Autonomic recovery following submaximal exercise in yoga practitioners versus aerobic and strength-trained individuals

Abstract

Background

Rapid autonomic recovery after physical stress is a hallmark of cardiovascular health. While both yoga and conventional exercise modulate autonomic function, direct comparisons of their effect on post-exercise recovery are scarce. This study compared autonomic recovery in yoga practitioners versus those in aerobic or resistance training.

Methods

We conducted a cross-sectional study of 51 healthy adults (18-35 years) in three long-term training groups: Yoga (n = 17), Aerobic (n = 17), and Resistance (n = 17). Participants performed a 5-minute submaximal Harvard step test. Heart rate variability (HRV) was analyzed from electrocardiograms recorded at baseline and during a 10-minute post-exercise recovery.

Results

After adjusting for baseline differences, the Yoga group showed a more efficient autonomic recovery profile. ANCOVA revealed a significant group effect on vagal reactivation, as measured by High-Frequency (HF) power (p = 0.001). Post-hoc tests confirmed that the Yoga group’s recovery was significantly greater than that of the Aerobic and Resistance groups. Similar significant effects favouring Yoga were found for pNN50, SDNN, LF power, and total power (all p < 0.05). No significant group differences were observed for pulse rate, blood pressure, or RMSSD recovery.

Conclusion

Regular yoga practice is associated with more efficient parasympathetic reactivation after physical exertion. This suggests yoga’s integrative nature is associated with unique advantages for autonomic strength compared to conventional aerobic and strength training.

The study was prospectively registered (CTRI/2023/06/054047).

Introduction

The well-regulated autonomic nervous system (ANS) is a cornerstone of physiological homeostasis and long-term health. An imbalance, typically presenting as elevated sympathetic activity and diminished parasympathetic (vagal) tone, is a well-documented precursor to chronic conditions such as hypertension and cardiovascular disease [1]. Heart rate variability (HRV) is a reliable, non-invasive substitution for this balance, with diminished variability serving as a potent predictor of adverse health outcomes and mortality [2,3].

Physical exercise is a primary nonpharmacological strategy for enhancing autonomic function [4], with systematic reviews confirming its efficacy in improving cardiovascular autonomic modulation even in at-risk populations such as postmenopausal women [5]. Conventional training modalities, including aerobic and resistance training, are recognized for their capacity to improve cardiovascular fitness and produce modest increases in vagal-mediated HRV. However, the ancient discipline of Yoga, which integrates physical postures (asanas) with deliberate breath regulation (pranayama) and meditation, is theorized to exert a more direct and profound influence on the ANS. This multifaceted approach is believed to enhance autonomic regulation through mechanisms that include improving baroreflex sensitivity and directly stimulating the vagus nerve, thereby promoting a state of parasympathetic predominance [6].

Although substantial evidence supports the benefits of aerobic exercise over strength training, the unique autonomic stability of yoga remains underinvestigated in direct comparative studies. While beneficial, aerobic training, particularly resistance training, is characterized by profound sympathetic activation, leading to significant suppression of vagal indices such as the high-frequency (HF) band and pNN50, which are direct markers of parasympathetic cardiac control [7]. The postexercise period thus presents a challenge for the autonomic nervous system, requiring the rapid withdrawal of sympathetic influence, as partially reflected in the low-frequency (LF) band, and the swift restoration of protective vagal activity.

This is where yoga’s methodology may offer a unique advantage. We hypothesize that yoga, by integrating deliberate, slow breathing (pranayama) that directly enhances HF vagal power [1], an effect demonstrated even in isolation by randomized controlled trials [8] and with mindfulness practices that refine the top-down regulation of the central autonomic network [9], trains the body for more efficient sympatho-vagal switching. Yoga interventions have already been shown to significantly improve HRV in clinical cohorts, such as individuals with hypertension, and to improve the quality of life of breast cancer survivors undergoing chemotherapy [6,10]. Even in healthy populations, specific yogic practices have led to marked increases in vagal-related HRV parameters. The landmark review by Tyagi and Cohen (2016) affirmed that yoga consistently promotes vagal dominance and underscored a critical need for more rigorous, comparative research that specifically tests this hypothesized recovery advantage.

