Can a Brain Scan Diagnose ADHD?
If ADHD is a neurodevelopmental condition with a biological basis, it seems reasonable to wonder why we can’t see it on a brain scan.
We can image the brain in remarkable detail. Researchers can measure aspects of its structure, observe patterns of activity and study how different regions communicate. Studies comparing people with and without ADHD have also identified average differences in a number of these measures.
So why can’t someone simply have a scan to find out whether they have ADHD?
The answer lies in an important distinction between identifying biological patterns across groups of people and being able to diagnose one individual accurately.
Brain imaging has taught us a great deal about ADHD. At present, though, it can’t reliably tell us whether the particular person sitting in front of a clinician has the condition.
What researchers can see
Several forms of brain imaging are used in ADHD research.
Structural magnetic resonance imaging, or MRI, can measure characteristics such as the volume and thickness of different areas of the brain. Functional MRI, or fMRI, can examine changes in blood flow associated with brain activity while someone is resting or completing a task. Other imaging methods allow researchers to investigate different aspects of brain function and chemistry.
When researchers compare sufficiently large groups of people, patterns can emerge. Studies have reported average differences associated with ADHD in aspects of brain structure, development, activity and connectivity, including systems involved in processes such as attention, cognitive control, movement and reward.
These findings form an important part of our understanding of ADHD as a neurodevelopmental condition. What they don’t provide is a diagnostic image of an individual ADHD brain.
To understand why, we need to look at what an average difference between groups actually means.
A group difference isn’t the same as a diagnostic marker
Imagine researchers measure one feature of the brain in a thousand people with ADHD and another thousand people without it. When they compare the results, they find that the average measurement differs between the two groups.
That difference can be scientifically meaningful even if the individual measurements overlap considerably.
Some people with ADHD may have measurements close to the average of the group without ADHD. Some people without ADHD may have measurements that fall well within the range seen among people with the condition. With sufficiently large samples, researchers can detect relatively small average differences even when looking at one person’s result wouldn’t tell you which group they belonged to.
A diagnostic test has a much more demanding job. It needs to help determine whether an individual person has the condition accurately and reliably enough to guide clinical decisions.
That requires more than evidence that two groups differ on average. A useful biomarker needs to distinguish individuals with and without the condition with sufficient accuracy, and ideally continue to do so across different ages, populations and clinical settings.
Current brain-imaging measures haven’t reached that point for ADHD.
There isn’t one ADHD pattern on a scan
Part of the difficulty is that ADHD itself is highly varied.
People can meet diagnostic criteria through different combinations of symptoms, and they differ in their cognitive profiles, development, genetics, environments and co-occurring conditions. It would therefore be surprising if every person with ADHD showed precisely the same biological pattern.
Neuroimaging research reflects that variation. Researchers haven’t identified one region that is consistently abnormal in every person with ADHD, a particular brain shape or size that appears only in people with the condition, or one pattern of activity that clinicians can reliably recognise as ADHD.
Current research is increasingly interested in whether multiple biological patterns or developmental pathways may contribute to the diagnosis. That may eventually help us understand why ADHD can look so different from one person to another.
For now, though, it also helps explain why the search for one universal ADHD scan has been unsuccessful. The condition itself isn’t biologically uniform enough to produce the simple signature we might hope to find.
What about images showing the “ADHD brain”?
If you’ve searched for information about ADHD online, you may have seen striking images showing an ADHD brain beside a non-ADHD brain. One might contain noticeably more colour than the other, or a particular region may appear smaller or less active.
Without context, these images can make the biological difference look unmistakable.
What you’re seeing, however, may be an average across many participants, a statistical comparison between groups or a particular measurement taken under specific research conditions. It isn’t necessarily the equivalent of placing one typical person with ADHD in a scanner, placing one typical person without ADHD beside them and photographing the difference.
The colours need interpretation too. They may represent statistical values, changes in blood flow or another measurement selected by researchers rather than simply showing parts of the brain that are on or off.
Research images can be enormously informative when read in the context of the study that produced them. Removed from that context, they can look much more diagnostically decisive than they really are.
A smaller region doesn’t mean a deficient person
Structural imaging findings can create a similar misunderstanding.
Large studies have identified average differences in the size or thickness of some brain regions when groups of people with ADHD are compared with groups without it. These findings can help researchers investigate how neurodevelopment may differ on average.
They don’t mean that a slightly smaller measurement represents a damaged or deficient part of an individual person’s brain.
Brain regions also don’t map neatly onto single behaviours. A region involved in aspects of attention will usually contribute to many other processes, while attention itself depends on activity and communication across multiple brain systems.
This is why we can’t take a research finding about an average structural difference and use it to explain why one particular person forgot an appointment or lost concentration during a meeting. The research finding and the individual experience sit at very different levels of explanation.
