Kumar Rajnish, Fu Junhan, Ortiz Bengie L, Cao Xiao, Shedden Kerby, Choi Sung Won
Department of Pediatrics, University of Michigan Medical School, Ann Arbor, MI 48109, USA.
Department of Statistics, College of Literature, Science, and the Arts, University of Michigan, Ann Arbor, MI 48109, USA.
Bioengineering (Basel). 2024 Jan 18;11(1):95. doi: 10.3390/bioengineering11010095.
Twenty-four-hour heart rate (HR) integrates multiple physiological and psychological systems related to health and well-being, and can be continuously monitored in high temporal resolution over several days with wearable HR monitors. Using HR data from two independent datasets of cancer patients and their caregivers, we aimed to identify dyadic and individual patterns of 24 h HR variation and assess their relationship to demographic, environmental, psychological, and clinical variables of interest.
a novel regularized approach to high-dimensional canonical correlation analysis (CCA) was used to identify factors reflecting dyadic and individual variation in the 24 h (circadian) HR trajectories of 430 people in 215 dyads, then regression analysis was used to relate these patterns to explanatory variables.
Four distinct factors of dyadic covariation in circadian HR were found, contributing approximately 7% to overall circadian HR variation. These factors, along with non-dyadic factors reflecting individual variation exhibited diverse and statistically robust patterns of association with explanatory variables of interest.
Both dyadic and individual anomalies are present in the 24 h HR patterns of cancer patients and their caregivers. These patterns are largely synchronous, and their presence robustly associates with multiple explanatory variables. One notable finding is that higher mood scores in cancer patients correspond to an earlier HR nadir in the morning and higher HR during the afternoon.
24小时心率（HR）整合了与健康和幸福相关的多个生理和心理系统，并且可以使用可穿戴心率监测器在几天内以高时间分辨率进行连续监测。利用来自癌症患者及其照顾者的两个独立数据集的心率数据，我们旨在识别24小时心率变化的二元和个体模式，并评估它们与感兴趣的人口统计学、环境、心理和临床变量之间的关系。
采用一种新颖的正则化高维典型相关分析（CCA）方法，以识别反映215对二元组中430人的24小时（昼夜节律）心率轨迹的二元和个体变化的因素，然后使用回归分析将这些模式与解释变量相关联。
发现了昼夜节律心率中四个不同的二元协变因素，约占昼夜节律心率总变化的7%。这些因素，连同反映个体变化的非二元因素，与感兴趣的解释变量呈现出多样且具有统计学稳健性的关联模式。
癌症患者及其照顾者在其全天24小时的心率模式中均存在二元和个体异常。这些模式在很大程度上是同步的，并且它们的存在与多个解释变量密切相关。一个值得注意的发现是，癌症患者较高的情绪得分对应于早晨较早出现的心率最低点以及下午较高的心率。
