Adolescence marks a critical window in brain development when neural systems governing reward and self-control are rapidly maturing. Understanding why some youth initiate substance use while others don’t has long been a focus in developmental neuroscience. In this study, researchers explored how subtle variations in brain structure may precede and predict early substance use initiation.
l
Researchers have long sought brain-based markers that might signal risk for later development of substance use disorder. These markers include factors like the thickness or volume of specific brain regions or broader measures such as total brain volume. Some previous investigations have linked such structural characteristics to later substance involvement. For instance, studies have found that children with thinner dorsolateral prefrontal cortices—a region critical for executive function and emotional regulation—are more likely to develop alcohol-related problems. However, these findings have often failed to replicate, largely because neuroimaging studies are typically costly and small, limiting their ability to detect effects and reproducibility.
To overcome these challenges, in 2015, the National Institute on Drug Abuse (NIDA) and the National Institute on Alcohol Abuse and Alcoholism (NIAAA) launched the Adolescent Brain Cognitive Development (ABCD) Study through a consortium spanning 22 United States research sites. This landmark longitudinal project is tracking brain, cognitive, and behavioral development from late childhood into early adulthood in nearly 12,000 youth who enrolled in this project around age 9.
The study summarized here drew on this rich ABCD dataset to examine how variations in brain structure may relate to the initiation of substance use during adolescence.
This was a study of 9,804 children ages 9 to 11 at the start of the study (born 2005-2009) recruited from 22 United States research sites, who were assessed every 6 months for 3 years, on any (yes/no) alcohol, cannabis, nicotine, or substance use more generally, including these and other substances. Participants’ brains were scanned using magnetic resonance imaging (MRI) at baseline, and substance use initiation and other psychosocial individual characteristics were assessed at each subsequent follow-up. Analyses controlled for prenatal exposure to alcohol and other drug, familial relationships (e.g., twins and siblings), and MRI scanner differences across sites, but not demographic factors.
The sample average age was 9.9 years at baseline, of whom 53% were boys. The sample was about three quarters White. By age 15, about 35% reported some form of alcohol or other drug use.
Children who initiated substance use showed subtle but identifiable brain differences
Compared with peers with no substance use, children who reported trying any substance before age 15 had larger overall brain volumes as well as greater cortical surface area (see graphic below). At the same time, they showed a thinner prefrontal cortex, particularly in the right rostral middle frontal gyrus, a region crucial for decision-making, planning, and impulse control.
The direction of associations varied by brain region
The direction of associations between observed brain differences and initiation of substance use varied by structure. In children who’d initiated substance use, the prefrontal regions (e.g., dorsolateral prefrontal cortex, superior frontal gyrus) were thinner. Conversely, the occipital, temporal, and parietal regions were thicker, suggesting imbalanced cortical development that might influence how youth process social and sensory cues. Subcortical areas such as the globus pallidus and hippocampus, which are involved in motivation, reward, and memory, were larger in initiators. Taken together, these patterns may suggest a form of neurobiological vulnerability for early substance experimentation marked by reduced inhibitory/regulatory control but heightened reward sensitivity.
These brain differences existed before substance use began
A critical strength of this study was its ability to separate features that predate (i.e., predispose) substance use from possible neurotoxic effects in the brain from alcohol and other drug use. When the researchers restricted analyses to children who had not yet used any substances at baseline (23% of participants had already initiated substance use before their baseline brain scans), the same associations, like a thinner prefrontal cortex and larger global brain volume, still predicted future initiation. This indicates that many of these brain differences were present prior to first use, making them more likely to be preexisting risk markers versus consequences of early alcohol or other drug use.
Substance-specific patterns were modest but distinct
While the overall profile was consistent across substances, some nuanced differences emerged (see graphic below). Alcohol initiation, the most common form of substance use, largely mirrored the general pattern of thinner prefrontal but larger posterior and subcortical regions. Nicotine initiation was associated with smaller right superior frontal gyrus volume and deeper folds in the orbitofrontal cortex, areas involved in craving and valuation of rewards. Cannabis initiation was associated with thinner motor-control areas (precentral gyrus) and smaller right caudate volume, a structure central to habit formation and reward learning. Although these effects were small, they hint at the possibility that differences in brain structure may confer risk for different substances.
