Mothers and Babies Show Heart Rhythm Synchrony With Surprising Time Lags

When a mother and her baby play together on a mat, something remarkable appears to be happening beneath the surface, in the quiet rhythm of their beating hearts. A new exploratory study from Tohoku University in Japan suggests that the heart rate variability of mothers and infants can fall into tight alignment during ordinary free play, and that this alignment almost never occurs at exactly the same moment. Instead, the physiological echoes between mother and child arrive with temporal offsets of up to tens of seconds, in either direction, a finding that could reshape how scientists measure the hidden physiology of human connection.

The research, published in Physiological Reports, focused on high-frequency heart rate variability, or HF-HRV, a measure derived from electrocardiogram recordings that reflects beat-to-beat fluctuations in heart rhythm within a specific frequency band. HF-HRV is closely associated with respiratory sinus arrhythmia, the natural speeding and slowing of the heart with breathing, which in turn is linked to parasympathetic, or vagal, nervous system activity. Because this branch of the autonomic nervous system is thought to support calmness, social engagement, and emotional regulation, HF-HRV has become a popular window onto the physiological states of caregivers and children during interaction.

Previous studies of physiological synchrony have typically averaged heart rate variability over fixed time windows and compared those averages between partners. Such approaches can detect gross similarity but may miss a crucial feature: the possibility that one person’s physiological rhythm precedes or follows the other’s by a meaningful interval. Eye contact, vocal exchanges, and physical touch between parents and children can change within seconds, and the physiological consequences of those moments may ripple forward or backward in time. The Japanese team, led by Aoi Sakurai, set out to build and test a method that explicitly accounts for these temporal shifts.

The study recruited 15 mother-infant dyads with healthy infants aged five to seven months, recruited between December 2022 and March 2023 through local childcare newspapers and social media in Sendai City. Each pair was scheduled for two recording sessions roughly two weeks apart. Mothers were asked to feed and change their infants beforehand to ensure calm, and then both wore small wireless ECG sensors, positioned at the upper sternum for mothers and near the clavicular region for infants, while spending five minutes in unrestricted free play on a yoga mat, talking, touching, and playing as they naturally would.

Turning raw infant ECG into usable data proved to be the study’s greatest technical hurdle. Of the 30 recorded sessions, 14 were excluded because of device failures, session interruptions, or electrode detachment and movement artifacts, leaving 16 analyzable sessions from 10 unique dyads. The researchers processed the surviving signals with a custom MATLAB pipeline that combined automated QRS detection with careful visual inspection and manual correction, flagging suspicious beats as missing values rather than interpolating them. Artifact rates in the retained data were remarkably low, at roughly 0.29 percent for mothers and 0.02 percent for infants, a testament to the labor-intensive cleaning process.

The core of the analysis was a lag-sensitive cross-correlation technique. The team computed HF-HRV power in 30-second windows advanced in half-second steps, using frequency bands of 0.15 to 0.40 hertz for mothers and 0.24 to 1.04 hertz for infants, reflecting developmental differences in breathing rate. For each maternal reference window, the infant’s signal was shifted across a search range spanning minus 30 to plus 30 seconds, and the correlation between the two traces was calculated at every offset. When a positive correlation coefficient of 0.8 or higher emerged as a local peak, that moment was marked as a synchronization point, classified as Lag-plus if the infant’s pattern aligned after a positive offset, Lag-minus if the mother’s pattern aligned after a negative offset, or Lag-zero for exact alignment.

The results were striking. Highly correlated intervals accounted for a mean of 74.5 percent of the total analyzed interaction time, a figure well above the variable or lower synchrony levels often reported in previous mother-infant studies. More intriguing still was the distribution of lags: within the highly correlated time, Lag-minus intervals made up about 54.4 percent and Lag-plus intervals about 44.6 percent, while exact zero-lag alignment accounted for a mere 1.1 percent. In other words, under this analytical definition, mother-infant heart rhythm synchrony was almost always offset in time, and the direction of the offset varied both across sessions and within them, with some dyads showing a bias toward one direction over limited periods.

Because high correlations between two slowly fluctuating biological signals can arise by chance from autocorrelation and shared measurement conditions, the team also compared their observations against a null model. For each session, they generated 100 surrogate datasets by randomly permuting the order of 30-second segments within the maternal and infant time series, destroying the original correspondence while preserving local structure. Observed synchrony exceeded the session-specific surrogate distribution in 13 of the 16 sessions, with a mean observed-minus-surrogate difference of about 0.205, and the positive difference held across all 10 dyads in a sensitivity analysis. This suggests that the alignment was more than coincidental temporal structure, though the authors are careful to frame the surrogate test as exploratory, since it relied on a single null model and a modest number of iterations.

The theoretical backdrop is Polyvagal Theory, which positions the vagal system as central to physiological co-regulation in close relationships. The findings are consistent with the idea that caregivers and infants mutually regulate one another’s autonomic states, and they echo earlier reports of heart rhythm synchrony between pregnant women and their fetuses, between marital partners, and even between dogs and their owners. Yet the authors caution against overinterpretation. The lag categories should not be read as evidence that one partner behaviorally leads or regulates the other, and the study measured HF-HRV without recording respiration, so the results reflect temporal physiological coordination rather than respiration-controlled estimates of vagal activity.

The limitations are candidly acknowledged. The sample was small and limited to mothers and five- to seven-month-old infants, with no fathers, no baseline or still-face conditions, and no non-partner dyad pairings to rule out shared environment or common movement. No synchronized behavioral coding was performed, so the physiological alignment cannot yet be tied to specific events like eye contact, vocalization, or touch. The predefined plus-or-minus 30-second lag range may itself have shaped the findings. Still, the study demonstrates the feasibility of a lag-sensitive approach that could, if replicated in larger and more diverse samples, eventually help researchers probe whether physiological synchrony differs in populations facing psychosocial risk, such as mothers with postpartum depression, offering an objective, minimally invasive complement to the behavioral observations that have long dominated attachment research.

Subject of Research: Lag-sensitive analysis of mother-infant heart rate variability synchrony during free interaction

Article Title: Lag‐sensitive cross‐correlation analysis of mother–infant HF‐HRV synchrony during free interaction: An exploratory feasibility study

Article References: Sakurai, A., Koide, K., Kasahara, Y., Nakamura, Y., & Yoshizawa, T. (2026). Lag‐sensitive cross‐correlation analysis of mother–infant HF ‐ HRV synchrony during free interaction: An exploratory feasibility study. Physiological Reports, 14(18), Article e71116. https://doi.org/10.14814/phy2.71116

Image Credits: AI Generated

DOI: 10.14814/phy2.71116

Keywords: heart rate variability, mother-infant synchrony, HF-HRV, cross-correlation, autonomic nervous system, respiratory sinus arrhythmia, polyvagal theory, attachment, physiological co-regulation, surrogate data, free play, infant development