This Week In HRV - Episode 45

[00:00:00] Welcome to this Week in Heart Rate Variability, the show where we take a close, careful look at the research shaping how clinicians, coaches and researchers think about autonomic function. I'm Matt Bennett. Before we get started, a quick note. Nothing in this episode is medical advice. We're here to discuss published research and its implications for practice, not to diagnose, treat or replace the judgment of a qualified healthcare provider. If something we discuss raises questions about your own health or the health of a client or patient, please talk with a licensed professional. This week is a big one. We've got seven studies on the docket and they span an unusually wide range of populations and contexts, from healthy young adults and office workers to youth with cerebral palsy, patients with severe brain injury, older adults at cardiovascular risk, a small methodological pilot and a group of professional footballers. If there's one thing that ties a lineup like this together, it's a reminder of just how far heart rate variability has traveled. As a research tool, it shows up just as comfortably in a sports science lab as it does in a neurological intensive care unit. We'll start with a look at how body mass index shapes autonomic recovery after a classic exercise test in young adults. From there, we'll move into the workplace, examining whether a co created intervention aimed at office workers with flexible schedules can improve sleep and recovery. Next, we'll look at physical activity and autonomic function in youth with cerebral palsy, followed by a fascinating clinical study using postural tilt testing to help distinguish between different disorders of consciousness. After our sponsor break, we'll dig into a machine learning study predicting cardiovascular disease risk in older adults using heart rate variability derived from photoplesmography. Then we'll spend some time on a small but important feasibility study exploring what it actually takes to run continuous 24 hour heart rate variability monitoring in the real world. And we'll close out with a study of resting and exercise heart rate variability in a group of footballers in Senegal, adding some much needed geographic diversity to the sport's science literature. As always, we'll try to give each study its full the design, the findings, what they mean in practice, and just as importantly, where the findings stop and speculation begins. Let's get into it. Our first study looks at how body composition shapes the body's autonomic response to a standardized bout of exercise using a tool that's been a staple of exercise physiology for decades, the Harvard Step Test. The authors are Viswadeepak Rajput, Vishavadeep Kaur and Ankaliyah Babhara. Here's why this matters. We already know broadly that carrying excess body weight is associated with a less favorable cardiovascular risk profile, but heart rate variability gives us a more granular window into that relationship, specifically into how well the autonomic nervous system is regulating the heart both at rest and in the crucial minutes after physical exertion. For clinicians and coaches working with clients across the weight spectrum, understanding whether body mass index tracks with meaningfully different autonomic profiles and how quickly those differences show up after exercise has real implications for how we interpret fitness assessments and structure recovery protocols. Rajput, Kaur, and Bobara designed a comparative analytical observational study to explore exactly this question. They recruited 200 young adults between the ages of 17 and 25, splitting them evenly into a normal body mass index group and an overweight group. 100 participants each every participant completed the Harvard Step Test, a standardized stepping protocol used to assess cardiovascular fitness. Before the test, each participant rested supine for five minutes while the team recorded a baseline electrocardiogram, which is simply a recording of the heart's electrical activity used to derive heart rate variability metrics. After the step test, the Researchers recorded another 15 minutes of electrocardiogram data, though for their primary analysis they focused on the first five minutes of that post exercise window, capturing the earliest phase of autonomic recovery. From these recordings, the team calculated a full panel of heart rate variability measures on the time domain side that included the mean RR interval, the average time between consecutive heartbeats, along with heart rate itself and several measures that capture beat to beat variability standard deviation of NN intervals, often abbreviated sdnn, which reflects overall variability root mean square of successive differences or rmssd, which is particularly sensitive to parasympathetic or vagal activity and tube related metrics NN50 and PNN50, which count how many pairs of consecutive heartbeats differ by more than 50 milliseconds, either as a raw number or percentage. On the frequency domain side, they looked at high frequency power, which also reflects vagal tone, and the ratio of low frequency to high frequency power, often used as a rough marker of the balance between sympathetic and parasympathetic activity. Alongside all of this, they calculated a physical fitness index from the step test itself, giving them an independent measure of cardiovascular fitness to relate back to the heart rate variability findings. The results were clear and consistent. At baseline, before anyone had done any exercise, they the overweight group already showed a lower mean RR interval, lower SDNN, lower RMSSD, fewer NN50 pairs, and a lower PNN50 alongside a higher resting heart rate in plain terms, the overweight participants hearts were beating faster and with less beat to beat variability even before any physical challenge was introduced, a pattern generally interpreted as reduced vagal or parasympathetic influence on the heart. After the step test that gap didn't close. If anything, it persisted in a way that mattered for recovery. The overweight group continued to show lower RMSSD, SDNN, NN50, PNN50 and high frequency power during the early post exercise window along with a lower physical fitness index score overall. When the researchers used mixed effects modeling, a statistical approach well suited to tracking how a variable changes over time within and between groups, they found significant interactions between group and time specifically for sdnn, RMSSD and and high frequency power. That interaction is important. It tells us the overweight group wasn't just starting from a lower baseline but was also changing differently over the course of exercise and recovery compared to the normal body mass index group. Taken together, these findings paint a picture of overweight young adults showing both reduced resting vagally mediated heart rate variability and a less favorable pattern of early autonomic recovery after exercise. RMSSD and SDNN emerged as the most consistent markers of this difference across both the resting and post exercise periods, which is a useful signal for anyone deciding which heart rate variability metrics to prioritize. When body composition is a variable of interest, it's worth pausing on why RMSSD and SDNN specifically are the metrics that separated these two groups so consistently. RMSSD is generally considered one of the cleanest available markers of parasympathetic or vagal influence on the heart because it's calculated from very short term beat to beat differences that are dominated by the fast acting vagus nerve rather than the slower acting sympathetic branch. Sdnn, by contrast, captures overall variability across the full recording window, blending contributions from multiple regulatory influences. Seeing both the metrics move together and move together specifically during the group by time interaction after exercise strengthens the case that what's being captured here is a genuine difference in autonomic recovery capacity rather than a statistical artifact showing up in just one measure. This also lines up with a broader body of literature connecting