Heartbeat-related spectral power as a Neural Marker of Meditative Depth
Abstract
Background. Meditation is thought to enhance well-being through sustained interoceptive attention and emotion regulation, yet neural markers tracking depth as it unfolds within a session remain elusive, limiting mechanistic insight and neurofeedback guidance. Heartbeat-evoked responses (HERs), cortical activity time-locked to each heartbeat, offer a beat-by-beat window onto interoceptive processing, but have been compared only between groups or conditions, not across graded depth within practice. We asked whether HERs track online depth reports in experienced Vipassana practitioners, a tradition coupling sustained bodily attention with non-reactivity. Methods. Thirty-nine experienced practitioners completed two sessions of 35-minute silent meditation with simultaneous 64-channel electroencephalography (EEG), electrocardiography (ECG) and respiration, reporting depth in real time on an ordinal scale (1–5). We derived three HER measures across the time (heartbeat-evoked potential, HEP), time-frequency (heartbeat-related spectral power, HRSP) and phase (heartbeat-related phase consistency, HRPC) domains, testing each in the scalp electrodes and independent component domains. Inverse variance-weighted, within-participant depth slopes were assessed with hierarchical general linear modeling, nonparametric statistics, robust correction for family-wise errors, surrogate heartbeats, leave-one-participant-out (LOSO) refits and confound models. Results. With increasing depth, HRSP decreased from 20 to 348 ms after the R-peak at 10–24 Hz in a posterior component (peak 188 ms, 15 Hz; p = .0018, Hedges’ g = −0.70). The effect exceeded all 50 surrogates, held in every LOSO refit, and survived adjustment for cardiac and respiratory activity and, in a smaller region, for ongoing beta power; autonomic measures did not change with depth. Neither the HEP nor HRPC tracked depth in the primary analysis, and at the scalp only small regions reached threshold, in three of nine settings. Participants with stronger decreases showed larger reductions in depressed mood and scored higher on trait observing (exploratory; q ≤ .011). Frontal-midline theta increased modestly with depth, at the scalp in the silent blocks (d = 0.37) and in a spatially filtered component when passive auditory oddball blocks were included (d = 0.48), opposite in direction to the heartbeat-related effect. Conclusions. HRSP offers a novel, ecologically valid group-level correlate of intra-session meditative depth that conventional HEP analyses did not detect. Rather than amplifying the heart’s cortical signal, deepening meditation appears to quiet the induced activity that elaborates on each beat, consistent with a relaxing of the self-located, predictive frame in which cardiac signals are processed.
Related
- Salience network expansion as a transdiagnostic neural marker for psychopathology risk in adolescence
- Deep phenomics of human PBMCs using a 31-marker spectral flow cytometry panel
- Learning Environment-Associated Marker Rankings with Attention-Based Graph Neural Networks
- Genome-Wide Marker-Trait Analysis of Agronomic and Biomass-Related Traits in an Interspecific Miscanthus F₁ Population Grown on Reclaimed Mine Land.
- Improved detection and spatiotemporal spectral analysis of neural traveling waves
BibTeX
@misc{cannard2026heartbeatrelated,
title = {{Heartbeat-related spectral power as a Neural Marker of Meditative Depth}},
author = {Cedric Cannard and Mihir Nath and Valeria Becattini and Ruben Eero Laukkonen and Saampras Ganesan and Adam Hanley and Andrew Holecek and Ninette Simonian and Michael Lifshitz and Felix Schoeller},
howpublished = {OSF Preprints},
year = {2026},
month = nov,
doi = {10.31219/osf.io/hpcgd_v3},
url = {https://osf.io/hpcgd_v3/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Heartbeat-related spectral power as a Neural Marker of Meditative Depth
AU - Cedric Cannard
AU - Mihir Nath
AU - Valeria Becattini
AU - Ruben Eero Laukkonen
AU - Saampras Ganesan
AU - Adam Hanley
AU - Andrew Holecek
AU - Ninette Simonian
AU - Michael Lifshitz
AU - Felix Schoeller
PY - 2026
DA - 2026/11/03
PB - OSF Preprints
DO - 10.31219/osf.io/hpcgd_v3
UR - https://osf.io/hpcgd_v3/
AB - Background. Meditation is thought to enhance well-being through sustained interoceptive attention and emotion regulation, yet neural markers tracking depth as it unfolds within a session remain elusive, limiting mechanistic insight and neurofeedback guidance. Heartbeat-evoked responses (HERs), cortical activity time-locked to each heartbeat, offer a beat-by-beat window onto interoceptive processing, but have been compared only between groups or conditions, not across graded depth within practice. We asked whether HERs track online depth reports in experienced Vipassana practitioners, a tradition coupling sustained bodily attention with non-reactivity. Methods. Thirty-nine experienced practitioners completed two sessions of 35-minute silent meditation with simultaneous 64-channel electroencephalography (EEG), electrocardiography (ECG) and respiration, reporting depth in real time on an ordinal scale (1–5). We derived three HER measures across the time (heartbeat-evoked potential, HEP), time-frequency (heartbeat-related spectral power, HRSP) and phase (heartbeat-related phase consistency, HRPC) domains, testing each in the scalp electrodes and independent component domains. Inverse variance-weighted, within-participant depth slopes were assessed with hierarchical general linear modeling, nonparametric statistics, robust correction for family-wise errors, surrogate heartbeats, leave-one-participant-out (LOSO) refits and confound models. Results. With increasing depth, HRSP decreased from 20 to 348 ms after the R-peak at 10–24 Hz in a posterior component (peak 188 ms, 15 Hz; p = .0018, Hedges’ g = −0.70). The effect exceeded all 50 surrogates, held in every LOSO refit, and survived adjustment for cardiac and respiratory activity and, in a smaller region, for ongoing beta power; autonomic measures did not change with depth. Neither the HEP nor HRPC tracked depth in the primary analysis, and at the scalp only small regions reached threshold, in three of nine settings. Participants with stronger decreases showed larger reductions in depressed mood and scored higher on trait observing (exploratory; q ≤ .011). Frontal-midline theta increased modestly with depth, at the scalp in the silent blocks (d = 0.37) and in a spatially filtered component when passive auditory oddball blocks were included (d = 0.48), opposite in direction to the heartbeat-related effect. Conclusions. HRSP offers a novel, ecologically valid group-level correlate of intra-session meditative depth that conventional HEP analyses did not detect. Rather than amplifying the heart’s cortical signal, deepening meditation appears to quiet the induced activity that elaborates on each beat, consistent with a relaxing of the self-located, predictive frame in which cardiac signals are processed.
ER -