Attending to the present: mindfulness meditation reveals distinct neural modes of self-reference

Study at a glance

• Authors: Norman A. S. Farb, Zindel V. Segal, Helen Mayberg, Jim Bean, Deborah McKeon, Zainab Fatima, Adam K. Anderson
• Year: 2007
• Journal: Social Cognitive and Affective Neuroscience (SCAN)
• DOI: 10.1093/scan/nsm030
• N: 36
• Design: fMRI, cross-sectional (pre-training waitlist novices vs post-MBSR trained), 2x2 group x self-focus

Key findings

• Two dissociable neural modes of self-reference: narrative focus (NF, self extended across time) recruits cortical midline structures (mPFC, PCC) and left-lateralized language areas; experiential focus (EF, momentary present-centred self) shifts toward a right-lateralized viscerosomatic network (lateral PFC, insula, SII, inferior parietal lobule).
• In novices the EF>NF dissociation is weak and left-lateralized (dlPFC/vlPFC/posterior parietal, read as linguistic/executive effort); the right-lateralized viscerosomatic EF network appears reliably only in the mindfulness-trained group (mPFC reductions and right lateral PFC / insula / SII-IPL increases significant in MT but not novices, ps > 0.3).
• Functional connectivity: in novices the right insula is strongly coupled to the vmPFC (R = 0.61); in the trained group this coupling is abolished (R = 0.06) and replaced by strong right insula-dlPFC coupling (R = 0.28 to 0.78). The two modes are habitually integrated and become dissociable through attentional training.
• EF in the trained group also reduced left dorsal amygdala activity, alongside the pervasive dorsal + ventral mPFC deactivation.

Limitations

• Cross-sectional, not pre/post: trained participants were scanned after MBSR and compared to a separate waitlist group, so the design cannot show that training per se produced the differentiation (the authors state this outright).
• The ‘experiential’ condition still reads trait adjectives, a narrative-friendly verbal stimulus; NF was rated easier to perform than EF.
• No behavioural responses collected during scanning (to protect momentary EF), so there is no trial-level index of what participants actually did; small n per cell (16 novice, 20 MT).
• The right insula/AIC is treated as the seat of viscerosomatic awareness on the Craig/Damasio/Critchley view of the day; the author’s own later work (Farb et al. (2013)) and Haruki & Ogawa (2023) show attention and arousal complicate that reading.

Connections

• Methods: body-focused contemplative training
• Concepts: Narrative and experiential self-reference, Embodied selfhood, Default mode network, Insular cortex, Decentering, Cognitive reactivity (to mood), Interoception
• Debates: contemplative training increase objective interoceptive accuracy?, Is more interoceptive awareness better?
• Related studies: Farb et al. (2010), Farb et al. (2011), Farb et al. (2022), Farb et al. (2013), Farb et al. (2013), Farb et al. (2015), Lutz et al. (2008)

The founding paper of the dual-mode framework the rest of this wiki’s Farb corpus rests on, and the one the Decentering, Default mode network, Embodied selfhood and every Farb clinical page (2010, 2011, 2022) had been citing secondhand. Read first-hand now the raw/papers/Farb/ exclusion is lifted. It is Farb’s first data paper, with Segal, Mayberg (of subgenual-cingulate/depression fame) and Adam Anderson senior.

The claim

There are two temporally distinct forms of self-reference, and they are neurally dissociable:

• Narrative focus (NF) — the self extended across time, William James’s explanatory “me”: reflecting on what a trait word says about you as an enduring person, allowing elaboration and getting “caught up in a train of thought.” This is the ruminative, self-evaluative mode.
• Experiential focus (EF) — the self in the present moment, James’s agentic “I”: monitoring moment-to-moment experience of thoughts, feelings and sensations without settling on or elaborating any one of them.

The hypothesis, drawn from Damasio, Craig and Critchley, was that momentary self-reference is not just narrative self-reference turned down — it is a distinct mode grounded in “neural markers of transient body states,” in particular right-lateralized interoceptive (insula) and exteroceptive (somatosensory) cortices. See Narrative and experiential self-reference, the concept page this study founds.

