Cognitive reactivity (to mood)

The construct that names the “Week 11: Absence of Emotion” theme: a recovered depressive can feel fine and still be at high relapse risk, because the vulnerability is not in their mood but in what a small drop in mood switches back on. Introduced to the wiki with Farb et al. (2011) and its senior author Segal, whose differential-activation lineage (Teasdale) it descends from.

The idea

In a person who has been depressed before, a mild, transient dysphoria does not stay mild. It reactivates the whole processing pattern of the original episode — ruminative self-focus, negative interpretive bias, a globally pessimistic reading of the self and the future. The reactivation is the mechanism of relapse; the triggering mood dip is merely its cue. Two people can have the same small sad mood and only one of them re-opens the depressive pattern. That difference — measurable by provoking a sad mood and seeing how much dysfunctional thinking comes back — is cognitive reactivity, and it predicts relapse up to 18 months out (Segal et al. 2006), independent of current symptoms.

What Farb et al. (2011) added

A neural signature, and a sharpening of what “reactivity” is:

• The relapse-predicting reactivity is in self-referential cortex (BA 24/32), reacting to a sad film — and it tracks trait rumination.
• Felt sadness does not predict relapse (mood ratings r = −.007, ns), but the mode of processing it does. Cognitive reactivity is therefore about elaboration, not intensity — the failure to just perceive an emotional stimulus without turning it into self-evaluation. In relapsers, dorsal (task) and ventral (self) mPFC — normally anticorrelated — become coupled, a “cannot disengage from self-referential processing” signature.
• It is a marker of risk, not regulation. The intuitive reading (frontal engagement = effortful coping) is wrong here: this frontal reactivity forecasts collapse. In predictive-coding terms it looks like active-inference-style elaboration/regulation that fails — the maladaptive twin of perceptual inference.

What Farb et al. (2022) added: it is the treatment-modifiable factor

The RCT sequel Farb et al. (2022) scanned 85 patients before and after prophylactic therapy and split relapse vulnerability into two (the two-factor model). Cognitive reactivity is the dynamic half: prefrontal (left-DLPFC) reactivity to sad film fell over the treatment period in patients who stayed well and did not fall in patients who relapsed (Relapse x Time p=.022), and failure to reduce it predicted relapse (HR 3.73). This is the sharpest evidence yet that cognitive reactivity is not a fixed scar but a movable target — and that lowering it is (part of) what prophylactic therapy does. Two refinements matter: the reactivity here is lateral PFC (elaboration/regulation), not only the medial self-referential mPFC of 2011; and it is trans-therapeutic — MBCT and a well-being reappraisal therapy reduced it equally, so cognitive reactivity is a common mechanism, not a mindfulness-specific one. The other (static) factor, sensory deactivation, treatment did not move at all.

What Wu et al. (2026) added: the reportable bias is not the prognostic one

Wu et al. (2026) scan the same trial cohort during a self-referential encoding task (rather than the sad film) and sharpen what “reactivity” forecasts. The behavioural negative self-referential bias — the conscious tendency to endorse negative over positive self-traits — predicted concurrent residual symptoms (b=0.10, p<.001) but not relapse. The prospective signal was neural: greater dysphoric frontal-DMN/salience activity was associated with relapse, and whole-brain markers (subgenual/lateral-occipital activation, somatosensory deactivation) predicted it. So the self-report bias indexes present burden; the relapse-forecasting reactivity is the brain activity underneath it, on the prefrontal-elaboration-vs-somatosensory-embodiment axis. This is consistent with 2011 (felt sadness didn’t predict relapse; the mode of processing did) — extended now to show that even a reportable self-schema bias is not the prognostic marker, only its neural substrate is.

The protective opposite

Cognitive reactivity has a counter-pole, and it is what the wiki’s contemplative material trains: decentering — seeing the dysphoric thought as a transient event rather than a cue to elaborate. In Farb et al. (2011) the protective mode is sensory (visual) reactivity tracking trait acceptance; MBCT (Segal) was built specifically to lower cognitive reactivity by cultivating decentering. This is where the depression-relapse literature meets mindfulness-interoceptive-training and mindfulness-meditation.

The earliest Farb paper on the wiki, Farb et al. (2010), supplies the training-side complement to the 2011 prediction: MBSR completers responding to a sad film activated the midline self-referential + language elaboration network less than untrained controls, at equal felt sadness, while keeping the interoceptive insula online. So the elaborative reactivity that 2011 shows forecasts relapse is, in 2010, exactly what mindfulness training turns down — the same construct read at both ends of the clinical arc (what predicts relapse, and what treatment moves). The design is cross-sectional, so it is an association, not a demonstration that the training caused the reduction.

Relation to the wiki’s other “reactivated prior” ideas

Cognitive reactivity is a clinical, mood-triggered instance of a pattern the wiki keeps meeting: a prior that reshapes present experience. It rhymes with background-somatic-states (a pre-existing somatic state filtering subsequent ones by congruence, so negative states breed pessimism — the somatic-marker framework’s model of mood) and with schema-guided-symptom-perception (perception driven by past experience rather than present physiology). All three are versions of the past overwriting the present read-out — cognitive reactivity is the version where the overwrite is a whole depressive processing style and the stakes are relapse.