Heart Rate Variability, Cognitive Performance and Cognitive Impairment in Older Adults: Cross‐Sectional Results From the Einstein Aging Study

Abstract

ABSTRACT

Aims

Although prior work has examined the relation of heart rate variability (HRV) to cognitive impairment, findings have been inconsistent. The association of cardiac vagal control with cognitive impairment remains unclear. Our goal was to examine the association of high frequency HRV (hf‐HRV) with mild cognitive impairment and global cognition in a community‐based sample of older adults.

Methods

84 participants (mean age 78.1 SD 5.2 years) wore single lead ECG devices for 6‐9 days. HRV in the high (0.15–0.40 Hz, [HF‐HRV]) frequency band was derived using power spectral analyses. The cognitive battery included the Montreal Cognitive Assessment (MoCA) to assess global cognition, and two tests per domain for memory, executive, language, visuo‐spatial and attention. Mild Cognitive Impairment (MCI) was defined using Jak‐Bondi criteria. Domain specific impairment was defined as scores > 1.0 SD below age, sex, education standardized norms on at least one test in a domain. Associations of HF‐HRV with cognition were examined using logistic and linear regression adjusted for demographics, diabetes, and hypertension.

Results

Participants were 82% female; 39% Non‐Hispanic White, 44% Non‐Hispanic Black, 25% had MCI. Within domains, impairment was present in 25% (memory), 28.6% (executive function), 21.4% (attention), 17.9% (language), and 27.4% (visuospatial). HF‐HRV was inversely associated with prevalent MCI (OR per 1 SD increase in ln HF‐HRV: 0.47, p = 0.02) and with memory impairment (OR per 1 SD increase in ln HF‐HRV: 0.52, p = 0.03). Higher HF‐HRV was associated with higher MoCA score (β for 1 SD increase in ln HF‐HRV = 0.65, p = 0.046).

Conclusion

Higher hf‐HRV, indicative of greater cardiac parasympathetic control is associated with lower odds of MCI, or memory impairment and with better global cognition after adjustment for cardiovascular risk factors. Future longitudinal studies are needed to confirm these associations.

Abbreviations

• EAS
• Einstein aging study
• HRV
• heart rate variability
• HF-HRV
• high frequency heart rate variability
• LF-HRV
• low frequency heart rate variability
• MCI
• mild cognitive impairment
• MoCA
• Montreal cognitive assessment
• RMSSD
• root mean square of successive differences
• SDNN
• standard deviation of all normal RR intervals

Introduction

The autonomic nervous system is the primary regulator of homeostasis, and its function has been linked to health outcomes [1]. There is increasing interest in better understanding the role of the autonomic nervous system in cognitive impairment [2, 3]. Heart rate variability (HRV), the variation in an RR interval time series derived from continuous electrocardiographic recordings, results from autonomic modulation of sinus node activity [4] and provides a feasible, noninvasive means of quantifying cardiac autonomic control in clinical and population based research settings [4–7].

Several lines of reasoning have led to interest in better understanding the role of autonomic function, particularly vagal control, in the development of cognitive impairment and dementia. Brain regions affected by Alzheimer's disease are also involved in autonomic function and autonomic control of cardiac function is involved in the regulation of cerebral perfusion [8]. In addition, autonomic dysregulation as indexed by HRV may be mechanistically linked to cognition via vascular or inflammatory processes. Low HRV is a marker of poor cardiovascular health [9–11] and cardiovascular disease is associated with dementia with vascular pathology contributing to at least half of all cases [12]. Alternatively, neuroinflammation is involved in the pathogenesis of Alzheimer's disease and vagus nerve dysfunction leads to increased peripheral inflammation [13]. Thus, low HRV may be related to poorer cognitive function via direct effects on brain perfusion or indirectly via vascular and inflammatory processes.

While a number of studies have reported on relationships between HRV and cognitive function there is heterogeneity regarding the assessment methods and specific HRV indices examined [3]. HRV may be quantified using either time or frequency domain approaches. Time domain measures include the standard deviation of all normal RR intervals, (SDNN), which reflects combined sympathetic and parasympathetic effects; and the root mean square of successive differences (RMSSD) which more specifically reflects cardiac parasympathetic modulation [14]. Frequency domain indices may be derived from spectral analysis of RR interval time series in which variation is decomposed into discrete frequency bands. HRV in the low frequency band (0.04–0.15 Hz, LF-HRV) reflects the joint effects of the sympathetic and parasympathetic nervous systems [4, 15]. High frequency HRV (0.15–0.40 Hz, HF-HRV) is modulated primarily by the parasympathetic nervous system that has the capacity to rapidly alter heart rate via acetylcholine input to the sinus node [4, 15]. In sum, while RMSSD and HF-HRV may be interpreted as reflecting vagal cardiac control, findings for SDNN or LF-HRV are more difficult to interpret given that they represent HRV stemming from multiple sources. Laborde et al. [14] have highlighted the importance of selecting HRV measures which reflect the mechanistic pathways of interest, noting that that the major theories linking HRV to psychological and physiological function focus specifically on the vagus.

Further, while work in various age groups has examined HRV concurrently with performance on a mental task [2], fewer studies have examined the relation of HRV to cognitive impairment within older, non-demented adults. Better understanding the association of HRV with mild cognitive impairment is important given that it is an intermediate state that precedes clinical dementia, and the critical need to identify factors that identify individuals at risk for dementia early in the disease process. Results have been inconsistent [3, 6, 16], and data regarding the association of HRV to cognitive impairment within specific cognitive domains are particularly limited [3]. Most prior studies have been limited to either a single clinic-based HRV assessment from a brief ( < 5 min) resting ECG [6, 17–21], or to ambulatory HRV measures over only two [22] or 24 h [23] that do not capture the dynamic nature of HRV in a real world setting where responses to naturally occurring challenges (e.g., daily stress, activity) may be relevant [23].

