Negative self-referential bias (NSB)

The construct at the centre of Wu et al. (2026): a hallmark of depression in which negative self-relevant information is preferentially endorsed, remembered, and elaborated over positive information. It is depression’s self-schema made measurable — the recovered depressive who, under a small drop in mood, more readily accepts worthless than worthy as a description of themselves.

How it is measured

Behaviourally, NSB is read off the Self-Referential Encoding Task: participants judge whether positive and negative trait adjectives describe them, and the ratio of negative to positive self-endorsement is the bias. Neurally (in the fMRI version), it is the Self-Negative minus Self-Positive contrast — greater dysphoric activity during negative self-encoding — measured in Wu et al. within a priori frontal DMN (dmPFC) and salience (ACC) regions and across the whole brain.

The dissociation Wu et al. (2026) found

The paper’s sharpest result is that the behavioural bias and the neural signature predict different clinical outcomes:

• Behavioural NSB → concurrent symptoms, not relapse. More negatively biased endorsement went with higher residual depressive symptoms (b=0.10 [0.06, 0.14], p<.001) and, longitudinally, with less decentering — but had no direct association with relapse. The reportable bias tracks how depressed you are now.
• Neural dysphoric NSB activity → relapse association. Greater frontal DMN and salience-network dysphoric self-referential activity was associated with relapse status (though it did not independently predict relapse timing in survival models), while the whole-brain markers accompanying self-referential processing (subgenual/lateral-occipital activation, somatosensory deactivation) did predict relapse.

The reading Wu et al. draw: behavioural NSB “may primarily reflect a conscious, reportable tendency that tracks concurrent symptom burden without necessarily driving episode recurrence,” whereas the neural activity patterns “may more closely reflect internal amplification and maintenance of negative self-relevant content that contributes more directly to relapse risk.” A self-report bias and the brain state underneath it are complementary, not redundant markers of depression vulnerability.

Relation to cognitive reactivity

NSB and cognitive-reactivity are cousins, not synonyms. Cognitive reactivity is the mood-triggered reactivation of an entire depressive processing style (rumination, negative interpretation, pessimism), measured by how much dysfunctional thinking a sad-mood provocation switches back on. NSB is the narrower self-schema bias — which trait words you accept as self-descriptive — measured on the SRET. In Farb (2011) the relapse-predicting reactivity was self-referential mPFC engagement to a sad film; Wu et al. sharpen this by showing that the behavioural self-referential bias is not itself the prospective marker, and by locating the prognostic signal in prefrontal activation and somatosensory deactivation instead.

Where it sits in the dual mode

NSB is the negative pole of self-referential valence, and in Wu et al.’s conjunction analysis it maps onto the elaborative side of the narrative-vs-experiential dyad: self-criticism recruited a broad prefrontal/posterior elaborative set (orbitofrontal/insula, DLPFC, PCC), while self-affirmation recruited posterior somatosensory regions. Negative self-evaluation is prefrontal elaboration; positive self-relation is somatosensory embodiment. This is why NSB belongs to the two-factor model: the protective counterweight to negatively biased self-elaboration is not merely less negative thinking but more embodied sensory self-representation — and decentering (which tracks lower NSB) is the trainable shift between them.