Despite the parallel streams of evidence supporting both yoga and traditional exercise, head–to-head comparisons remain surprisingly uncommon. One recent study revealed that a yoga breathing program yielded greater benefits in terms of autonomic function than did aerobic exercise[11]; however, few investigations have focused specifically on the crucial period of post-exercise recovery. This phase is a critical window into autonomic resilience, the system’s ability to efficiently transition from a state of stress back to homeostasis. This gap raises a pivotal question: Does the holistic methodology of yoga promote more efficient autonomic recovery, as measured by a faster rebound in pNN50 and HF power, after a physical challenge compared to conventional exercise?

To address this, our study conducted a direct, cross-sectional comparison of autonomic recovery profiles in healthy adults regularly engaged in yoga, aerobic, or resistance training. We hypothesized that the unique integration of breath and the body during yoga would manifest as a most efficient and rapid restoration of vagal-mediated HRV indices following a standardized exercise stressor.

Methodology

Study design and ethical considerations

This investigation was conducted as a cross-sectional, comparative study, with reporting adherent to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. The study was conducted in accordance with the Declaration of Helsinki, following full approval from the Kasturba Medical College & Kasturba Hospital Institutional Ethics Committee (IEC1: 301/2022, dated June 3, 2023). All individuals provided written informed consent before participating, acknowledging their understanding of the procedures and their voluntary involvement. The protocol was prospectively registered in the Clinical Trial Registry of India (CTRI) with the reference number CTRI/2023/06/054047, and recruitment of participants commenced on June 23, 2023, and continued until April 18, 2024.

Sample size

The sample size was calculated using G*Power (version 3.1.9.7) to ensure adequate statistical power for detecting differences among the three groups. Based on previous comparative studies of HRV and exercise modalities, we anticipated a medium to large effect size. For a one-way ANOVA comparing three groups, with an alpha (α) of 0.05, a desired power (1-β) of 0.80, and a projected large effect size (Cohen’s f = 0.40), the calculation indicated a required total sample size of 51 participants (17 per group).

Inclusion and exclusion criteria

Inclusion criteria: Participants were eligible if they:

1. Were healthy adults aged 18–35 years.

2. Reported no known cardiovascular, metabolic, neurological, or respiratory disease.

3. Engaged in a consistent primary training modality for ≥ 6 months.

4. Accumulated ≥ 600 met-min/week of aerobic or resistance activity (for the respective groups); or Practiced yoga 5–6 days per week following the standardized university curriculum.

Exclusion criteria: Individuals were excluded if they:

1. Used medications known to influence autonomic or cardiovascular function.

2. Had any chronic illness affecting autonomic regulation.

3. Reported acute illness on the day of testing.

4. Consumed caffeine, nicotine, alcohol, or stimulants prior to assessment; or

5. Unable to complete the study protocol.

Participant cohort

A cohort of 51 healthy, physically active adults (18–35 years old) of both sexes was recruited from the Department of Yoga and the Sport Science & Exercise Department on a multifaceted university campus. Eligibility was determined through a comprehensive International Physical Activity Questionnaire (IPAQ) (Supplemental File), a standardized and validated instrument designed for population surveillance of physical activity among adults aged 15 to 69 years [12]. Participants in the aerobic and strength training groups were required to accumulate at least 600 MET-minutes of activity per week and maintain consistent activity levels. For the Yoga group, eligibility required consistent practice as detailed below. Individuals with any exclusion criteria were not enrolled. Based on their primary and consistent mode of physical activity for at least six months, the participants were categorized into one of three distinct groups based on their primary training modality:

1. Aerobic exercise group (n = 17): Individuals who reported regularly engaging in activities such as brisk walking, jogging, cycling, or swimming for at least 50 minutes, 3–4 times per week.[13]

2. Resistance exercise group (n = 17): Individuals who reported following a structured resistance training program of 3 sets at 40% 1RM (repetition maximum) with 45-s rest periods or 10 reps at 80% 1RM with 60-s rest periods, 3-4 times per week, targeting major muscle (chest/shoulder/legs/back/arm/calf) groups [14]

   As this was a cross-sectional design, aerobic and resistance training sessions were not supervised by the investigators. Adherence was based on self-reported training frequency (3–4 days/week). Objective monitoring (e.g. wearable devices) was not employed.

