Brain and cardiovascular-related changes are associated with aging, hypertension, and atrial fibrillation

The neural pathways in which the brain regulates the cardiovascular system is via sympathetic and parasympathetic control of the heart and sympathetic control of the systemic vasculature. Various cortical and sub-cortical sites are involved, but how these critical brain regions for cardiovascular control are altered in healthy aging and other risk conditions that may contribute to cardiovascular disease is uncertain. Here we review the functional and structural brain changes in healthy aging, hypertension, and atrial fibrillation – noting their potential influence on the autonomic nervous system and hence on cardiovascular control. Evidence suggests that aging, hypertension, and atrial fibrillation are each associated with functional and structural changes in specific areas of the central nervous system involved in autonomic control. Increased muscle sympathetic nerve activity (MSNA) and significant alterations in the brain regions involved in the default mode network are commonly reported in aging, hypertension, and atrial fibrillation. Further studies using functional and structural magnetic resonance imaging (MRI) coupled with autonomic nerve activity in healthy aging, hypertension, and atrial fibrillation promise to reveal the underlying brain circuitry modulating the abnormal sympathetic nerve activity in these conditions. This understanding will guide future therapies to rectify dysregulation of autonomic and cardiovascular control by the brain.


Introduction
Cardiovascular disease (CVD) contributes to more than 32% of deaths worldwide [152]. With the increasing burden of CVD on individuals and the healthcare system, there is a growing need to understand the pathophysiology and the effects of different risk conditions that may contribute to cardiovascular diseases. For example, the risk of developing cardiovascular disease is notably higher in older age, with hypertensive status, and with atrial fibrillation [32, 99,100,114].
The burden of CVD on an individual's health increases dramatically with age, and more than 86% of hospitalizations due to CVD occur in people over 65 years old [46]. A shift in population demographics towards the older age group in the coming decades is likely, with projections that close to 20% of the population will be older than 65 years by 2030, posing a significant health burden worldwide [99].
Other risk factors for CVD include hypertension, which contributes approximately 35% of the total CVD burden [46]. Hypertension is associated strongly with intracerebral hemorrhage, subarachnoid hemorrhage, and stable angina [108]. The onset of CVD is associated with the age at which hypertension develops; consistent with this, individuals who survive longer generally develop hypertension later in life, and the onset of CVD is delayed [3,32].
Atrial fibrillation (AF) is the most frequently observed cardiac arrhythmia associated with an increased risk of 1 3 cardiovascular mortality, major cardiovascular events, and stroke [100]. It is estimated that 46.3 million people had AF in 2016 globally [9]. In addition, hypertensive status is strongly associated with the development of AF [73,143]. Also, increasing age is one of the most critical risk factors for AF [70,105], with the risk of hospitalization due to AF increasingly higher in individuals aged 65 years or above [46]. Moreover, AF patients are three times more likely to develop heart failure than individuals free of AF [134].
These conditions can be readily detected within the general population using simple, non-invasive measures [7,59]. Nevertheless, if left untreated, they pose a significant burden on the healthcare system. One common feature observed in these conditions is the alteration in the autonomic nervous system. While the autonomic nervous system can be regulated via neural or endocrine pathways, we focus our attention on the neural path to explore how changes in the brain contribute to the autonomic nervous system, particularly the sympathetic branch and cardiovascular control.