excess adiposity with low grade systemic inflammation and and altered baroreflex sensitivity, both of which are plausible biological pathways through which body composition could blunt vagal reactivation after physical exertion. It's also worth thinking about what a physical fitness index reduction alongside these heart rate variability differences adds to the picture. The physical fitness index derived directly from the Harvard step test is a purely mechanical, performance based measure how well someone's heart rate profile holds up across a standardized bout of stepping. Seeing it move in the same direction as the autonomic markers suggests that the overweight group's altered heart rate variability or wasn't simply a lab curiosity disconnected from functional capacity. It appears to be paired with genuinely reduced exercise tolerance as well. [00:07:03] Now, a few important caveats. This is a comparative cross sectional study. The two groups were assessed at a single point in time and compared to one another, not followed longitudinally. That means we can describe a strong association between body mass index category and autonomic profile, but we cannot say that excess weight directly causes reduced heart rate variability or vice versa. There could be shared underlying factors in their physical activity habits, cardiorespiratory fitness independent of weight, genetics, or other health conditions that explain part of this relationship. It's also worth noting that this sample was restricted to young, presumably otherwise healthy adults between 17 and 25. Whether these same patterns and effect sizes would hold in middle aged or older populations or in people with diagnosed metabolic or cardiovascular conditions is an open question. [00:07:49] And while a five minute post exercise analysis window is a reasonable and common approach, it does mean the study speaks to early recovery specifically, not the full trajectory of how these two groups might differ 10, 20 or 30 minutes after exercise. Finally, body mass index itself is a fairly blunt classification tool. It doesn't distinguish between fat mass and muscle mass or account for how body fat is distributed, both of which could independently influence autonomic function for practitioners. The clinical and coaching takeaway here is fairly practical. And if you're using post exercise heart rate variability recovery as part of a fitness or wellness assessment, body composition is a variable worth accounting for since it may meaningfully shift both baseline values and the shape of early recovery. But treat body mass index as one input among several not a complete explanation on its own. It's also worth thinking about the developmental angle here. Because this sample was restricted to young adults, the autonomic differences observed can yet be attributed to years of accumulated cardio metabolic strain, the way they might be in an older population with a longer history of excess weight. That actually makes the findings somewhat more striking. In one sense, measurable autonomic differences were already detectable in a group whose average age was still in the early 20s. Which raises the question of how much earlier in the lifespan these patterns might be identifiable, and whether early intervention during adolescence or young adulthood could meaningfully shift the trajectory before it becomes more deeply entrenched. Our second study shifts from the exercise lab to the office, looking at whether a workplace based intervention can improve recovery including sleep and heart rate variability during sleep for employees. Navigating flexible work arrangements. The authors are Johann Edmondson, Sven, Eric Matiasyn, and David M. Hallman. This is a timely question as flexible work arrangements have become the norm rather than the exception. In many organizations, there's been growing concern that the blurred boundaries between work and personal time can erode recovery, the physical and psychological processes that let the body and mind bounce back from daily demands. Heart rate variability during sleep is one of the more established physiological markers of that recovery process, since parasympathetic activity typically increases during restorative sleep. If workplace interventions can meaningfully shift sleep duration or sleep related heart rate variability, that has implications not just for individual well being but for how organizations think about supporting employees under flexible work models. Edmondson, Matthiasson, and Hallman ran a controlled intervention study within a large governmental organization that offers its employees flexible work arrangements. Employees from one unit, 27 participants in total, took part in a two part intervention, an individual level course focused on strategies for managing work under flexible conditions and a work group level workshop in which teams collaboratively developed shared rules and routines for how flexible work would function within their group. A comparable unit within the Same organization with 21 employees served as a control group and continued working as usual without the intervention. The researchers assessed physical behaviors at baseline and again at a 12 month follow up using 24 hour accelerometry wearable devices that track movement over 3 days at each time point alongside heart rate variability recorded during sleep as an indicator of nighttime recovery. [00:10:33] To analyze how time was distributed across physical activity, inactivity and sleep, the team used a compositional data analysis framework, which is a statistical approach designed specifically for data where the parts must add up to a whole like the 24 hours in a day. Group differences over time were then tested using repeated measures multivariate analysis of variance, a statistical method that examines how multiple related outcomes change together across groups and time points. The headline finding was about sleep duration and it was a substantial one. On average, the intervention group gained 36 additional minutes of sleep per night compared to their baseline, which while the control group actually lost 23 minutes of sleep per night over the same period. That's a meaningful divergence between the two groups, and the statistical test confirmed it was unlikely to be due to chance with a fairly large effect size by conventional standards. However, the intervention's effects were much more limited elsewhere. The balance between physical activity and inactivity during waking hours didn't shift meaningfully between groups and just perhaps most relevant to our show, heart rate variability during sleep didn't show a significant intervention effect either. So how should we interpret a study that found a large effect on one outcome and and essentially no effect on the physiological outcome we care most about here? The author's own framing is a useful and honest one. An intervention combining individual level strategies with work group level collaboration led to meaningfully longer sleep duration, which is itself a plausible pathway toward better recovery and health over time, even though it didn't translate into a measurable change in waking physical activity patterns or in heart rate variability during sleep, at least not within the 12 month window studied. It's worth spending a moment on why the researchers chose a compositional data analysis framework in the first place, because the choice itself reflects a genuinely important methodological point. Time spent in physical activity inactivity and sleep across a 24 hour day isn't independent. More time in one category mathematically requires less time in another. Treating these as ordinary independent variables in a standard statistical model can produce misleading results since a change attributed to one behavior might really just reflect a trade off with another. By analyzing the full daily composition together, the researchers were able to more accurately isolate where the additional sleep time was actually coming from and and to more fairly test whether the intervention shifted the balance of the whole day, not just a single slice of it. The finding