Design

36 participants recruited on enrolment in an 8-week MBSR course, scanned either before training (16 novices, waitlist) or after (20 trained) — a cross-sectional split, not a pre/post design. In the scanner they read intermixed positive and negative personality-trait adjectives (from Anderson’s 1968 likeableness norms) under one of two instructions: NF (“what does this say about you as a person”) or EF (“monitor your moment-to-moment experience”). 36-second blocks, no overt responses collected during EF (to avoid disrupting momentary awareness).

Results

Narrative self-focus (both groups collapsed) reproduced the standard self-reference map: rostral and dorsal mPFC (BA 8/9/10/32), posterior cingulate, and left-hemisphere language areas (inferolateral PFC, middle temporal gyrus) — the midline self-referential network, consistent with a linguistically-mediated, higher-order mode.

Experiential vs narrative focus is where the groups came apart:

• Novices showed only weak, restricted midline reductions during EF (subgenual cingulate, PCC, a small rostral-mPFC cluster) and a left-lateralized increase in dlPFC/vlPFC/posterior parietal — read by the authors as linguistic and executive effort (resisting mind-wandering), not the neural signature of present-centred awareness.
• Trained participants showed a pervasive mPFC deactivation (both dorsal and ventral) plus reduced left dorsal amygdala, and a marked shift toward a right-lateralized network: lateral PFC, insula, secondary somatosensory cortex (SII), and inferior parietal lobule — the hypothesized viscerosomatic representations of the momentary body.

ROI analyses confirmed these EF effects were significant in the trained group but not the novices (ps > 0.3), so the right-lateralized experiential network is not a subthreshold version of what novices do — it is qualitatively theirs.

The functional-connectivity result is the mechanistic heart of the paper. In novices the right insula was strongly coupled to the vmPFC (R = 0.61) — viscerosomatic signals by default viewed “through the lens of” the medial prefrontal self-model. In the trained group that coupling was abolished (R = 0.06, a large and significant drop) and replaced by strong right insula-to-dlPFC coupling (R = 0.28 → 0.78). Present-moment awareness, on this reading, requires decoupling the automatic vmPFC response to insular activation — freeing the interoceptive signal from the narrative self-model and routing it instead to a lateral, more detached analytic mode.

Why it matters to the wiki

This is the paper that first states, in neural terms, the sensory/experiential vs elaborative/narrative dyad that every later Farb page turns on:

• Farb et al. (2010) shows the same dyad under an emotional challenge (sad film): training reduces midline elaboration and keeps the interoceptive insula online, with insula recruitment tracking lower depression.
• Farb et al. (2011) and Farb et al. (2022) give it a hard clinical outcome: elaborative mPFC reactivity predicts depressive relapse; sensory reactivity protects. The two-factor model is this 2007 dissociation with survival curves attached.
• Decentering is the psychological name for what the insula-vmPFC decoupling does: experiencing a bodily/mental event as a transient object rather than as a fact about the enduring self.

The honest boundaries

The authors are unusually candid about the design. Because it is cross-sectional, “the evidence does not support a claim that MT per se allowed for the increased differentiation” — only that trained people differentiate the modes where novices do not; the trait-like-vs-state-like question is left explicitly for a pre/post study (which the later papers partly supply). The “experiential” condition still reads verbal trait adjectives, the least present-centred stimulus imaginable, and NF was rated easier than EF, so effort is not perfectly matched despite the ROI controls.

One evolution worth flagging rather than filing as a contradiction: the 2007 paper treats the right insula/AIC as the seat of viscerosomatic awareness, following Craig, Damasio and Critchley. The author’s own 2013 attention study later shows the insula is a graded posterior→anterior tuning field and the anterior insula is “not an area of pure body awareness,” and Haruki & Ogawa (2023) reach the same deflation — the right AIC may be about arousal rather than the body in general. The 2007 finding stands; its interpretation of which insular subregion does what is the earlier, coarser reading.