higher compared to PHA⁰ (t = -4.19, β = −0.62, SE = 0.15, p = 0.0004, p^FDR =0.0036) and PHA¹ (t = -3.44, β = −0.54, SE = 0.16, p = 0.0031, p^FDR =0.01); PHA⁰ did not significantly differ from PHA¹. PHD. PHD¹ (t = -3.19, β = −0.24, SE = 0.07, p = 0.006, p^FDR =0.03) and PHD² (t = -4.29, β = −0.60, SE = 0.14, p = 0.001, p^FDR =0.01) scored significantly higher than PHD⁰; PHD² was not significantly different from PHD¹.

4. Discussion

Parental history of problematic substance use (PH) is a known risk factor for adolescent substance use, and evidence suggests that premorbid differences in striatal activation—a key brain region involved in reward processing—may contribute to increased vulnerability. The PH density (i.e., number of affected biological parents) and the SU type may further affect these outcomes, though these relationships have not been fully explored longitudinally in a preadolescent/early adolescent cohort while accounting for parenting factors that are linked with PH. This study investigated the relationship between PH density and SU type (overall SU, alcohol, and drug) and brain activation patterns in the striatum during reward anticipation while controlling for relevant parental and youth substance use risk factors in a large and diverse longitudinal early adolescent sample. Our primary findings showed that youth with one parent with alcohol (PHA¹) demonstrated significantly greater NAcc BOLD activation in both the right (β = 0.01, p^FDR = 0.03) and left (β = 0.01, p^FDR = 0.03) hemispheres compared to those without PHA (PHA⁰). These results remained significant after controlling for sociodemographic and parental factors, survived FDR correction, and were consistent across both time periods, suggesting that preadolescents and young adolescents with PHA¹ may experience a sensitized brain response to reward anticipation. Notably, the heightened NAcc activation observed in PHA¹ youth remained significant after accounting for externalizing symptoms, suggesting that these neural differences may reflect a specific sensitivity to reward anticipation rather than a broader behavioral dysregulation. Further work is warranted to understand the role reward sensitivity plays in unique and shared susceptibility for substance use and other externalizing behaviors (Tervo-Clemmens et al., 2020). A similar pattern was observed in the PH group, though these findings did not remain significant after FDR correction. These findings broadly align with previous research demonstrating that adolescents with PH¹ exhibit significantly greater BOLD signaling during reward anticipation in the striatum compared to their peers without such histories (Kwarteng et al., 2021, Stice and Yokum, 2014). The observed effect sizes were modest but consistent with previous studies, which reported small to moderate effects (e.g., Kwarteng et al. 2021, Cohen’s d = 0.11, 95 % CI [0.00, 0.20]; Stice and Yokum, 2014, r = 0.45–0.57). Similarly, Martz et al. (2022) found a trend toward greater NAcc activation in PHA¹ youth compared to PHA⁰, though their results did not reach statistical significance.

Behaviorally, we found that youth with PHA¹ had faster MID RT for large rewards and neutral trials compared to youth with PHA⁰ and PHA², while reaction times did not differ between PHA² and PHA⁰. However, despite greater NAcc activation, PHA¹ youth did not significantly differ from PHA⁰ on behavioral measures of reward motivation and processing, though they did report significantly lower scores on the Fun Seeking scale compared to PHA². This suggests that this hypersensitive BOLD NAcc response to reward anticipation in youth with PHA¹ is not reflected in downstream reward-motivated behavior at this developmental stage. One possible explanation for this discrepancy is broader inconsistencies in how reward sensitivity is measured. Prior research has shown weak convergence between self-report and cognitive task measures, with shared variance accounting for less than 2 %, suggesting these methods may capture distinct psychological constructs rather than a unified mechanism of reward sensitivity (Demidenko et al., 2019). Further research is needed to examine how these activation patterns and brain-behavior relationships evolve as youth transition into middle and late adolescence.

These findings demonstrate that youth with PHA¹ exhibit heightened NAcc activation, a brain region involved in processing rewarding stimuli and integrating information from limbic and cortical areas to facilitate motivated, goal-directed behavior (Morrison et al., 2017). Preadolescence into adolescence is characterized by increases in reward sensitivity, and the NAcc is particularly responsive to rewards, likely due to increased receptor expression and peak dopamine release (Ernst and Luciana, 2015, Galvan et al., 2006, Tarazi and Baldessarini, 2000). Longitudinal studies indicate that NAcc signaling and structural volume exhibit an inverted U-shaped trajectory, peaking in mid-adolescence before declining in late adolescence and early adulthood (Dennison et al., 2013, Schreuders et al., 2018, Urošević et al., 2012, Braams et al., 2015, Van Leijenhorst et al., 2010). The neurobiological mechanisms underlying NAcc differences in PHA youth may involve genetic variations in COMT and DAT1, which may be more pronounced or prevalent in youth with a parental substance use history of alcohol (PHA). For example, the DAT1 9-repeat allele and the COMT met/met genotype have been associated with increased NAcc activation during reward anticipation in healthy adults (n = 27, M age = 27.3, SD = 5.7). Additionally, an interaction between COMT and DAT1 genes is linked to heightened activation in the ventral striatum, with carriers of both variants showing the highest activation levels (Dreher et al., 2009). These genetic variations could mediate the relationship between a parental substance use history and altered NAcc activation, potentially increasing vulnerability to reward-driven behaviors such as substance use in adolescents.