3. Yoga Group (n = 17): Individuals reporting from the university’s yoga program who followed a structured, unsupervised daily practice based on a consistent curriculum. Adherence to the 5-6 day per week schedule (except on menstruation days for females) was a prerequisite for inclusion and was confirmed via self-report at the beginning of the study, as participants were recruited from a population with an already established daily practice as follows:

   1. Structure and Duration: Each 45-minute session consisted of three components: initial warm-up and dynamic postures (10 minutes), a sequence of static asanas (25 minutes), and a concluding pranayama and relaxation phase (10 minutes)
   2. Content and Sequence:

      • Warm-up: Sessions began with 5-7 rounds of Surya Namaskar (Sun Salutation).
      • Asanas: The core practice included a consistent sequence of postures: Trikonasana (Triangle Pose), Parshvakonasana (Side Angle Pose), Paschimottanasana (Seated Forward Bend), Purvottanasana (Upward Plank Pose), Janushirshasana (Head-to-Knee Pose, Pavanamuktasana (Wind-Relieving Pose), Bhujangasana (Cobra Pose), Shalabhasana (Locust Pose), Paryankasana (Couch Pose), and Padottanasana (Wide-Legged Forward Bend), with each held for 30-60 seconds. [15]
      • Pranayama & Relaxation: The session concluded with 5 minutes of Anulom Vilom (Alternate Nostril Breathing), followed by 5 minutes of Savasana (Corpse Pose) for relaxation.
   3. Intensity Characterization and Adherence: The practice was self-paced. Based on the prescribed protocol, which emphasizes static postures held for extended periods and controlled breathing exercises (pranayama), the intended intensity of the sessions was characterized as low to moderate activity. As this was an observational study of an existing routine, daily training logs were not collected.

Experimental study design

All assessments were conducted in a quiet, temperature-controlled environment (22–24 °C) to minimize the influence of environmental factors on autonomic function. Upon arrival, participants acclimatized by resting in a supine position for 15 min. They were fitted with electrocardiogram (ECG) electrodes, and a 15-minute baseline ECG was then recorded. The participants subsequently performed a 5-minute submaximal exercise test with continuous ECG monitoring. Immediately upon cessation of exercise, they returned to the supine position for a 10-minute recovery phase, during which the ECG recording continued (Figure 1).

Figure 1.

Flow chart of the study design.

Measurements

Submaximal exercise stressor: the Harvard step test

To induce a standardized and significant cardiovascular response, the Harvard step test was employed as the submaximal exercise stressor [16]. This test was specifically chosen for its established validity, ease of administration in a non-clinical setting, and the absence of specialized calibration requirements. Crucially, its fixed duration and intensity are known to elicit a strong and consistent sympatho-excitatory response, providing the necessary physiological prerequisite from which to measure the efficiency of subsequent parasympathetic recovery.

To perform the Harvard step test, the participants are made to perform a periodic motion of stepping up onto and down from a 50.8 cm (20-inch) platform. An analogue metronome was used to maintain a constant pace of 30 step cycles per minute (a four-beat cycle of up with the left foot, up with the right foot, down with the left foot, down with the right foot), which was equal to one step every two seconds. The participants performed the test for a total duration of 5 min. The test was terminated prior to 5 min only if the participant could not maintain the prescribed pace for 15 consecutive seconds or upon reaching volitional exhaustion (exhaustion was defined as the athlete’s inability to maintain the stepping movement for a duration of 15 s). This standardized physical challenge significantly elevates heart rate and sympathetic activity, providing a robust physiological stressor from which the efficiency of autonomic recovery can be measured.

Heart rate variability (HRV) data acquisition and analysis

Data acquisition

A continuous lead II ECG signal was acquired throughout the baseline, exercise, and recovery periods via a Power Lab 8/35 data acquisition system (ADInstruments Pty Ltd, Australia). Three pre-gelled disposable electrodes were placed in a standard lead II configuration to optimize the detection of the R-wave. The analogue ECG signal was digitized at a sampling rate of 1000 Hz to ensure high temporal resolution for subsequent R-R interval detection.

