Current theory has it that panic disorder, instead of being a conditioned response or a cognitive feedback loop, is actually biological in origin. The idea is that the brain stem's respiratory control center is oversensitive to an excess of carbon dioxide, attributing it to a lack of oxygen. The respiratory control center interprets the excess as extremely serious, like you're drowning or otherwise completely deprived of oxygen, and sends the rest of the mind into panic mode. Perceptibly, it feels like something is terribly terribly wrong, but you are at a loss to figure out quite what that is.

Before this biological explanation, it was assumed that panic attacks came from cognitive feedback, like so: You feel your heart beat being a bit irregular, and are a little concerned about it. Your body senses your concern, and your metabolism speeds up, speeding up your heart. You sense your heart becoming even more irregular, so your anxiety increases, and so forth. The learning theory states that panic attacks happen when the mind associates a particular environment or set of stimuli with the fear response, and thus triggers it when exposed to similar stimuli. A conditioned response to the environment, if you will. An even earlier theory was that panic attacks were caused by "repressed memories" coming to the surface level of consciousness that are so horrible the mind goes into a kind of shock. This was what Freud thought, but most researchers today believe the very notion of repressed memories (and pretty much the rest of Freud) is bunk.

Here are four points that make the biological explanation look like the best one:

  1. Panic attacks often come on during periods of relaxation. During relaxation, one's heart is working less, and is thus is less likely to have perceptible arrhythmias. Also, when one is relaxed, one is usually in a comfortable environment unlikely to cause a panic stimulus. In a relaxed state breathing is slowed, leading to increased levels of CO2, and therefore a (biological) panic attack.
  2. Inhaling CO2 can cause panic attacks in people prone to them, or with a family history of panic disorder. Obviously, breathing a bit of CO2 would cause neither a feedback loop nor a conditioned response in otherwise normal people.
  3. In 80% of patients with panic disorder, sodium lactate given intravenously will bring on a panic attack, as it's converted to carbon dioxide by the body. The same reasons given above apply to this.
  4. Panic disorder can be completely eliminated with an SSRI, and serotonin is responsible for inhibition of the respiratory control center. Panic attacks can be eliminated to the point that neither inhaling CO2 nor injecting sodium lactate will bring on any kind of response at all. If the panic response were somehow cognitively or conditionally related to those stimuli, the response wouldn't (necessarily) be eliminated by these drugs.

It has been theorized that night terrors -- and perhaps (though it's unlikely) even simple nightmares -- are caused by panic attacks that occur during sleep. The brain, given the ordinarily psyche-crushing panic response, but having no cognitive way to deal with it, keeps producing horrible dreams. This goes on until the panic induced neural activity becomes so strong that the dreamer is forced awake, where she finds herself paralyzed with inexplicable fear. This may be the connection between sleep apnea and night terrors.

None of this is to say that cognitive effects cannot bring on a panic attack, just that they aren't the cause of true, DSM-specified panic disorder. Hyperventilation while under extreme stress is an example of cognitive effects precipitating a full-blown panic attack.