Given that the one way in which the brain controls cardiac output is via changes in sympathetic and parasympathetic outflow to the heart and total peripheral resistance through changes in sympathetically mediated increases or decreases in sympathetic vasoconstrictor drive, it is possible that many cardiovascular diseases feature changes in the brain-either as a consequence or, potentially, a cause. In 2010, our lab conducted the first study in which we performed microelectrode recordings of muscle sympathetic nerve activity (MSNA) while performing functional magnetic resonance imaging of the brain (fMRI) intermittently [78]. This work contributes to the growing research to simultaneously measure changes in the brain and measures of the autonomic nervous system to see how these two systems may interact. These include studies looking at heart rate variability induced by handgrip exercise and fMRI [93], high-frequency heart rate variability with structural MRI [155], lower body negative pressure to induce autonomic changes while performing fMRI [82], concurrent measurement of EEG and fMRI [103], and performing physiological maneuvers such as apnea during fMRI [64]. Our study was the first to use an invasive direct recording of the sympathetic nervous system and fMRI of the brain in a closed-loop manner, whereby no external stimulus was used to elicit an autonomic response but relied on the autonomic regulation of spontaneous fluctuations in blood pressure. This study focused on the brainstem regions responsible for generating sympathetic outflow to the muscle vascular bed. We found that BOLD (blood oxygen level-dependent) signal changes covaried with muscle sympathetic activity in the rostral ventrolateral medulla, caudal pressor area, the dorsal motor nucleus of the vagus, and medullary raphé. We also found that signal intensity decreased with increased sympathetic activity in the nucleus tractus solitarius and caudal ventrolateral medulla.
A subsequent study identified areas above the brainstem, including specific nuclei of the hypothalamus, the insula cortex, precuneus, and the dorsolateral prefrontal cortex [56].
Adding support to the findings from functional brain imaging studies, evidence of anatomical neural connections between these areas can be seen. The central autonomic network can be divided into three hierarchical levels: brainstem, subcortical, and cortical regions [130]. The currently proposed sympathetic connectome is insula-hypothalamus-PAG-RVLM circuity, which is tightly coupled to MSNA at rest [80]. It has been observed from animal studies that rostral ventrolateral medulla (RVLM) receives projections from periaqueductal gray (PAG) [66]. In addition, there are direct projections from the hypothalamus to PAG in rats [20,51]. MRI studies in humans also support the connection between the hypothalamus and PAG [27,48]. The insula has direct connections with the hypothalamus in rats and monkeys [101,119]. Moreover, the medial prefrontal cortex has projections to the insula [101], hypothalamus [101,113], PAG [29,31], and RVLM [52,141]. These anatomical connections demonstrated in other animals give support to there being similar connections in humans.
Further details on the functionally identified central autonomic network can be found in recent reviews [79,80,130], but here we shall simply highlight some of these areas. Figure 1 shows a schematic representation of cortical and subcortical regions of the brain known to be involved in cardiovascular control. This review will consider how the brain changes with healthy aging and two other risk conditions that may contribute to cardiovascular diseases.

The burden of aging
According to the World Population Aging report by the United Nations, the number of individuals over 65 years old reached 703 million in 2019. Furthermore, it is estimated that this population will reach more than 1.5 billion by 2050 in the next three decades [140]. There will be a shift in population demographics towards the older age group, with estimates that 20% of the population will be over 65 by 2030 [99]. The burden of CVD on an individual's health increases dramatically with age, and more than 86% of hospitalizations due to CVD occur in people over 65 years old [46]. The world's aging population trend will continue to increase CVD incidence across different nations [47,63].

Cardiovascular changes with aging
It is well known that aging leads to many functional and structural changes in the body, including in the heart. Aging reduces high-and low-frequency heart rate variability [4,121] and is associated with more frequent arrhythmias [75]. The prevalence and incidence of arrhythmic disorders, including AF, supraventricular tachycardia, and ventricular premature complexes, are elevated amongst the older population [75].
Cardiac MRI studies in more than 5000 patients revealed an increased left ventricular (LV) mass-to-volume ratio with aging [14], a finding consistent with studies indicating that the LV wall thickens with age [18]. Moreover, studies also suggest left ventricle (LV) mass increases with aging [34,104], although there is some disagreement around this change since it has been shown that LV mass changes differ with sex and ethnic groups [94]. Nevertheless, overall, it is evident that alteration in LV structure with increasing age is due to changes in LV mass-to-volume ratio, LV mass, or LV mass index.
