Exploring the Interplay Between Mechanisms of Neuroplasticity and Cardiovascular Health in Aging Adults: A Multiple Linear Regression Analysis Study

Abstract

Background:

Neuroplasticity and cardiovascular health behavior are critically important factors for optimal brain health.

Objective:

To assess the association between the efficacy of the mechanisms of neuroplasticity and metrics of cardiovascular heath in sedentary aging adults.

Methods:

We included thirty sedentary individuals (age= 60.6±3.8y; 63% female). All underwent assessments of neuroplasticity, measured by the change in amplitude of motor evoked potentials elicited by single-pulse Transcranial Magnetic Stimulation (TMS) at baseline and following intermittent Theta-Burst (iTBS) at regular intervals. Cardiovascular health measures were derived from the Incremental Shuttle Walking Test and included Heart Rate Recovery (HRR) at 1-min/2-min after test cessation. We also collected plasma levels of brain-derived neurotrophic factor (BDNF), vascular endothelial growth factor (VEGF), and c-reactive protein.

Results:

We revealed moderate but significant relationships between TMS-iTBS neuroplasticity, and the predictors of cardiovascular health (∣r∣= .38 to .53, p< .05). HRR1 was the best predictor of neuroplasticity (β= .019, p= .002). The best fit model (Likelihood ratio= 5.83, p= .016) of the association between neuroplasticity and HRR1 (β= .043, p= .002) was selected when controlling for demographics and health status. VEGF and BDNF plasma levels augmented the association between neuroplasticity and HRR1.

Conclusions:

Our findings build on existing data demonstrating that TMS may provide insight into neuroplasticity and the role cardiovascular health have on its mechanisms. These implications serve as theoretical framework for future longitudinal and interventional studies aiming to improve cardiovascular and brain health. HRR1 is a potential prognostic measure of cardiovascular health and a surrogate marker of brain health in aging adults.

1. Introduction

Neuroplasticity can be broadly defined as a change in neural structure and function in response to experience or environmental stimuli (Pascual-Leone et al., 2005). Neuroplasticity is a critically important mechanism for brain health, as it underlies the capacity of the nervous system to change and adapt to the ever-changing demands of the human experience (Mora et al., 2007; Phillips, 2017). Neuroplasticity is the underlying mechanism for cognitive and motor learning and is critically involved in the ability to maintain cognitive health and independence during aging. For example, a reduction in the efficacy of the mechanisms of neuroplasticity is tied to a higher risk of developing age-related cognitive decline (Freitas et al., 2013; Hullinger & Puglielli, 2017; Pascual-Leone et al., 2011b), and to poorer recovery from other insults to the nervous system, such as TBI (Tomaszczyk et al., 2014), stroke (Mang et al., 2013), and dementia (Kumar et al., 2017; Shigihara et al., 2020).

Much of what is known about neuroplasticity comes from animal models utilizing invasive methods (Barnes, 2003; Farmer et al., 2004). However, advances in neurophysiological assessments using transcranial magnetic stimulation (TMS) have enabled the non-invasive assessment of the mechanisms of neuroplasticity in humans (Huang et al., 2005). Corticospinal excitability can be indexed by motor evoked potentials (MEPs) from stimulation of the motor cortex. Then, a train of repetitive TMS in the form of intermittent theta-burst stimulation (iTBS) can be delivered to the motor cortex, and the effects assessed again recording MEPs. The iTBS-induced modulation of MEPs shares important similarities with mechanisms of synaptic plasticity (Long Term Potentiation, LTP-like mechanism), including a reciprocal time course of modulation after stimulation, and the reliance on both GABA (Stagg et al., 2009) and NMDA receptor activity (Huang et al., 2007). This TMS-iTBS paradigm for assessment of the mechanisms of neuroplasticity (Rossi et al., 2009; Rossini et al., 2015) has been used to evaluate age-related changes in healthy individuals (Freitas et al., 2011) and modifications in various diseases, including traumatic brain injury (Tremblay et al., 2015) and type-2 diabetes mellitus (Fried et al., 2016a).

Optimized mechanisms of neuroplasticity and cardiovascular health behaviors (e.g. nonsmoking status, physical activity at goal levels, body mass index <25 kg/m², untreated blood pressure <120/<80 mmHg, and physical activity and healthy diet) are critically important for brain health (Gorelick et al., 2017). Brain health is defined by the American Heart Association as “an optimal capacity to function adaptively in the environment”. Accordingly, investigating the behavioral relevance of the TMS-iTBS neuroplasticity response in relation to cardiovascular health is a worthwhile scientific endeavor.

Another important aspect of cardiovascular health is cardiorespiratory fitness (CRF), broadly defined as the ability of the body to transport and consume oxygen during sustained exercise (Farnsworth & Cannon, 2008). High CRF is a strong protecting factor to deter the risk of cognitive decline and neurodegenerative disorders (Dougherty et al., 2021). For instance, heart rate recovery (HRR) is a well-established and robust measure of CRF in aging adults. HRR is a valid and reliable index (Fecchio et al., 2019), and attenuated HRR is an independent risk factor for cardiovascular events and all-cause mortality (Cole et al., 1999; Qiu et al., 2017a), and diabetes (Qiu et al., 2017b). HRR is also strongly associated with cognitive impairment in aging (Intzandt et al., 2020; Mehta, 2012). Defined as the change in the heart rate from the peak of exercise to the heart rate after 1-min and 2-min cessation (Lamberts et al., 2004), HRR reflects the dynamic mechanisms of autonomic heart rate regulation and can be easily measured through a submaximal exercise fitness assessment (Oppewal et al., 2014), which is generally recommended due to challenges with the gold standard CRF assessment – the maximal oxygen uptake (VO_2max) in the aging population.

Autonomic heart rate regulation is dependent of the homeostasis among the body systems along the brain-heart pathway. A balance of endothelial function, growth factor induction and signaling cascade, are – in combination – critical mediators of brain health and function. Circulating markers of endothelium function, herein investigated by the growth factors (aka angiogenic factors) brain-derived neurotrophic factor (BDNF) and vascular endothelial growth factor (VEGF), and C-reactive protein (CRP), a sensitive marker of inflammation, may modify the brain-heart pathway homeostasis. BDNF and VEGF are the two growth factors that are most consistently associated with angiogenesis and synaptic neuroplasticity (Ding et al., 2006; Zhang et al., 2015a). Despite the inconsistent findings, upregulation of VEGF that is not related to angiogenesis or neuroprotective effects, and greater levels of CRP may be a red flag for the endothelium dysfunction that can precede a disruption of the blood-brain barrier and are associated with stroke and neurodegenerative diseases (Janelidze et al., 2017; Poggesi et al., 2016; Zhang et al., 2015b). Thus, these studies provide sufficient data linking the effect of BDNF, VEGF, and CRP levels in the brain-heart pathway. However, it is not known whether baseline circulating markers of endothelial function can modify the association between neuroplasticity and cardiovascular health in aging adults.