Data processing and segment selection

The raw ECG signal was processed via Lab Chart Pro software (version 8, ADInstruments Pty Ltd, Australia). The software’s built-in algorithm detected the R-peak of each QRS complex, generating a continuous R-R interval (the time elapsed between two successive R-waves) tachogram. This R-R interval series was then inspected for any movement-based or electrical noise-based artifacts and ectopic beats. Any identified artefacts, defined as R-R intervals deviating by more than 20% from the local average of the preceding five intervals, were corrected via the software’s cubic spline interpolation method. To ensure the highest data quality, only segments with less than 5% of beats requiring correction were included in the final analysis.

For the HRV analysis, two distinct, artifact-free 5-minute (300-second) segments were extracted for each participant, in accordance with established guidelines for short-term HRV assessment [2]: The first segment was obtained from the initial resting period and served as the baseline value, and the second segment was taken from the first 5 min of the post-exercise recovery period. Autonomic recovery was operationally defined as the absolute HRV values measured during this recovery segment, not as a change score from baseline.

Calculation of HRV parameters

From the cleaned R–R interval segments, LabChart Pro calculates standard time-domain and frequency-domain HRV parameters.

Time-domain analysis

These metrics quantify the amount of variability in the R–R intervals.

• RMSSD (root mean square of successive differences): The square root of the mean of the squared differences between successive R–R intervals. It is a primary indicator of short-term, high-frequency variations and is considered a strong and reliable marker of vagally mediated cardiac control (parasympathetic activity).

• pNN50 (percentage of successive R–R intervals that differ by more than 50 ms): This parameter is also a strong marker of parasympathetic activity and is highly correlated with the RMSSD.

• SDNN (Standard Deviation of all Normal-to-Normal intervals): The standard deviation of all R–R intervals in the segment. It reflects overall variability and is influenced by both sympathetic and parasympathetic inputs.

Frequency-Domain Analysis: This method utilizes a fast Fourier transform (FFT) to decompose the R-R interval series into its underlying frequency components, thereby quantifying the power within specific frequency bands.

• High-frequency (HF) power (0.15–0.40 Hz): This band reflects respiratory sinus arrhythmia (RSA) and is widely accepted as a marker of parasympathetic (vagal) modulation of the heart.

• Low-frequency (LF) power (0.04–0.15 Hz): The interpretation of this band is more complex, reflecting a combination of sympathetic and parasympathetic influences, as well as baroreflex activity.

• Total power (TP): represents the total variance in the R–R intervals and reflects overall autonomic activity.

Statistical analysis

Descriptive statistics summarized demographic and baseline variables. Continuous data are presented as mean ± standard deviation and categorical data as frequencies and percentages. Heart rate variability (HRV) outcomes included time-domain (SDNN, RMSSD, pNN50) and frequency-domain measures (LF, HF, total power, LF/HF ratio).

Group differences in post-exercise autonomic recovery were assessed using Analysis of Covariance (ANCOVA). For each HRV parameter, the corresponding baseline value was included as a covariate, along with age, sex, BMI, diet, and physical activity (MET-min/week). Although the sample size was modest, ANCOVA was selected because these covariates are established determinants of HRV, and adjustment reduces bias; results were interpreted cautiously given the potential for over-adjustment. Statistical significance was set at p < 0.05.

Post-hoc comparisons were performed using Tukey’s HSD test, and effect sizes with 95% confidence intervals were reported. All analyses were conducted using the Jeffreys’ Amazing Statistics Package (JASP, v0.19.0).

Results

Baseline participant characteristics

A total of 51 healthy adults (17 in each of the yoga, Aerobic, and Resistance training groups) completed the study. Groups were comparable in terms of age and BMI, although the Yoga group had a higher proportion of female participants. Physical activity levels (MET-min/week) were highest in the resistance group, followed by aerobic and yoga groups (Table 1).

Table 1.

Baseline characteristics of study participants across exercise groups.