The objective of this analysis is to examine the association of frequency domain measures of HF-HRV with mild cognitive impairment (MCI), impairment within specific cognitive domains, and with global cognition based on the Montreal Cognitive Assessment (MoCA) [24] within older adults enrolled in the Einstein Aging Study (EAS). Ambulatory continuous ECG recordings over 7 days facilitated assessment of average HRV levels in real world settings. We focus specifically on HF-HRV given that it may be interpreted as an indicator of parasympathetic control and the relevance of vagal cardiac control to previously hypothesized mechanisms that may underlie associations between HRV and cognition. The primary hypothesis was that lower HF-HRV would be associated with worse cognitive performance and with MCI.

Methods

Study Population

The EAS is a longitudinal population-based study of older adults in Bronx County, NY [25]. Since 1993, participants age 70 and above have been systematically recruited using sampling frames generated from Health Care Financing Administration/Centers for Medicaid and Medicare Service rosters, or since 2004, New York City Board of Elections registered voter lists for Bronx County [25]. To be eligible, participants were required to be age 70 years or older, fluent in English, and to not meet diagnostic criteria for dementia at enrollment [25]. Informed consent was obtained following a protocol approved by the IRB of the Albert Einstein College of Medicine, Bronx, NY.

Between February 2018 and March 2020, an ambulatory HRV protocol was completed by 132 EAS participants. For the present analyses, we excluded individuals if they reported using medications which may impact the autonomic nervous system: (mirtazapine, trazodone, digoxin, ACE inhibitors, benzodiazepines, nortriptyline; N = 19). Also excluded were individuals who provided less than 6 days of continuous ECG data, or for whom the data processing algorithm identified less than 100 valid 5-min ECG epochs per day (N = 29, see methods below). This resulted in an analysis sample of 84 individuals (Figure 1). Demographic characteristics of this sample at enrollment were similar to those in the overall EAS cohort with the exception that the analysis sample had a higher percentage of females (82.1% vs 65.7% overall).

[IMAGE OMITTED. SEE PDF]

HRV Assessments

Participants wore a small, single lead ECG device (MOX3 ECG and Physical Activity Recording System; Maastricht Instruments, Maastricht NL) which was applied at an in-person study visit by a certified medical technician. Individuals were instructed to wear the device continuously for 7 days. This occurred within 1 week of the participant completing in-person neuropsychological testing as described below. Participants returned devices at a second clinic visit within the next 2 weeks as part of the overall EAS protocol. ECG data were digitized at 1000 Hz. The continuous ECG signal was downloaded and transferred to the reading lab at Columbia University (R. Sloan PI). ECG waveforms were submitted to custom-written software that detected the time of each R wave, resulting in RR interval (RRI) time series. Ectopic beats were identified using algorithms available on Physionet (). Values corresponding to ectopic beats or to noisy signals due to poor electrode contact and or movement artifact were corrected by interpolation. HF-HRV (0.15-0.40 Hz) was computed based on 300-second epochs, using an interval method for computing Fourier transforms similar to that described by DeBoer, Karemaker and Strackee [26]. Before computing Fourier transform, the mean of the RRI series was subtracted from each value in the series. The series was filtered using a Hanning window [27] and the power over the HF band was summed. Estimates of spectral power were adjusted to account for the resultant attenuation [27]. For each person a summary HF-HRV estimate was obtained by averaging across all 300-second epochs over all days. All HF-HRV values were natural log transformed for analysis, as in prior studies [14] for normalization and variance stabilization purposes.

Cognitive Assessments

Dementia is an exclusion for MCI which is an intermediate state that precedes clinical dementia. Thus, dementia defined at consensus case conference using DSM-IV criteria [28] was an exclusion for this analysis. DSM-IV rather than DSM-V criteria were used given that the EAS is a longitudinal cohort, and consistency in diagnostic criteria over time is critical. Cognitive function was assessed using two neuropsychological tests within each of five domains: (1) Memory: Free and Cued Selective Reminding Test-Free recall [29] and Benson Complex Figure Delayed [30]; 2) Executive Function: Trail Making Test Part B (limit time 300 s) [31] and Letter Fluency (Letters F, and L for 1 min each) [32, 33]; (3) Attention: Trail Making Test Part A (limit 150 s) [31] and Number Span (forward and backward) [34]; (4) Language: Multilingual Naming Test (total score) [35] and Category Fluency (Animals, Vegetables: 1 min each) [36] Visual-spatial: Benson Complex Figure (Immediate) [30], WAIS III Block Design [34]. MCI was defined according to Jak/Bondi criteria [37] using the following actuarial formula: (1) impaired scores, defined as > 1 SD below the age, gender, and education adjusted normative means, on both measures within at least one cognitive domain (i.e., memory, language, or speed/executive function); or (2) one impaired score in at least three of the five cognitive domains; or (3) functional dependence based on the Lawton Brody Instrumental Activities of Daily Living scale [38]. Otherwise, an individual was classified as being cognitively unimpaired. Secondary analyses considered impairment within each cognitive domain. Impairment within a domain was defined as present if the participant scored > 1 SD below the age, gender and education adjusted normative mean for one, or both tests within that domain. These diagnostic criteria, which require two impaired scores for MCI, were chosen given prior work demonstrating that these criteria improve specificity while maintaining high sensitivity [39].

Global cognition was assessed using the Montreal Cognitive Assessment (version 7.2; MoCA-30) [24]. The MoCA-30 measures aspects of memory, executive function, attention, concentration, language, abstract reasoning, and orientation, with a maximum score of 30, such that higher score reflects better cognitive performance.

