Lazar et al. (2000) — Functional brain mapping of the relaxation response and meditation
Not an interoception study, kept as a historical precursor to the wiki’s contemplative-neuroscience thread. Published in NeuroReport in 2000 — seven years before Farb (2007), eight before Lutz (2008) — this is among the earliest fMRI studies of meditation, and it reaches its conclusions with none of the framing (interoceptive attention, predictive coding, the sensory-vs-elaborative dyad) the wiki’s later meditation sources take for granted. It is filed here the way Zeharia and Moffat are: for what it marks, not for a result the wiki relies on.
What was done
Five experienced Kundalini practitioners (≥4 years daily practice, four male, aged 22–45) meditated in a 3T scanner by passively observing the breath and silently repeating a mantra (“sat nam” on inhalation, “wahe guru” on exhalation). The control condition was silently generating a random list of animals without attending to breathing. Two 12-min meditation epochs followed a 6-min control epoch in each 42-min scan; most subjects were scanned twice. Two analyses were run — meditation vs control, and late vs early meditation (steady state vs induction) — and cardiorespiratory measures were recorded in two subjects to check whether physiological change was driving the BOLD signal.
What it found
The primary contrast (meditation vs control) gave increases in pregenual anterior cingulate, putamen, midbrain, and hippocampus/parahippocampus. The late-vs-early contrast added a broad attentional set — dorsolateral and medial prefrontal cortex, superior/inferior parietal lobule (BA7/40), temporal cortex, and pre-/post-central gyri. Amygdala, hypothalamus, septum and caudate were significant in at least three subjects but dropped from the tables because they sat in susceptibility-artifact zones. The authors’ summary reading: meditation recruits attention (frontal, parietal) and autonomic/arousal control (pregenual ACC, midbrain, hypothalamus, amygdala) systems — see central-autonomic-network.
Activation grew with practice: more foci, larger signal change, and more subjects reaching significance in the second scan and in late vs early meditation. The authors read this as a slowly evolving dynamic state, consistent with meditators’ reports that the subjective quality keeps changing throughout a sit.
Two things worth carrying
The insula is absent — and that is not a null. The wiki’s entire interoceptive-neuroimaging tradition centres the insula; this early meditation study does not report it at all. But the paper says explicitly that medial-temporal and subcortical foci near susceptibility-artifact zones could not be localized, and the insula shares that vulnerability at 2000-era resolution. So this is silence, not evidence of absence — a reminder that the insula-centric picture was built with later methods, and that a breath-focused practice can look, through an early scanner, like an ACC-and-autonomic story rather than an insular one. (Compare Farb 2013, which needed attention-controlled contrasts and better resolution to pull the posterior insula out during breath attention.)
The cardiorespiratory-confound caution. The single most-monitored subject showed large global fMRI signal decreases (p<10⁻³⁰) that tracked a meditation-linked drop in respiration rate and end-tidal CO₂ (with rising heart rate and O₂ saturation). Regional increases had to be detected against this whole-brain decrease. Because BOLD is by construction sensitive to cardiorespiratory state, and because meditation reliably changes breathing (see slow-breathing, respiratory-interoception), every meditation-fMRI result inherits this confound — the paper’s argument that its increases are regionally specific (not global), replicate a PET study with no CBF change (Lou et al. 1999), and hold across two subjects with different physiological responses is the template for how the later literature had to defend itself.
Relation to the wiki’s meditation thread
This is the relaxation-response — Herbert Benson’s construct — imaged for the first time. Benson (senior author) had spent three decades characterizing meditation as inducing an integrated hypometabolic state opposite to fight-or-flight; this paper is the attempt to give it a neural map. Read against the wiki’s later, richer sources it is thin (no interoceptive measure, tiny sample, a confounded control), and it makes none of the self-reference or interoceptive-attention distinctions the field later found load-bearing. Its value is as a baseline: the questions the contemplative-neuroscience literature spent the next twenty years sharpening — which system does meditation train, is it interoceptive, how do you rule out the breathing confound — are all already visibly in play here, unsettled. See does-mindfulness-enhance-interoceptive-accuracy, mindfulness-meditation.
(Note: “Gurucharan Khalsa” here is a Kundalini teacher and is not Sahib Khalsa, the interoception methodologist elsewhere in the wiki.)