ADHD Isn’t One Condition—Brain Scans Suggest Three Types
Brain imaging study identifies three biologically distinct ADHD types linked to genetics and treatment response.
If every child with attention-deficit/hyperactivity disorder (ADHD) presents differently, why does the medical community continue to treat the diagnosis as a single category?
A new study, conducted at West China Hospital, Sichuan University, used brain network modeling on over 1,150 participants and identified three specific ADHD biotypes linked to different genetic risks and treatment outcomes.
The biological limits of the current ADHD diagnosis
ADHD is a common neurodevelopmental diagnosis in children; however, its presentation is notoriously inconsistent. Two children with the same label may have entirely different daily experiences, with one struggling with focus and another with hyperactivity. This variability, or heterogeneity, suggests that a single category is too broad.
Doctors currently rely on the DSM-5, which uses behavioral checklists to make a diagnosis. These lists track symptoms such as inattention, but they rarely reflect the biological processes happening in the brain.
“The current diagnostic framework assigns a single diagnostic label to what is fundamentally a heterogeneous syndrome that likely arises from diverse neural mechanisms,” said the authors of the new study.
Previous research has often relied on group averages, which mask individual differences that make each child's experience unique. Earlier attempts to group children by symptom clusters, such as “Inattentive” vs “Combined” types, did not solve the problem. These categories often change over time and do not accurately predict if a child will respond to a specific treatment.
By using normative modeling and morphometric similarity networks, the new study aimed to move beyond behavioral checklists. The team worked to identify stable biological profiles to see how a child's brain structure deviates from a typical developmental path.
Normative modeling
A framework that allows researchers to understand atypical brain features by comparing an individual's data to population expectations, quantifying how much an individual deviates from the normative variation of a specific phenotype.
Morphometric similarity networks
Individualized models used to map how different regions of the brain share physical traits.
Using brain network modeling to map ADHD biotypes
The researchers looked at structural MRI scans from over 1150 children to create a growth chart for brain connectivity, similar to how a pediatrician tracks height and weight. Instead of looking at one brain region at a time, they mapped “morphometric similarity”—a technique that measures how different brain areas resemble each other in physical traits such as thickness and volume.
By using an AI algorithm called HYDRA clustering, the team grouped participants based on how their brain networks deviated from neurotypical scans.
“We identified three distinct topology-derived biotypes, each characterized by unique clinical-neural profiles, longitudinal trajectories, and spatial molecular signatures,” said the authors.
- Biotype 1: showed widespread, global changes in brain networks and was linked to the most severe symptoms. These children also showed a higher likelihood of symptoms persisting as they aged.
- Biotype 2: showed more localized changes, especially in deep brain areas called subcortical regions. These children had moderate symptoms.
- Biotype 3: had minimal changes and cognitive profiles that looked much like those of children without ADHD.
The team validated these types by associating them with genetic risk factors, brain chemicals, and medication responses, confirming that the three groups represented real biological differences.
Ask the Expert: Next-Generation Sequencing
In this interactive Ask the Expert session, Dr. Deborah Cragun will answer your questions on how Next-Generation Sequencing is advancing research and clinical applications.View Ask The Expert
Brain patterns were linked to specific gene expression through spatial transcriptomics, and the team compared the biotypes to neurotransmitter maps, finding correlations with chemicals such as serotonin and dopamine.
Different biotypes also responded differently to stimulant medications, supporting the idea that the categories are based on biology instead of only behavior.
Improving clinical outcomes through personalized ADHD care
Moving toward precision psychiatry changes how ADHD is treated. Instead of examining every diagnosis the same way, doctors can look at the specific biology of a child’s brain, moving ADHD care away from a one-size-fits-all diagnosis. The research suggests that a brain scan could one day help predict which children need the most support.
“Our comprehensive approach offers a promising framework for parsing the inherent ADHD heterogeneity in a clinically valuable way, which may ultimately create a path toward developing personalized therapeutic strategies,” said the authors.
Longitudinal studies are needed to follow these children as they grow into adults to see if these brain categories remain the same and whether the biotypes stay stable over many years.
Future research could also test if these maps can help pick the best medication for a child; for example, doctors might use this data to decide between stimulants or non-stimulant options. This evidence-based approach aims to make treatment more effective and less of a guessing game.
Reference: Pan N, Long Y, Qin K, et al. Mapping ADHD heterogeneity and biotypes by topological deviations in morphometric similarity networks. JAMA Psychiatry. 2026. doi: 10.1001/jamapsychiatry.2026.0001
Add Technology Networks as a preferred Google source to see more of our trusted coverage.