Therefore, we aimed to unravel the mechanisms of neuroplasticity through the TMS/iTBS assessment, cardiovascular health assessment, and circulating markers of endothelium function (BDNF, VEGF, and CRP) in sedentary aging adults. First, we assessed the association between TMS/iTBS neuroplasticity and measures of cardiovascular health (primarily assessed by HRR) in sedentary aging adults. We hypothesized that greater efficacy of the mechanisms of neuroplasticity would be associated with better cardiovascular health. Second, we explored if the association between TMS/iTBS neuroplasticity and measures of cardiovascular health would be modified by BDNF, VEGF, and CRP in sedentary aging adults. We hypothesized that including circulating markers of endothelium function predictors would lead to a better fit model and modify the association between neuroplasticity and cardiovascular health response in aging adults.

2. Methods

2.1. Study Design and participants

We carried out a cross-sectional analysis of baseline data of a clinical study conducted at the University of Miami registered on ClinicalTrials.gov with the identifier NCT03804528. The elaboration and reporting of this original article is following the STROBE (The Strengthening the Reporting of Observational Studies in Epidemiology) checklist of items that should be included in cross-sectional study reports (von Elm et al., 2007).

We used a multifaceted recruitment strategy by advertising recruitment flyers in the local community, and potential participants were contacted from the University Research Informatics Data Environment via the Consent to Contact Initiative from February 2019 to February 2020. We included thirty (n = 30) individuals aged 55 and above, with sedentary status (low physical activity level as per the International Physical Activity Questionnaire) (Craig et al., 2003), and no clinically detectable cognitive impairment (Montreal Cognitive Assessment score ≥ 24) (Trzepacz et al., 2015). Major exclusions were any unstable medical condition, including contraindication to physical exercise, and contraindications to undergo TMS (Rossi et al., 2021; Rossini et al., 2015). All participants provided written informed consent, and all procedures were approved by the University of Miami institutional review board. Refer to Figure 1 for study design and conceptual framework.

Figure 1.

Study Design and Conceptual Framework.

2.2. Data collection

Demographics, physical measures (e.g., vital signs), medical history and etiology, and clinical data were obtained at baseline. All participants underwent a TMS/iTBS neuroplasticity assessment, a CRF assessment, and a blood draw collection. All assessments were conducted by a study member who received specific training for each test.

2.3. Endpoints

2.3.1. TMS measures of neuroplasticity

Neuroplasticity assessment using TMS was the primary endpoint. All study parameters followed the current guidelines for the safe application of TMS recommended by the International Federation of Clinical Neurophysiology (Rossi et al., 2021; Rossini et al., 2015) and all involved technicians and scientists met the recommended training criteria (Fried et al., 2021). Safety recommendations and TMS precautions involved but were not limited to the screening of central nervous system active medications, history of neurological/psychiatric disorders, implanted medical devices, hospitalization and surgery history, seizure and epilepsy history, and smoking and other recreational drugs, sleep pattern, caffeine and alcohol consumption. TMS adverse events were also assessed prior to and after the TMS procedures.

TMS was delivered using a static-cooled handheld MagPro MCF-B65 figure-of-eight coil (outer diameter: 75 mm/3.0 in) connected to a MagPro X100 stimulator (MagVenture A/S, Farum, Denmark). The TMS coil was held tangentially to the participant’s head, with the handle oriented 45° relative to their mid-sagittal axis to induce a biphasic (anterior-posterior—posterior-anterior) intracranial current. TMS/iTBS neuroplasticity targeted the hand representation of the primary motor cortex in the dominant hemisphere and peak-to-peak MEPs were recorded using surface electrodes applied to the first dorsal interosseous muscle. TMS pulses were applied to identify the motor ‘hotspot’—the coil position that consistently elicited the largest MEPs with a visible contraction of the FDI—and determines motor thresholds. Following, resting motor threshold (RMT) and active motor threshold (AMT) were collected and used to set the intensity of subsequent stimulation. Further details of TMS parameters of baseline corticomotor reactivity (single-pulse-TMS), iTBS, and variance-normalization procedures can be found elsewhere (Fried et al., 2016b; Gomes-Osman et al., 2017). Briefly, the neuroplasticity assessment was measured by the change in amplitude of MEPs elicited by sets of single-pulse TMS delivered at 120% of resting motor threshold at baseline and following TMS-iTBS at regular intervals (5, 10, 20, and 30 minutes) and expressed in percent change (%Δ). To reduce the influence of extremadee values, individual MEPs were log10-transformed, averaged across each timepoint (Baseline, Post5, Post10, Post20, Post30), and back-transformed into geometric means. All subsequent analyses were performed using these values.

2.3.2. Cardiorespiratory fitness

The Incremental Shuttle Walking Test (ISWT) was performed to determine aerobic capacity by evaluating maximal walking distance during the test (Singh et al., 1992). Cardiovascular health measures were derived from the ISWT and included: resting heart rate, heart rate recovery at 1-min (HRR1) and 2-min (HRR2) after test cessation with the participant still in the standing position, rate-pressure product (RPP), and estimated VO₂ peak. RPP, the product of heart rate and systolic blood pressure, was measured at rest, and immediately after exercise completion, and after 5 minutes of recovery. HRR1 was selected as the primary measure of cardiovascular health. We also calculated the predicted ISWT distance for each individual based on a previous publication (Dourado et al., 2011).

The ISWT is a simple, valid, and safe measure of cardiorespiratory fitness in healthy aging adults that strongly relates to gold-standard measures including the VO₂max during cardio-pulmonary exercise testing on a treadmill and the 6-minute walking test (Dourado et al., 2010; Dourado et al., 2013; Harrison et al., 2013; Neves et al., 2015). Maximum walking distance was used to estimate the VO_2peak as a measure of aerobic capacity (Dourado et al., 2013) Participants were instructed to walk a distance of 10 meters around a marking between two cones, keeping to the speed indicated by the bleeps on the audio recording up to the point of exhaustion. The walking speed increased by 0.17 m/s, with an initial speed of 0.5 m/s. The test was finished when: (1) the participant was not able to maintain the required speed (more than 0.5 m from the cone when the bleep sounds on a second successive 10 length), (2) at the request of the participant, or for some other reported symptom (e.g., extreme discomfort, pain, dyspnea, dizziness, vertigo, and angina), and (3) Participant reaches the age-predicted maximum heart rate (220 - age) or operator determined that the participant was not fit to continue. During the test, the laps were recorded to calculate the distance and gait speed reached at the last full level. Additional measures were conducted as cardiovascular outcomes and to assure the safety of participants. Participants were fitted with a heart rate monitor (Polar H10, Polar Electro Inc., Lake Success, NY, USA) and were closely monitored before, during, and after the physical test. Blood pressure was measured by a mercury sphygmomanometer cuff and a stethoscope at rest (5 minutes before the test), immediately after the test, and 5-minute upon completion. Additionally, we use the Borg scale assess perceived exertion (RPE, Borg scale, range 6–20, where 6 represents rest (no effort), and 20 represents maximal effort) (Borg, 1982). The test and all measures were conducted by a trained physical therapist.