that sleep increased substantially while waking activity and sleep heart rate variability stayed essentially flat also raises an interesting question about the sequencing of recovery related change. It's plausible that behavioral shifts, like carving out more protected time for sleep, happen relatively quickly once a work group agrees on shared norms and routines, while the downstream physiological adaptations are like measurable changes in autonomic activity during sleep may take considerably longer to emerge, if they emerge at all within a 12 month window. This kind of staggered timeline, where behavior shifts before physiology catches up, shows up in other areas of health behavior change research as well, and it's a useful frame for interpreting a result like this one. There are some important limitations to sit with here. The sample sizes 27 in the intervention group and 21 in the control group are modest and both were drawn from a single governmental organization, which limits how confidently we can generalize these findings to other workplaces and industries or cultures of work. This also wasn't a randomized controlled trial. The intervention and control groups were different existing units within the organization, not randomly assigned individuals, which opens the door to unmeasured differences between the units that could have influenced the results independent of the intervention itself. Participation in the workshops was also likely to some degree self selected, meaning the employees who engaged most actively with the program may have been more motivated to change their habits in the first place. And on the null finding for heart rate variability specifically, it's worth being cautious about concluding the intervention had no physiological effect at all. With a sample this size, the study may simply have lacked the statistical power to detect a smaller but still meaningful change in sleep heart rate variability for anyone working in occupational health or organizational wellness. The practical takeaway is that behavioral workplace level interventions can move sleep duration in a meaningful way relatively quickly, even when downstream physiological markers like heart rate variability haven't yet caught up. That's a useful reminder to set realistic timelines and outcome expectations when designing or evaluating similar programs and to consider sleep duration itself as a legitimate, measurable early win rather than waiting solely for autonomic markers to shift. It's also worth noting what this study implies about the value of workgroup level rather than purely individual level intervention design. Many workplace wellness programs focus exclusively on equipping individual employees with better personal strategies, but this intervention explicitly paired that individual component with a collective workshop in which teams negotiated shared norms together. The fact that sleep duration improved so substantially suggests that giving teams the social and structural permissions to actually use flexible work arrangements as intended, rather than simply handing individuals more autonomy and hoping they use it well, maybe a meaningfully different and possibly more effective lever for change. That's a distinction worth considering for anyone designing similar programs in their own organization, particularly if early attempts at purely individual, focused coaching haven't produced the desired results. Our third study takes us into pediatric and adaptive rehabilitation, examining how different levels of physical activity relate to physical performance and heart rate variability in youth with cerebral palsy. The authors are Sunny Rickelmi, Maria Isabel Cornejo, Rosemary Arenas, Javier Russell Guzman, Alexis Espinoza Salinas, Rafael Lima Cans, and Matias Henriquez. Cerebral palsy is a group of movement and posture disorders arising from a non progressive injury to the developing brain. Individuals with cerebral palsy often experience impaired autonomic nervous system regulation of the heart, but exactly how that impairment relates to physical activity levels in this population has remained relatively unclear, particularly in higher functioning ambulatory youth. Given that physical activity generally supports healthier autonomic profiles in the broader population, understanding whether that same relationship holds and to what degree in youth with cerebral palsy has direct implications for how clinicians and rehabilitation specialists counsel families in design activity. Recommendations this research team conducted an exploratory cross sectional study involving 18 youth with cerebral palsy, 15 boys and three girls with an average age of roughly 15 and a half years. All participants were classified as level one on the Gross Motor Function Classification System, a widely used scale that describes functional mobility and cerebral palsy, with level one representing the most independently mobile end of the spectrum. Using the Physical Activity Questionnaire for Adolescents, a self report tool, the researchers grouped participants into low, moderate and high physical activity categories. They then assessed physical performance using two the shuttle run test, which evaluates cardiorespiratory endurance and and the muscle power sprint test, which captures short duration explosive performance. Heart rate variability was recorded at rest using wearable heart rate monitors comparing the low and high activity group. Specifically, the researchers found meaningful differences in both cardiorespiratory endurance and sprint performance, with the high activity group outperforming the low activity group on both measures. On the autonomic side, the high physical activity group showed a lower ratio of low frequency to high frequency power compared to the low activity group, a pattern generally interpreted as relatively less sympathetic dominance or put another way, a more parasympathetically balanced profile. Interestingly, there was no significant difference between groups in the standard deviation of NN intervals, a common overall marker of heart rate variability. The study also found non significant trends suggesting that higher self reported physical activity was associated with lower body mass index, greater shuttle run distance, faster sprint times, and higher standard deviation of NN intervals, values that were themselves associated with lower body mass index. Put together, the author's interpretation is that higher self reported physical activity levels were were associated with better physical performance and a potentially more favorable autonomic profile in youth with cerebral palsy. An encouraging if preliminary signal for a population where structured exercise recommendations are still evolving, it's helpful to place this study within the broader context of what's already known about autonomic function and cerebral palsy. Because cerebral palsy arises from injury to the developing brain and because the brain regions involved in motor control often overlap or interact with those involved in autonomic regulations. Some degree of altered cardiac autonomic control has been observed across prior work in this population independent of physical activity level. That backdrop makes the specific comparison in this study activity level within a cerebral palsy population rather than cerebral palsy versus typically developing peers, a genuinely useful contribution since it starts to tease apart how much of the autonomic picture in cerebral palsy might be modifiable through behavior as opposed to being a fixed feature of the underlying neurological injury itself. The fact that the low frequency to high frequency ratio differed between activity groups while the standard deviation of NN intervals did not, is also worth sitting with these two metrics don't always move together, and a dissociation like this suggests the activity related difference here may be more about the balance between sympathetic and parasympathetic influence specifically rather than an overall increase in total autonomic variability. For a population where global variability measures are sometimes reduced simply