Aging leads to changes in the vasculature, including arterial thickening, vascular endothelial dysfunction, and arterial stiffening [72,123]. These vascular changes increase the risk of hypertension and cardiovascular diseases, including atherosclerosis and stroke [72]. Within the wall of large elastic arteries such as the aorta, a significant increase in the intimal media's thickness (IM) is observed with age [71,145]. Also, an age-related linear increase in the carotid artery intimamedial thickness has been reported [142]. Moreover, with aging, pulse wave velocity, which is indicative of vascular stiffening, increases in both sexes [72]. Another feature seen with aging is the dysfunction in the vascular endothelium. This change is attributed to the decreased nitric oxide bioavailability due to oxidative stress and inflammation with aging, which leads to dysfunctional endothelium-dependent dilation [123]. Overall, it can be seen that disruption in vasculature occurs with aging, and these changes may contribute to an increased risk of CVD.

Autonomic changes with aging
A number of subtle cardiovascular measures have been reported to change with aging. For example, an assessment of cardiac baroreflex sensitivity revealed that parasympathetic (vagal) sensitivity is reduced while the cardiac sympathetic component is unchanged with aging [88]. This change in parasympathetic baroreflex sensitivity could result from vascular stiffening with aging and reduced beta-adrenergic function [87]. Plasma noradrenaline also increases with aging, with cardiac noradrenaline spillover rising to twice the average concentration level and hepatomesentric spillover rate increasing by around 50% [115]. However, the renal spillover rate remains unchanged with aging, suggesting regional variability in the noradrenaline spillover rate [26]. It is suggested that the changes occur due to increased sympathetic outflow from the central nervous system and reduced noradrenaline clearance with advancing age [137]. Approximately a 10-15% increase in noradrenaline concentration is found in the postganglionic sympathetic nerve terminals with age [36]. Higher plasma noradrenaline in older adults is linked with decreased autonomic support of arterial blood pressure [60]. It has been shown that ganglionic blockade induced by trimethaphan results in a more significant decrease in mean blood pressure in older subjects compared to their younger counterparts [60]. Moreover, the baseline plasma noradrenaline level (higher in older adults) was significantly related to the decreased mean blood pressure caused by trimethaphan [60], indicating altered autonomic support of blood pressure regulation with aging.
Increased overall MSNA levels with advancing age have been observed in many research groups [25,61,136,153]. An increasing level of MSNA is observed in normotensive, healthy individuals with aging, indicating that increasing MSNA with age results from alterations in physiological processes rather than other complications that may occur with aging, such as hypertension and atrial fibrillation [21,97]. Recent evidence suggests the MSNA increase with age follows a sigmoidal pattern, particularly in females; they reach the lowest point at 30 years before rapidly increasing until menopause, after which the rate of increase becomes similar in both sexes [61]. The age-related increase in MSNA does not evenly affect the sympathetic nervous system, with regional variation in the sympathetic outflow -including increases in a sympathetic drive to muscle vasculature and gut, but no change in sympathetic supply to the kidney [122].
A possible mechanism for elevated MSNA with aging could be altered baroreflex, which features increased central drive but decreased peripheral transduction of sympathetic activity to blood pressure. Sympathetic transduction, the measure of the pressor response following a spontaneous MSNA burst, was reduced with aging, particularly in postmenopausal women [144]. Recently, D'Souza and colleagues used a wavelet-based approach to reveal increased firing probability of medium-sized action potential clusters in the older age group compared to younger adults. As well as this, they reported an upper-right shift in the MSNA baroreflex threshold, which indicates a higher set-point for diastolic blood pressure and MSNA burst incidence in older adults [156]. Moreover, the authors suggested these changes in the central baroreflex arc may reflect a compensatory mechanism to rectify decreased sympathetic transduction that occurs with aging [156]. While the correlation between the changes in the central baroreflex arc and the peripheral sympathetic transduction remains to be investigated, one hypothesis that may be derived from this finding is that structural and functional correlates of the autonomic nervous system in the brain may be affected. Hence, future studies should address whether changes in the central arc of the baroreflex seen from D'Souza et al. [156] may associate with structural and functional changes in the brain.