2.3.4. Blood draw collection

Blood samples were collected from all patients at baseline on the same day and right before the neuroplasticity assessment. We collected peripheral circulating markers of endothelium function: BDNF levels (pg/mL), VEGF (pg/mL), and high-sensitivity CRP-HS (mg/L) concentrations.

BDNF and VEGF were measured in platelet-poor plasma, since activated platelets (e.g., during clot formation or as a result of sample freeze/thaw) releases BDNF and VEGF. Blood was collected in sodium citrate tubes and plasma was collected by centrifugation at 1000g at 4°C for 15 minutes. The plasma was then centrifuged for an additional 15 minutes at 4000g and 4°C and the supernatant containing the platelet-poor plasma was stored frozen at −80°C until assay.

Hr-C-reactive protein (CRP) was measured by latex-enhanced immunoturbidimetry (Roche Diagnostics, Indianapolis, IN; cat # 04628918 190) on a Roche Cobas 6000 analyzer following all manufacturer’s instructions. For CRP intra- and inter-assay coefficients of variation (CV) were 2.1% and 3.7%, respectively. BDNF and VEGF were measured by ELISA (R&D Systems, Minneapolis, MN, cat# DBNT00, and DVE00, respectively) following all manufacture instructions. For BDNF, the lower limit of detection was 20 pg/mL, and the inter- and intra-assay CVs were 5.0% and 9.3% respectively. For VEGF, the lower limit of detection was 30 pg/ml and the intra-assay CVs were 4.4% and 8.1% respectively.

2.4. Statistical analyses

All statistical analyses were performed using JMP Pro (v.15.0, SAS Institute Inc, USA) and StataCorp 2021 (v. 17, StataCorp LLC, USA) using a two-tailed 95% confidence interval (a=0.05). Data were presented as means ± SD and percentage (%) of the total. To test the homoscedasticity assumption, we use the default test in the Stata Breusch-Pagan/Cook-Weisberg test. To test the normality of the studentized residuals, we visually inspected using the Q-Q plots and tested using the Shapiro-Wilks test of normality. The models were corrected by bootstrap standard errors to handle departure from normality, and robust standard errors to account for heteroscedasticity. Regression models were inspected to avoid departure from collinearity using the variance inflation factor assuming mean VIF < 10. Multiple comparisons were corrected for using Benjamini and Hochberg’s false discovery rate, at a q value of 0.05, after pooling the P values from the regression analyses for each predictor model. Our sample of 30 participants provided a two-tailed 80% power to detect at least a moderate correlation (∣r∣≥.46), and to detect an adjusted r-squared of 0.29.

To quantify the mechanisms of neuroplasticity, within-group differences were assessed using average MEP amplitudes (mV) for each timepoint into random-effects linear models with the within-subject factor Time (Post0, Post5, Post10, Post20, Post30). Planned t-tests computed pairwise comparisons of baseline (Post0) and each time point (Post5, Post10, Post20, Post30) and intervals (Post 0-10, Post 10-20, Post 20-30, and Post0-30). We also assessed within-subject factor Time using post-iTBS timepoints expressed as the percent change (%Δ) from baseline (Post5, Post10, Post20, Post30). A subgroup exploratory random-effects analysis evaluated MEPs modulation within ‘Excitatory Responders’ (positive MEP%Δ) and ‘Inhibitory Responders’ (negative or zero MEP%Δ), and mixed-effects analysis compared those who exhibited an excitatory or inhibitory response.

To test our primary hypothesis, we defined, a priori, the period 10-20 minutes post-iTBS, the peak effect of iTBS in healthy adults (Wischnewski & Schutter, 2015), as our neuroplasticity measure (Post10-20_%Δ). Concomitant, in this present study, the period of 10-20 minutes post-iTBS was a period with larger effect (η²= .19, p= .009) and significant change in neuroplasticity response (24.3% ± 53.8). First, we calculated Pearson product-moment correlation coefficient (r) (or Spearman correlation when normality has not assumed [r_s]) to estimate the correlation of two continuous variables between the interval Post10-20_%Δ and: (1) cardiovascular health measures (HRR1, HRR2, resting HR, resting RPP, recovery RPP, estimated VO2peak), (2) and circulating markers of endothelium function (BDNF, VEGF, and Hs-CRP). Second, simple linear regression models were fitted to select the best cardiovascular health predictor of neuroplasticity. Then, four adjusted multiple linear regression models were fitted between the outcome neuroplasticity and the cardiovascular health predictor HRR1 to control for baseline data of age, gender, race/ethnicity, body mass index, education level, physical activity level, number of comorbidities, and smoking history. The likelihood ratio (LR) tests were used to evaluate the difference between nested models and select the best fit model to predict neuroplasticity. The LR test compared regression models that included one model with a set of parameters (variables), and a second model with all the parameters from the first, plus one or more other variables. Third, to test our secondary hypothesis, adjusted multiple linear regression models were fitted to include the modifiers variables of circulating markers of endothelium function (BDNF, VEGF, Hs-CRP) into the model that accounted for the neuroplasticity outcome variable “Post10-20_%Δ” and the predictor variable of cardiovascular health, HRR1. The LR test compared regression models, and the best-adjusted fit model including modifiers was reported.

3. Results

3.1. Participants

A total of 125 individuals were screened for eligibility. Of these, 41 were eligible and enrolled in the study. Figure 2 shows the study participants’ flowchart. Due to withdrawal and loss to follow-up (n= 11), the final data available for analysis included 30 participants (age = 60.3 ± 3.8; 63.3% females). Participants’ demographics and global health status are presented in Table 1. It is noteworthy that 56.7% were obese and 62% reported being diagnosed with hypertension.

Figure 2.

Study Flowchart.

Table 1.