as a function of the underlying condition, a shift specifically in the balance between branches, even without a shift in the total amount of variability, could still represent a meaningful and encouraging signal. As the authors themselves flag, and as we want to be careful to underscore, this is a cross sectional study, which means it can only describe an association between physical activity level and physical performance or heart rate variability at a single point in time. It cannot establish that engaging in more physical activity causes these more favorable outcomes. Since the reverse is equally plausible, youth with inherently better cardiovascular or autonomic function might simply find it easier to be more physically active in the first place. [00:19:18] Beyond the causal question, the sample size here is Quite small, just 18 participants split across three activity groups, which sharply limits statistical power and the confidence we can place in any single comparison, including the finding that didn't reach conventional significance. Like the group difference in standard deviation of Indian intervals, the physical activity measure itself was self reported, which introduces the possibility of recall bias or participants over or underestimating their own activity levels. And because the sample was restricted to level one on the Gross Motor Function classification system, meaning the most independently ambulatory youth with cerebral palsy, these findings may not extend to youth with more significant motor impairment who represent a substantial portion of the broader cerebral palsy population. Finally, when multiple comparisons are run on a sample this size, there's always an elevated risk that some findings reaching statistical significance, like the low frequency to high frequency ratio difference, could reflect chance rather than a true underlying effect. The practical takeaway for clinicians and rehabilitation specialists working with ambulatory youth with cerebral palsy is that this exploratory data lends some support to encouraging physical activity and as part of a broader care approach consistent with what we already know from the general pediatric population. But given the small sample and cross sectional design, the specific autonomic findings here are best treated as hypothesis, generating a starting point for larger, ideally longitudinal studies rather than as a settled basis for specific clinical recommendations. It's also worth noting how this study fits into a broader pattern in pediatric rehabilitation research, where recruiting adequately powered samples is often genuinely difficult and given how heterogeneous conditions like cerebral palsy are across severity levels, ages, and functional profiles. 18 Participants may sound small in isolation, but studies like this one frequently represent years of patient recruitment within a specialized clinical population, and their real value often lies less in any single statistical result and more in building the foundation and the recruitment infrastructure that eventually makes larger multi site studies possible. Our fourth study moves into a high stakes clinical context using heart rate variability under postural challenge to help distinguish between different disorders of consciousness, a distinction that carries enormous weight for prognosis and rehabilitation planning. The authors are Weiqiangkai Xu, Han, Zhuojun, Chao Shijiang, Naren, Xiang, Tongji, Ziyu, Jun, Faw, Yi, Wu, Yuanyuan, Chu, Xin, Weitang, and Hong Yujie. It's worth noting that Kai, Han, Chao, and Zhang are credited as equally contributing first authors on this work. Distinguishing between a prolonged disorder of consciousness, which includes both vegetative state and minimally conscious state, and a patient who has emerged from a minimally conscious state is one of the more difficult and consequential judgments in neurorehabilitation. These distinctions are typically made using bedside behavioral scales. But behavioral assessment has real limitations, particularly in patients whose ability to demonstrate awareness is inconsistent or subtle. Because the autonomic nervous system is closely tied to brain regions involved in arousal and consciousness, there's growing interest in whether autonomic reactivity, how the body's heart rate and blood pressure regulation respond to a physical challenge could serve as an objective physiological complement to behavioral assessment. To test this, Kai and colleagues used a 75 degree head up tilt test, a well established method for challenging the cardiovascular system's radio regulatory responses by shifting a patient from lying flat to a near standing angle. 51 participants took part 18 patients with a prolonged disorder of consciousness broken down into seven in a vegetative state and 11 in a minimally conscious state 10 patients who had emerged from a minimally conscious state and 23 healthy controls. For comparison, the team recorded linear heart rate variability metrics during both the supine or lying flat position and during the tilt itself, analyzing the data using linear mixed effects models, a statistical approach well suited to handling repeated measurements taken from the same patients under different conditions. They also use Spearman's correlation, a statistical test for examining relationships between two variables that don't necessarily follow a straight line pattern, to see how heart rate variability related to scores on the Coma Recovery Scale Revised, a standardized behavioral assessment tool widely used to evaluate level of consciousness in these patients. The findings were layered. First, there was a significant main effect of time, meaning position itself lying flat versus tilted, significantly affected heart rate variability across the board, which is exactly what you'd expect physiologically. [00:23:16] More specifically, standard deviation of nnn intervals, low frequency power, high frequency power, and normalized low frequency power all showed significant group effects, with healthy controls showing significantly higher standard deviation of NN intervals, low frequency power and high frequency power than patients with a prolonged disorder of consciousness. When the researchers narrowed their comparison to just the two clinical groups, prolonged disorder of consciousness versus emergence from minimally conscious state, those group differences held up for standard deviation of NN intervals, low frequency power and high frequency power, with patients who had emerged from a minimally conscious state showing higher values across these measures than those still in a prolonged disorder of consciousness. And critically, the tilt induced changes in low frequency power and standard deviation of NN intervals were positively correlated with total scores on the Coma Recovery Scale Revised, meaning patients with more autonomic reactivity to the postural challenge also tended to show more behavioral evidence of consciousness on standard clinical assessment. The author's interpretation is that patients with a prolonged disorder of consciousness show a measurably weaker autonomic response to postural stimuli compared to both patients who have emerged from a minimally conscious state and healthy controls, and that a head up tilt based heart rate variability assessment holds promise as a non invasive tool to help differentiate between these conditions and potentially guide rehabilitation planning. It's worth unpacking why a postural challenge like head up tilt makes physiological sense as a probe of consciousness level in the first place. Regulating blood pressure and heart rate in response to a sudden shift toward upright posture requires rapid coordinated signaling between the brainstem, higher autonomic control centers and the cardiovascular system. Many of the same brain networks implicated in generating and sustaining conscious awareness overlap with or feed into this autonomic regulatory circuitry. The shared neural architecture is the theoretical basis for expecting that patients with more severely disrupted consciousness would also show more blended autonomic responses