Findings held even after accounting for prenatal substance exposure and family factors
Importantly, the associations detected by the researchers remained when analyses controlled for prenatal exposure to alcohol and other drug, familial relationships (e.g., twins and siblings), and MRI scanner differences across sites. This strengthens confidence that the observed brain differences were not artefacts of measurement or family background.
Traditionally, differences in brain structure observed in people who use alcohol or drugs have been interpreted as consequences of substance exposure. However, the results of this large, longitudinal study suggest a more complex picture.
In this research, brain differences linked to early substance use were already present before any use began, implying that certain aspects of brain development may predispose some youth to experiment earlier than others. For instance, thinner prefrontal brain regions—which play a central role in impulse control, planning, and emotion regulation—could make it harder for some children to inhibit impulses or delay gratification in social contexts where substances are available. At the same time, larger subcortical structures and overall brain volumes may reflect developmental patterns associated with greater sensation seeking or reward sensitivity.
It is important to note that these patterns do not mean that brain structure is destiny. The effects for the correlations between these brain characteristics and substance use initiation were quite small, suggesting that some, but not all, youth with these profiles went on to use substances. Also, in terms of the utility, these results cannot yet be applied in clinical settings. For example, we do not yet know the threshold for these neurobiological characteristics (e.g., thinner vs. thicker brain regions) that would accurately classify a child as being at risk for substance use initiation and downstream consequences. Rather, they highlight that differences in neurodevelopment might influence vulnerability, which in turn interacts with other biological, psychological, and social factors. Indeed, this study could not determine why these brain characteristics are present – i.e., developed during childhood as a function of life experience (e.g., trauma, stressors, etc.). That said, the findings held even when controlling for familial risk for substance use – suggesting the risk goes beyond genetic predisposition alone. Supportive family environments, positive peer relationships, and opportunities to develop self-regulatory skills may help buffer against these risks.
For prevention science, these findings may underscore the need for earlier, developmentally informed interventions that target self-regulation, stress management, and healthy risk-taking before adolescence. They also hint that brain-based measures, in combination with behavioral and genetic data, could one day help providers identify individuals most in need of early support.
Finally, the study challenges us to rethink how we interpret neuroimaging findings in addiction research. Not all brain differences seen in youth who use alcohol and other drugs are the result of damage; some may reflect preexisting variations in brain maturation that set the stage for how individuals respond to risk, reward, and social influences during adolescence. Recognizing these developmental nuances can help refine both prevention strategies and public messaging around youth substance use.
This uniquely large national study of brain imaging suggests that subtle differences in brain structure—particularly in regions responsible for impulse control and planning—may predispose some children to some degree to try substances earlier in life. While the thresholds at which these subtle differences predict initiation are not yet clear, ultimately, recognizing and supporting these youth could reduce the likelihood of later substance-related problems and improve developmental outcomes.
Miller, A. P., Baranger, D. A. A., Paul, S. E., Garavan, H., Mackey, S., Tapert, S. F., LeBlanc, K. H., Agrawal, A., & Bogdan, R. (2024). Neuroanatomical variability and substance use initiation in late childhood and early adolescence. JAMA Network Open, 7(12). doi: 10.1001/jamanetworkopen.2024.52027.
l
Researchers have long sought brain-based markers that might signal risk for later development of substance use disorder. These markers include factors like the thickness or volume of specific brain regions or broader measures such as total brain volume. Some previous investigations have linked such structural characteristics to later substance involvement. For instance, studies have found that children with thinner dorsolateral prefrontal cortices—a region critical for executive function and emotional regulation—are more likely to develop alcohol-related problems. However, these findings have often failed to replicate, largely because neuroimaging studies are typically costly and small, limiting their ability to detect effects and reproducibility.
To overcome these challenges, in 2015, the National Institute on Drug Abuse (NIDA) and the National Institute on Alcohol Abuse and Alcoholism (NIAAA) launched the Adolescent Brain Cognitive Development (ABCD) Study through a consortium spanning 22 United States research sites. This landmark longitudinal project is tracking brain, cognitive, and behavioral development from late childhood into early adulthood in nearly 12,000 youth who enrolled in this project around age 9.
The study summarized here drew on this rich ABCD dataset to examine how variations in brain structure may relate to the initiation of substance use during adolescence.