Demographics and global health status.

Demographics and global health status mean ± SD	Sample (n = 30)
Age, years	60.6 ± 3.8
Age, range	56-70
Gender, n (%)
Male	11 (36.7)
Female	19 (63.3)
Ethnicity, n (%)
Asian	2 (6.65)
Black	5 (16.7)
Hispanic	15 (50.0)
Hispanic/White	2 (6.65)
White	6 (20)
Handedness, n (%)
Right	26 (86.7)
Left	4 (13.3)
Education level, n (%)
Incomplete College	5 (16.7)
Complete College	15 (50.0)
Technical/Associate degree	2 (6.7)
Graduate degree	8 (26.6)
BMI, n (%)
Normal (18,5 – 24,9)	4 (13.3)
Overweigh (25 – 29,9)	9 (30)
Obese (≥ 30)	17 (56.7)
BMI, kg/m2	30.6 ± 5.5
Physical activity level
IPAQ total, METs	346.3 ± 361.2
IPAQ weekday sitting, hours	6.5 ± 3.1
Global cognition
MoCA	25.3 ± 2.1
Reported Comorbidities and disorders, n (%)
Hypertension	18 (62)
Heart problem/disease	6 (20.7)
Lung or breathing problem	8 (31)
Diabetes	5 (17.2)
Kidney disease	2 (7)
Thyroid disease	7 (24)
Arthritis or joint pain	16 (55)
Tumor or Cancer	4 (13.8)
Smoking history, n (%)
Yes	6 (20)
No	24 (80)

Abbreviations. n = sample; % = percentage; SD top= standard deviation; kg = kilogram; kg / m² = kilograms per square meter; BMI = Body Mass Index; METs = metabolic equivalent; IPAQ = International Physical Activity Questionnaire; MoCA = Montreal Cognitive Assessment.

3.2. TMS-iTBS neuroplasticity results

Table 2 displays the results for the TMS-iTBS neuroplasticity assessment. All subjects tolerated TMS-iTBS and no adverse events were reported. Concerning the effects of TMS-iTBS within-group, the random-effects models indicated that MEPs did not vary significantly by Time (F_4,106= 1.3, p= .28,η²= .05). This finding indicates that MEPs amplitudes were not significantly potentiated by TMS-iTBS in the sedentary aging individuals. Exploratory planned t-tests comparisons showed a significant change effect in the group at Post10 (t= 2.10, p= .03), intervals Post0-10 (t= 2.26, p= .025), Post10-20 (t= 2.64, p= .009), and Post20-30 (t= 2.58, p= .011). Figure 3 shows the Post-iTBS MEPs response at interval times and highlights the largest effect size in the range interval time of Post 10-20 (η²= .19). Summarizing both analyses, we conclude that the MEPs modulation was not consistent over time, but when considering isolated time points, it is likely that TMS-iTBS are potentiated. This may be due to inter- and intra- variability of TMS response that may be increased by the sedentary status of the individuals.

Table 2.

TMS baseline and post-TMS-iTBS data.

TMS neuroplasticity	Sample (n = 29)	Statistical Analysis	df	F/t	p-value	∣η2∣
Motor threshold (%)
RMT	59.3 ± 13.4	-
AMT	47.7 ± 11.3
Baseline MEPs (mV)
Post0	0.76 ± 0.47	-
Post-iTBS MEPs (mV)		Random-effects linear model (Time factor)	4,106	1.3	0.28	.05
Post5	0.82 ± 0.55	Planned t-test (Baseline vs. time-points)	29	0.68	0.50	.02
Post0-10	0.89 ± 0.55			2.26	0.025*	.15
Post10	0.98 ± 0.65			2.10	0.03*	.13
Post10-20	0.93 ± 0.59			2.64	0.009*	.19
Post20	0.92 ± 0.61			1.42	0.15	.07
Post20-30	0.91± 0.65			2.58	0.011*	.18
Post30	0.85 ± 0.88			0.69	0.49	.02
Post0-30	0.86 ± 0.56			1.46	0.14	.07
Post-iTBS MEPs (%Δ)		Random-effects linear model (Time factor)	3,78	0.85	0.47	.03
Post5 %Δ	17.4 ± 79.1
Post0-10 %Δ	23.7 ± 66.6
Post10 %Δ	35.0 ± 72.8
Post10-20 %Δ	24.3 ± 53.8
Post20 %Δ	20.2 ± 53.0
Post20-30 %Δ	17.0 ± 45.0
Post30 %Δ	10.5 ± 66.5
Post0-30 %Δ	15.1 ± 46.4

Abbreviations: TMS = Transcranial Magnetic Stimulation; RMT = Resting Motor Threshold; AMT = Active Motor Thresholds; MEPs = Motor Evoked Potentials; iTBS = Intermittent Theta-Burst Stimulation; mV = millivolts; %Δ = percent change; * = statistically significant.

Figure 3.

TMS-iTBS induced modulation of MEPs. Shade highlights the interval Post 10-20_%Δ.

Abbreviations. M = mean; SE = standard error; iTBS = intermittent Theta-Burst; MEPs = motor evoked potentials; %Δ = percent change.

Despite nonsignificant within-subject factor Time changes (F_3,78= .85, p= .47,η²= .03), the neuroplasticity assessment demonstrated, an overall potentiation of MEPs post-iTBS. Specifically, the standardized mean change of MEP amplitudes showed a mean increase of 35% at Post10_%Δ, 24.3% at the interval of Post 10-20_%Δ (Figure 3), and an overall 15.1% accounting for the interval Post0-30_%Δ.

3.2.1. Excitatory vs. Inhibitory responders

A subgroup exploratory analysis revealed that 21 (72.4%) of the participants showed a Post10-20_%Δ excitatory response and only 8 (27.6%) an inhibitory or no response to TMS-iTBS neuroplasticity assessment (Figure 4). Concerning the effects of TMS-iTBS within-subgroups, the random-effects models indicated MEPs varied significantly by Time in both the ‘Excitatory Responders’ (F_4,79= 2.52, p= .047, η²= .11), and the ‘Inhibitory Responders’ (F_4,28= 3.89, p= .012, η²= .36). The ‘Excitatory Responders’ demonstrate an overall MEPs potentiation of 42.5%, while the ‘Inhibitory Responders’ a decrement of −31.4% (between-group effect of F_1,25= 21.5, p< .0001, η²= .46).

Figure 4.

Subgroup analysis of ‘Excitatory’ and ‘Inhibitory’ TMS-iTBS induced modulation of MEPs.

Abbreviations. M = mean; SE = standard error; MEPs = motor evoked potentials; %Δ = percent change.