to a challenge like tilt. And it's exactly the pattern this study observed, which lends the findings a reasonable degree of biological plausibility beyond the statistics alone. The correlation between tilt induced changes in low frequency power and standard deviation of NN intervals on one hand and Coma Recovery Scale Revised scores on the other is arguably the most clinically interesting piece of this study. The Coma Recovery Scale Revised is already the gold standard behavioral tool and clinicians use to track a patient's trajectory over time, so finding that a physiological autonomic measure tracks alongside it opens the door to using heart rate variability as a complementary, more objective data point, one that doesn't depend on a patient's ability to reliably demonstrate a behavioral response, which can be inconsistent even in patients who do retain some level of awareness, there are meaningful limitations worth sitting with here. The subgroups involved are quite small, especially the vegetative state group at just seven patients, which sharply limits the statistical power available for the finer grained comparisons within the prolonged disorder of consciousness category. This is fundamentally a cross sectional comparison across groups defined at a single point in time. So while the associations between autonomic reactivity and consciousness level are compelling, this study alone can tell us whether improving autonomic reactivity would help drive recovery of consciousness, or whether it's simply a passive marker that tracks alongside recovery driven by other neurological processes. The the study also doesn't detail the underlying causes of the disorders of consciousness in its participants, whether traumatic brain injury, anoxic injury from cardiac arrest or another cause. [00:26:23] Different etiologies are known to carry different prognoses and potentially different autonomic profiles, which this analysis doesn't appear to separate out. And practically speaking, performing a 75 degree tilt test in medically fragile, often bedbound patients with disorders of consciousness carries logistical and safety considerations that could limit how widely this kind of testing is gets adopted outside of specialized centers. For clinicians working in neurorehabilitation, the takeaway is that autonomic reactivity to postural challenge looks like a promising low burden physiological signal that correlates meaningfully with established behavioral consciousness scales. And it's worth watching as this line of research matures. But at this stage it should be understood as a complement to careful behavioral assessment, not a replacement for it. There's also a broader point here about the value of bringing objective physiological measurement into a clinical space that has traditionally relied so heavily on behavioral observation. [00:27:19] Families and care teams making decisions about rehabilitation intensity, discharge planning or long term care often want more than a single bedside behavioral score to lean on, especially given how much day to day variability can occur in a patient's demonstrated responsiveness. A measure like tilt induced heart rate variability, particular precisely because it doesn't depend on the patient actively producing an observable behavior, could eventually serve as a useful second data stream to sit alongside the Coma Recovery Scale revised, particularly in ambiguous cases where behavioral scoring alone leaves real uncertainty about a patient's true level of awareness. That brings us to our sponsor break for this episode. This week's episode is sponsored by Optimal hrv, whose app supports a structured morning measurement protocol, longitudinal tracking of your heart rate variability trends over time, and built in biofeedback tools to help translate that data into practice. We also want to highlight two upcoming training opportunities for practitioners. The first is a BCIA aligned heart rate variability biofeedback training led by Dr. Ina Kazan, offering 16 APA continuing education credits. The second is a training on ethical principles and practice standards in clinical biofeedback. Also, BCIA align, led by Dr. Donald Moss, offering three APA continuing education credits. Full details and registration links for both trainings, along with more on the Optimal HRV app are available in this episode's show. Notes now back to the research Our fifth study takes us into the world of machine learning using heart rate variability derived from photoplethysmography to predict cardiovascular disease risk in older adults. The authors are Kouat Abzaliev, Agbota Bugubayeva, Simbad Abzaliyeva, Gulsimachmetova, Gulzira Balkanay, Aliya Omarbayeva, Sakhen Anartayev, Nazima Zarubekova and Medina Suleymanova. Cardiovascular disease remains the leading cause of illness and death among older adults worldwide, and there's enormous clinical interest in finding accessible, non invasive tools that can flag elevated risk before a major event occurs. Photoplethysmography, often abbreviated ppg, is the same basic optical technology used in many consumer wearables. It measures blood volume changes at the skin surface using light, and from that signal heart rate variability can be derived without the need for a full electrocardiogram setup. Combining that accessible measurement approach with machine learning raises the possibility of scalable, relatively low cost cardiovascular risk screening, which is exactly what this study set out to explore. Abzalief and the research team enrolled 100 participants aged 65 and older, dividing them into those identified as having Cardiovascular disease risk, 54 participants and those without, 46 participants. Photoplethysmography Recordings were collected using a dedicated device and software system, and from those recordings the team derived a range of time domain and spectral heart rate variability parameters. They then built and evaluated machine learning models, specifically logistic regression, a standard statistical classification method, and a random forest algorithm, which combines many decision trees to make more robust predictions using a nested cross validation framework, a rigorous approach designed to reduce the risk of a model appearing more accurate than it truly is by testing it on data it hasn't seen during training. To understand which features were driving the model's predictions, they used SHAP or Shapley additive explanations, an interpretability technique that assigns each input variable a quantified contribution to a given prediction. The differences between the two groups were substantial. Participants with cardiovascular disease risk showed a standard deviation of NN intervals, roughly half that of the no risk group. A Percentage of NN 50 intervals about 3 1/2 times lower and overall heart rate variability around two and a half times lower. On the frequency side, very low frequency power was reduced roughly one fold, low frequency power was reduced about four fold and high frequency power was reduced about five fold, all statistically significant differences. At the same time, the ratios of very low frequency to high frequency power and low frequency to high frequency power were both significantly elevated in the at risk group, a pattern generally interpreted as a shift towards sympathetic predominance and away from parasympathetic influence. When it came to prediction, the best performing model was the random forest, using a combined set of clinical features and heart rate variability signs, specifically high frequency power and root mean square of successive differences, achieving a receiver operating characteristic area under the curve, a standard measure of how well a model distinguishes between two groups of 0.9988, which is remarkably close to perfect discrimination. The authors interpret these results as demonstrating a close association between cardiovascular disease risk and autonomic nervous system dysfunction in older adults, and is evidence that combining machine learning with photoplithysmography derived heart rate variability features meaningfully improves diagnostic and predictive