Functional scans don’t solve the problem either
If structure can’t give us a diagnostic signature, it might seem as though watching the brain in action should.
Functional imaging has identified differences between ADHD and comparison groups during tasks involving attention, inhibition, working memory and other cognitive processes. Researchers have also examined patterns of connectivity while people are resting.
This work has contributed significantly to models of the neural systems associated with ADHD, but functional activity is influenced by many things. The task being performed matters, as do age, medication status and characteristics such as sleep and alertness. Studies also differ in the people they recruit, the tasks they use, the way scans are collected and the methods used to analyse the data.
All of this makes it difficult to turn a pattern detected in a research study into a stable signature that can identify ADHD reliably in one individual.
The scan can tell researchers something valuable without being able to answer the clinical question, Does this person have ADHD?
Could artificial intelligence change this?
This is an active and interesting area of research.
Machine-learning methods can analyse very large amounts of imaging data and look for combinations of features that might be difficult for a human researcher to detect. Some studies have used these methods to investigate whether computers can distinguish people with ADHD from those without it, with promising results reported in particular datasets.
The difficult part is making sure those results continue to work outside the dataset in which the model was developed.
A system might perform well with people recruited for one study and then become less accurate when tested with scans from a different hospital, scanner, age group or population. There is also an important difference between distinguishing a carefully selected ADHD research group from healthy participants and working in an ordinary clinic, where the real question may be whether someone’s difficulties are better explained by ADHD, anxiety, autism, depression, sleep problems, another condition, or some combination of them.
For an AI-supported imaging test to become clinically useful, it would need to remain sufficiently accurate across those much messier real-world situations.
That research is continuing, but we don’t currently have such a test.
A scan can still be useful during an ADHD assessment
None of this means brain imaging has no place in medicine when someone is being assessed for ADHD.
There may occasionally be a separate reason for a clinician to recommend imaging. Particular neurological symptoms or aspects of someone’s medical history, for example, might warrant further investigation.
In that situation, the scan is being used to investigate another possible condition or explanation. It isn’t being used to confirm ADHD itself.
For most people undergoing an ADHD assessment, routine brain imaging isn’t required because there is currently no established imaging result that can confirm or exclude the diagnosis.
So how is ADHD actually diagnosed?
ADHD remains a clinical diagnosis, which means the clinician builds a picture from several kinds of information rather than relying on one scan, blood test or laboratory result.
They consider the person’s current symptoms and the way those symptoms affect everyday functioning, including whether relevant difficulties occur across more than one setting. Because ADHD is neurodevelopmental, developmental history is also important, including evidence that relevant symptoms were present during childhood.
The clinician also considers whether another condition or circumstance might better explain what is happening and whether other conditions may be present alongside ADHD.
Questionnaires and structured interviews can contribute to the assessment, while information from parents, partners, school reports or other sources may sometimes provide useful context.
No single piece of information is expected to prove ADHD. The clinician is looking at whether the overall developmental and functional pattern meets established diagnostic criteria.
That can feel less concrete than a scan
For some people, this part can be uncomfortable.
A brain scan feels objective in a way that a conversation about childhood, attention, forgotten responsibilities and everyday functioning may not. If you’ve spent years questioning your own experience, the possibility of an image that could finally settle the matter can be particularly appealing.
But clinical doesn’t mean arbitrary.
ADHD assessment uses established diagnostic criteria and draws on developmental history, evidence of symptoms and impairment, information across different areas of life and clinical judgement. Evidence-based guidelines set out how assessments should be conducted and the different explanations clinicians need to consider.
Like many areas of medicine, ADHD assessment has limitations. That is different from saying the diagnosis rests simply on someone’s impression.
A scan might feel more concrete, but a test isn’t more scientifically useful simply because it produces an image. It has to answer the diagnostic question accurately.
What if someone offers a scan that claims to diagnose ADHD?
This is where it’s worth being cautious.
Commercial services sometimes promote brain imaging as a way of identifying ADHD, determining someone’s ADHD type or selecting treatment. The claims can feel persuasive precisely because the technology is sophisticated and the images look compelling.
The existence of sophisticated technology, though, doesn’t establish that every proposed use of it has been scientifically validated.
Before a biological test becomes part of routine diagnosis, researchers need evidence that it can accurately distinguish individuals with the condition, adds useful information beyond established assessment methods and continues to perform reliably across different clinical populations.
Current ADHD guidelines don’t recommend routine neuroimaging as a diagnostic test for ADHD. A service claiming that a scan can definitively show whether an individual has ADHD is therefore offering a level of certainty that the current evidence doesn’t support.
The absence of a scan doesn’t make ADHD less real
Sometimes there is no brain scan for ADHD becomes the starting point for a very different claim: If ADHD were really biological, surely we would be able to see it.