3.3. Cardiorespiratory fitness and circulating markers of endothelium function results

Table 3 displays the results of the CRF assessment. All subjects tolerated the ISWT, and no adverse events were reported during or after the tests. One participant presented a high systolic blood pressure (>180 mmHg) prior to commencement of the physical tests and was excluded from the sample. The sample, on average, exhibited cardiovascular outcomes within normal ranges or slightly abnormal for aging adults. Among the heart rate measures, the population exhibited a normal mean resting heart rate (73.3 ± 11.4 bpm) and an optimal HRR1 (28.1 ± 11.5 bpm, optimal mean HRR1 > 12 bpm) (Cole et al., 1999). In accordance with the American Heart Association’s newest guidelines, the mean resting BP was elevated (parameters of SBP 120-129mmHg and DBP <80 mmHg) (Whelton et al., 2018). Immediately after the test, the expected elevation on SBP was observed, and a slight elevation of DBP. BP returned to baseline levels after the recovery phase. The mean predicted ISWT distance observed was similar to the real mean distance of the sample (ISWT distance % of predicted 102.9 ± 22.2). The estimated VO₂peak for the current sample was below normative values for the age group demonstrating poor-to-fair aerobic capacity (Loe et al., 2016).

Table 3.

Cardiorespiratory fitness assessment and circulating markers of endothelium function data.

Cardiorespiratory function testing, ISWT (n = 29)		Mean ± SD	95% CI
Heart Rate, bpm	Resting HR	73.3 ± 11.4	68.9 - 77.6
	Max HR	134.9 ± 19.3	127.5 – 142.3
	HR Reserve	61.6 ± 18.1	54.7 – 68.5
	HR Recovery post 1 min	28.1 ± 11.5	23.7 – 32.5
	HR Recovery post 2 min	41.2 ± 14.4	35.7 – 46.7
Blood Pressure, mmHg	Resting Systolic BP	124.8 ± 15.2	119.1 - 130.6
	Resting Diastolic BP	77.7 ± 8.7	74.3 – 80.9
	Immediately after SBP	157.2 ± 27.4	145.9 – 168.6
	Immediately after DBP	80.9 ± 11.1	76.4 – 85.5
	Recovery SBP	128.3 ± 17.0	121.6 – 135.0
	Recovery DBP	77.1 ± 9.2	73.5 – 80.7
Rate-pressure product	Resting RPP	9163 ± 1880	8448 - 9879
	Immediately after RPP	21491 ± 5846	19078 - 23078
	Recovery RPP	12137 ± 3011	10945 - 13328
ISWT distance and VO2peak	ISWT distance, m	493.2 ± 124.2	445.0 – 541.4
	Predicted ISWT distance, m	481.6 ± 75.5	453.4 – 509.7
	% of predicted	102.1 ± 22.1	93.5 – 110.7
	Estimated VO2peak, ml/kg/min	21.9 ± 4.8	20.1 – 23.8
Circulating Markers of endothelium function (n = 24)*		Mean ± SD	95% CI
	BDNF, pg/mL	492.6 ± 348.1	345.6 – 639.6
	VEGF, pg/mL	12.3 ± 5.4	10.0 – 14.6
	Hs-CRP, mg/L	1.9 ± 1.3	1.3 – 2.4

Abbreviations: ISWT = Incremental Shuttle Walking Test; SD = Standard Deviation; CI = Confidence Interval; HR = Heart Rate; BP = Blood Pressure; SBP = Systolic Blood Pressure; DBP = Diastolic Blood Pressure; RPP = Rate-Pressure Product; VO₂peak = peak oxygen consumption during exercise; BDNF = brain-derived neurotrophic factor; VEGF = vascular endothelial growth factor; Hs-CRP = High-sensitivity C-reactive protein; bpm = beats per minute; mmHg = Millimeter of mercury; ml/kg/min = milliliters per minute per kilogram; pg/mL = Picograms per milliliter; mg/L = milligrams per liter.

^*Two individuals were excluded from further analysis due to outlier values above 2.5 SD.

Currently, there is no consensus about normative levels of BDNF and VEGF for aging adults. For BDNF plasma levels, our sample reported a mean (492.6 ± 348.1 pg/ml) below the range when compared to reference values from mid-older healthy individuals (reference values of 780-819 pg/ml) (Driscoll et al., 2012; Lee et al., 2007). Our sample showed VEGF plasma levels (12.3 ± 5.4 pg/ml) within normal ranges for mid-range adults (0–115 pg/ml) (Raimondo et al., 2001; Smets et al., 2016). Hs-CRP mean levels (1.9 ± 1.4 mg/L) were within a moderate range (1.0 – 3.0 mg/L) of cardiovascular disease risk (Cozlea et al., 2013).

3.4. Correlation analysis

3.4.1. Neuroplasticity vs. cardiovascular health measures

Pearson product-moment correlation coefficient between neuroplasticity and cardiovascular outcomes revealed that greater neuroplasticity (Post10-20_%Δ) was moderately significant associated with better HRR-1min (r= .41, p= .03, Figure 4A), HRR-2min (r= .40, p= .04, Figure 4B), and HR reserve (r= .37, p= .05, Figure 4D). Additionally, greater neuroplasticity was significantly associated with lower resting HR (r= −.38, p= .04, Figure 4C), resting RPP (r= −.42, p= .03, Figure 4E), recovery RPP (r= −.53, p= .005, Figure 4F). No other significant correlation between the two factors was identified. Together, these findings indicate that cardiovascular performance is relevant for cortical neuroplasticity modulation in aging adults. Those with better cardiovascular performance demonstrated greater efficacy of mechanisms of TMS-iTBS neuroplasticity.

3.4.2. Neuroplasticity vs. circulating markers of endothelium function

Spearman correlation coefficient between neuroplasticity and circulating markers of endothelium function revealed that greater neuroplasticity (interval Post10-20_%Δ) was significantly associated with lower VEGF levels (r_s=−.41, p=.048, Figure 5A), and lower CRP concentration (r_s=−.43, p=.04, Figure 5B). No correlation was found between BDNF levels and neuroplasticity (r_s=−.26, p=.23, Figure 5C).

Figure 5.

Neuroplasticity vs. Cardiovascular health measures.

Note. Post10-20%Δ represents the percentage change (%Δ) in peak-to-peak MEP amplitude from baseline to the period 10-20 minutes post-iTBS. The values are represented in the decimal form where −1.0 represents a decrement in neuroplasticity of 100% and +1.0 represents an increase in neuroplasticity of 100%.