Moreover, studies suggest that increased MSNA with aging is associated with an increasing level of leptin and abdominal body fat that often occurs with age [89]. Also, it was found that both norepinephrine and insulin levels are positively associated with systolic and diastolic blood pressure in the aged cohort, which may reflect the interactive effect of the sympathetic nervous system and insulin on elevated blood pressure with aging [147]. Indeed, hyperinsulinemia in rats significantly increased lumbar sympathetic nerve activity, and this activation was achieved via a melanocortin-dependent pathway to the hypothalamic paraventricular nucleus [148]. Therefore, it is possible that increased MSNA is associated with altered structural and functional changes in the brain areas that mediate sympathetic outflow. Given this, the cortical and subcortical structures involved in MSNA generation that have been documented in young, healthy individuals [78,122] are most likely to be associated as the primary driver of increased MSNA in the elderly. However, currently, no study links direct measurement of the MSNA to functional or structural changes in the brain with aging.

Changes in the brain with aging
Significant structural changes in the brain occur with aging [43], with both grey and white matter affected [38,110]. Grey matter volume in most brain regions displays linear declines, while white matter volume is altered in a U-shaped pattern, where the volume reduces with age but, after reaching a certain point, increases again in the late stage of life [35,37]. The reductions in grey matter volumes are thought to result from a decrease in synapses rather than the decline of neuronal density, as experimental animal studies report no change in neural numbers with increasing age [91,107,117].
While grey matter volumes in frontal, parietal, and occipital regions show a linear reduction with age [1], the prefrontal cortex appears to be most significantly affected by aging [138]. It is believed that there is a posterior to anterior shift with aging, i.e., in younger adults, posterior regions are more active, but with increasing age, anterior regions are recruited more prominently [19]. Other significant structural changes occur in the hippocampus [30,77], with the reduction in the hippocampus grey matter volume accelerated with aging compared to other surrounding areas [30,81]. While the hippocampus is most often associated with memory functions, numerous studies indicate it is involved in cardiovascular regulation, including control over the hypothalamic neuroendocrine system [28, 93,127,135].
Grey matter volume in various functional networks is also altered with aging [43,77]. For example, the medial visual cortical network, sensorimotor network, default mode network, and executive control network all show a decline in grey matter volume with aging [43,77]. Multivariate independent components analysis of magnetic resonance imaging (MRI) data has identified 15 networks that followed the linear reduction of grey matter volume with age and one network with U-shaped variation [77]. One identified network by Liu et al. [77] includes the default mode network, which comprises the precuneus and posterior cingulate cortex. Given that these regions are associated with the generation of spontaneous MSNA in young, healthy individuals [56], it may be hypothesized that activity in these brain regions, and hence control of sympathetic outflow, is altered with aging.
In addition to changes in neural numbers with age, there is also evidence that non-neural cells in the brain change.
Astrocytes are glial cells that contribute to the cellular homeostasis of the brain, and there is growing evidence that astrocytes' phenotypic and genetic expression changes significantly with aging. Astrocytes show enlargement of their cell bodies and processes with age, which resembles the changes seen following a lesion in the brain and subsequent reactive astrogliosis [116]. Furthermore, the genetic expression of astrocytes shifts to an inflammatory state with aging, secreting pro-inflammatory factors such as CXCL10 [16]. Aging astrocytes upregulate the expression of phagocytic markers involved in synapse elimination, such as MEGF10 and Mertk [15,16]. These astrocytic changes may impact the activity of the sympathetic nerves, as astrocytes are closely associated with RVLM neurones that innervate sympathetic preganglionic neurones located in the intermediolateral cell column in the thoracic and upper lumber spinal cord [42]. Indeed, Marina and colleagues (2013) showed that optogenetic activation of astrocytes in the RVLM leads to increased heart rate, arterial pressure, and renal sympathetic nerve activity [83]. Overall, it can be argued that astrocytes become dysfunctional with aging, affecting their capacity to maintain normal functioning of neurons within the central nervous system. Importantly, alterations in aging astrocytes that lead to changes in their genetic expression are observed in brain regions important for regulating resting sympathetic nerve activity, such as the dorsolateral prefrontal cortex and hypothalamus [11]. MSNA-coupled fMRI studies have revealed that the dorsolateral prefrontal cortex and the hypothalamus, specifically the dorsomedial and ventromedial hypothalamic regions, contribute to regulating resting MSNA in healthy young adults [79]. However, whether the functional changes in the dorsolateral prefrontal cortex and hypothalamus contribute to increased MSNA observed in older adults remains to be investigated.