This was a study of 9,804 children ages 9 to 11 at the start of the study (born 2005-2009) recruited from 22 United States research sites, who were assessed every 6 months for 3 years, on any (yes/no) alcohol, cannabis, nicotine, or substance use more generally, including these and other substances. Participants’ brains were scanned using magnetic resonance imaging (MRI) at baseline, and substance use initiation and other psychosocial individual characteristics were assessed at each subsequent follow-up. Analyses controlled for prenatal exposure to alcohol and other drug, familial relationships (e.g., twins and siblings), and MRI scanner differences across sites, but not demographic factors.
The sample average age was 9.9 years at baseline, of whom 53% were boys. The sample was about three quarters White. By age 15, about 35% reported some form of alcohol or other drug use.
Children who initiated substance use showed subtle but identifiable brain differences
Compared with peers with no substance use, children who reported trying any substance before age 15 had larger overall brain volumes as well as greater cortical surface area (see graphic below). At the same time, they showed a thinner prefrontal cortex, particularly in the right rostral middle frontal gyrus, a region crucial for decision-making, planning, and impulse control.
The direction of associations varied by brain region
The direction of associations between observed brain differences and initiation of substance use varied by structure. In children who’d initiated substance use, the prefrontal regions (e.g., dorsolateral prefrontal cortex, superior frontal gyrus) were thinner. Conversely, the occipital, temporal, and parietal regions were thicker, suggesting imbalanced cortical development that might influence how youth process social and sensory cues. Subcortical areas such as the globus pallidus and hippocampus, which are involved in motivation, reward, and memory, were larger in initiators. Taken together, these patterns may suggest a form of neurobiological vulnerability for early substance experimentation marked by reduced inhibitory/regulatory control but heightened reward sensitivity.
These brain differences existed before substance use began
A critical strength of this study was its ability to separate features that predate (i.e., predispose) substance use from possible neurotoxic effects in the brain from alcohol and other drug use. When the researchers restricted analyses to children who had not yet used any substances at baseline (23% of participants had already initiated substance use before their baseline brain scans), the same associations, like a thinner prefrontal cortex and larger global brain volume, still predicted future initiation. This indicates that many of these brain differences were present prior to first use, making them more likely to be preexisting risk markers versus consequences of early alcohol or other drug use.
Substance-specific patterns were modest but distinct
While the overall profile was consistent across substances, some nuanced differences emerged (see graphic below). Alcohol initiation, the most common form of substance use, largely mirrored the general pattern of thinner prefrontal but larger posterior and subcortical regions. Nicotine initiation was associated with smaller right superior frontal gyrus volume and deeper folds in the orbitofrontal cortex, areas involved in craving and valuation of rewards. Cannabis initiation was associated with thinner motor-control areas (precentral gyrus) and smaller right caudate volume, a structure central to habit formation and reward learning. Although these effects were small, they hint at the possibility that differences in brain structure may confer risk for different substances.
Findings held even after accounting for prenatal substance exposure and family factors
Importantly, the associations detected by the researchers remained when analyses controlled for prenatal exposure to alcohol and other drug, familial relationships (e.g., twins and siblings), and MRI scanner differences across sites. This strengthens confidence that the observed brain differences were not artefacts of measurement or family background.
Traditionally, differences in brain structure observed in people who use alcohol or drugs have been interpreted as consequences of substance exposure. However, the results of this large, longitudinal study suggest a more complex picture.
In this research, brain differences linked to early substance use were already present before any use began, implying that certain aspects of brain development may predispose some youth to experiment earlier than others. For instance, thinner prefrontal brain regions—which play a central role in impulse control, planning, and emotion regulation—could make it harder for some children to inhibit impulses or delay gratification in social contexts where substances are available. At the same time, larger subcortical structures and overall brain volumes may reflect developmental patterns associated with greater sensation seeking or reward sensitivity.
It is important to note that these patterns do not mean that brain structure is destiny. The effects for the correlations between these brain characteristics and substance use initiation were quite small, suggesting that some, but not all, youth with these profiles went on to use substances. Also, in terms of the utility, these results cannot yet be applied in clinical settings. For example, we do not yet know the threshold for these neurobiological characteristics (e.g., thinner vs. thicker brain regions) that would accurately classify a child as being at risk for substance use initiation and downstream consequences. Rather, they highlight that differences in neurodevelopment might influence vulnerability, which in turn interacts with other biological, psychological, and social factors. Indeed, this study could not determine why these brain characteristics are present – i.e., developed during childhood as a function of life experience (e.g., trauma, stressors, etc.). That said, the findings held even when controlling for familial risk for substance use – suggesting the risk goes beyond genetic predisposition alone. Supportive family environments, positive peer relationships, and opportunities to develop self-regulatory skills may help buffer against these risks.