Notably, during this developmental period (ages 9–13), we found no interactions between PH and time in predicting NAcc BOLD response, possibly due to the limited age range and developmental stages. These trajectories may shift as the neural reward system continues to develop. The observed increase in NAcc BOLD response in PHA¹ youth raises questions about whether it serves as a risk or resiliency biomarker, given their lack of differences in reward-motivated behaviors compared to PHA⁰ youth. Research on the ventral striatum (VS) challenges the notion that heightened activation in these regions always leads to maladaptive behavior (Pfeifer and Allen, 2012). Longitudinal research has shown that increased VS activation in response to emotional facial expressions is associated with greater resistance to peer influence and decreased risky behavior, suggesting that heightened sensitivity can support positive developmental outcomes (Pfeifer et al., 2011). Longitudinal analyses examining the impact of PHA density on reward processing from preadolescence into later stages of adolescence, when risk-taking behaviors significantly increase, will better inform whether these activation patterns represent a risk or resiliency biomarker.

One prior study reported that youth with PHA² showed reduced right NAcc activation compared to peers with PHA¹ (Martz et al., 2022), suggesting a potential dose-dependent PH relationship. These differences may be attributed to the current study’s control of both youth and parental-level substance use risk factors (i.e., prenatal SU exposure and parental monitoring) in the models. It is notable that behaviorally, youth with PHA² scored significantly higher than PHA¹ on Reward Responsiveness and Fun Seeking, though only the latter survived statistical thresholding. Similarly, PH²/PHA²/PHD² scored significantly higher than PH⁰/PHA⁰/PHD⁰ on Fun Seeking, and PHD¹ also scored higher than PHD⁰ on Fun Seeking. Taken together, this indicates that youth with two parents with PH reported greater behavioral signs of reward sensitivity and motivation, specifically a heightened tendency to seek out novel and exciting experiences.

Interestingly, self-reported reward sensitivity is a stronger predictor of real-world risk behaviors (e.g., binge drinking, marijuana use, risky driving) than cognitive tasks (Demidenko et al., 2019). Prior research has linked these reward-seeking traits to an increased likelihood of engaging in risk-taking behaviors and substance use onset during adolescence (Shao et al., 2025, Urošević et al., 2012) as well as among college students (O’Connor et al., 2009). These reward-seeking behaviors are also associated with increased nucleus accumbens (NAcc) activation in response to rewards in older adolescents (Braams et al., 2015). Taken together, the increased self-reported reward motivation with similar striatal neural activation during reward anticipation for youth with PH² compared to the youth without suggests early evidence of reward hyposensitivity, suggesting individuals with PH need more intense or frequent rewards to achieve the same level of engagement or satisfaction. Additional research is needed to examine whether having two parents with PH/PHA is linked with BOLD striatal response to reward anticipation as the youth ages into later adolescence. Further investigation is warranted to assess whether other aspects of reward processing, such as processing reward gains or losses, are differentially impacted by the density of PH/PHA.

Parental history of problematic drug use (PHD) was not associated with significant differences in BOLD activation between PHD⁰, PHD¹, and PHD² across ROIs. Interestingly, our hypothesis concerning differences in putamen activation across parental history of drug use analyses was not supported, contrasting with the findings of Kwarteng et al. (2021). These inconsistencies may be due to differential measurement of PH (dichotomous vs. density measure) and inclusion of parent and youth SU risk factors, which were independently linked with NAcc, putamen, and caudate BOLD activation.

Although not the primary focus, several parent- and youth-level covariates—including prenatal SU exposure, parental monitoring, and internalizing symptoms—emerged as independent predictors of BOLD signaling in the PH, PHA, and PHD analyses. Higher levels of youth’s perception of parental monitoring were positively associated with BOLD signaling in the bilateral putamen and caudate, as well as the right NAcc. This is the first study to our knowledge linking reward anticipation with parental monitoring. Increased parental monitoring may provide adolescents with an environment that allows for safe engagement with rewarding stimuli, though future studies are needed to examine how aspects of parental monitoring relate to adolescent reward-seeking behaviors. Higher internalizing symptoms were associated with reduced BOLD signaling in the NAcc, consistent with previous findings indicating that adolescents with internalizing disorders exhibit reduced ventral striatum activation and reward sensitivity compared to healthy youth (Stringaris et al., 2015). Lastly, our results suggest prenatal SU exposure is associated with decreased BOLD activation in the right putamen, which is consistent with research indicating such exposure can interfere with the normal development of the brain's reward circuitry in both animal and human studies (Eiden et al., 2023). Finally, while sociodemographic factors were included as covariates, future research should examine their role as potential moderators to clarify how they interact with PH effects on neural activation and how these relationships change over time. Understanding these parental and youth risk and resilience factors is crucial, as they may represent modifiable targets for prevention efforts.

