Neurohemodynamic correlates of 'OM' chanting: A pilot functional magnetic resonance imaging study

Deep Investigation

Context

Before this 2011 study, claims about OM chanting and "vagal stimulation" were largely theoretical — rooted in the anatomical observation that the vagus nerve has auricular branches that pass near the ear canal, and that vibrations from chanting are felt around the ears. But nobody had actually looked inside the brain during OM chanting to see if vagus nerve-like effects were occurring.

Kalyani's team at NIMHANS (National Institute of Mental Health and Neurosciences, Bangalore — India's premier neuroscience institution) used fMRI to do exactly that. Their hypothesis was specific: if OM chanting stimulates the vagus nerve via its auricular branches, then the brain's hemodynamic response should mirror what's seen during clinical transcutaneous vagus nerve stimulation (tVNS). They designed a clever control: the sound "ssss" — which matches the expiratory duration of OM but produces no vibration around the ears.

The result was striking: OM chanting produced significant bilateral limbic deactivation including the right amygdala (the brain's threat detector), while "ssss" produced absolutely nothing. The deactivation pattern was remarkably similar to what Kraus et al. (2007) had documented during tVNS — a clinical procedure used for treatment-resistant depression and epilepsy.

For NeuroNest Research Hub, this is the neuroimaging anchor. When we discuss vagal tone and parasympathetic activation across other posts, this paper provides the "here's what it actually looks like inside the brain" evidence. The clinical VNS comparison lets you draw the line between a $30,000+ implanted medical device and a practice that costs nothing.

Methodology deep-dive

• Design: Pilot fMRI study, within-subjects design with active control condition
• Subjects: 12 right-handed healthy volunteers (9M, 3F), mean age 28±6 years. Screened by TWO independent psychiatrists to exclude: psychiatric diagnoses, family history of serious mental illness (first-degree), pregnancy/postpartum, substance abuse, neurological disorders, and MRI contraindications.
• Protocol: 10 consecutive cycles of: REST (15s) → OM chanting (15s) → REST (15s) → "ssss" pronunciation (15s). Total scan ~10 minutes.
• OM chanting specification: Vowel "O" for 5 seconds, consonant "M" for 10 seconds. Subjects trained by experienced yoga teacher. Loud chanting chosen deliberately to (a) objectively confirm task performance via audio monitoring and (b) provide vibration sensation around ears for auricular vagal stimulation.
• Control condition ("ssss"): Continuous production of "sssss..." for 15 seconds — matched for expiratory duration but WITHOUT vibratory sensation around the ears. This is the key design decision — it isolates vibration from simple exhalation.
• Imaging: fMRI with BOLD contrast. Analysis via Statistical Parametric Mapping 5 (SPM5).
• Statistical approach: Fixed-effects analysis with a priori regions of interest (limbic structures). Family-wise Error Correction (FWE) at P<0.001 — stringent threshold.
• ROIs: Amygdala, hippocampus, parahippocampal gyrus, insula, orbitofrontal cortex, anterior cingulate cortex, thalamus. Masks created using WFU Pickatlas.

Sound protocol specifics

• OM structure: 5 seconds "O" (vowel/open sound) + 10 seconds "M" (nasal consonant/humming). Total 15 seconds per cycle.
• Breathing rate: Each OM cycle = one breath (inhale before OM, exhale through OM). At 15 seconds per cycle, this approximates 4 breaths/min — even slower than the 6/min in Bernardi's rosary study.
• Frequency/pitch: NOT specified. No acoustic measurement of the OM frequency. This is a notable gap — was it a low hum (~100-150 Hz) or higher chanting (~200-300 Hz)? The frequency would affect which structures vibrate and potentially which vagal branches are most stimulated.
• Intensity (dB SPL): NOT specified. "Loud" chanting was used but no decibel measurement.
• Posture: Supine (lying in MRI scanner) — different from normal seated or standing chanting position, which could affect respiratory mechanics and vagal tone.
• From Dion's sound engineering perspective: The "M" phase is essentially a 10-second sustained hum — the same basic acoustic event as in Bhramari pranayama and the same type of vibration that Weitzberg & Lundberg showed pumps NO from the sinuses. The vibratory sensation at the ears is created by bone conduction of the humming sound through the skull. The "ssss" control is acoustically brilliant — it's a fricative that creates airflow noise but no tonal vibration, so no bone conduction to the ear region.