Covariates

Covariates were included based on prior work showing that they may be potential confounders [14]. All study visits included standardized assessments of demographics, smoking, medical history, physical activity and anthropometrics [25]. Years of education and race/ethnicity were self-reported. History of myocardial infarction, stroke, hypertension, and diabetes were determined using self-report of ever having a physician diagnosis for the condition. Medications were brought to the study visit and reviewed by the trained interviewers. Drugs were coded using the World Health Organization Anatomical Therapeutic Chemical Classification system ().

Statistical Analysis

Sample characteristics were summarized and compared to the overall EAS cohort at baseline. Associations of HRV with cognition were examined using logistic (for binary outcomes of MCI and domain specific impairment) and linear (for the continuous outcome of MoCA) regression controlling for (1) age, sex, race/ethnicity and education (Model 1), (2) further controlling for history of diabetes, and history of hypertension (Model 2). Histories of myocardial infarction and stroke were not controlled due to their low prevalence in this sample. Only 2 individuals in the sample were current smokers, and thus smoking was not included as a covariate. All tests were two-sided, and statistical significance was set as p < 0.05. All analyses were conducted using R version 4.1.2 [R Core Team (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL ].

Results

Characteristics of the 84 individuals in the analysis sample are described in Table 1. The mean age was 78 (SD 5.2) years, and 82% were female. Mean MoCA score was 24 (SD 3.5), and a quarter (N = 21) met criteria for MCI. All individuals classified as MCI met criteria based on cognitive test performance, and none had functional impairment for instrumental activities of daily living. Impairment within cognitive domains was present in 25% (memory), 28.6% (executive function), 21.4% (attention), 17.9% (language), and 27.4% (visuospatial). Mean days of HRV data was 7.7 (SD 0.71).

Table 1 Descriptive characteristics of analysis sample.

	Analysis sample (N = 84) Mean (SD) or (N) %	EAS cohort at baseline (N = 446) Mean (SD) of (N) %
Age, years, mean (SD)	78 (5.2)	78.3 (5.3)
Female: (N) %	(69) 82.1%	(293) 66%
Race/Ethnicity (N) %: Non-Hispanic White	(33) 39.3%	(207) 46%
Non-Hispanic Black	(37) 44%	(167) 37%
Other	(14) 16.7%	(72) 16%
Years education, mean (SD)	14.7 (3.3)	15.1 (3.5)
History of stroke (N) %	(3) 3.6%	(40) 9.2%
History of myocardial infarction (N) %	(1) 1.2%	(33) 7.6%
History of heart failure (N) %	(2) 2.4%	(16) 3.7%
History of hypertension (N) %	(56) 66.7%	(309) 70%
History of diabetes (N) %	(15) 17.9%	(109) 25%
BMI kg/m2, mean (SD)a	28.4 (4.6)	29.1 (5.7)
Days ECG mean (SD)	7.7 (0.71) (Range 6–9)	---
Mean in HF-HRV, mean (SD)b	4.5 (0.8)	---
Mean HF-HRV msec.2 median (IQ Range)b	91 (57–146)	---
MoCA Score, mean (SD)	24 (3.5)	23.4 (3.7)
MCI: (N) %	(21) 25%	(152) 34%
Impairment memory (N) %	(21) 25%	(142) 32%
Impairment executive (N) %	(24) 28.6%	(169) 38%
Impairment attention (N) %	(18) 21.4%	(137) 31%
Impairment language (N) %	(15) 17.9%	(119) 27%
Impairment visuospatial (N) %	(23) 27.4%	(136) 30%
Memory domain
Free and cued selective reminding test- free recall, mean (SD)	33.6 (5.7)	32.2 (5.8)
Benson complex figure delay, mean (SD)	9.4 (2.8)	9.3 (3.2)
Executive function domain
Trail making test part B, time (seconds), mean (SD)	131.6 (74.1)	146.1 (80.2)
Letter fluency, mean (SD)	25.8 (9.0)	24.8 (8.7)
Attention domain
Trail making test part A, time (seconds), mean (SD)	43.5 (18.5)	49.5 (24.3)
Number span (forward and backward), mean (SD)	13.7 (3.8)	13.2 (3.8)
Language domain
Multilingual naming test, mean (SD)	27.5 (3.8)	26.8 (4.3)
Category fluency, mean (SD)	29.2 (8.1)	27.6 (7.7)
Visual-spatial domain
Benson complex figure immediate, mean (SD)	14.8 (1.4)	14.7 (1.5)
WAIS III block design, mean (SD)	24.4 (8.8)	24.3 (9.2)

Logistic regression models demonstrated that HF-HRV was inversely associated with prevalent MCI (OR per 1 SD increase in ln HF-HRV: 0.47, p = 0.02) (Table 2). Analyses of impairment within cognitive sub-domains showed that HF-HRV was associated with impairment in memory (OR per 1 SD increase in ln HF-HRV: 0.52, p = 0.03).

Table 2 Results of logistic regression (MCI and domain specific cognitive impairment) and linear regression (MoCA score) for cross-sectional associations with log-transformed HF-HRV, Einstein aging study, N = 84.

Outcome	Model 1	Model 2
Estimatea	95% CI	p-value	Estimatea	95% CI	p-value
MCI
HF-HRV	0.49	(0.27, 0.92)	0.03	0.47	(0.24, 0.90)	0.02
Impairment memory
HF-HRV	0.51	(0.28, 0.92)	0.03	0.52	(0.29, 0.94)	0.03
Impairment executive function
HF-HRV	0.85	(0.50, 1.44)	0.55	0.86	(0.51, 1.45)	0.56
Impairment attention
HF-HRV	0.65	(0.37, 1.14)	0.14	0.62	(0.33, 1.14)	0.12
Impairment language
HF-HRV	1.04	(0.56, 1.92)	0.90	1.05	(0.56, 1.99)	0.87
Impairment visuospatial
HF-HRV	0.80	(0.48, 1.34)	0.40	0.81	(0.48, 1.35)	0.41
MoCA score
HF-HRV	0.64	(0.01,1.27)	0.05	0.65	(0.01,1.28)	0.05

In linear regression models, HF-HRV was associated with global cognition. After adjustment for demographics, hypertension and diabetes, each 1 SD increase in ln HF-HRV was associated with a 0.65 point increase in MoCA score (p = 0.05), (Table 2). Sensitivity analyses excluding the three individuals using Beta-blockers did not change the results.