3.5. Linear Regression Analysis

3.5.1. Unadjusted Regression Analysis

Simple linear regressions were fitted to identify the strongest association between the outcome neuroplasticity (T10-20%Δ) and cardiovascular measure predictors (HRR1, HRR2. Resting HR, HR reserve, resting RPP, recovery RPP and estimated VO2peak). Unadjusted models are represented in Table 4. Both HRR1 (β= .019, p= .002), and HRR2 (β= .014, p= .018), demonstrated a significant and positive association between with neuroplasticity, which only remained significant for the HRR1 after FDR corrections (p= 0.026).

Table 4.

Unadjusted models of association between neuroplasticity and cardiovascular health.

Unadjusted Models	β Coefficients (95% CI)	SE	Adj R- squared	p- value
T10-20%Δ + HRR1	.019 (.007, .031)	.007	.134	.002*
T10-20%Δ + HRR2	.014 (.001, .028)	.006	.122	.018
T10-20%Δ + Resting HR	−.015 (−.031, .001)	.009	.051	.055
T10-20%Δ + HR Reserve	.007 (−.003, .017)	.005	.066	.133
T10-20%Δ + Resting RPP	−.0001 (−.0002, −.00001)	.0001	.092	.049
T10-20%Δ + Recovery RPP	−.0001 (−.0001, −.00001)	.0001	.097	.057
T10-20%Δ + Estimated VO2 Peak	.018 (−.027, .064)	.024	.013	.425

Abbreviations: T10-20%Δ= percent change in the neuroplasticity index measure from baseline to interval time of 10-20 minutes; HRR1= heart rate recovery at minute 1; HRR2= heart rate recovery at minute 2; HR= heart rate; RPP= rate pressure product.

Notes.

1) * = p-value remained significant after false rate discovery correction.

3.5.2. Adjusted Regression Analysis for demographics and health status

Four multiple linear regressions were fitted to analyze the association between neuroplasticity and cardiovascular health adjusting for baseline demographic and health status data. All adjusted models are presented in table 5. Model 0 (T10-20%Δ + HRR1) was selected as the best model fit from the previous unadjusted model of table 4. Model 1 only adjusted for age and gender. Model 2 adjusted for age, gender, race/ethnicity, body mass index, education level, and physical activity level. Model 3 included all variables from model 2 adding the number of comorbidities. Lastly, Model 4 included all variables from model 3 and added smoking history. The results showed the model 3 (β= .043, p= .002) adjusted for age, gender, race/ethnicity, body mass index, education, physical activity level, and the number of comorbidities as predictor variables together resulted in a statistically significant improvement in model fit (LR= 5.83, p-value = .016) which remained significant after FDR corrections (p= 0.043). Thus, model 3 shows that for every unit increase in HRR1, neuroplasticity also increases by .043 (p= .002), accounting for baseline demographic and health status.

Table 5.

Adjusted models of association between neuroplasticity and cardiovascular health.

Adjusted Models	β Coefficients (95% CI)	SE	Adj R- squared	p- value	LR test
Model 0: T10-20%Δ + HRR1	.019 (.007, .031)	.007	.134	.002*	-
Model 1: T10-20%Δ + HRR1 + Age + Gender	.023 (.010, .036)	.007	.164	.001*	M0 > M1, p= .20
Model 2: T10-20%Δ + HRR1 + Demographic variables	.035 (.015, .055)	.009	.391	.003*	M2 > M0, p= .006*
Model 3: T10-20%Δ + HRR1 + Comorbidities + Demographic variables	.043 (.021, .065)	.010	.454	.002*	M3 > M2, p= .016*
Model 4: T10-20%Δ + HRR1 + Comorbidities + Smoking history + Demographic variables	.045 (.021, .069)	.011	.440	.002*	M3 > M4, p= .16

Abbreviations. T10-20%Δ= percent change in the neuroplasticity index measure from baseline to interval time of 10-20 minutes; HRR1= heart rate recovery at minute 1.

Notes.

1. All β Coefficients represents the association between Neuroplasticity and Cardiovascular health controlled for the other variables.

2. Demographics variables included age, gender, race/ethnicity, body mass index, education, and physical activity level.

3. Comorbidities included current of past health condition history including hypertension, heart disease, lung disease, diabetes, kidney disease, thyroid disease, arthritis, cancer, and others.

4. Model 0 was selected the best fit unadjusted model.

5. Likelihood ratio tests compared two models at a time and the best fit model was represented by the equation “Mx > My” meaning Mx was a better fit when compared to My. 3

6. * = p-value remained significant after false rate discovery correction

3.5.3. Adjusted Regression Analysis accounting for the circulating markers of endothelium function

Model 3 (T10-20%Δ, HRR1, Comorbidities, and Demographic variables) was selected from the previous models as the best fit comparator model. Four additional multiple linear regressions models were fitted to analyze if circulating markers of endothelium function (BDNF, VEGF, hs-CRP) would modify the association between neuroplasticity and cardiovascular health. Three models (Models 6, 7, and 8) were fitted to included circulating markers individually. Model 7, that included VEGF as a modifier, was the only individual model that showed a better fit and significantly improved the association of neuroplasticity and cardiovascular health (LR= 9.65, p-value = .0019). Model 7 showed that for every unit increase in HRR1, neuroplasticity also increases by .038 (p= .036), when accounting for VEGF (β= −.075), baseline demographic and health status.

When combining the modifiers BDNF (β= .0006) and VEGF (β= −.120) in model 9, the new model showed a better fit and significantly improved the association of neuroplasticity and cardiovascular health when compared to model 7 that only included VEGF (LR= 3.85, p-value = .0496). Model 9 showed that for every unit increase in HRR1, neuroplasticity also increases by .041 (p= .043), when accounting for VEGF and BDNF baseline levels, demographics, and health status.

4. Discussion

The present findings supported our primary hypothesis that greater neuroplasticity response would be associated with better cardiovascular health. Our findings revealed significant associations between TMS-iTBS neuroplasticity and the predictors of cardiovascular health and circulating markers of endothelium function in sedentary aging adults. Particularly, we found that the best cardiovascular health predictor of neuroplasticity is HRR1, and a better fit model was selected when controlling for demographics (age, gender, race/ethnicity, body mass index, education) and health status (physical activity level and the number of comorbidities). We also were able to partially support our secondary hypothesis. We demonstrated that including the VEGF plasma levels in combination with BDNF plasma levels predictors augmented the association between neuroplasticity and cardiovascular health. These findings also add to the behavioral relevance of TMS/iTBS as an assessment of mechanisms of neuroplasticity that is associated with well-established cardiovascular measures driven by the crucial role of autonomic changes on both nervous and cardiovascular systems.