Key findings (beyond the headline)

• OM vs REST: Significant bilateral deactivation in orbitofrontal cortex, anterior cingulate cortex (ACC), parahippocampal gyri, thalami, and hippocampi. Right amygdala also significantly deactivated. No significant activation observed anywhere.
• "ssss" vs REST: Neither activation NOR deactivation in any of these regions. This is the critical null finding — it proves the effect is specific to OM's vibratory component.
• Pattern match to clinical VNS: The deactivation pattern (amygdala, hippocampus, parahippocampal gyrus, ACC) closely mirrors findings from Kraus et al. (2007) who studied transcutaneous vagus nerve stimulation using fMRI, and Henry et al. (2004) who studied cervical VNS in epilepsy patients via PET imaging.
• Right amygdala specificity: Only the right amygdala showed significant deactivation. This lateralization is interesting — the right amygdala is more associated with rapid threat detection and negative emotional processing.

What the authors didn't say

• Fixed-effects analysis is a major limitation: It means results apply only to THIS specific group of 12 individuals and cannot be generalized to the broader population. A random-effects analysis would be needed for group-level inference. This is a pilot study, not a definitive finding.
• N=12 with 9 males: Severely underpowered by modern fMRI standards. Most current neuroimaging studies require N≥20 minimum. Sex ratio is skewed.
• No autonomic measures alongside fMRI: They hypothesise vagal stimulation but don't simultaneously measure HRV, skin conductance, or any peripheral autonomic marker. The "vagal" interpretation is inferred from the brain pattern, not directly measured.
• No experienced vs naive comparison: All subjects were yoga-naive. Would experienced practitioners show stronger or different patterns?
• Supine posture: Lying in an MRI scanner is not how people normally chant. Supine position itself affects autonomic balance (increases vagal tone relative to standing), which could confound the baseline.
• No frequency/pitch data: Without knowing what frequency the subjects hummed at, the vibration transmission pathway through the skull to the auricular vagal branches remains somewhat speculative.
• The causal chain is incomplete: They show OM → limbic deactivation and note it LOOKS LIKE VNS → limbic deactivation. But "looks like" is not "works through the same mechanism." The vibrations could affect the brain through other pathways (e.g., interoceptive processing of the vibratory sensation, respiratory coupling, or cognitive/attentional effects of the chanting task).
• Journal impact: International Journal of Yoga is not a high-impact journal. However, the paper's citation record (~250+ citations) and its consistent appearance in subsequent neuroimaging reviews validates its influence despite the venue.

Cross-references in NeuroNest Research Hub

• Autonomic complement: Inbaraj et al. 2022 — provides the HRV data that this paper lacks (increased HF-HRV power after OM chanting)
• Acoustic mechanism: Weitzberg & Lundberg 2002 — the NO pathway (what the "M" phase does to the sinuses)
• Cardiovascular synchronization: Bernardi et al. 2001 — the respiratory slowing and baroreflex pathway
• HRV during humming: Trivedi et al. 2023 — Bhramari (essentially the "M" phase sustained) produces lowest stress index
• Clinical VNS comparison papers: Kraus et al. 2007 (fMRI + tVNS) and Henry et al. 2004 (PET + cervical VNS)
• Combined narrative: This is the "brain side" of the vocal sound story. Weitzberg gives you the respiratory/NO side, Bernardi gives you the cardiovascular/baroreflex side, Trivedi gives you the HRV side, and Kalyani shows what's happening centrally — limbic quieting via vagal afferents.

7-Dimension score

Note on tier: Like Weitzberg & Lundberg, the scoring reflects methodological limitations of a pilot study. Content treatment is Gold-tier because the neuroimaging evidence is irreplaceable — no other paper shows this specific brain pattern during vocal chanting with an appropriate control condition.