Given that age-related changes in heart structure may affect blood flow to important organs such as the brain, it is essential to review whether there are associations between changes in the heart with functional changes in the brain. Indeed, there is an association between LV structural changes and cognitive decline with aging, with Razavi and colleagues reporting that increased LV mass index is associated with decreased cognitive function [109]. In addition, LV mass has also been reported to be associated with higherorder brain region structure changes, with LV increases correlated with microstructural changes in the medial prefrontal and orbitofrontal cortices [90]. Furthermore, in older individuals with mild cognitive impairment, increased LV mass is associated with altered white matter microstructure [90]. Consistent with these findings, Haring and colleagues reported a correlation between LV mass and cognitive performance (0.63 per 25 g difference LV mass; 95% CI, − 1.124 to − 0.14; P = 0.012) and an association between increased LV mass with lower hippocampal volume (0.009% per 25 g difference LV mass; 95% CI, − 0.014 to − 0.003%; P = 0.003) [44]. This finding is significant since the hippocampus has been implicated with autonomic modulation assessed with the HRV-fMRI approach [93] and MSNA-fMRI method [28]. Moreover, we know that spontaneous bursts of MSNA are coupled with activity in the dorsolateral prefrontal cortex [56], a region involved in executive function, and we recently showed that sinusoidal electrical stimulation of this region causes a cyclic modulation of MSNA, blood pressure, and heart rate [124]. Given that altered sympathetic nervous system activity is associated with LV mass (r = 0.86, P < 0.0001; r = 0.50; P < 0.01) [12, 120], cognitive function (time × group interaction; P < 0.05) [39], and regional grey matter volume (RVLM; r = 0.33, P = 0.01) [68], it is possible that the sympathetic nervous system plays a role in the relationship between LV mass and cognitive performance, however, further investigations are required to understand this relationship fully.

The burden of hypertension
According to the guidelines from American Heart Association, stage 1 hypertension is defined as systolic blood pressure equal to 130-139 mmHg or diastolic blood pressure equal to 80-89 mmHg, and stage 2 hypertension as systolic blood pressure equal to or higher than 140 mmHg or diastolic blood pressure equal to or higher than 90 mmHg [151]. Most hypertension is caused by genetic and environmental factors and is referred to as essential or primary hypertension [102]. Secondary hypertension is caused by other conditions, such as primary aldosteronism, renal artery stenosis, or pregnancy-induced hypertension. The prevalence of hypertension has reached more than 30% across the globe for over a decade [86], with a higher prevalence in low and middle-income countries than in high-income countries [154]. These prevalence differences are thought to result from differences in access to health care, pharmacological treatment, and lifestyle factors [86]. Hypertension is still one of the most prevalent risk factors for CVD, and thus investigations into the pathophysiology of hypertension and the potential for the development of novel effective treatments will have a significant impact on health worldwide.

Cardiovascular changes with hypertension
In addition to the increase in blood pressure and organ perfusion pressure, hypertension can lead to various cardiovascular changes, including LV hypertrophy and systolic, diastolic, and endothelial dysfunction. It is believed that elevated blood pressure can result in compensatory thickening of the LV wall to alleviate high wall stress [22,23]. The underlying structural changes include hypertrophy of cardiomyocytes and vascular smooth muscle cells. Also, the conversion of fibroblasts to myofibroblasts is seen in some cases, resulting in arrhythmias of the heart [23]. Moreover, these hypertrophic changes could reduce diastolic filling and ejection fraction that underlies other cardiovascular diseases such as heart failure [22,23].