Our findings also revealed relevant data on TMS-iTBS neuroplasticity response in aging adults. Although our findings revealed inconsistent MEPs modulation at a whole group level, the subgroup analyses indicate that TMS-iTBS neuroplasticity assessment response may act differently in sedentary aging adults, promoting not only excitatory response in the participants but also an inhibitory response in others. These results further emphasize the inter-and intra- variability TMS response in aging adults reported in other studies (Dickins et al., 2015; Puri et al., 2016), and also drive the importance to discuss physiological factors (e.g., cardiovascular health and autonomic function) that may be influencing the underlying mechanisms of neuroplasticity responses (Makovac et al., 2017; Schestatsky et al., 2013). Other studies have also reported that BDNF polymorphism plays a crucial role in neuroplasticity response, where Val66Met carriers (heterozygotes) individuals show a reduced modulation after an exercise program (Gomes-Osman et al., 2017).

Although there are reports linking CRF and cardiovascular risk factors to cortical plasticity measures and brain health in aging and clinical populations (Colcombe, Kramer, Erickson, et al., 2004; España-Irla et al., 2021; Fried et al., 2016a), this study further adds to the existing body of literature by demonstrating that a non-invasive measure of neuroplasticity with TMS-iTBS is associated with HRR1. Importantly, HHR1 is modifiable by means of lifestyle changes such as increased physical activity levels. This raises the intriguing possibility that training-induced changes in HHR1 could also influence TMS-iTBS neuroplasticity, and lead to beneficial functional, cognitive, and behavioral consequences.

In previous studies, cardiovascular fitness was reported to drive improvements in cerebral perfusion, brain structure, brain connectivity, neural efficiency, and trophic factors (Cabral et al., 2019) that have the potential to improve neuroplasticity mechanisms and ultimately lead to improvements in cognitive brain health (Colcombe, Kramer, McAuley, et al., 2004; España-Irla et al., 2021; Hötting & Röder, 2013). An important implication of these findings is the potential utility of HRR to estimate cardiorespiratory performance in sedentary aging adults. While CRF, as measured by maximal oxygen consumption, is the gold standard protocol, one disadvantage of this method is that it can be easily altered by the amount of effort an individual is putting in during the test. Alternatively, HRR avoids this bias because HRR is a physiological cardiovascular measure that seems to work as a function of the autonomic nervous system (ANS) adaptation and is thus more sensitive and accurate than maximal oxygen consumption.

The demonstrated association between cardiovascular health and neuroplasticity in this study can augment the relevance of current findings in animal studies and serve as a theoretical framework for future translational longitudinal and interventional studies in humans. For instance, within the hypothalamus and brainstem circuitry, animal models showed that exercise-induced brain plasticity augments the positive linking of cardiovascular function and ANS adaptations mainly due to sustained activation and increased efficacy of the connections within the NTS-PVN (Nucleus Tractus Solitarius and Hypothalamic Paraventricular Nucleus) reciprocal network (Dampney, 2016; Michelini et al., 2015; Michelini & Stern, 2009). This pattern of interconnectivity is also relevant for the body’s homeostasis in pathological conditions (e.g. hypertension and diabetes), as exercise can overcome neuroplastic maladaptive response by attenuating NTS-PVN neuronal excitability and regulating ANS response. In humans, cardiovascular diseases are frequently related to the overactivation of the ANS (Malpas, 2010). Thus, improving cardiovascular health, including HRR as demonstrated in this study, may improve the ANS, and be associated with greater neuroplasticity response (Mueller, 2007). Other implication of these findings is that HRR is modifiable via exercise and that future studies may further examine these underlying physiological mechanisms in the human body systems.

In this context, crucial circulating markers of endothelium function, including trophic factors BDNF and VEGF, seem to also influence the relationship between cardiovascular health and neuroplasticity. In our exploratory multiple linear regression, VEGF was the best independent predictor among the circulating markers analyzed in this study and negatively modified the association (β= −.075) between neuroplasticity and the main predictor HRR1. This result was expected since our sample demonstrated a high degree of cardiovascular risk factors (sedentary, majority obese [56.7%], 62% hypertension, 21% with a history of cardiovascular disease, and 17% diabetic). Although the importance and benefits of VEGF in angiogenesis and neuroplasticity, studies have observed higher levels of circulating VEGF in persons with chronic conditions including obesity, hypertension, metabolic syndrome, cancer, and markers of atherosclerosis, that may be due to compensation to their subclinical disease (Kusumanto et al., 2007; Zachary, 2005). Thus, lower neuroplasticity observed in individuals with higher VEGF levels could be associated with endothelium dysfunction and thus influence a maladaptive ANS response, as we further explained above.

We failed to observe a correlation between neuroplasticity and BDNF as an independent predictor. However, our exploratory analysis revealed that BDNF plasma levels improved the model combined with VEGF plasma levels as modifiers of the association between neuroplasticity and cardiovascular health represented by HRR1. This is consistent with previous findings that demonstrated BDNF and VEGF levels give insights into endothelial function (Trigiani & Hamel, 2017), and the interaction of both factors contribute to LTP neuroplastic changes (Aicardi et al., 2004; Zhang et al., 2015b). Further studies should investigate the association of neuroplasticity response in the presence of BDNF polymorphism. It is already known that BDNF Val66Met polymorphism attenuates plasticity responses (Andrews et al., 2020), and high BDNF level was only associated with greater cardiorespiratory fitness gains only on Val/Val carriers individuals and those with greater intracortical inhibition (Nicolini et al., 2019). Although we did not observe polymorphisms in this population, we believe this could be one reason for not showing the expected independent relationship between BDNF and neuroplasticity.

Elevated Hs-CRP levels are a conditional risk factor for cardiovascular disease. Our study showed that higher levels of Hs-CRP are associated with lower neuroplasticity. Considering that the present study population reported a variety of cardiovascular risk factors, this finding is consistent with studies that concluded abnormal mechanisms of neuroplasticity response in individuals associated with a high rate of cardiovascular diseases and risk factors (Fried et al., 2016a; Johnson & Xue, 2018). However, this significant correlation did not survive when adjusted for baseline and health status in the multiple regression analysis. We also failed to show Hs-CRP as an independent modifier predictor of the association between neuroplasticity and cardiovascular health. Although we were not able to establish an adjusted regression model linking Hs-CRP with neuroplasticity and cardiovascular, we believe this association should be incremented in further studies with larger samples and in combination with other inflammatory markers of endothelium dysfunction predictors. Future studies should further investigate the detrimental effect of higher levels of Hs-CRP catabolic response (e.g., inspiratory muscle weakness/dysfunction and poorer respiratory capacity) (Schaap et al., 2006; Tay et al., 2019; Wåhlin-Larsson et al., 2017) in the body of aging adults, and its effect on neuroplasticity and brain health.