For prevention science, these findings may underscore the need for earlier, developmentally informed interventions that target self-regulation, stress management, and healthy risk-taking before adolescence. They also hint that brain-based measures, in combination with behavioral and genetic data, could one day help providers identify individuals most in need of early support.
Finally, the study challenges us to rethink how we interpret neuroimaging findings in addiction research. Not all brain differences seen in youth who use alcohol and other drugs are the result of damage; some may reflect preexisting variations in brain maturation that set the stage for how individuals respond to risk, reward, and social influences during adolescence. Recognizing these developmental nuances can help refine both prevention strategies and public messaging around youth substance use.
This uniquely large national study of brain imaging suggests that subtle differences in brain structure—particularly in regions responsible for impulse control and planning—may predispose some children to some degree to try substances earlier in life. While the thresholds at which these subtle differences predict initiation are not yet clear, ultimately, recognizing and supporting these youth could reduce the likelihood of later substance-related problems and improve developmental outcomes.
Miller, A. P., Baranger, D. A. A., Paul, S. E., Garavan, H., Mackey, S., Tapert, S. F., LeBlanc, K. H., Agrawal, A., & Bogdan, R. (2024). Neuroanatomical variability and substance use initiation in late childhood and early adolescence. JAMA Network Open, 7(12). doi: 10.1001/jamanetworkopen.2024.52027.
l
Researchers have long sought brain-based markers that might signal risk for later development of substance use disorder. These markers include factors like the thickness or volume of specific brain regions or broader measures such as total brain volume. Some previous investigations have linked such structural characteristics to later substance involvement. For instance, studies have found that children with thinner dorsolateral prefrontal cortices—a region critical for executive function and emotional regulation—are more likely to develop alcohol-related problems. However, these findings have often failed to replicate, largely because neuroimaging studies are typically costly and small, limiting their ability to detect effects and reproducibility.
To overcome these challenges, in 2015, the National Institute on Drug Abuse (NIDA) and the National Institute on Alcohol Abuse and Alcoholism (NIAAA) launched the Adolescent Brain Cognitive Development (ABCD) Study through a consortium spanning 22 United States research sites. This landmark longitudinal project is tracking brain, cognitive, and behavioral development from late childhood into early adulthood in nearly 12,000 youth who enrolled in this project around age 9.
The study summarized here drew on this rich ABCD dataset to examine how variations in brain structure may relate to the initiation of substance use during adolescence.
This was a study of 9,804 children ages 9 to 11 at the start of the study (born 2005-2009) recruited from 22 United States research sites, who were assessed every 6 months for 3 years, on any (yes/no) alcohol, cannabis, nicotine, or substance use more generally, including these and other substances. Participants’ brains were scanned using magnetic resonance imaging (MRI) at baseline, and substance use initiation and other psychosocial individual characteristics were assessed at each subsequent follow-up. Analyses controlled for prenatal exposure to alcohol and other drug, familial relationships (e.g., twins and siblings), and MRI scanner differences across sites, but not demographic factors.
The sample average age was 9.9 years at baseline, of whom 53% were boys. The sample was about three quarters White. By age 15, about 35% reported some form of alcohol or other drug use.
Children who initiated substance use showed subtle but identifiable brain differences
Compared with peers with no substance use, children who reported trying any substance before age 15 had larger overall brain volumes as well as greater cortical surface area (see graphic below). At the same time, they showed a thinner prefrontal cortex, particularly in the right rostral middle frontal gyrus, a region crucial for decision-making, planning, and impulse control.
The direction of associations varied by brain region
The direction of associations between observed brain differences and initiation of substance use varied by structure. In children who’d initiated substance use, the prefrontal regions (e.g., dorsolateral prefrontal cortex, superior frontal gyrus) were thinner. Conversely, the occipital, temporal, and parietal regions were thicker, suggesting imbalanced cortical development that might influence how youth process social and sensory cues. Subcortical areas such as the globus pallidus and hippocampus, which are involved in motivation, reward, and memory, were larger in initiators. Taken together, these patterns may suggest a form of neurobiological vulnerability for early substance experimentation marked by reduced inhibitory/regulatory control but heightened reward sensitivity.