That assumes biological conditions must produce one distinctive marker that can be detected in every person who has them. Biology is often considerably more complicated.
ADHD involves multiple genetic and developmental influences rather than one discrete abnormality. The brain differences identified in research tend to be distributed, variable and overlapping rather than forming one diagnostic signature shared by everyone with the condition.
Whether we can diagnose something using a biological test therefore depends on more than whether biology is involved. We also need to have identified a marker, or combination of markers, sufficiently specific and reliable to distinguish individual people accurately.
For ADHD, we haven’t yet done that. This tells us about the current limits of diagnostic technology, not about whether ADHD has biological foundations.
What brain imaging is actually good for
Brain imaging remains enormously valuable precisely because research and diagnosis aren’t the same job.
Neuroimaging allows researchers to investigate questions that clinical observation alone can’t answer. It can help us understand how aspects of brain development differ on average, examine networks associated with particular cognitive processes and explore how genetic variation might relate to neurodevelopment. It may also help researchers understand why ADHD is so heterogeneous and whether biologically meaningful patterns exist within the broad clinical diagnosis we use today.
Over time, combinations of biological, cognitive and clinical information may also contribute to better assessment or more individually tailored treatment. Those possibilities are worth investigating.
The important distinction is that a research method capable of teaching us about ADHD doesn’t automatically become a clinical test capable of diagnosing it.
You don’t need to see ADHD on a scan for your experience to count
If you’re exploring whether ADHD might be part of your experience, it’s understandable to want something more tangible than a history made up of attention difficulties, restlessness, forgotten responsibilities, childhood patterns and the ways everyday life has been affected.
But those aren’t second-rate pieces of evidence that clinicians use only because neuroscience hasn’t yet produced something better. They are central to what ADHD is and how it is currently diagnosed.
Whether ADHD is the right explanation for your experience requires careful assessment. A brain image can’t currently settle that question for you. If you already have a diagnosis, the fact that there is no scan displaying your ADHD doesn’t make that diagnosis less substantial.
Neuroscience and clinical assessment are simply looking at ADHD at different levels. Brain imaging helps researchers understand patterns across brains and populations. Clinical assessment asks whether the developmental pattern experienced by this particular person meets the criteria for ADHD.
They can both contribute to our understanding without needing to do the same job.
A moment to wonder
When you imagine proof that ADHD is real, what kind of evidence feels most convincing to you?
Perhaps a brain image feels more certain than a history made up of experiences, patterns and memories. You don’t need to talk yourself out of wanting something concrete.
For now, you might simply notice whether understanding the difference between research evidence and an individual diagnostic test changes what you expect a brain scan to be able to tell you.
Keep exploring
There’s more to explore across Through the Disconnect, from understanding ADHD and life after diagnosis to masking and making sense of your experience.
References
American Psychiatric Association. (2022). Diagnostic and statistical manual of mental disorders (5th ed., text rev.; DSM-5-TR). American Psychiatric Association Publishing.
Australian ADHD Professionals Association. (2022). Australian evidence-based clinical practice guideline for attention deficit hyperactivity disorder (ADHD).
Faraone, S. V., Bellgrove, M. A., Brikell, I., Cortese, S., Hartman, C. A., Hollis, C., Newcorn, J. H., Philipsen, A., Polanczyk, G. V., Rubia, K., Sibley, M. H., & Buitelaar, J. K. (2024). Attention-deficit/hyperactivity disorder. Nature Reviews Disease Primers, 10, 11.
Koirala, S., Grimsrud, G., Mooney, M. A., Larsen, B., Feczko, E., Elison, J. T., Nelson, S. M., Nigg, J. T., Tervo-Clemmens, B., & Fair, D. A. (2024). Neurobiology of attention-deficit hyperactivity disorder: Historical challenges and emerging frontiers. Nature Reviews Neuroscience, 25, 759–775.
Rana, B., et al. (2025). A systematic literature review of machine learning techniques for the detection of attention-deficit/hyperactivity disorder using MRI and/or EEG data. Neuroscience.
Faraone, S. V., Banaschewski, T., Coghill, D., et al. (2021). The World Federation of ADHD International Consensus Statement: 208 evidence-based conclusions about the disorder. Neuroscience & Biobehavioral Reviews, 128, 789–818.
Hoogman, M., Bralten, J., Hibar, D. P., et al. (2017). Subcortical brain volume differences in participants with attention deficit hyperactivity disorder in children and adults: A cross-sectional mega-analysis. The Lancet Psychiatry, 4(4), 310–319.
Hoogman, M., Muetzel, R., Guimaraes, J. P., et al. (2019). Brain imaging of the cortex in ADHD: A coordinated analysis of large-scale clinical and population-based samples. American Journal of Psychiatry, 176(7), 531–542.
