Generalized Anxiety Disorder in the Brain

Generalized anxiety disorder is worry that runs most days for six months or more, across several parts of life, with physical signs such as restlessness, fatigue, poor concentration, irritability, muscle tension and disturbed sleep.

Also called: GAD, Generalised anxiety disorder, Chronic anxiety.

ICD-10 F41.1 · ICD-11 6B00

A brake problem, not an engine problem.

Almost everyone has been told the same story. Anxiety is an overactive alarm. The amygdala has a hair trigger, and it fires at things that are not really threats.

It is a good story. It is also not what the brain scans show in this particular condition. What separates generalized anxiety from ordinary fear is not how loud the alarm is. It is that the alarm never gets switched off.

Here are five findings, from the least settled to the most useful.

The overactive alarm

Evidence: contested

This is the explanation everyone has been given. In GAD, it does not hold up.

The amygdala is the brain's threat detector. In most anxiety disorders it clearly reacts too strongly. Show someone with a phobia or post-traumatic stress disorder, or PTSD, a frightening face and it lights up harder than it should.

In generalized anxiety disorder, the finding will not sit still. Some studies find the amygdala overactive. Some find it overactive even to ordinary, neutral images, which is a different claim. Some find it UNDERactive to fearful faces. Some find no difference at all.

That mess is itself the result. It suggests GAD is not simply a stronger version of the fear disorders. This is anxiety rather than fear. It is waiting for something rather than reacting to it. Those two things run on different machinery.

This does not mean the amygdala is irrelevant. It means we should stop offering "your amygdala is overactive" as the explanation. Researchers have not even agreed that it is true here.

The wrong dial

You can spend a long time turning the volume down on a radio. It will not help if the real fault is that the radio will not switch off.

Why the myth is sticky

It is the only part of this that makes a clean picture. A bright spot on a scan is something you can point at. "Waiting for something that never arrives" is not.

Anticipation that never resolves

Evidence: proposed

This is the finding that really separates this condition from fear.

Threat responses come in two kinds. One is short and sharp, and it answers something that is here now. The other is long and low, and it answers something that might be coming. They run on different parts of the brain.

The amygdala handles the short one. A different area handles the long one. It is called the BNST, short for bed nucleus of the stria terminalis. It is a small region sitting next to the amygdala. It deals with steady, uncertain threat rather than immediate danger.

People with GAD show raised BNST activity while waiting for a threat. The threat has not arrived, and it may never arrive. This is not a spike of activity. It is a hold.

This is why "calm down" fails so badly. There is nothing here to calm down from. The system is braced against something that has not happened. So nothing that happens can stand it down.

A sprint and a heavy load

Fear is a sprint. It is hard, and then it is over. Anxiety is carrying something heavy with nowhere to set it down.

The brake that does not hold

Evidence: established

This is the most repeatable finding in GAD. It is about a connection, not about a part.

The prefrontal cortex controls the threat system from above. It is the part that can look at an alarm and decide it is not worth acting on.

In GAD, the link between the amygdala and the prefrontal cortex is weaker. This is found in adults and in teenagers. The weaker the link, the higher the anxiety scores.

Be clear about what that means. No region is damaged. Both regions work. The connection between them is weak. That is why it shows up as a control problem, and why you cannot see it on an ordinary scan of brain structure.

This weak connection is currently the most promising biological marker for GAD. That sentence is worth reading twice. The best biological handle anyone has on this condition is a measure of how badly two parts are talking to each other.

Why this is the hopeful one

Connections can be trained. Lost tissue cannot. This is the finding that makes therapy a real mechanism rather than just a comfort.

Excitation over inhibition

Evidence: proposed

This is the same failure, one level further down.

GABA, or gamma-aminobutyric acid, is the brain's main quietening signal. Glutamate is the main signal that fires things up. A stable brain state is a balance between the two.

In anxious states that balance is tipped. The signals coming up from below beat the control coming down from above. Where there is less GABA compared with glutamate in the prefrontal cortex, anxiety runs higher.

This is the level that benzodiazepines act on. That is why they work within the hour. It is also why they are a problem rather than a solution. They restore the quietening artificially. The brain then adapts, so the same dose does less. Stopping becomes much harder than starting.

Worry as the strategy

Evidence: proposed

This explains why someone in pain keeps doing the thing that hurts.

Worry does not feel good, so it looks like a fault. One model argues the opposite. It is called the contrast avoidance model, and it says worry is doing a job.

Worrying holds you in a steady bad state. From down there, bad news is a much smaller drop. Nothing can knock you down if you never got up. The cost is that you never got up.

The body agrees. In GAD, heart-rate variability is reduced, and returning to a normal resting state is measurably hard. The real sign is not agitation. It is a loss of range. The system has stopped flexing because it is holding.

This is also why reassurance fails. Reassurance removes the uncertainty for a few minutes. That feels like a reward, and the reward trains the asking. The worry then returns, and the habit is slightly stronger than before.

Bracing for the pothole

Brace for every pothole and no pothole will ever jolt you. You will also arrive aching, having braced for a hundred that were never there.

The same thing, without the vocabulary

The alarm is probably not louder than anyone else's. It simply never switches off. It is set for something that might happen, not for something that is happening.

The part of the brain that should switch it off is connected too weakly to do that reliably. And the chemical balance that does the quietening is tipped the wrong way.

Worry keeps going because it is working. It holds you low, so nothing can drop you. That is a real benefit. You pay for it with everything else.