Further adjustment of the linear and logistic regression models for hours per week of moderate or heavy physical activity or for BMI did not impact the results of the more parsimonious model 2. As HRV may be impacted by sleep, we conducted sensitivity analyzes among the subset of participants who also had actigraphy data (N = 67). Results remained similar after adjustment for either wake after sleep onset, sleep duration or self-reported sleep apnea although statistical significance was not achieved due to the reduced sample size. We also tested for a potential nonlinear trend by adding a quadratic term of HF-HRV in the models which were found not significant for any of the outcomes studied, which suggested that the linear trend of HF-HRV used was acceptable.

Discussion

We applied Fourier-based spectral analysis to ambulatory ECG tracings obtained over an average of 7.7 days within the Einstein Aging Study cohort to examine associations of HRV with cognitive impairment and global cognitive performance. To examine the role of parasympathetic activity, we focused specifically on a frequency domain measure of HF-HRV.

In this cross-sectional analysis, lower average level of HF-HRV was associated with increased odds of prevalent MCI, worse global cognitive performance based on the MoCA and greater odds of memory impairment. These results add to prior work suggesting that the ANS, specifically parasympathetic function, is related to cognitive function and cognitive impairment in older adults [4].

Prior studies regarding the contributions of specific HRV indices to cognition have yielded mixed results, particularly regarding the relative contributions of parasympathetic versus sympathetic influences [40]. Others have reported that no HRV indices were associated with cognitive performance across domains of verbal memory, reasoning, vocabulary, or verbal fluency [20]. Some of the inconsistencies across studies may be related to variations in HRV measures used including reliance on global HRV indices that are not specific to vagal control. Further, some of the seemingly contradictory evidence may stem from the use of LF-HRV, which unlike HF-HRV, is influenced by both the parasympathetic and sympathetic nervous systems depending on measurement conditions. In the supine position, LF-HRV is primarily parasympathetic in origin with little sympathetic contributions. In the upright position, in contrast, both the SNS and PNS contribute to LF power [40]. Studies manipulating physical position by graded head-up tilt suggest an increasing SNS contribution with increasing tilt [41]. Studies that examine relationships between LF-HRV and cognitive indices must control for these important positional effects.

Our observed positive association between HF-HRV and global cognition is in contrast to findings from some prior studies. In the Irish Longitudinal Study, HF-HRV was not associated with MoCA score after adjustment for cardiovascular risk factors and medications [17]. Further, in the MESA cohort, RMSSD, a time domain measure equivalent to HF-HRV, was not related to global cognition [30]. Both studies derived HRV from short recordings over 5 min or less. In contrast, in the EAS, HRV was derived from 7-day monitoring, which should reduce measurement error and enhance the ability to detect associations. In addition, given that both cognitive performance and HRV are related to age, it is important to note that these differences may be related to the fact that the mean age of our cohort was older by at least a decade.

In our cohort, prevalent MCI also was significantly associated with lower levels of HF-HRV. This is consistent with a prior study of women over age of 65 years which demonstrated that lower HF-HRV and lower RMSSD were associated with cognitive impairment on the Mini Mental Status Exam [19]. It should be noted that this study defined cognitive impairment solely on the basis of the MMSE global cognition assessment while the present analysis applied standard criteria to a comprehensive neuropsychological battery to define MCI [37]. Similarly, Collins et al [42] observed that HF-HRV was lower among MCI cases compared with cognitively normal controls, and Xavier et al examined very brief, 10 s ECG recordings and noted lower RMSSD among older indiviurals with MCI comnpared to those with normal cognition [43]. In contrast, a small cross-sectional study based on 32 MCI cases and 36 cognitively normal controls did not identify group differences in either time or frequency domain measures of HF-HRV although the power to detect differences was limited by the small sample [44].

The literature regarding the relation of HRV to function within specific cognitive domains has been limited [3]. The present analyses revealed a significant association between lower HF-HRV with impairment in the memory domain. Previous studies regarding the role of HF-HRV in memory have yielded mixed results. Our findings contrast with findings from other cohorts where memory was not associated with either frequency or time domain measures of vagally mediated HRV [16–18]. Further work is needed to confirm the association of HF-HRV and memory and to elucidate the specific brain regions and neural networks that underlie associations of HRV with different domains of cognition.

Finally, we did not find evidence that HF-HRV was associated with impairment in processing speed. Previous work, using time domain measures of parasympathetic function has shown mixed results, with faster processing speed associated with higher RMSSD in the MESA study [18], but not related to RMSSD in the CARDIA cohort after adjusting for covariates [16].

One reason for the lack of clarity regarding the specific cognitive domains associated with related to HRV may be the variable cognitive batteries used across studies, with cognitive domains defined broadly and various tests used in different studies [2]. Future work, based on standard definitions of cognitive impairment and standardized test batteries, is required to evaluate the differing results across cohorts. Another factor may be the different ages at which these relationships have been studied. The relationship of HRV to cognitive impairment within specific domains may confounded by age. Age is inversely associated with HRV [45], and the temporal sequence of cognitive decline may vary by cognitive domain [46]. Further, the sensitivity of specific cognitive tests may vary by the age of the individuals studied.

