Page 110 - AIH-2-4

Artificial Intelligence in Health Autonomic nervous system patterns in men

diagnosis and prevention of CVDs, ultimately reducing for categorizing and analyzing variations in autonomic
morbidity and mortality. 2 function. Importantly, it investigates heterogeneity within a
Heart rate variability (HRV) measures the beat-to- healthy population, rather than attempting to differentiate
beat fluctuations in R-R intervals (RRI) as recorded by between healthy individuals and those with disease. It is
an electrocardiogram (ECG). As a non-invasive metric, hypothesized that, even among asymptomatic young men,
HRV provides quantitative insight into the activity of there are distinct patterns of autonomic regulation that are
the autonomic nervous system (ANS) by reflecting its physiologically meaningful and may represent different
modulation of cardiac function. Elevated HRV is generally trajectories of cardiovascular health long before clinical
3
associated with enhanced cardiovascular adaptability disease manifests.
and greater resilience to stress, whereas reduced HRV This stratification can be achieved through an
may reflect underlying autonomic dysfunction. Due to integrated analytical framework combining principal
4
its sensitivity, HRV has emerged as a valuable biomarker component analysis (PCA) and K-means clustering. PCA,
for assessing autonomic regulation and cardiac function, an unsupervised dimensionality reduction technique, is
playing a vital role in the early detection and prevention used during data pre-processing to maximize variance
of CVDs. 5 preservation within a reduced two-dimensional space,
However, the practical application of HRV monitoring, thereby facilitating improved visualization and subsequent
particularly in field settings such as athletic training, clustering. 15,16 The K-means algorithm is then applied to
is often limited by the cost, complexity, and bulk of these reduced dimensions to group individuals according
conventional ECG equipment. Wearable technologies, to their position in the new analytical space. As an iterative
6
®
such as the Polar Heart Rate Monitor Interface Board, partitioning method, K-means require a pre-specified
provide an affordable, high-quality alternative for acquiring number of clusters (k) and operate by minimizing intra-
beat-to-beat RRI data in real-world environments. These cluster distance while maximizing inter-cluster distance
7
17,18
devices enable athletes and coaches to access objective to form distinct, coherent groups. This study aims
physiological feedback, thereby enhancing performance to develop and validate a method that integrates PCA
monitoring and supporting personalized, data-driven and K-means clustering to identify distinct patterns of
training adjustments. 8 autonomic regulation in healthy men using HRV data.
The stratification of cardiovascular autonomic 2. State of the art
function, as assessed by HRV in healthy men, provides 2.1. HRV
valuable insights into ANS regulation and classification
of cardiovascular health. As a non-invasive indicator of HRV refers to the variation in the time intervals between
9
autonomic control, HRV quantifies the constant interplay successive heartbeats. As a well-established indicator
19
between the sympathetic and parasympathetic nervous of ANS activity, HRV offers a non-invasive means of
20
system’s inputs to the heart. From a clinical perspective, a assessing cardiac autonomic regulation. Historically,
high level of HRV suggests that the cardiovascular system accurately evaluating autonomic function has presented
is resilient and able to cope with stress. Furthermore, a significant challenge for cardiologists. A breakthrough
10
21
understanding the stratification of ANS regulatory occurred in 1981 when Akselrod et al. demonstrated
patterns may provide valuable insights into the individual that specific components of HRV correspond directly
cardiovascular health profiles and inform targeted to parasympathetic and sympathetic nervous system
prevention strategies. activity. Building on this, a 1987 study by Kleiger et al.
22
The integration of artificial intelligence (AI) with a focus established the clinical prognostic value of HRV, showing
on machine learning offers significant potential to advance that it serves as a predictor of mortality risk following
the analysis of HRV and deepen our understanding myocardial infarction. In recent years, HRV analysis has
of the regulatory mechanisms of the ANS. This data- been widely recognized as a reliable, non-invasive method
23
driven approach is crucial for shifting the focus from for evaluating ANS modulation at the heart’s sinus node.
population-level analyses to more precise and personalized Time-domain analysis of HRV quantifies the variability
evaluations of autonomic health. 4,5,11-14 which underpins in RRI using direct statistical measurements. Key
the rationale for this research. The present study aims to parameters include the mean normal-to-normal (NN)
explore ANS regulation a critical aspect of cardiovascular mean R-R interval (MRR), standard deviation of NN
health through the application of AI-based methods. By intervals (SDNN), root mean square of successive RRI
stratifying individuals into distinct groups based on their differences (RMSSD), and the proportion of adjacent NN
HRV patterns, this study proposes a novel framework intervals differing by more than 50 ms (pNN50). The
24

Volume 2 Issue 4 (2025) 104 doi: 10.36922/AIH025050006