Variables	Aerobic training (n = 17)	Resistance training (n = 17)	Yoga (n = 17)
Age (years)	19.00 (4.00)	21.00 (3.00)	22.00 (3.00)
Female participants (n, %)	6 (35%)	6 (35%)	12 (71%)
BMI (kg/m2)	22.50 (4.00)	23.50 (6.00)	21.06 (4.60)
Physical Activity levels (MET-min/week)	3417.50 (3331.00)	5076.00 (3201.00)	2079.00 (1047.00)

Values for continuous variables are presented as median (interquartile range) as data were not normally distributed. Values for categorical variables are presented as count (percentage).

BMI: body mass index; MET: metabolic equivalent.

Post-exercise autonomic recovery

The central finding of this study is that the Yoga group exhibited a greater autonomic recovery profile, and this finding held true even after accounting for baseline differences between the groups.

Cardiovascular parameters

Pulse Rate (PR) and Systolic Blood Pressure (SBP) recovery did not differ significantly among the groups (PR: F = 0.21, p = 0.815, ηp² = 0.01; SBP: F = 0.20, p = .818, ηp² = 0.01), indicating similar cardiovascular responses post-exercise (Table 2).

Table 2.

Comparison of the parasympathetic recovery after a Sub-maximal exercise among the groups.

Variables		Unadjusted#			Adjustedɸ			ANCOVA (F, p)#	Effect Size (ηp²)
		Aerobic training (n = 17)	Resistance training (n = 17)	Yoga (n = 17)	Aerobic training	Resistancetraining	Yoga
PR (bpm)		43.24 ± 9.85	43.47 ± 15.37	40.27 ± 15.28	43.1 ± 2.4	43.5 ± 3.7	40.3 ± 3.9	0.21 (0.815)	0.01
SBP (mmHg)		43.53 ± 11.78	42.18 ± 13.70	41.67 ± 10.30	43.5 ± 2.9	42.2 ± 3.3	41.7 ± 2.7	0.20 (0.818)	0.01
Heart Rate Variability	SDNN (ms)	172.3 ± 157.9	148.4 ± 148.2	287.6 ± 153.8	174.2 ± 38.3	150.1 ± 36.0	289.0 ± 39.7	2.15 (0.026)*	0.09
	RMSSD (ms)	257.0 ± 240.8	240.5 ± 210.8	420.7 ± 227.2	259.0 ± 58.4	242.0 ± 51.1	422.0 ± 58.7	0.97 (0.387)	0.04
	pNN50(%)	21.32 ± 33.35	27.67 ± 27.00	47.56 ± 32.60	21.3 ± 8.1	27.7 ± 6.5	47.6 ± 8.4	2.40 (0.012)*	0.10
	LF (ms2)	10946 ± 15757	9985 ± 20220	25306 ± 27767	10946 ± 3822	9985 ± 4904	25306 ± 7169	2.23 (0.035)*	0.09
	HF (ms2)	38753 ± 48683	19269 ± 29463	76139 ± 68893	38753 ± 11807	19269 ± 7146	76139 ± 17788	2.67 (0.001)**	0.11
	LF-HF ratio	−2.51 ± 4.37	−1.28 ± 1.74	−2.07 ± 2.18	−2.51 ± 1.06	−1.28 ± 0.42	−2.07 ± 0.56	0.68 (0.512)	0.03
	Total power (ms2)	55044 ± 67324	54785 ± 103980	128224 ± 115719	55044 ± 16329	54785 ± 25219	128224 ± 29879	2.37 (0.016)*	0.10

Values are mean ± Standard deviation.

#Unadjusted values = raw descriptive mean ± SD.

ɸAdjusted values = ANCOVA estimated means ± SE, controlling for age, gender, BMI, diet, and physical activity levels (MET-min/week).

Significance: *p < 0.05, **p < 0.01.

BMI: body mass index; HF: high frequency power; LF: low frequency power; MET: metabolic equivalent; pNN50: proportion of successive NN intervals differing by more than 50 ms; PR: pulse rate; RMSSD: root mean square of successive differences between normal heartbeats; SBP: systolic blood pressure; SDNN: standard deviation of all NN intervals for each 5 min segment.