performance compared to models built on clinical data alone. It's worth spending a little more time on why photoplithysmography specifically matters for the future of this kind of screening. Unlike a full electrocardiogram setup, which typically requires adhesive electrodes placed on the chest or limbs, photopletysmography can be captured from a fingertip, wrist or earlobe using nothing more than a light source and a sensor, the same basic technology already embedded in many consumer smartwatches and fitness trackers. If heart rate variability features derived from this kind of low friction measurement can genuinely support cardiovascular risk prediction, the implications for scalable, low cost real world screening in older adult populations are considerable, particularly in settings where routine electrocardiogram based screening isn't practical or accessible. That said, the specific machine learning finding here also offers a useful teaching moment about how to read model performance metrics. Critically, a receiver operating characteristic area under the curve is bounded between 0.5 representing no better than chance, and 1.0 representing perfect discrimination between groups. Genuinely excellent clinical prediction models in cardiovascular medicine typically land somewhere in the range of 0.75 to 0.85. Values above 0.95 are unusual even in large, well validated data sets, and a value as high as 0.9988 in a sample of only 100 participants is the kind of result that experienced researchers tend to treat with real caution rather than celebration precisely because it's more consistent with a model having learned quirks specific to this particular sample than with a truly robust generalizable signal. Now that near perfect accuracy figure deserves some direct scrutiny, and we'd be doing you a disservice if we didn't flag it. Clearly, a receiver operating characteristic area under the curve of 0.9988 in a sample of only 100 participants is an extraordinarily high number and results this strong in a relatively small data set that can sometimes reflect the model overfitting to particular patterns in that specific sample rather than capturing a truly generalizable signal. Nested cross validation is a genuine safeguard against some forms of overfitting, but it doesn't eliminate every risk, especially with a modest sample size split across two groups. This was also a single center study using one specific photoplithysmography device and software system, which means we don't yet know whether these exact results would replicate with different hardware, different populations or larger multi site samples. [00:33:28] As with several studies this week, the design here is cross sectional. Participants were classified into risk categories and compared at a single point in time, so we can describe a strong association between autonomic markers and cardiovascular disease risk status. But we can't say from this data alone whether autonomic dysfunction precedes and contributes to rising cardiovascular risk or whether it's a downstream consequence of early subclinical cardiovascular changes that were already underway. The sample was also restricted to adults 65 and older, so these these findings shouldn't be assumed to extend to younger populations, and the study doesn't fully detail the clinical criteria used to classify participants into the cardiovascular disease risk category in the first place, which matters for anyone trying to reproduce or build on this work. The practical takeaway is that this is a genuinely exciting proof of concept for accessible, wearable compatible cardiovascular risk screening in older adults, and the underlying pattern of autonomic changes is entirely consistent with what we already understand about cardiovascular disease and autonomic function. But given the sample size and the almost suspiciously perfect accuracy figure, this is a result that calls for replication in larger independent multi site samples before anyone should consider translating it into clinical screening tools for practitioners who work with older clients and are considering wearable based screening tools built on similar technology, the sensible posture right now is measured enthusiasm. The underlying physiological story, reduced overall variability, reduced parasympathetic markers and and a shift towards sympathetic predominance in those at elevated cardiovascular risk is well supported by decades of prior heart rate variability research independent of this particular machine learning model's accuracy figure that underlying physiology is trustworthy. It's specifically the exact predictive performance reported here that deserves a healthy dose of skepticism until it's been replicated elsewhere. Our sixth study is a change of pace, a small, deliberately modest feasibility study focused not on testing a clinical hypothesis, but on figuring out what it actually takes to run continuous, continuous 24 hour heart rate variability monitoring in practice. The Authors are Charles N.R. henderson, Monica Smith, Dale F. Johnson and Phyllis K. Stein. This kind of groundwork often doesn't get much attention, but it's genuinely valuable, especially for a show like ours where longitudinal real world heart rate variability tracking is a recurring theme. Before any research team or any practitioner recommending extended monitoring to clients can meaningfully interpret 24 hour heart rate variability data, someone has to work out the practical logistics. Does the equipment hold up over a full day and night? Does it interfere with normal activity? And what kinds of data quality problems are likely to crop up? Henderson and colleagues set out with three specific to evaluate heart rate variability monitoring across morning, midday, evening and sleep periods to assess the study procedures, personnel and physical resources available at their own institution for running this kind of protocol and to identify potential problems along with possible solutions associated with 24 hour heart rate variability monitoring. Six participants were fitted with a wearable continuous heart rate variability recording device and monitored across three normal daily activity treadmill walking and paced deep breathing, a controlled breathing exercise often used to elicit a clear, reliable respiratory influence on heart rate variability. Throughout the monitoring period, participants also kept written event diaries to log their activities. The device performed well overall, providing good quality data across the full 24 hour period without meaningfully interfering with participants normal daily activities. But the study surfaced some genuinely useful practical problems. There was roughly an hour of data loss following device exposure to water during showering, a straightforward but important reminder that not all wearable monitoring equipment is fully shower proof. The team also identified a need to better stabilize a specific electrode wire, referred to as the V5 lead, to reduce disconnections that were occurring during sleep when participants naturally move around more and are less able to notice or correct a loose sensor. And on the human side, the usefulness of participants written diaries varied considerably. Some captured detailed, meaningful context about their day, while others provided much thinner information, which matters because diary data is often what researchers use to interpret unusual patterns in the physiological recording. The author's own conclusion is appropriately modest. Based on these six participants, 24 hour heart rate variability assessment appears feasible for use in research studies at their institution and the specific problems identified here around water exposure, electrode stability and diary compliance can inform how future studies in this space are designed it's worth making the case directly for why a study like this deserves airtime on a show that usually focuses on hypothesis testing research Continuous real world heart rate variability monitoring has become a central feature of both clinical research and consumer wellness technology, but the vast majority of published work simply