Despite the contributions of this cross-sectional analysis on the importance of cardiovascular health on neuroplasticity, future analyses should incorporate the role of pre to post changes in neuroplasticity and cardiovascular health measures after a structured physical exercise program. Furthermore, other factors and behaviors influencing brain health should be explored, including nutrition and diet habits, sleep habits, mental health, genetic and epigenetic factors, social interaction, and other social determinants of health.

Additional limitations were noted in this cross-sectional analysis and findings should be interpreted with caution and further explored in future studies. A recent study has shown substantial inter-and intra-individual variability for the modulatory effects of TMS theta-burst in humans (Corp et al., 2020; Ozdemir et al., 2021). This highly variable response was also pointed out in our sample, and in combination with the sedentary and aging status of the sample may explain the inconsistent MEPs modulatory response over time in this population. Future studies should seek to explore the association between cardiovascular health and the direct assessment of brain cortical neuroplasticity. For non-invasive cortical excitability evaluation, the combination of TMS and simultaneous electroencephalography (TMS-EEG) may be more suitable for non-motor circuitries and could provide a global measure of cortical reactivity, inhibition, and interconnectivity using TMS-EEG neuromodulatory techniques (Pascual-Leone et al., 2011a; Tremblay et al., 2019). Additionally, the limited sample size may have affected the fulfillment of the proposed hypothesis, and further research with a larger sample is warranted to clarify some of the mechanisms proposed in our study. Furthermore, two additional TMS applications may help elucidate supplementary mechanisms between cardiovascular health and neuroplasticity. For instance, TMS-induced long-term depression (LTD)-like plasticity changes may decrease corticospinal excitability using the protocol of continuous theta-bust stimulation (cTBS). This may be particularly relevant to target ANS overactivation in cardiovascular pathological conditions as discussed above. Second, TMS modulation is a promised therapeutic tool and may be clinically relevant to mitigate cardiovascular symptoms (e.g., higher heart rate and lower heart rate variability) present in psychiatric conditions such as major depressive disorder (MDD). For instance, researchers have recently proposed the Neuro-Cardiac-Guided TMS (NCG-TMS) technique, which revealed a heart rate deceleration when targeting the frontal-vagal pathway in MDD and healthy young-middle-aged adults (Iseger et al., 2020, 2021).

Conclusion

Taken together, our findings strengthen the theoretical framework linking the brain-heart pathway, as we demonstrate that neuroplasticity, as assessed by TMS/iTBS, was associated with well-established clinical outcomes of cardiovascular health. This finding builds on existing data demonstrating that TMS may provide insight into neuroplasticity and the role of exercise, and cardiovascular health appears to have on its mechanisms. The data can also contribute to the field by providing an easy, safe, and well-established cardiovascular health measure, the HRR1, as a predictor of mechanisms of neuroplasticity in aging adults. We also were able to explore preliminary results on the role of circulating markers of endothelium function on neuroplasticity mechanisms and cardiovascular health. These implications can serve as theoretical framework for future longitudinal and interventional studies aiming to influence attitudes towards to change behavior aiming to improve cardiovascular health, and subsequently improve brain health in this population. Clinicians would benefit from this study by including HRR1 as a potential screening and prognostic measure of cardiovascular health and as a surrogate marker of brain health.

Figure 6.

Neuroplasticity vs. Circulating markers of endothelium function

Note. Post10-20%Δ represents the percentage change (%Δ) in peak-to-peak MEP amplitude from baseline to the period 10-20 minutes post-iTBS. The values are represented in the decimal form where −1.0 represents a decrement in neuroplasticity of 100% and +1.0 represents an increase in neuroplasticity of 100%.

Table 6.

Adjusted models accounting for circulating markers modifiers of the association between neuroplasticity and cardiovascular health.

Adjusted Models	β Coefficients (95% CI)	SE	Adj R- squared	p- value	LR test
Model 3: T10-20%Δ + HRR1 + Comorbidities + Demographic variables	.043 (.021, .065)	.010	.454	.002*	-
Model 6: T10-20%Δ + HRR1 + BDNF + Comorbidities + Demographic variables	.044 (.002, .086)	.011	.153	.044	M3 > M6, p= .12
Model 7: T10-20%Δ + HRR1 + VEGF + Comorbidities + Demographic variables	.038 (.004, .073)	.011	.399	.036	M7 > M3, p= .0019*
Model 8: T10-20%Δ + HRR1 + Hs-CRP + Comorbidities + Demographic variables	.032 (−.031, .095)	.015	.007	.227	M3 > M8, p= .084
Model 9: T10-20%Δ + HRR1 + BDNF + VEGF + Comorbidities + Demographic variables	.041 (.002, .080)	.012	.374	.043	M9 > M7, p= .0496

Abbreviations. T10-20%Δ= percent change in the neuroplasticity index measure from baseline to interval time of 10-20 minutes; HRR1= heart rate recovery at minute 1; BDNF= brain derived neurotrophic factor; VEGF = vascular endothelial growth factor; Hs-CRP= high-sensitivity c-reactive protein.

Notes.

1. All β Coefficients represents the association between Neuroplasticity and Cardiovascular health controlled for the other variables.

2. Demographics variables included age, gender, race/ethnicity, body mass index, education, and physical activity level.

3. Comorbidities included the number of current or past health condition history including hypertension, heart disease, lung disease, diabetes, kidney disease, thyroid disease, arthritis, cancer, and others.

4. Model 3 was selected from previous analysis as the best fit adjusted model of demographics and health status.

5. Likelihood ratio tests compared two models at a time and the best fit model was represented by the equation “Mx > My” meaning Mx was a better fit when compared to My.

6. * = p-value remained significant after false rate discovery correction

Highlights.

• We provide new insights regarding the association of TMS/iTBS measures of neuroplasticity and cardiovascular health in sedentary aging adults.

• Heart rate recovery at minute 1 is the best cardiovascular health predictor of the efficacy of the mechanisms of neuroplasticity.

• Heart rate recovery at minute 1 is a potential surrogate marker of brain health.

• VEGF and BDNF levels improve the model of the association between neuroplasticity and cardiovascular health.

Abbreviations

%Δ

Percent change

HRR

Heart Rate Recovery

ISWT

Incremental Shuttle Walking Test

iTBS

Intermittent Theta-Burst Stimulation

LTP

Long Term Potentiation

TMS

Transcranial Magnetic Stimulation