Narrative and experiential self-reference
The distinction the whole Farb corpus is built on, given its own page because it is cited on more pages than any single Farb study and is the conceptual root of the wiki’s sensory-vs-elaborative dual mode. Introduced in Farb et al. (2007) and carried forward, under shifting names, through every later Farb paper.
The two modes
The framing goes back to William James’s split between the “me” (the self as a known object, a source of permanence woven from narrative across time) and the “I” (the agentic self acting in the present moment):
- Narrative self-reference (narrative focus, NF) — evaluating experience against an enduring self-concept, elaborating on what an event means about you as a person, linking experiences across time. Neurally: the cortical midline structures — medial prefrontal cortex and posterior cingulate — plus left-lateralized language areas. This is the default-mode-network’s self-referential mode, and the seat of ruminative cognitive-reactivity.
- Experiential self-reference (experiential focus, EF) — attending to moment-to-moment experience (thoughts, feelings, bodily sensations) as transient events, without elaborating any one into a story about the self. Neurally: a right-lateralized network of lateral PFC, insula, secondary somatosensory cortex and inferior parietal lobule — the viscerosomatic representation of the present body.
The mechanism: coupled by default, decoupled by training
The load-bearing finding of the founding study is not the two maps but the connectivity between them. In untrained novices the right insula is strongly coupled to the vmPFC (R = 0.61): interoceptive signals are, by default, viewed through the lens of the narrative self-model, which is why the two modes are hard to separate. Mindfulness training abolishes that coupling (R = 0.06) and re-routes the insula toward lateral PFC — freeing the bodily signal from the self-story. So the modes are “habitually integrated but can be dissociated through attentional training.” This decoupling is the neural form of what decentering names.
How the distinction travels through the corpus
The names change but the dyad is constant:
| paper | narrative / elaborative pole | experiential / sensory pole | what’s added |
|---|---|---|---|
| 2007 | mPFC narrative self-reference | right insula/SII/IPL momentary self | the dissociation itself + insula-vmPFC decoupling |
| 2010 | midline + language elaboration of sadness | right insula kept online under sad film | emotional challenge; insula tracks lower depression |
| 2011 | mPFC reactivity predicts relapse | sensory (visual) reactivity protects | prospective clinical outcome |
| 2022 | dynamic prefrontal reactivity biomarker | static sensory-deactivation biomarker | RCT; two-factor-model-of-relapse-vulnerability |
| 2026 | self-criticism → prefrontal/OFC/DLPFC elaboration | self-affirmation → posterior somatosensory embodiment | the dyad found inside self-referential valence; somatosensory pole is the protective relapse marker |
The through-line: elaboration is pathogenic, present-centred observation is protective. The 2007 paper is where that shape is first drawn, in healthy meditators with no clinical outcome — the proof-of-concept that the two modes are neurally real before they are shown to matter for depression. Wu et al. (2026) find the same split within the valence of self-reference: using the SRET, self-criticism recruited the elaborative pole (orbitofrontal/insula, left DLPFC, PCC) while self-affirmation recruited the somatosensory embodiment pole (postcentral, supramarginal, precuneus, superior parietal) — and it is that somatosensory pole whose deactivation forecasts relapse and whose treatment-related recovery protects. The dyad is no longer only narrative task vs experiential task; it is also negative vs positive self-relation, drawn on the same elaboration-vs-embodiment axis.
The interoceptive reading, and its honest boundary
Farb’s later interoception work generalizes the experiential pole to interoceptive awareness — attending to the body as the exemplary present-moment object. The boundary the wiki keeps visible (see decentering, norman-farb) is that the protective sensory pole is only sometimes interoceptive: it is the right insula in 2007 and 2010, but exteroceptive vision in 2011 and somatosensory/visual in 2022. The dual-mode structure is robust across the corpus; whether its sensory pole is specifically interoceptive is the generalization, not the finding.
One evolution to hold: the 2007 identification of the right anterior insula as the seat of viscerosomatic awareness is the Craig/Damasio-era reading, later complicated by Farb (2013) (the insula is a graded attentional tuning field) and Haruki & Ogawa (2023) (the right AIC may track arousal rather than the body in general). The two-mode dissociation does not depend on that identification.
The dyad, decoded rather than contrasted
Weng et al. (2020) arrive at the same dyad from the opposite methodological direction. Their EMBODY framework (embody-framework) trains classifiers on three meditation-relevant states — breath (the experiential/interoceptive pole), self-referential processing (operationalized as thinking about the past week / present moment regarding oneself, i.e. the narrative pole of this concept), and mind-wandering (the DMN-associated movement between states) — and decodes which one a meditator occupies, second by second. Where Farb establishes the dissociation by contrast (narrative vs experiential task blocks) and by insula–vmPFC decoupling, Weng treats the two poles as decodable brain states and reads the traffic between them during uninstructed practice. Weng et al. cite Farb’s dual-mode and interoception work as load-bearing; the two lines are measuring one dyad from two ends — one by design, one by decoder. The caveat is the decoder’s: its self-referential class is defined by the experimenter’s training labels, and the meditation-period read-out has no ground truth.