Mood and anxiety run on thin blood flow.

For the panic and the standing kind, anxiety’s link to blood flow is easy to measure. Over-breathing in a panic attack drops blood CO2 and clamps the brain’s arteries; flow velocity in the main cerebral artery has been clocked falling a fifth to a quarter during a lab-provoked attack, and the drop comes before the racing heart, not after.¹⁸ The same loop runs in orthostatic intolerance: standing cuts brain blood flow twice as much as normal, the shortfall arriving about seventeen seconds ahead of the adrenaline surge.¹⁹ That is why most of these patients end up labelled with an anxiety disorder before the cause is found.¹ Much of what gets called anxiety here is the brain reacting, correctly, to being briefly starved of blood.

That the body can raise the alarm on its own is not a metaphor. Three people whose amygdala had been destroyed, who feel no fear of snakes or haunted houses, still had full panic attacks the moment they breathed in air rich in CO2: the alarm can fire from the chemistry of the blood, below the part of the brain that handles fear.²⁰

There is also a resting version. Trait anxiety does not change whole-brain flow as a single number, but region by region a signal shows: the threat circuitry, the amygdala and anterior insula, runs over-perfused. The largest test scanned 875 young people at rest and found trait anxiety, not momentary state anxiety, tracked elevated amygdala and insula perfusion. The effect was strongest in postpubertal girls, and because resting-flow scans of this kind are sensitive to blood-composition differences between the sexes, that part is best read as one strong study rather than a settled adult map.²¹ Whether the arrow also runs the other way, with chronic low brain blood flow itself feeding anxiety, is plausible but not yet established: in cerebral small-vessel disease, depression and apathy are the robust associations while anxiety is still flagged as needing further study.²²

Depression is a real condition, and most of it has nothing to do with what follows. Alongside that, researchers now read part of depression as a disorder of vessels and inflammation. In older adults, small-vessel brain damage tracks new depression closely enough that researchers call it the vascular depression hypothesis, though the link is so far associative rather than proven cause, and depressed adults, including medication-naive young ones, show dysfunction of the vessel lining that eases when oxidative stress is scavenged, alongside raised inflammation in body and brain.²³²⁴ Blood flow is altered too, region-specific, some areas higher and some lower, and the imaging methods disagree on the overall direction; antidepressants and current symptom severity both move perfusion, so part of what is measured is state, not a fixed injury.²⁵

One famous claim about depression is simply wrong, though: that it is a serotonin chemical imbalance. That was never a blood-flow story, and a 2022 umbrella review found no consistent serotonin-to-depression link.²⁶ The theory failing does not mean antidepressants fail; current models lean on stress, inflammation and neuroplasticity.

Anger is less a surge than a brake slipping. The prefrontal cortex, the orbitofrontal part in particular, holds the amygdala in check, and that brake is metabolically expensive and easily under-resourced. In impulsive aggression the pattern is consistent: the amygdala over-reacts to a provocation while orbitofrontal control drops away, and the normal reciprocal coupling between them is lost.²⁷ Pooled across 43 brain-imaging studies of antisocial and violent people, prefrontal structure and function run measurably lower, concentrated in the orbitofrontal cortex and anterior cingulate. That is the far end of a curve everyone sits on, not the everyday range.²⁸ Pull resources from the same brake and ordinary irritability climbs. One night without sleep makes the amygdala over 60 percent more reactive and cuts its prefrontal regulation,²⁹ and even mild acute stress weakens prefrontal control while strengthening the primitive alarm.³⁰ Irritability is also a real, severity-marking feature of depression³¹ and a common face of bipolar mania,³² and sudden uncontrolled tears or laughter follow stroke in roughly one in six survivors, where the vascular damage itself unsteadies the controls.³³ If anger is taking over, the same things that steady brain blood flow help here too: sleep, easing chronic stress, and treating any mood disorder or vascular risk with a clinician.

Schizophrenia does not lower brain blood flow evenly. The frontal and limbic cortex tends to run underperfused, the hypofrontality that tracks negative symptoms like flat affect, a medium-sized effect rather than a fixed textbook fact, while deeper structures such as the striatum tend to run overperfused, tracking positive symptoms such as hallucinations and delusions.³⁴ There is also molecular evidence of a leakier blood-brain barrier: the tight-junction protein claudin-5 is patchy or broken in most postmortem cases, and the 22q11 deletion that removes one copy of its gene carries a roughly 30-fold rise in risk. The specific gene-variant link, though, is weak.³⁵ Microglial neuroinflammation is plausible but the in-vivo imaging is mixed, so these flow and barrier changes are one strand of a complex disorder. One confound keeps them honest: dopamine-blocking antipsychotics themselves shift blood flow in the very striatal and frontal regions these findings sit in, and most patients scanned were medicated.³⁶ The perfusion changes track metabolism, symptom state and treatment, not a single upstream vascular cause.³⁷