These brain differences existed before substance use began
A critical strength of this study was its ability to separate features that predate (i.e., predispose) substance use from possible neurotoxic effects in the brain from alcohol and other drug use. When the researchers restricted analyses to children who had not yet used any substances at baseline (23% of participants had already initiated substance use before their baseline brain scans), the same associations, like a thinner prefrontal cortex and larger global brain volume, still predicted future initiation. This indicates that many of these brain differences were present prior to first use, making them more likely to be preexisting risk markers versus consequences of early alcohol or other drug use.
Substance-specific patterns were modest but distinct
While the overall profile was consistent across substances, some nuanced differences emerged (see graphic below). Alcohol initiation, the most common form of substance use, largely mirrored the general pattern of thinner prefrontal but larger posterior and subcortical regions. Nicotine initiation was associated with smaller right superior frontal gyrus volume and deeper folds in the orbitofrontal cortex, areas involved in craving and valuation of rewards. Cannabis initiation was associated with thinner motor-control areas (precentral gyrus) and smaller right caudate volume, a structure central to habit formation and reward learning. Although these effects were small, they hint at the possibility that differences in brain structure may confer risk for different substances.
Findings held even after accounting for prenatal substance exposure and family factors
Importantly, the associations detected by the researchers remained when analyses controlled for prenatal exposure to alcohol and other drug, familial relationships (e.g., twins and siblings), and MRI scanner differences across sites. This strengthens confidence that the observed brain differences were not artefacts of measurement or family background.
Traditionally, differences in brain structure observed in people who use alcohol or drugs have been interpreted as consequences of substance exposure. However, the results of this large, longitudinal study suggest a more complex picture.
In this research, brain differences linked to early substance use were already present before any use began, implying that certain aspects of brain development may predispose some youth to experiment earlier than others. For instance, thinner prefrontal brain regions—which play a central role in impulse control, planning, and emotion regulation—could make it harder for some children to inhibit impulses or delay gratification in social contexts where substances are available. At the same time, larger subcortical structures and overall brain volumes may reflect developmental patterns associated with greater sensation seeking or reward sensitivity.
It is important to note that these patterns do not mean that brain structure is destiny. The effects for the correlations between these brain characteristics and substance use initiation were quite small, suggesting that some, but not all, youth with these profiles went on to use substances. Also, in terms of the utility, these results cannot yet be applied in clinical settings. For example, we do not yet know the threshold for these neurobiological characteristics (e.g., thinner vs. thicker brain regions) that would accurately classify a child as being at risk for substance use initiation and downstream consequences. Rather, they highlight that differences in neurodevelopment might influence vulnerability, which in turn interacts with other biological, psychological, and social factors. Indeed, this study could not determine why these brain characteristics are present – i.e., developed during childhood as a function of life experience (e.g., trauma, stressors, etc.). That said, the findings held even when controlling for familial risk for substance use – suggesting the risk goes beyond genetic predisposition alone. Supportive family environments, positive peer relationships, and opportunities to develop self-regulatory skills may help buffer against these risks.
For prevention science, these findings may underscore the need for earlier, developmentally informed interventions that target self-regulation, stress management, and healthy risk-taking before adolescence. They also hint that brain-based measures, in combination with behavioral and genetic data, could one day help providers identify individuals most in need of early support.
Finally, the study challenges us to rethink how we interpret neuroimaging findings in addiction research. Not all brain differences seen in youth who use alcohol and other drugs are the result of damage; some may reflect preexisting variations in brain maturation that set the stage for how individuals respond to risk, reward, and social influences during adolescence. Recognizing these developmental nuances can help refine both prevention strategies and public messaging around youth substance use.
This uniquely large national study of brain imaging suggests that subtle differences in brain structure—particularly in regions responsible for impulse control and planning—may predispose some children to some degree to try substances earlier in life. While the thresholds at which these subtle differences predict initiation are not yet clear, ultimately, recognizing and supporting these youth could reduce the likelihood of later substance-related problems and improve developmental outcomes.
Miller, A. P., Baranger, D. A. A., Paul, S. E., Garavan, H., Mackey, S., Tapert, S. F., LeBlanc, K. H., Agrawal, A., & Bogdan, R. (2024). Neuroanatomical variability and substance use initiation in late childhood and early adolescence. JAMA Network Open, 7(12). doi: 10.1001/jamanetworkopen.2024.52027.
SIGN UP FOR THE BULLETIN
151 Merrimac St., 4th Floor. Boston, MA 02114
