ELI5 If ADHD means your brain is different, why can’t it be tested for using things like functional MRIs?

Read the Story

Show Top Comments

#1 reason that we don’t use MRIs and fMRIs diagnositically is money. It costs researchers like $600 minimum to run an MRI (I don’t know what patients or insurers get charged, but in my lab which owns its own scanner we still have to pay hundreds of dollars per hour to run it). That’s just the cost for the scanner, not including pay for the tech who runs it or the doctor who reads it. But beyond that, there are scientific reasons too. Right now, scientists are trying to find signature “biomarkers” that could be used to diagnose, but we can’t assign humans to be ADHD, so the best we can do is compare groups. Group differences are meaningful, but don’t actually tell you what individuals will look like. For example, on average, men are taller than women. But if that’s all I know, and you tell me that patient A is 5’7″ tall, I can’t tell you 100% whether that person is a man or a woman. Additionally, measuring the brain comes with a lot of caveats. The most common fMRI technique is Blood Oxygen Level Dependent (BOLD) signal. This doesn’t directly measure brain function. Instead, it measures where the most oxygen is being used in the brain. We assume the brain wouldn’t send oxygen to a region that isn’t active, and that the proportion of oxygen is proportional to the activity, but we really have no proof of that. There’s also been some research showing that the difference between patient A and patient A on another day can in some cases be more different than patient A vs patient B, so there’s still a lot of debate about how reliable fMRIs are. Structurally, there’s nothing in the brain where 100% of people with have some trait but 0% of people without ADHD have that trait. Remember, we can’t see the actual cell types/shapes/sizes without cutting out a piece of the brain, so the variables we’re working with are mostly region volume, cortical thickness, and some white matter measures that can be difficult to interpret even for white matter researchers.


fMRI is a surprisingly unreliable technique. It looks for extremely tiny changes in the image, and makes a lot of assumptions about how the body and blood flow works. Basically, fMRI marks a region of the brain as active, if it detects more oxygen in that area of the brain. Now, you’d think that if the brain was active, it would be using more oxygen, so there would be less oxygen left in that bit of brain. But what it seems happens is that the blood vessels overcompensate for the oxygen use, and they deliver more oxygen than is needed. So, fMRI assumes that the blood flow is normal and overcompensating. The second issue is that the changes are so tiny that they are very difficult to measure – it requires taking hundreds of scans during some sort of task (e.g. saying rhyming words), hundreds of scans when resting (doing nothing), and then merging all those hundreds of scans, and looking for differences which are consistent across hundreds of scans. For research purposes, it is quite common to scan dozens of people, each having hundreds or thousands of individual scans – and then comparing “group 1” against “group 2” as a whole. This allows you to detect even weaker and fainter signals – because you can average across dozens of volunteers. For diagnosis purposes, you can only detect things which are strong enough to be detected in a single person, in a sensible period of time (you couldn’t do thousands of scans, it would take too long). This means that for diagnosis purposes even “obvious” things can’t be reliably detected. For example, I regularly do fMRI to tell a brain surgeon where not to operate (because that bit of brain is too important). However, fMRI is not reliable enough to do things like locate the parts of the brain which handle speech. Speech areas do show up on the fMRI, but the edges are often wrong, so the surgeon needs to double check during the operation (the surgery is done awake, and the surgeon gives electric shocks to an area of the brain before cutting it to check what it does). However, the fMRI is reliable enough to tell if someone has the speech parts of the brain on the left, right or both. So, if the patient needs right sided surgery, but the fMRI shows left sided speech, the surgeon can generally operate without double checking.


I read an article recently arguing whether or not ADHD and some other disorders are disorders at all. It’s possible that it is not your brain that is unsuitable for this environment, but the other way around, that the environment that we have in part created is unsuitable for the many types of cognitions that humans have. The main reason I see benefit in this approach is that you wouldn’t have to be labelled “sick/disordered” in order to receive personalized accomodations and support as needed. That being said, coming back to your original question, if there is in fact nothing explicitly “wrong/disordered” with your brain, then there would be nothing to look for in various medical imagery one could perform. P.S. I really don’t want to antagonize anyone who feels like this is a minimization of their disorder, if so, I apologize


So something that I want to parrot from what my Psychiatrist told me that really helped me understand the same thing (why can’t we just do a brain scan). Obviously the comments about the technology not being reliable are absolutely true but another reason is this: Nearly everywhere in medicine you’re usually more interested in the cause, and the symptoms help diagnose that cause, for example: My leg hurts, I just fell over, there is swelling, I may have broken a bone. For treatment the last part is the most important thing, so they know how to approach. However, the brain is extremely complex and we don’t really understand it all too well even now. There are so many more layers of complexity in what can change, a bone will never break because you’ve failed to reward it for not breaking as an example. But with a brain, that can absolutely be the case, even things like drugs can work better/worse depending on what your brain thinks that they are going to do etc etc. Because our understanding of the brain is so primitive, when it comes to a lot of issues we can’t really talk about the cause (either because we don’t know, or because it can be caused by a whole lot of different issues). So we have to talk about the symptoms, you don’t have ADHD because your brain has issues with Dopamine uptake – you have ADHD because you have a list of symptoms that fall within what we understand to be ADHD such as: – Lack of executive function (basically being unable to start/continue tasks) – Hyper fixation – General issues with focus There are a ton more (obviously) but you get the point. The interesting thing about this is that to use ADHD as an example, some people have nothing wrong with how their brain processes or provides Dopamine, but still have ADHD, because they still have all the symptoms. That is STILL ADHD. Same story with depression (it can be chemical, it can be life events, it can just happen), but it is all valid depression. This is also why most neurodivergent cases can be even just hard to treat. For example, some people can resolve their symptoms with nothing but cognitive therapy, some gain little/no benefit from cognitive therapy at all, but can take Dex for a year and then have no problems. Most people benefit from both and it is often lifelong (for those who are still showing signs in adulthood). Rather thematically, this is all over the place but here is the TL;DR: There are tons of reasons that neurodivergency can present, ranging from no reason at all, to chemical, to life situation (and everything in between). Our understanding is so bad that we have to talk about these vastly different medical issues based on the symptoms that they present, because we can’t diagnose the exact cause with any accuracy (and every brain is different). This makes any scanning (as of writing) not very useful, and it even makes treatment hard.


I think the ELI5 is – they could, but it would cost a lot of money and they’re not sure if it would work so they don’t.