PTSD perfusion is region-specific and reactivity-driven rather than a whole-brain drop. When a trauma cue hits, the most replicated picture is an amygdala that over-responds while the medial prefrontal cortex that should calm it goes quiet, with the size of that reciprocal imbalance tracking symptom severity in the original PET work.³⁸ The prefrontal regulatory failure sets PTSD apart from the other anxiety disorders, hypoactivation of the anterior cingulate and ventromedial prefrontal cortex, rather than a louder alarm.³⁹ At rest, perfusion is mostly preserved and in places even locally elevated rather than uniformly low.⁴⁰ The robust evidence is about reactivity and regional balance, drawn from small, mostly combat-veteran samples in which the amygdala finding was strongest in men, so the perfusion changes follow from an over-reactive threat circuit, not a primary vascular cause.

Dissociation leaves a fingerprint on brain activity. The most replicated finding is a corticolimbic flip: during dissociative states the prefrontal cortex and anterior cingulate ramp up and damp down the emotional limbic system, so the amygdala and insula go quiet. This overmodulation is the signature of the dissociative subtype of PTSD, the mirror image of the hyperarousal pattern.⁴¹ Depersonalization tracks the same frontal and cingulate circuitry, with altered activity in the sensory and body-map regions.⁴² In dissociative identity disorder, a controlled arterial-spin-labelling study found separate identity states carry genuinely different resting perfusion, and an earlier PET study found instructed simulators could not reproduce the identity-state patterns, though those simulators had only brief practice, so this argues against but does not fully exclude role-enactment. Nearly all imaged patients were medicated, which on its own shifts perfusion.⁴³⁴⁴ DID is recognised in DSM-5-TR and ICD-11, and its origins are debated; the imaging does not settle that. The blood-flow data here are sparse, mostly from one research group, and from small medicated samples, so read them as state-markers, not proof that flow drives dissociation. Persistent dissociation is worth raising with a clinician.

Bipolar disorder does change brain blood flow, but mostly during mood episodes: imaging shows altered perfusion in cingulate, frontal and temporal regions during depression and mania, while between episodes the signal is inconsistent rather than reliably normal, so flow looks like a marker of mood state, not a fixed injury.⁴⁵ The sturdier vascular link is what comes with bipolar disorder: small-vessel white-matter lesions are nearly twice as common, and cardiovascular disease strikes about a decade early.⁴⁶⁴⁷ Even vessel-lining function flips with mood, worse in depression and better in hypomania, though that comes from a single preliminary study of teenagers, so the vascular link here is real but secondary and mood-modulated, not the steady, causal small-vessel decline you see in the dementias.⁴⁸

OCD is a circuit problem, not a perfusion one. The leading model is an overactive loop through the orbitofrontal cortex, cingulate, caudate and thalamus, though that single circuit does not capture how varied OCD is. Blood flow does shift there, and the most replicable version is the provoked one: trigger the symptoms and relative flow rises in those regions. Resting baseline differences are far patchier across studies.⁴⁹ But it is a state marker, not a cause. A meta-analysis of fourteen studies and 188 patients found that successful treatment, an SSRI or exposure therapy alike, brings that activity back down as people improve; the effects are small but the caudate drop holds up across the studies.⁵⁰ Two popular claims need a caveat: the serotonin chemical-imbalance story is incomplete, and PANDAS, the strep-triggers-OCD idea, stays scientifically contested. The imaging itself is medication-confounded: in the largest brain-structure study to date, patients and controls separated cleanly only once medication status was accounted for.⁵¹ If it disrupts your life, exposure and response prevention is first-line; see a clinician.

Panic earns its own entry, and not for blood flow. The body-not-mind angle is sharpest here: an attack can start before a single frightened thought, because the brain’s chemical sensors misread an ordinary breath as suffocation. Inhaled carbon dioxide lowers brain pH, and the amygdala detects that acidity through an ion channel and fires the alarm.⁵² The tell that it is its own condition: people with panic spike to a carbon-dioxide breath while people with generalized anxiety react like healthy controls.⁵³ Resting-state imaging exists but is scattered and inconsistent, and the clearer signals are provoked, arousal-driven and downstream of the alarm rather than its cause. If attacks recur, see a clinician; this is highly treatable.