This study has several limitations worth noting. We relied on retrospective parent-report measures of PH, which may have resulted in underreporting. Self-reported data on SU can often be biased by factors such as social desirability, stigma, and fear of intervention from child protection services (Johnson and Fendrich, 2005). These biases may have led to misclassification of PH, potentially weakening the strength of the observed associations. Future research should consider incorporating multiple methods of data collection, such as collateral reports from other family members or caregivers, to validate self-reported measures of parental SU. In addition, we did not account for whether youth with PH lived with that parent. As a result, we could not examine how the presence or absence of a parent with current substance use problems affects outcomes (although other current parent factors were in the models as covariates).

Although several associations reached statistical significance, standardized beta coefficients were small, indicating small effect sizes, which is expected for large-scale datasets such as ABCD (Owens et al., 2021). However, even small effect sizes may have clinical significance, as true causal associations in nature are often modest (Dick et al., 2021), and brain correlates tend to have limited predictive power in isolation (Feng et al., 2022). While the effect sizes in our study are small, they likely represent more accurate estimates than those reported in smaller studies, which are often subject to effect size inflation due to publication bias and selective reporting (Ioannidis, 2008, Button et al., 2013). Despite their small magnitude, these findings provide insight into early neural sensitivity to reward anticipation in youth with parental substance use history, which may contribute to long-term differences in reward processing and risk-related behaviors. Youth with PH¹/PHA¹ also had consistently and significantly faster reaction times (RT) than PH⁰/PHA⁰ and PH²/PHA², suggesting that an increased behavioral approach and faster RT during reward anticipation may contribute to the observed BOLD patterns.

Several methodological limitations should also be acknowledged. This study used pre-determined ROIs, limiting our understanding of how PH impacts whole-brain activity (van den Heuvel and Hulshoff Pol, 2010). Future research should use longitudinal whole-brain analysis techniques to better clarify the links between PH and activation patterns in different brain regions. Stringent head motion thresholds also resulted in the exclusion of key findings, particularly in the PH² group, where motion differences were most pronounced; given that head motion covaries with parental history, these stringent thresholds may inadvertently obscure meaningful associations, highlighting the need for more refined QC methods to better balance data retention and artifact reduction. Furthermore, the ABCD pre-processing pipeline has known timing issues with GE scanners (Task-Based Functional Magnetic Resonance Imaging, 2025); however, we appropriately accounted for scanner type in our models and re-ran additional analyses, excluding observations acquired on GE scanners, and observed consistent results. Lastly, while we followed the standard ABCD modeling approach for the Monetary Incentive Delay (MID) task available in the data release, its exclusion of trial-level RT variability and certain task events (i.e., probe, fixation, and feedback) may introduce bias in BOLD estimates, especially in regions like the striatum (Mumford et al., 2025). Future studies should consider more comprehensive modeling approaches (e.g., saturated models) to improve the accuracy and interpretability of task-related neural signals.

In conclusion, our study reveals that preadolescent and early adolescent youth with PHA¹ demonstrated subtly greater BOLD activation in reward-related brain regions, specifically the NAcc. Although striatal reward response was similar to controls, youth with PHA² and PHD² demonstrated significantly greater levels of self-reported reward-motivated behaviors. Taken together, these findings suggest that PHA plays a subtle but significant role in shaping youths’ reward sensitivity and behavioral tendencies, which has implications for understanding their risk for developing future SU problems or other reward-motivated behaviors. Interestingly, our results indicate that greater parental monitoring was independently associated with greater BOLD responses in reward centers (bilateral putamen and caudate and right NAcc). These findings underscore the importance of considering environmental factors when examining the impact of PH on reward processing, including the number of parents with substance use issues, the type of substance used, and other parenting behaviors that may shape this relationship. Further, parental monitoring is a modifiable risk factor that warrants further research to clarify its potential in mitigating the effects of PH on adolescent brain development and mental health outcomes. Future research should also investigate how anticipatory reward responses in the NAcc predict substance use initiation and trajectories among youth with PHA they transition into later stages of adolescent development.

CRediT authorship contribution statement

Elizabeth A. Stinson: Writing – review & editing, Writing – original draft, Validation, Formal analysis, Data curation, Conceptualization. Ryan M. Sullivan: Writing – review & editing, Writing – original draft, Visualization, Validation, Formal analysis, Data curation. Krista M. Lisdahl: Writing – review & editing, Writing – original draft, Validation, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Gabriella Y. Navarro-Love: Writing – review & editing, Writing – original draft, Visualization, Validation, Formal analysis, Data curation, Conceptualization.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The authors do not have permission to share data.