assumes the underlying data collection process works smoothly and jumps straight to analyzing the resulting number numbers. Very little published research actually documents in careful detail what goes wrong along the way, where devices fail, where data gets lost, and where the human side of the protocol, like participant diary keeping breaks down. That's precisely the gap this study fills, and findings like these tend to quietly shape the design of dozens of later studies that build on the same monitoring approach, even though they rarely get the same visibility as a large clinical trial. It's also worth noting how directly relevant these specific findings are to anyone using or recommending wearable heart rate variability devices in practice well beyond the research context. Water exposure during showering or bathing is one of the most common real world scenarios in which continuous monitoring gets interrupted, and the electrode stability issue the authors identified during sleep speaks to a challenge that shows up repeatedly across the wearable monitoring literature. Nighttime is often when the richest recovery related data is captured, but it's also when equipment is most prone to shifting or loosening without the wear noticing. The limitations here are largely built into the study's own design and purpose and worth stating plainly. With only six participants, this was never intended to test a clinical or physiological hypothesis, and it shouldn't be read as evidence for or against the validity of 24 hour heart rate variability as a meaningful wellness or clinical biomarker. That's simply not what this study was designed to evaluate. It's also a single institution study using one specific monitoring device, so the practical problems identified and the solutions the authors propose may not generalize to other devices, other research environments or non research real world monitoring contexts. The evaluation period, while spanning 24 hours, represents a single day and night for each participant rather than a sustained multi day monitoring window, which is often the more clinically or practically relevant use case. That said, for anyone piloting extended real world heart rate variability monitoring protocols, whether in a research setting or with clients in practice, this is a genuinely practical roadmap. Pay close attention to how your monitoring device handles water exposure, invest time in properly securing electrodes before overnight wear, and think carefully about how you're going to capture contextual information from participants. Since the raw physiological data alone often isn't enough to interpret what happened during the day, there's a broader lesson here too, for practitioners who may be tempted to skip straight to interpreting 24 hour heart rate variability trends without first stress testing their own measurement setup. A single hour of lost data from an unexpected shower or a night of unreliable readings from a loosely placed electrode can meaningfully distort a day's summary statistics if it isn't identified and accounted for, building in even a brief feasibility or pilot phase before committing to a larger monitoring protocol. Exactly the kind of exercise this study represents can save considerable time and prevent misleading conclusions further down the line. Our seventh and final study for this week brings us to the world of competitive athletics, looking at resting and activity related heart rate variability in a group of footballers in Senegal, a population and geographic context that's historically underrepresented in the sport science literature. The authors are Abdu Kadir, Soh, Sharif Hussain, Ulay Thiom, Mor, Diao Fatou Kinendoye, Mami Salam, Kali Awaba, Salamata de Agnihanjo, Maimouna, Tour Aisatucek, Fabien Brejon, Stephane Dalio and Abdoulaye Bah. Heart rate variability is well established as a marker of cardiac autonomic control and it's known to be shaped by physical training. Well trained athletes typically show a distinct autonomic signature characterized by strong resting parasympathetic tone and an appropriately responsive shift towards sympathetic activation under physical or orthostatic challenge. Much of the foundational research establishing these patterns, though, has come from a fairly narrow set of countries and competitive contexts, which is part of what makes this study valuable. It extends that evidence base to a West African professional football setting that hasn't been well represented previously so, and the research team conducted a cross sectional descriptive study of 32 male footballers from AS Camborine, a club competing in Senegal's second tier football league. Using a Holter electrocardiogram monitor, a portable device that continuously records the heart's electrical activity, the team assessed heart rate variability across four quiet rest, orthostatic positioning, meaning a shift from lying down to standing, the rough ear test, a standardized step based cardiovascular fitness assessment and a subsequent recovery period. They analyzed both time domain and frequency domain heart rate variability parameters alongside indices of cardiac adaptation derived from the rough ear test itself. [00:41:59] The findings were consistent with what we'd expect from a well trained athletic population and reassuring from a cardiovascular health standpoint. The majority of players showed resting sinus bradycardia, meaning a resting heart rate slower than the general population average originating from the heart's normal pacemaker. A common and generally healthy finding in trained endurance and team sport athletes. According to the Rough Year Index, cardiovascular adaptation was considered normal across the entire group, with no players flagged as showing a concerning response to the fitness test. [00:42:25] Looking at the autonomic patterns, specifically markers of parasympathetic tone root mean square of successive differences, the percentage of NN50 intervals and high frequency power expressed in normalized units were normal to elevated at rest, decreased during both the orthostatic challenge and the rough ear test, and then increased again during the recovery period, tracing exactly the kind of responsive curve you'd want to see. Meanwhile, markers of sympathetic tone, specifically low frequency power and normalized units in the low frequency to high frequency ratio moved in the opposite direction throughout rising during the challenges and falling during recovery. The researchers also found that good cardiovascular adaptation as measured by the Rough Year Index was significantly linked to good resting overall heart rate variability measured as total power with a correlation coefficient of -0.35. Put together, the authors conclude that these football players demonstrated good total heart rate variability at rest and responded normally to both physical and orthostatic stress, with recovery patterns consistent with healthy autonomic function and that overall cardiovascular adaptation was meaningfully linked to resting variability. It's worth reflecting on why geographic and competitive diversity matter so much in sports science research. Specifically, training loads, climate, nutrition access, travel demands, and even cultural approaches to rest and recovery can all plausibly shape an athlete's autonomic profile. Yet a large share of the foundational heart rate variability research in sport has historically come from a fairly narrow band of well resourced football and endurance programs in Europe, North America and parts of Asia. A study like this one conducted within a Senegalese second tier football club doesn't just add another data point. It tests whether the physiological patterns researchers have come to expect in trained athletes actually hold up outside of that narrow band, which matters both scientifically and practically. For sports scientists and practitioners working with athletes in underrepresented regions, the specific combination of assessments used rest, orthostatic challenge, the rough ear test, and recovery also mirrors a broader trend in athlete monitoring toward multi condition heart rate