Attentional bias to threat

The construct that gives Ehlers’s second characteristic of panic patients a home, and the mechanism by which “how well you perceive your body” and “how much you attend to your body” come apart — which is exactly the seam the heartbeat-counting validity debate runs along.

What it is, and why it is not the same as accuracy

Anxious patients across diagnoses — generalized anxiety, OCD, PTSD, simple phobia — preferentially process threat cues related to their disorder, often outside awareness: they respond differently to threatening stimuli without being able to recognize them better than controls (Mathews & MacLeod 1986, via Ehlers 1993). In panic disorder the biased-toward class is physical threat — bodily-harm words, and the body’s own signals.

The distinction the wiki cares about: an attentional bias is not a perceptual ability. Two people with identical cardiac afferent signal can differ in how much attention they spend on it, and the one who attends more will detect more — will cross the threshold into “I feel my heart” more often. Ehlers makes exactly this move to explain why panic patients score better on the Schandry counting task: the group difference “may not reflect differences in the subjects’ general ability to perceive their heartbeats, but differences in the habit of attending to bodily cues.” Attention as the confound between sensitivity and score.

The evidence Ehlers brought

• Modified Stroop: panic patients (and nonclinical panickers) show greater colour-naming interference for physical-threat words than controls — less able to ignore the threatening content.
• Tactile reaction-time paradigm (Ehlers 1989, adapting MacLeod, Mathews & Tata 1986): 61 panic patients, 45 infrequent panickers, 24 specific phobics, 40 normal controls. A finger was vibrated (target); on half the trials a mild shock preceded it (threat). Panickers and phobics — not controls — detected the target faster when it followed a shock, i.e. attention had shifted toward the threat cue. The first direct demonstration of a physical-threat attentional bias in panic — but not specific: phobics showed it too.

Two things it buys the cognitive model

1. It explains the “spontaneous” attack. The cognitive model’s standing embarrassment is that it posits triggers for attacks patients cannot identify. If attention detects an internal trigger without conscious memory of it (Mathews & MacLeod’s out-of-awareness bias), the trigger exists and the patient’s inability to report it is expected rather than anomalous.
2. It closes the loop with anxiety-sensitivity. Believing arousal is dangerous is precisely the kind of standing concern that would direct attention toward bodily and physical-threat cues. Bias and belief are the same disposition seen from the attentional and the evaluative side.

Open

• The bias is not specific to panic — phobics show it, and the primary reviews (Mathews & Eysenck 1987) place it across the anxiety disorders. It characterizes clinical anxiety generally, not the disorder.
• Whether the interoceptive version (attending to the body) and the exteroceptive version (attending to physical-threat words/shocks) are one mechanism or two is untested in anything the wiki has read.
• The construct has an obvious neural candidate — the salience-network, which arbitrates what interoceptive and exteroceptive events get attended — but no source here connects Ehlers’s behavioural bias to it. That bridge is unbuilt.
• Ehlers herself flags that these effects “tended to be small compared to the error variance,” and proposes the paradigms as diagnostic tools (assessing panic triggers or treatment outcome independent of self-report) only after methodological improvement.