Time-domain HRV measures

SDNN (Standard deviation of NN intervals) showed a significant group effect (F = 2.15, p = .026, ηp² = 0.09), with yoga practitioners demonstrating higher recovery values (287.6 ± 153.8 ms) compared to aerobic (172.3 ± 157.9 ms) and resistance (148.4 ± 148.2 ms) groups. RMSSD did not differ significantly among groups (F = 0.97, p = .387, ηp² = 0.04), although yoga participants had numerically higher values. pNN50 (% of successive RR intervals differing by >50 ms) showed a significant difference (F = 2.40, p = .012, ηp² = 0.10), with yoga practitioners exhibiting greater vagal reactivation (47.56 ± 32.60%) than aerobic (21.32 ± 33.35%) and resistance (27.67 ± 27.00%) groups (Table 2).

Frequency-domain HRV measures

Low-frequency (LF) Power was significantly higher in the yoga group (F = 2.23, p = .035, ηp² = 0.09), suggesting enhanced sympathetic-parasympathetic modulation. High Frequency (HF) Power, a marker of parasympathetic activity, was significantly elevated in yoga practitioners (F = 2.67, p = .001, ηp² = 0.11), indicating robust vagal recovery. Total Power, reflecting overall autonomic activity, was significantly greater in the yoga group (F = 2.37, p = .016, ηp² = 0.10), supporting enhanced autonomic resilience. LF/HF Ratio did not show significant differences (F = 0.68, p = .512, ηp² = 0.03), suggesting similar sympathovagal balance across groups. (Table 2; Figure 2).

Figure 2.

Comparison of heart rate variability parameters across exercise groups. Raincloud plots display the distribution (half-violin), individual data points (dots), and summary statistics (box plot) for time-domain (a, b) and frequency-domain (c-e) HRV parameters. The box plots indicate the median (center line) and interquartile range (box). Brackets indicate pairwise comparisons from post-hoc tests following a one-way ANCOVA, controlling age, gender, Body Mass Index (BMI), and physical activity levels (MET-min/week). *p < 0.05.

After controlling for key confounders, post-hoc analysis confirmed that the yoga group was the primary driver of these significant effects, particularly in SDNN, pNN50, LF, HF, and Total Power indices. Thus, the yoga group demonstrates a significantly greater restoration in these key HRV parameters compared to both the aerobic and resistance training groups (all p-values < 0.05) (Table 2). In contrast, no significant group differences were observed in the recovery of pulse rate, systolic blood pressure, or the broader time-domain HRV metric of RMSSD (Table 2). These results confirm that the greater autonomic recovery in the yoga group is a genuine effect of the training modality, independent of the potential confounders. Figure 2 illustrates the mean change in HRV (time and frequency domain variables) among participants in aerobic, resistance, and yoga groups.

Figure 2 illustrates that yoga practitioners consistently demonstrated higher recovery values across key HRV parameters (SDNN, pNN50, LF, HF, and total power) compared to aerobic and resistance groups. The raincloud plots also show larger effect sizes favouring yoga in pairwise comparisons.

Discussion

Our experimental study aimed to explore the differences in autonomic recovery among practitioners of yoga, aerobic training, and strength training. Our study revealed that regular yoga practice is associated with significantly more efficient and beneficial autonomic recovery following submaximal exercise stressors than conventional aerobic or resistance training. The central finding, a greater restoration of vagal-mediated and overall heart rate variability in yoga practitioners, suggests that the integrative methodology of yoga is associated with a unique form of physiological flexibility.

The differing recovery profiles observed may be rooted in the fundamental physiological demands of each training modality. Both aerobic and resistance training are characterized by periods of high-intensity, predominantly sympathetic-driven effort designed to create significant physiological stress and adaptation. In contrast, the practice of yoga, as reported by our participants, integrates periods of low-to-moderate physical exertion with conscious breathing (pranayama) and meditative focus, which are known to directly engage and enhance parasympathetic tone. This inherent difference in training focus, stress induction versus homeostatic regulation, likely contributes to the enhanced autonomic recovery seen in the yoga group.