A critical component of the homeostasis of vascular tone is nitric oxide (NO). Endothelial cells release NO in response to mechanical shear induced by blood flow in the vascular wall. NO acts via guanylate cyclase, which leads to the generation of cyclic GMP, resulting in the relaxation of vascular smooth muscle cells [131]. In experimental models in animals and humans, restriction of NO production by inhibiting NO synthase (eNOS) resulted in progression into a hypertensive state [5]. Furthermore, decreased NO production was observed in patients with hypertension, indicating the importance of endothelial dysfunction in hypertension [5]. Cardiovascular changes with hypertension pose a significant risk for disease progression.

Autonomic changes with hypertension
Numerous investigations have demonstrated other pathophysiological determinants of hypertension, such as the influence of the autonomic nervous system. A meta-analysis of 1216 patients aggregated from 63 studies found a significant elevation in MSNA in prehypertensive, borderline essential hypertensive and established essential hypertensive patients [41]. Interestingly, an increase in MSNA was observed in both untreated and those taking antihypertensive medications. These results show that increased sympathetic outflow is consistently observed in all types of essential hypertension regardless of the severity of hypertension and treatment status [41]. In addition, it has been demonstrated that carotid sinus stimulation in patients with resistant arterial hypertension can decrease blood pressure and MSNA [49]. Furthermore, the changes in diastolic blood pressure due to the stimulation were correlated to the changes in MSNA [49], indicating that MSNA contributes to the increased blood pressure in hypertension.
Increases in MSNA can also be observed in secondary hypertension. Johansson and colleagues compared MSNA and noradrenaline spillover in individuals with renovascular hypertension. In both of these measurements, hypertensives showed an increase in sympathetic activity compared to the controls [58]. Increased MSNA was also observed in pregnancy-induced hypertension. A study on healthy women without a history of hypertension found that MSNA levels increased significantly from pre-pregnancy to early pregnancy [57] and continued to increase until late pregnancy before dropping back to the pre-pregnancy level in the post-pregnancy phase [50]. Moreover, it has been shown that patients with coarctation of the aorta who develop hypertension have significantly increased MSNA compared to patients with normal BP [74]. Overall, these results demonstrate a strong relationship between MSNA and hypertension, showing increased activity in various types of hypertension.
A recently proposed mechanism contributing to increased sympathetic outflow in hypertension is via immune system activation. It has been argued that increased sympathetic nerve activity and the resultant elevation in blood pressure lead to tissue stress and increased reactive oxidant species (ROS) [85]. Increased tissue stress and ROS can generate new antigens and damage-activated molecular patterns. Innate immune cells such as monocytes, macrophages, and natural killer cells can recognize damage-activated molecular patterns via Toll-like receptors (TLRs) and become activated [85]. In addition, activated innate immune cells and increased sympathetic outflow may activate adaptive immune cells, such as T effector lymphocytes, producing pro-inflammatory molecules that increase blood pressure and end-organ damage [53]. Also, pro-inflammatory cytokines such as interferon 1β and tumor necrosis factor-alpha (TNFα) can induce sympathetic activation [62]. Therefore, it is possible that, in hypertensive patients, the interaction of the sympathetic nervous system and the immune system may create a positive feedback loop that contributes to the progression of the disease. Future studies will be required to investigate whether these interactions between the autonomic and immune systems bring about changes in the functional activation of brain regions that modulate MSNA.
Similar to aging, one potential mechanism of increased MSNA in hypertension is the altered baroreflex mechanism. It has been shown that hypertensives have lower sympathetic transduction than normotensives [67]. As seen in aging, decreased peripheral sympathetic transduction may lead to compensatory increases in the baroreflex's central arc. Indeed, in women with gestational hypertension, an increased frequency of action potential firing was observed [6], representing a potential alteration in the central sympathetic outflow from the brain. However, the association with the changes in the structural and functional changes in the brain and altered central baroreflex arc remains to be investigated by further research.