variability testing. Rather than a single resting measurement, resting heart rate variability alone can be a somewhat blunt instrument since two athletes with similar resting values might respond quite differently once challenged. Capturing the full arc from rest through challenge to recovery, as this study did, gives a much richer picture of autonomic flexibility, which is arguably a better proxy for both fitness and readiness than any single static number. A few caveats are worth naming. Clearly, this is a cross sectional single time point study, meaning it captures a snapshot of these players autonomic profiles at one moment, rather than tracking how heart rate variability shifts across a training cycle, a competitive season or in response to accumulated fatigue, all of which are highly relevant in athletic populations and would require a longitudinal design to properly capture. The sample is also relatively small and drawn entirely from a single club, all male, all from one competitive level, which limits how confidently these specific findings generalize to female and athletes, other sports, higher or lower levels of competition, or footballers training under different climates and conditions. This was a descriptive study without a non athlete comparison group, which means that while the autonomic pattern observed here is consistent with what's typically seen in trained athletes, this particular study can't directly demonstrate that these findings are attributable specifically to athletic training as opposed to other factors like age, genetics or lifestyle, since there was no untrained comparison group included. And on the correlation finding linking cardiovascular adaptation to resting variability, a coefficient of minus 0.35, while statistically significant, reflects a fairly modest relationship. It explains only a limited portion of the variation in cardiovascular adaptation across players, so it shouldn't be overstated as a strong or dominant predictor. The broader takeaway here is an encouraging one for the field. The classic autonomic signature of trained athletic fitness strong resting vagal tone, appropriate sympathetic engagement under challenge and clean recovery holds up in a population and geographic context that hasn't previously received much research attention, which is a meaningful contribution in its own right. At the same time, this cross sectional snapshot leaves the door wide open for future longitudinal work tracking these players across a full season. That brings us to the end of our study rundown for this week. But before we close, it's worth stepping back and noticing a few threads that ran across all seven of these studies. The first is just how consistently a small handful of heart rate variability markers, standard deviation of NN intervals, root mean square of successive differences, and the low frequency to high frequency ratio kept showing up as meaningful discriminators whether the comparison was overweight versus normal body mass index, young adults, patients with different levels of consciousness, older adults with and without cardiovascular disease risk, or trained athletes under physical challenge. That's a reminder of just how much information these core relatively simple metrics continue to carry across wildly different populations. [00:46:51] Clinical Questions the second thread is one we found ourselves repeating almost every single study this week. Nearly all of today's research was cross sectional or observational in design, comparing groups at a single point in time rather than following the same individuals as they change. That's not a criticism of the researchers. Cross sectional designs are often the practical, sensible starting point for a new question. But it does mean that across the board, what we're looking at are associations, not proof of cause and effect. Whether it's body weight and autonomic recovery, physical activity and autonomic balance in youth with cerebral palsy or autonomic reactivity and level of consciousness, the honest scientific position is that these variables travel together, not that we've proven one drives the other. The third thread is about scale and technology. This week gave us a nice cross section of where the field is headed methodologically, from a relatively low tech six person feasibility study working out the basic logistics of continuous monitoring to a sophisticated machine learning model built on photoplithysmography data achieving eyebrow raising, predictive accuracy. Both ends of that spectrum matter. The unglamorous feasibility work is what eventually makes larger, more ambitious studies possible. And the machine learning work shows us where the ambition can lead so long as we stay appropriately skeptical of results that look almost too good to be true, as we discussed with that cardiovascular disease prediction model. And finally, sample size varied enormously across today's studies, from just six participants in our feasibility study, up to 200 in our opening study on body mass index and exercise recovery. That range is worth keeping in mind anytime you're reading a new heart rate variability paper. A striking finding from a small exploratory sample deserves genuine interest and further investigation, but not the same confidence you'd place in a well powered replicated result. Holding both of those things at once curiosity and appropriate caution is really what good scientific literacy in this field looks like. [00:48:40] There's one more pattern worth naming before we close, and it's less about the statistics and more about who these studies were actually about. This week's lineup covered healthy young adults, office workers, youth with a developmental motor condition, patients with severe brain injury, older adults at cardiovascular risk, research volunteers helping validate a monitoring protocol, and professional athletes on another continent. That's an unusually broad cross section of humanity for a single episode, and it's a good illustration of just how versatile heart rate variability has become as a shared scientific language across specialties that otherwise have very little in common. A pediatric rehabilitation specialist and a sports scientist working with elite footballers are in a real sense speaking the same physiological dialect when they talk about RMSSD or the low frequency to high frequency ratio, even though their patients and athletes could hardly be more different. That shared vocabulary is part of what makes a show like this possible in the first place, and and it's part of why cross pollination between these specialties, borrowing methods and interpretive frameworks from each other, tends to move the whole field forward faster than any one subfield working in isolation. It's also worth reflecting on what today's studies collectively suggest about where autonomic nervous system research still has the most ground to cover. We heard again and again this week about the need for larger samples, longitudinal follow up and replication across sites and devices. Not because today's researchers did anything wrong, but because that's simply the nature of a maturing field working through its early evidence space. The studies that will matter most over the next several years are likely to be the ones that take today's promising, often small scale associations and test them properly, following the same overweight young adults over time to see whether autonomic recovery actually predicts later cardiovascular outcomes, tracking the same footballers across a full competitive season, or validating that machine learning model on cardiovascular risk in a data set 10 or 20 times larger. None of that diminishes the value of the work we covered today, though. [00:50:22] Every mature field needs exactly this kind of exploratory groundwork. But it does help set expectations for what comes next, and it gives us a good preview of what future episodes of this show are likely to be reporting on. That's all for this week's research roundup. Until next week, keep measuring, keep questioning and keep learning. This has been this week in heart rate variability.