We hypothesize that these effects are driven by both peripheral and central mechanisms. Peripherally, the deliberate practice of slow, controlled breathing (pranayama) is considered a powerful tool for direct vagal stimulation, which is known to augment respiratory sinus arrhythmia and enhance baroreflex sensitivity [1]. Centrally, it is plausible that the meditative and mindfulness components integral to yoga likely exert a top-down influence on the central autonomic network. Seminal work has shown that interoceptive awareness, a skill refined in yoga, is processed in key brain regions, such as the anterior insular cortex, which integrates visceral signals and modulates autonomic outflow [9]. This finding aligns with research by Streeter et al. who postulated that yoga increases the activity of the inhibitory neurotransmitter GABA, thereby reducing sympathetic drive [17]. This combination of bottom-up (breath-to-vagus) and top-down (mind-to-brain) regulation offers a plausible framework for explaining the enhanced autonomic flexibility observed in our yoga cohort. This integrated effect is so profound that studies in clinical populations have demonstrated that yoga can modulate the entire psycho-neuroimmune axis, leading to reduced disease activity in conditions such as rheumatoid arthritis [18].

In contrast, while aerobic and resistance exercise modalities are undeniably beneficial, their primary impacts on the ANS may be mechanistically different. It is essential to compare our findings on acute recovery with the extensive literature demonstrating that chronic aerobic training is highly effective in improving long-term resting vagal tone [5]. Our study does not challenge this established benefit; rather, it highlights a distinct phenomenon that occurs in the immediate post-exercise window. During this acute phase, intense exercise, particularly resistance or high-intensity interval training, is often characterized by prolonged suppression of vagal activity and delayed recovery of HRV, which can last for hours [7]. Conversely, yoga’s methodology, which integrates periods of parasympathetic activation via controlled exhalations and restorative postures, may uniquely ‘train’ the nervous system for more rapid transitions between sympathetic and parasympathetic states. This distinct advantage is what becomes clearly visible in the accelerated post-exercise recovery profile measured in our study.

A particularly insightful aspect of our results is the clear divergence between different HRV parameters. While the greater recovery in HF power (Figure 2d) and pNN50 (Figure 2b) in the yoga group provides a clear signal of enhanced parasympathetic reactivation, the concurrent significant recovery in SDNN (Figure 2a) indicates a more robust and globally responsive autonomic system. Interestingly, the LF/HF ratio, often interpreted as a marker of sympatho-vagal balance, did not differ significantly between groups. This suggests that the primary advantage conferred by yoga in the immediate post-exercise period is a more potent and rapid reactivation of the parasympathetic branch (as shown by absolute HF power and pNN50) rather than a differential shift in the overall balance relative to the sympathetic branch. This pattern reinforces that the unique benefit of yoga lies in its powerful augmentation of vagal tone, a finding clearly captured by specific time-domain and spectral parameters of parasympathetic function [19].

An intriguing finding was the complex pattern of results among the time-domain HRV parameters. We observed a significant group difference in the recovery of pNN50 and SDNN, but not in RMSSD (Table 2). The divergence between pNN50 and RMSSD is particularly unexpected, given that both are considered strong markers of cardiac vagal outflow. One possible explanation is statistical; with a modest sample size, our study may have been underpowered to detect a true difference in RMSSD, which can sometimes be less sensitive to change than pNN50. Alternatively, this may reflect a subtle physiological distinction. RMSSD quantifies the root mean square of all successive beat-to-beat differences, while pNN50 is a count of only large (>50 ms) changes. It is possible that the nature of post-exercise parasympathetic reactivation in the yoga group is characterized by more frequent, large, phasic bursts of vagal activity (captured by pNN50 and contributing to overall SDNN) rather than a smoother, overall increase in beat-to-beat variance (which would be more strongly reflected in RMSSD). This possibility warrants further investigation in future studies.

This study contributes a novel perspective by extending observations from clinical populations to a healthy, physically active cohort. The improvements in parasympathetic reactivation observed in our yoga participants align with these findings [20], which was recently demonstrated in a randomized controlled trial, where a yoga intervention fostered a significant shift towards parasympathetic dominance even under conditions of high occupational stress. Much of the literature has focused on the restorative effects of yoga in individuals with stress-related or cardiovascular risk factors [21]. Our research demonstrated that even among individuals who are already fit, yoga provides an additive benefit to autonomic resilience. This aligns with large-scale evidence synthesis, which increasingly recognizes mind-body practices as valuable components of integrative cardiology for both primary and secondary prevention [22]. By conducting a direct comparison, our work addresses a critical gap identified by previous reviews [1] and provides a clear, comparative depiction of post-exercise physiology.