FND is the one people most often get wrong: genuine, involuntary symptoms, weakness, tremor, seizure-like attacks, with no lesion on routine clinical MRI. The modern model calls it a software problem, not a hardware one; the networks that generate, attend to and own movement misfire. The most consistently reported signal is overactive coupling between emotion circuits and motor control, with a disturbed sense of agency; the field’s own reviewers note the studies are still small and inconsistent, and the emotion-circuit signal overlaps with medication and the high comorbidity load.⁵⁴ Blood flow is not the story, and the field’s own reviewers say no scan is diagnostic yet.⁵⁵ Two old ideas are dead: FND is not a diagnosis of exclusion or all in the head, it is confirmed by positive signs like Hoover’s sign, and DSM-5-TR no longer requires a hidden trauma. See a neurologist; physiotherapy and tailored therapy help.

Borderline personality disorder is not really a perfusion story. The established picture is frontolimbic: a reactive amygdala under weak prefrontal control, on a gene-by-trauma foundation. There is a real blood-flow signal, reduced prefrontal and orbitofrontal flow and a blunted prefrontal response to a serotonin challenge,⁵⁶ but it is modest, from small samples, and tracks impulsivity more than the diagnosis. The largest functional meta-analysis, fifty-two studies and over two thousand people, found no robust, consistent brain signature at all.⁵⁷ Confident single-scan explanations, including commercial SPECT pitches, overstate the evidence. If symptoms disrupt your life, structured therapies like DBT have the strongest support.

Anorexia produces some of the most dramatic brain changes in psychiatry: widespread cortical thinning and reduced volume while acutely starved. Blood flow drops too, but it is the smaller, blurrier part of the picture; some of the apparent drop is the atrophic brain itself, less tissue per voxel reading as less flow. Here is the honest part: these are consequences of starvation, not its cause, and blood flow is the downstream marker, not the driver. The reversal is the well-evidenced part: a landmark study found thinning across most of the cortex that normalized after weight restoration, which is why clinicians call it pseudoatrophy.⁵⁸ Across consortium data the cortical effects are large while acutely ill and shrink with recovery. The structural rebound is the confident part; the blood-flow picture is thinner and mixed, so flow is best read as a downstream marker, not a clean on-off switch.⁵⁹ The roots of the illness are genetic and metabolic, not vascular, and recovery is slower in older patients and not always complete. If you or someone you know is struggling, treatment works, and early refeeding protects the brain.

Body dysmorphic disorder is not a perfusion story, it is a perception one. The leading model is perceptual: the brain appears to over-weight fine detail and under-weight the whole face. Medication-free fMRI finds under-activation of the holistic visual system and over-activation of the same orbitofrontal-caudate machinery seen in OCD,⁶⁰ which is why BDD sits in the obsessive-compulsive family and why SRIs help. Blood flow itself is barely studied; the only perfusion data come from a single uncontrolled six-person SPECT scan whose findings were discrepant and non-diagnostic.⁶¹ Despite that, some clinics sell SPECT scans to diagnose BDD. There is no validated perfusion signature, and diagnosis stays clinical. If preoccupation with a perceived flaw is taking over, a clinician can help; BDD is treatable.

Addiction is the clearest case where the honest answer is no, this is not a perfusion story. The established model is neuroadaptive: repeated use floods the mesolimbic dopamine pathway, then blunts it, striatal receptors drop and the prefrontal cortex loses its grip on craving.⁶² That is a wiring-and-motivation problem, not a blood-supply one, and partly a predisposition the drug deepens as much as it creates. Perfusion changes do appear, a dose of heroin lowers flow in the cingulate and insula of people already in heroin treatment,⁶³ but they read as drug effects and markers. The real vascular exception is cocaine, which directly constricts brain vessels and causes strokes, covered under what each substance costs you. Even there, the wanting lives in dopamine circuits, not the bloodstream. If use is harming you, addiction medicine helps.

Seasonal affective disorder is a light-and-clock condition, not a perfusion one. DSM-5-TR does not even make it its own disorder; it is a seasonal-pattern specifier on recurrent depression. The two leading mechanisms are circadian misalignment in short winter days and a seasonal rise in serotonin-transporter binding that runs higher in affected people on PET, though that PET signal comes from one small unreplicated study of twenty patients.⁶⁴ Blood flow is downstream: the one direct perfusion study, just ten people, found frontal and cingulate flow rising only in the light-therapy responders, a sign of getting better rather than a cause.⁶⁵ The popular more-the-farther-north claim is shaky. Light therapy is still a standard first-line option, though the controlled evidence is thinner than its popularity; see a clinician before self-treating.