Several mechanisms have been postulated to underlie the association of HRV to cognitive performance and impairment. As the autonomic nervous system regulates blood pressure, reduced HRV may be linked to cognition via mechanisms related to blood pressure dysregulation including cerebral hypoperfusion [47] and increased blood pressure variability [48] which has been associated with incident dementia risk and cognitive decline [49] and with cognitive structural brain changes including microvascular injury and presence white matter hyperintensitie [50, 51]. Further, reduced levels of HF-HRV have been associated with white matter hyperintensity burden in MCI patients [52]. In addition, low HRV has been linked to increased risk for cardiovascular events, hypertension, diabetes and increased inflammation which are each potent risk factors for cognitive impairment [2, 11, 18, 53]. Our cross-sectional analyses indicate that associations between HF-HRV and cognitive impairment and global cognitive performance were independent of diabetes and hypertension. In sensitivity analyses, additional adjustment of these models for sleep quality and quantity or sleep apnea, BMI or physical activity did not modify the results (data not shown). This is in agreement with several other studies which have reported that HRV is related to cognitive performance after adjusting for cardiovascular comorbidities and risk factors [17–19, 21]. A study by Grassler et al. [54] demonstrated that during performance of cognitive tasks, individuals with MCI had a greater decrease in HRV during performance of a cognitive task than did cognitively normal individuals. While this suggests that modulation of parasympathetic activity may be associated with cognitive impairment, conclusions from this study are limited as the within group changes in RMSSD among MCI and normal cognition groups were not significant, and the finding was not present for HF-HRV. Finally, we cannot rule out the possibility of reverse causality explaining our findings. Reduced HRV may be indicative of AD related neuropathological changes given evidence that the central autonomic network is impacted early in the course of AD [8, 42] and that autonomic function may be impacted before the onset of clinical cognitive symptoms [8]. Further work based on longitudinal data is required to determine the temporal sequence of HRV cognition associations and to establish whether modifiable cardiovascular risk factors meditate these associations.

Our study has some significant strengths. First, HRV was assessed using ambulatory assessments over multiple days in real-world settings in a race/ethnically diverse population-based cohort. Most prior studies have relied on measures of HRV from brief clinic-based ECG recordings which may not accurately reflect average HRV as individuals negotiate activities of daily life [23]. In addition, reliance on brief ECG intervals in studies of older adults runs the risk that ectopic beats or noisy signals will result in fewer valid epochs in the RR interval time series. The result of this would be a more unstable HF-HRV estimate, compared to estimates from the greater number of epochs available from longer ECG recordings.

Second, the choice of an HRV index should be consistent with the nature of the associations under study [14]. We focused specifically on HF-HRV to study the specific associations of parasympathetic function to cognition given that vagal control is related to many of the mechanisms by which HRV may be associated with cognition. Further, our definitions of mild cognitive impairment were based on widely accepted criteria based on a robust cognitive battery that included two tests within each of five cognitive domains [37]. Our measure of global cognition, the MoCA, is also a widely used test which has been applied in numerous aging cohorts [24]. Finally, our study expands prior work by examining the impact of HRV on cognition within an older, community-based cohort that is racially/ethnically diverse.

There are also several limitations to the present study. The analysis was cross-sectional and thus we are not able to determine the causal relation between HRV and cognitive impairment and thus cannot rule out the possibility of reverse causality or the effect of a third factor influencing both HRV and cognition. Future analyses will explore how these initial measures of HF-HRV predict change in cognitive performance. Another limitation is the small sample, which limits study power particularly for the analyses of domain specific impairment. Although the cohort is racially diverse, the small sample may limit the generalizability of results. However, to a significant degree, the small sample limitation is offset by the 7-day length of the ECG recordings which, relative to short-term assessments, provide stable estimates of an individual's average HRV in a real-world setting. Further, our analysis population had a low prevalence of prior MI, heart failure or stroke, likely attributable to the fact that the participants in the community-based cohort are relatively healthy and due to exclusions from the analysis of those using drugs that impact the autonomic nervous system. Thus, we were not able to explore fully whether adjustment for prevalent cardiovascular disease attenuated the observed associations. History of MI and stroke were based on self-report and we did not have information regarding peripheral arterial disease or chronic kidney disease. However, the cohort consists of relatively healthy, community residing adults who able to attend in person clinic visits. While we adjusted for level of physical activity, we did not have a measure of aerobic fitness to include as a covariate. Further, our analysis controlled for self-reported history of hypertension and diabetes. Supplemental analyses based on the subset of individuals with objectively defined hypertension and diabetes showed similar results. Finally, a number of transient factors have been noted to impact the assessment of HRV including coffee consumption, intense physical activity, alcohol use or poor sleep in proximity to HRV testing [14]. While we did not have daily information on these factors during the HRV protocol, it should be noted that since our HRV assessment is based on continuous 24 h ECG monitoring over multiple days the impact of any transient variables should be minimized relative to studies where HRV is examined over a single, brief interval.

In summary, the present study adds to the growing body of literature regarding the relationship of an index of cardiac vagal control to cognitive impairment. Analyses demonstrate that higher HF-HRV is associated with better global cognition and inversely associated with presence of mild cognitive impairment in an older community-based cohort. This study extends prior findings suggesting that reduced HRV may be an early indicator of risk for subsequent dementia onset [43]. Future work based on longitudinal data is required to confirm these findings.

Author Contributions

Carol A Derby: conceptualization, funding acquisition, writing—original draft, writing—review and editing, project administration, methodology, investigation. Jiyue Qin: formal analysis, writing—review and editing, methodology. Grace Liu: writing—review and editing, methodology, formal analysis, data curation, investigation. Cuiling Wang: visualization, methodology, formal analysis. Richard P Sloan: investigation, writing—review and editing, writing—original draft, methodology, formal analysis, data curation.