Strengths and implications

The primary strength of this investigation lies in its direct, head-to-head comparison of three well-defined, distinct training modalities within a single, standardized protocol. The use of robust nonparametric statistics and adherence to the STROBE guidelines further bolsters the validity of our conclusions. For athletes in other disciplines, our findings suggest that incorporating a yoga practice could serve as a potent cross-training strategy to accelerate recovery, enhance autonomic resilience, and potentially mitigate the risk of overtraining. For public health, our results position yoga as an accessible practice for cultivating a healthy autonomic profile that is foundational for preventing chronic disease.

Limitations and future directions

We acknowledge several limitations that frame the interpretation of our findings and guide future research. First and foremost, the cross-sectional design of this study is a primary limitation. While our findings demonstrate a strong association between yoga practice and more efficient autonomic recovery, this design precludes any inference of causality. We cannot definitively conclude that yoga causes better recovery; however, it is associated with it in this cohort.

Second, limitations related to our participant sample must be considered. Our sample size (n = 51), although calculated to have adequate power for a large effect size, remains modest. This may have limited our statistical power to detect smaller, yet potentially meaningful, differences between groups, which could explain the non-significant finding for RMSSD. Furthermore, our participants were a homogenous group of young, healthy adults from a single university campus, which limits the generalizability of these findings to other populations, such as older adults, individuals with chronic health conditions, or elite athletes.

Third, potential confounding variables related to group composition and training were present. We observed a notable imbalance in the sex distribution across groups, with a higher proportion of females in the Yoga group. Although we included sex as a covariate in our statistical models, the possibility of residual confounding cannot be entirely ruled out. Similarly, the groups differed in their average weekly training volume (MET- min/week). While we also statistically controlled for this variable, it represents a potential confounder that could contribute to the observed differences. The reliance on self-reported training data, without direct supervision of exercise intensity and volume, is another inherent limitation that future studies should address with objective monitoring.

Finally, specific aspects of our experimental protocol and statistical analysis warrant discussion. Our analysis was limited to the initial 10-minute post-exercise recovery window, which does not capture the full recovery trajectory that may unfold over a longer period. The Harvard Step Test, while standardized, uses a fixed platform height that may have imposed different relative exercise intensities for participants of varying heights, subtly influencing recovery kinetics. From a statistical standpoint, employing an ANCOVA model with multiple covariates in a modest sample carries a risk of over-adjustment, which can reduce statistical power. Therefore, our results should be interpreted with this consideration in mind.

These limitations underscore the clear need for future research. Longitudinal and, ideally, randomized controlled interventional studies are required to establish a causal relationship. Such studies should employ larger, more diverse, and sex-matched cohorts; directly supervise and quantify training loads to avoid confounding; and utilize longer post-exercise monitoring periods to map the complete recovery curve. Furthermore, collecting detailed data on the total duration of training (in years) would allow for robust correlational analyses to explore potential dose-response relationships, building upon the exploratory analysis in our dataset which suggested consistent practice may be more critical than ­duration alone.

Conclusion

Healthy individuals who regularly practice yoga exhibit a significantly enhanced autonomic recovery profile following a standardized exercise challenge. Compared with conventional aerobic and resistance training, the enhanced and more rapid restoration of key heart rate variability indices underscores the unique capacity of yoga to promote parasympathetic reactivation. These findings suggest that incorporating yoga into a regular fitness regimen may offer distinct and complementary benefits for cardiovascular health and autonomic flexibility, distinguishing it from conventional forms of exercise. However, these implications should be considered preliminary given the study’s cross-sectional nature and modest sample size. While these results are promising, future longitudinal studies with larger cohorts are needed to confirm these findings and to inform public health and athletic training recommendations.

Funding Statement

The author(s) received no specific funding for the work presented in this study. The open-access publication was funded by the Manipal Academy of Higher Education (MAHE), Manipal.

Disclosure statement

No potential conflicts of interest were reported by the authors.

Data availability statement

The data supporting the findings of this study are not publicly available due to privacy restrictions outlined in the participant-informed consent and institutional ethics approval. However, de-identified data can be made available upon reasonable request to the corresponding author.