Brain imaging has found real average differences between people with depression and people without it, and it still cannot tell you whether any particular person is depressed. Both halves of that sentence are true at once, and holding them together is the whole of what this article is for.
The gap between them is not a technology problem waiting to be solved next decade. It is a statistical fact about what a small average difference can and cannot do.
What the scans actually show
The most consistent structural finding is a slightly smaller hippocampus. A coordinated analysis by the ENIGMA consortium pooled MRI data from thousands of people with major depression and thousands of controls across fifteen research samples, and found lower hippocampal volume in patients, with the effect driven by those with recurrent depression rather than those in a first episode. [schmaal-enigma-mdd]
Other findings appear across the literature: differences in cortical thickness in some regions, and altered activity in circuits involving the prefrontal cortex and the amygdala in functional studies. They are less consistent than the hippocampal result, and their size varies with age, medication, illness duration and how the samples were assembled.
Notice what the strongest finding is. Not a marker present in depression and absent otherwise, but a shift in the average of a measurement that varies enormously between healthy people anyway.
Why an average difference cannot diagnose you
Because the two distributions overlap almost entirely. This is the point where intuition fails people, so it is worth making concrete.
Men are taller than women on average, by a substantial margin, and that difference is far larger than anything imaging has found in depression. You still could not look at one person’s height and state their sex with confidence, because plenty of women are taller than plenty of men. A group difference tells you about groups.
A schematic of the overlap described in this article, drawn to show the shape of two distributions rather than measured volumes.
The second view is the one that matters. The overlap is so large that for the great majority of individuals the measurement adds nothing to what the clinical interview already established, which is why no diagnostic manual or treatment guideline lists a scan among the criteria for depression. [dsm-depression-scans]
Why clinics sell them anyway
Because a picture of a brain is far more persuasive than a questionnaire, and persuasion is what sells. This is a documented concern rather than a suspicion: commentary in the psychiatric literature has criticised the marketing of functional imaging such as SPECT for diagnosing and treating psychiatric conditions, on the basis that the claims made for it are not supported by the evidence. [adinoff-unfounded-scans]
Two things are worth knowing before spending money on one.
| What is offered | What it actually is |
|---|---|
| ”An objective test for your depression” | No such test exists, and none appears in any guideline |
| ”See what is really happening in your brain” | A group-level research measure applied to one person |
| ”A personalised treatment plan from the scan” | Treatment selection is not guided by imaging in any protocol |
| ”Rule out other causes” | Blood tests and a history do this; imaging is for suspected neurological disease |
The genuine medical uses of brain imaging in this area are real and different: a scan is appropriate where a neurological cause is suspected, such as sudden personality change, new seizures, or first-onset psychiatric symptoms in later life. That is ruling something else out, not diagnosing depression.
The same thing happened to “chemical imbalance”
That phrase followed exactly this pattern, and watching it happen is the fastest way to understand why the scan claim should be treated carefully.
It began as a serviceable shorthand for explaining why a medication acting on serotonin or noradrenaline might help. It was never a measurement: no test has ever established an individual’s neurotransmitter levels for the purpose of diagnosing depression, and the condition involves stress, circumstance, sleep, inflammation, genetic liability and learning history alongside anything happening at a synapse. The shorthand outran its evidence, got repeated until it sounded like a finding, and eventually had to be walked back in public.
The important part is what the walking back did not change. Antidepressants did not stop working when the slogan was retired, because their evidence never rested on it, and depression did not become less real. A simplified explanation being wrong is not the same as the thing it was explaining being imaginary, and a great many people were unnecessarily unsettled by that conflation.
Apply the same test to any scan you are offered. Ask what the claim is, what measurement supports it, and whether that measurement was taken on groups or on people. Most of what gets sold in this area fails at the third question.
What the finding does change
Two things, and neither is diagnostic. It is evidence that depression is not a failure of character, which matters to a great many people even though it should not have needed proving. And the recurrence-linked pattern is one of several reasons clinicians take relapse prevention seriously rather than treating each episode as a standalone event.
What it does not change is the treatment. The psychological and pharmacological treatments with the best evidence were tested and shown to work without reference to anybody’s hippocampal volume, and no imaging result currently tells a clinician which of them to pick. Our guides to what depression actually is and CBT for depression cover what does inform that choice, and the depression guide covers the wider picture.
When to seek help
See a doctor if low mood has lasted more than two weeks, if sleep or appetite have changed, if you have lost interest in things you used to enjoy, or if you cannot work or care for people who depend on you. Bring the history rather than an image: when it started, what changed around then, and what you have stopped doing. A completed depression self-assessment is more useful at a first appointment than anything a private scan would produce.
Ask about psychological therapy and medication by name. Both have a substantial trial record, and neither requires imaging to be chosen or to work. Go sooner if the low mood arrived with a sudden change in personality, new physical neurological symptoms, or for the first time in later life, since those are the circumstances where a doctor may genuinely want to image something.
Contact your local emergency services or a crisis helpline if you feel unsafe or have thoughts of harming yourself.
How MyFreud can help
MyFreud is a mobile app that helps you find solutions to problems that have affected your mind and productivity. Daily tracking gives a clinician the thing an image cannot, which is a record of how your mood, sleep and activity have actually moved over weeks rather than a reconstruction produced under pressure in a ten-minute appointment. Live coaching sessions end with an actionable plan rather than an explanation of your brain, which is the part that changes anything. The notepad is where the history for that first appointment goes, in the order it happened.
Download MyFreud and start today: App Store or Google Play.