Acknowledgments

This article was prepared in accordance with STROBE guidelines. All authors have read and approved the final version of the manuscript. Dr. Derby had full access to all of the data in this study and takes complete responsibility for the integrity of the data and the accuracy of the data analysis. This study was funded by the National Institutes of Health NIA-AG003949. The funding source (National Institutes of Health NIA-AG003949) had no involvement in study design; collection, analysis, interpretation of data; writing of the report; or the decision to submit the report for publication.

Conflicts of Interest

The authors have nothing to disclose. All authors have read and approved the final version of the manuscript. Dr. Derby had full access to all of the data in this study and takes complete responsibility for the integrity of the data and the accuracy of the data analysis.

Data Availability Statement

Data used in this manuscript may be available by request to the author.

Transparency Statement

The lead author Carol A. Derby affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

References

J. F. Thayer, A. L. Hansen, E. Saus‐Rose, and B. H. Johnsen, “Heart Rate Variability, Prefrontal Neural Function, and Cognitive Performance: The Neurovisceral Integration Perspective on Self‐Regulation, Adaptation, and Health,” Annals of Behavioral Medicine 37, no. 2 (2009): 141–153, https://doi.org/10.1007/s12160-009-9101-z.

M. F. Elias and R. V. Torres, “The Renaissance of Heart Rate Variability as a Predictor of Cognitive Functioning,” American Journal of Hypertension 31, no. 1 (2017): 21–23, https://doi.org/10.1093/ajh/hpx150.

G. Forte, F. Favieri, and M. Casagrande, “Heart Rate Variability and Cognitive Function: A Systematic Review,” Frontiers in Neuroscience 13 (2019): 710, https://doi.org/10.3389/fnins.2019.00710.

P. Nicolini, M. M. Ciulla, C. D. Asmundis, F. Magrini, and P. Brugada, “The Prognostic Value of Heart Rate Variability in the Elderly, Changing the Perspective: From Sympathovagal Balance to Chaos Theory,” Pacing and Clinical Electrophysiology 35, no. 5 (2012): 621–637, https://doi.org/10.1111/j.1540-8159.2012.03335.x.

D. Kimhy, O. V. Crowley, P. S. McKinley, et al., “The Association of Cardiac Vagal Control and Executive Functioning–Findings From the Midus Study,” Journal of Psychiatric Research 47, no. 5 (2013): 628–635, https://doi.org/10.1016/j.jpsychires.2013.01.018.

A. Zeki Al Hazzouri, M. N. Haan, Y. Deng, J. Neuhaus, and K. Yaffe, “Reduced Heart Rate Variability is Associated With Worse Cognitive Performance in Elderly Mexican Americans,” Hypertension 63, no. 1 (2014): 181–187, https://doi.org/10.1161/HYPERTENSIONAHA.113.01888.

P. Nicolini, M. M. Ciulla, G. Malfatto, et al., “Autonomic Dysfunction in Mild Cognitive Impairment: Evidence From Power Spectral Analysis of Heart Rate Variability in a Cross‐Sectional Case‐Control Study,” PLoS One 9, no. 5 (2014): e96656, https://doi.org/10.1371/journal.pone.0096656.

G. D. Femminella, G. Rengo, K. Komici, et al., “Autonomic Dysfunction in Alzheimer's Disease: Tools for Assessment and Review of the Literature,” Journal of Alzheimer's Disease 42, no. 2 (2014): 369–377, https://doi.org/10.3233/JAD-140513.

D. Liao, J. Cai, W. D. Rosamond, et al., “Cardiac Autonomic Function and Incident Coronary Heart Disease: A Population‐Based Case‐Cohort Study. the ARIC Study,” American Journal of Epidemiology 145, no. 8 (1997): 696–706, https://doi.org/10.1093/aje/145.8.696.

J. P. Singh, M. G. Larson, H. Tsuji, J. C. Evans, C. J. O'Donnell, and D. Levy, “Reduced Heart Rate Variability and New‐Onset Hypertension: Insights Into Pathogenesis of Hypertension: The Framingham Heart Study,” Hypertension 32, no. 2 (1998): 293–297, https://doi.org/10.1161/01.hyp.32.2.293.

M. R. Carnethon, S. H. Golden, A. R. Folsom, W. Haskell, and D. Liao, “Prospective Investigation of Autonomic Nervous System Function and the Development of Type 2 Diabetes: The Atherosclerosis Risk in Communities Study, 1987–1998,” Circulation 107, no. 17 (2003): 2190–2195, https://doi.org/10.1161/01.CIR.0000066324.74807.95.

M. M. B. Breteler, “Vascular Risk Factors for Alzheimer's Disease: An Epidemiologic Perspective,” Neurobiology of Aging 21, no. 2 (2000): 153–160. doi: S0197‐4580(99)00110‐4 [pii].

K. J. Tracey, “Physiology and Immunology of the Cholinergic Antiinflammatory Pathway,” Journal of Clinical Investigation 117, no. 2 (2007): 289–296, https://doi.org/10.1172/JCI30555.

S. Laborde, E. Mosley, and J. F. Thayer, “Heart Rate Variability and Cardiac Vagal Tone in Psychophysiological Research ‐ Recommendations for Experiment Planning, Data Analysis, and Data Reporting,” Frontiers in Psychology 8 (2017): 213, https://doi.org/10.3389/fpsyg.2017.00213.

L. J. Capuana, J. Dywan, W. J. Tays, J. L. Elmers, R. Witherspoon, and S. J. Segalowitz, “Factors Influencing the Role of Cardiac Autonomic Regulation in the Service of Cognitive Control,” Biological Psychology 102 (2014): 88–97, https://doi.org/10.1016/j.biopsycho.2014.07.015.

A. Zeki Al Hazzouri, T. Elfassy, M. R. Carnethon, D. M. Lloyd‐Jones, and K. Yaffe, “Heart Rate Variability and Cognitive Function in Middle‐Age Adults: The Coronary Artery Risk Development in Young Adults,” American Journal of Hypertension 31, no. 1 (2017): 27–34, https://doi.org/10.1093/ajh/hpx125.

J. Frewen, C. Finucane, G. M. Savva, G. Boyle, R. F. Coen, and R. A. Kenny, “Cognitive Function is Associated With Impaired Heart Rate Variability in Ageing Adults: The Irish Longitudinal Study on Ageing Wave One Results,” Clinical Autonomic Research 23, no. 6 (2013): 313–323, https://doi.org/10.1007/s10286-013-0214-x.

C. L. Schaich, D. Malaver, H. Chen, et al., “Association of Heart Rate Variability With Cognitive Performance: The Multi‐Ethnic Study of Atherosclerosis,” Journal of the American Heart Association 9, no. 7 (2020): e013827, https://doi.org/10.1161/JAHA.119.013827.

D. H. Kim, L. A. Lipsitz, L. Ferrucci, et al., “Association Between Reduced Heart Rate Variability and Cognitive Impairment in Older Disabled Women in the Community: Women's Health and Aging Study I,” Journal of the American Geriatrics Society 54, no. 11 (2006): 1751–1757.

A. Britton, A. Singh‐Manoux, K. Hnatkova, M. Malik, M. G. Marmot, and M. Shipley, “The Association Between Heart Rate Variability and Cognitive Impairment in Middle‐Aged Men and Women,” Neuroepidemiology 31, no. 2 (2008): 115–121, https://doi.org/10.1159/000148257.

S. Mahinrad, J. W. Jukema, D. van Heemst, et al., “10‐second Heart Rate Variability and Cognitive Function in Old Age,” Neurology 86, no. 12 (2016): 1120–1127, https://doi.org/10.1212/WNL.0000000000002499.

R. Zulli, F. Nicosia, B. Borroni, et al., “QT Dispersion and Heart Rate Variability Abnormalities in Alzheimer's Disease and in Mild Cognitive Impairment,” Journal of the American Geriatrics Society 53, no. 12 (2005): 2135–2139.

A. M. Dalise, R. Prestano, R. Fasano, A. Gambardella, M. Barbieri, and M. R. Rizzo, “Autonomic Nervous System and Cognitive Impairment in Older Patients: Evidence From Long‐Term Heart Rate Variability in Real‐Life Setting,” Frontiers in Aging Neuroscience 12 (2020): 40, https://doi.org/10.3389/fnagi.2020.00040.

Z. S. Nasreddine, N. A. Phillips, V. Bédirian, et al., “The Montreal Cognitive Assessment, MoCA: A Brief Screening Tool for Mild Cognitive Impairment,” Journal of the American Geriatrics Society 53, no. 4 (2005): 695–699.

M. J. Katz, R. B. Lipton, C. B. Hall, et al., “Age‐Specific and Sex‐Specific Prevalence and Incidence of Mild Cognitive Impairment, Dementia, and Alzheimer Dementia in Blacks and Whites: A Report From the Einstein Aging Study,” Alzheimer Disease and Associated Disorders 26, no. 4 (2012): 335–343, https://doi.org/10.1097/WAD.0b013e31823dbcfc.

R. W. DeBoer, J. M. Karemaker, and J. Strackee, “Comparing Spectra of a Series of Point Events Particularly for Heart Rate Variability Data,” IEEE Transactions on Biomedical Engineering 31, no. 4 (1984): 384–387, https://doi.org/10.1109/TBME.1984.325351.

F. J. Harris, “On the Use of Windows for Harmonic Analysis With the Discrete Fourier Transform,” Proceedings of the IEEE 66 (1978): 51–83.

Diagnostic and Statistical Manual of Mental Disorders, DSM‐IV. American Psychiatric Association; 1994.

H. Buschke, “Cued Recall in Amnesia,” Journal of Clinical Neuropsychology 6, no. 4 (1984): 433–440, https://doi.org/10.1080/01688638408401233.

K. L. Possin, V. R. Laluz, O. Z. Alcantar, B. L. Miller, and J. H. Kramer, “Distinct Neuroanatomical Substrates and Cognitive Mechanisms of Figure Copy Performance in Alzheimer's Disease and Behavioral Variant Frontotemporal Dementia,” Neuropsychologia 49, no. 1 (2011): 43–48, https://doi.org/10.1016/j.neuropsychologia.2010.10.026.

R. Reitan, “Validity of the Trail Making Test as an Indicator of Organic Brain Damage,” Journal of Comparative and Physiological Psychology 48 (1958): 474–477.

A. L. Benton and K. Hamsher Multilingual Aphasia Examination. 1989.

A. Rofes, M. Beran, R. Jonkers, M. I. Geerlings, and J. M. J. Vonk, “What Drives Task Performance in Animal Fluency in Individuals Without Dementia? The SMART‐MR Study,” Journal of Speech, Language, and Hearing Research 66 (2023): 3473–3485.

D. Wechsler, Wechsler Adult Intelligence Scale–Third Edition (The Psychological Association, 1997).

A. Stasenko, D. M. Jacobs, D. P. Salmon, and T. H. Gollan, “The Multilingual Naming Test (MINT) as a Measure of Picture Naming Ability in Alzheimer's Disease,” Journal of the International Neuropsychological Society 25, no. 8 (2019): 821–833, https://doi.org/10.1017/S1355617719000560.

A. U. Monsch, M. W. Bondi, N. Butters, D. P. Salmon, R. Katzman, and L. J. Thal, “Comparisons of Verbal Fluency Tasks in the Detection of Dementia of the Alzheimer Type,” Archives of Neurology 49, no. 12 (1992): 1253–1258, https://doi.org/10.1001/archneur.1992.00530360051017.

A. J. Jak, M. W. Bondi, L. Delano‐Wood, et al., “Quantification of Five Neuropsychological Approaches to Defining Mild Cognitive Impairment,” American Journal of Geriatric Psychiatry 17, no. 5 (2009): 368–375, https://doi.org/10.1097/JGP.0b013e31819431d5.

M. P. Lawton and E. M. Brody, “Assessment of Older People: Self‐Maintaining and Instrumental Activities of Daily Living,” Gerontologist 9, no. 3 (1969): 179–186.

K. H. Chang, C. Wang, J. Qin, et al., “Comparing the Predictive Validity of Four MCI Definitions for Incident Dementia in Demographically Diverse Community‐Dwelling Individuals: Results From the Einstein Aging Study (EAS),” Journal of the International Neuropsychological Society 31, no. 1 (2025): 32–41.

B. Pomeranz, R. J. Macaulay, M. A. Caudill, et al., “Assessment of Autonomic Function in Humans by Heart Rate Spectral Analysis,” American Journal of Physiology 248, no. 1 Pt 2 (1985): 151–153, https://doi.org/10.1152/ajpheart.1985.248.1.H151.

S. Mukai and J. Hayano, “Heart Rate and Blood Pressure Variabilities During Graded Head‐Up Tilt,” Journal of Applied Physiology 78, no. 1 (1995): 212–216, https://doi.org/10.1152/jappl.1995.78.1.212.

O. Collins, S. Dillon, C. Finucane, B. Lawlor, and R. A. Kenny, “Parasympathetic Autonomic Dysfunction Is Common in Mild Cognitive Impairment,” Neurobiology of Aging 33, no. 10 (2012): 2324–2333, https://doi.org/10.1016/j.neurobiolaging.2011.11.017.

A. Xavier, S. Noble, J. Joseph, A. Ghosh, and T. G. Issac, “Heart Rate and its Variability From Short‐Term ECG Recordings as Potential Biomarkers for Detecting Mild Cognitive Impairment,” American Journal of Alzheimer's Disease and Other Dementias 39 (2024): 1–8.

M. S. Kim, J. H. Yoon, and J. M. Hong, “Early Differentiation of Dementia With Lewy Bodies and Alzheimer's Disease: Heart Rate Variability at Mild Cognitive Impairment Stage,” Clinical Neurophysiology 129, no. 8 (2018): 1570–1578.

R. E. De Meersman and P. K. Stein, “Vagal Modulation and Aging,” Biological Psychology 74, no. 2 (2007): 165–173.

P. Mistridis, S. Krumm, A. U. Monsch, M. Berres, and K. I. Taylor, “The 12 Years Preceding Mild Cognitive Impairment Due to Alzheimer's Disease: The Temporal Emergence of Cognitive Decline,” Journal of Alzheimer's Disease 48, no. 4 (2015): 1095–1107, https://doi.org/10.3233/JAD-150137.

C. Qiu, B. Winblad, M. Viitanen, and L. Fratiglioni, “Pulse Pressure and Risk of Alzheimer Disease in Persons Aged 75 Years and Older: A Community‐Based, Longitudinal Study,” Stroke 34, no. 3 (2003): 594–599.

R. P. Sloan, R. E. Demeersman, P. A. Shapiro, et al., “Cardiac Autonomic Control is Inversely Related to Blood Pressure Variability Responses to Psychological Challenge,” American Journal of Physiology 272, no. 5 Pt 2 (1997): 2227–2232, https://doi.org/10.1152/ajpheart.1997.272.5.H2227.

M. E. Ernst, J. Ryan, E. K. Chowdhury, et al., “Long‐Term Blood Pressure Variability and Risk of Cognitive Decline and Dementia Among Older Adults,” Journal of the American Heart Association 10, no. 13 (2021): e019613, https://doi.org/10.1161/JAHA.120.019613.

P. J. Tully, Y. Yano, L. J. Launer, et al., “Association Between Blood Pressure Variability and Cerebral Small‐Vessel Disease: A Systematic Review and Meta‐Analysis,” Journal of the American Heart Association 9, no. 1 (2020): e013841, https://doi.org/10.1161/JAHA.119.013841.

Y. Ma, P. Yilmaz, D. Bos, et al., “Blood Pressure Variation and Subclinical Brain Disease,” Journal of the American College of Cardiology 75, no. 19 (2020): 2387–2399.

S. Galluzzi, F. Nicosia, C. Geroldi, et al., “Cardiac Autonomic Dysfunction Is Associated With White Matter Lesions in Patients With Mild Cognitive Impairment,” Journals of Gerontology Series A: Biological Sciences and Medical Sciences 64, no. 12 (2009): 1312–1315, https://doi.org/10.1093/gerona/glp105.

A. Sajadieh, “Increased Heart Rate and Reduced Heart‐Rate Variability are Associated With Subclinical Inflammation in Middle‐Aged and Elderly Subjects With no Apparent Heart Disease,” European Heart Journal 25, no. 5 (2004): 363–370, https://doi.org/10.1016/j.ehj.2003.12.003.

B. Grässler, M. Dordevic, S. Darius, et al., “Is There a Link Between Heart Rate Variability and Cognitive Decline? A Cross‐Sectional Study on Patients With Mild Cognitive Impairment and Cognitively Healthy Controls,” Arquivos de Neuro‐psiquiatria 81, no. 1 (2023): 009–018, https://doi.org/10.1055/s-0042-1758862.