Changes in the brain with hypertension
Significant functional and structural changes in the brain occur with hypertension. Early studies investigating the association between the changes in the brain with hypertension focused on cognitive performance measures. For example, Harrington and his colleagues found that older people with hypertension showed slower reaction time, immediate recognition of words, spatial memory, and number vigilance [45]. There is also evidence suggesting that significant structural changes occur in brain white matter in hypertensives. A longitudinal study comparing the risk of brain structural changes between normotensive and hypertensive individuals found a significantly increased risk of developing white matter hyperintensities, indicating brain structural alterations, in hypertensives after a 4-year follow-up [24].
More recently, studies have attempted to uncover the underlying changes in brain circuitry responsible for altered cognition in hypertension, utilizing functional and structural MRI techniques. Naumczyk and colleagues [96] found that hypertensives show significantly increased functional activation in frontal, parietal, occipital, and limbic cortical regions. They also observed significant clusters of activation in the posterior cingulate cortex and precuneus, which-as noted above-are areas that we identified to be associated with the generation of spontaneous MSNA at rest in healthy young individuals [56,96]. Surprisingly, this study did not find white matter differences in hypertensives [96]. This finding may be because the mean age of recruited participants in the study was relatively young (~ 40 years) compared to other studies that reported differences in structural changes in the hypertensive brain. This suggests that structural changes seen in other studies could be confounded by other risk factors, such as older age or disease duration. Therefore, the contribution of hypertension to the structural changes in the brain should be carefully delineated from those changes associated with aging or disease duration. Nevertheless, this previous study highlights significant functional alterations in the brain with hypertension.
A recent paper published by Carnevale and colleagues supports this idea of altered neural function in the brain of hypertensives. Using resting-state fMRI, Carnevale [13] found increased connectivity in the dorsal attention network with occipital regions and decreased connectivity with the lateral prefrontal cortex. Furthermore, the precuneus and anterior cingulate cortex, which are associated with the generation of MSNA [80], had altered functional connectivity. A more recent study showed the association between white matter lesions and functional connectivity in the neural networks of hypertensive patients [125]. In non-diabetic hypertensives, decreased connectivity in the default mode network, which includes medial prefrontal, posterior cingulate, and precuneus cortices, was negatively correlated with white matter lesion volumes. In addition, a negative correlation between functional connectivity and white matter lesion volume was found in the insular and cingulate cortices, areas that modulate MSNA in the resting state in healthy young adults [56]. One hypothesis generated from these findings is that functional and structural changes in the hypertensive could drive altered sympathetic outflow, resulting in increased BP. Alternatively, altered cardiovascular changes such as high BP due to other factors in hypertension may be causing changes in sympathetic nerve activity that consequently changes the function and structure of the brain areas responsible for generating sympathetic nerve activity.

The burden of atrial fibrillation
Atrial fibrillation is a type of cardiac arrhythmia characterized by heart palpitations. Around the globe, approximately a 30% increase in incidence has been reported compared to 20 years ago [76]. As a result, it is becoming one of the significant health burdens worldwide. Furthermore, the hospitalization rates show a 295% increase over the 20 years from 1993 to 2013, compared to only 73 and 39% with myocardial infarction and heart failure, respectively [33]. These data highlight the need for further investigation into the pathophysiology of AF.

Cardiovascular changes with AF
AF is characterized by high-frequency excitation of the atria, which results in dyssynchronous atrial contraction and irregular patterns of ventricular excitation [132]. It is considered that AF is a final endpoint of atrial remodeling that occurs due to various cardiovascular diseases. In addition, AF itself can cause remodeling that contributes to the maintenance of AF [95]. Two main pathophysiological mechanisms proposed are focal ectopic firing and the reentry of electrical conduction [55,95,132]. Ectopic firing occurs from the myocyte sleeves in the pulmonary veins that propagate into the left atrium, and it is suggested to initiate AF [132]. Re-entry occurs via vulnerable atrial substrates such as abnormal atrial cardiomyocytes and fibrotic changes within the heart. The re-entry of electrical signals is known to contribute to the persistence of AF [146].
Numerous factors can influence these two pathological mechanisms in AF. Electrical remodeling can occur, which creates a re-entry-prone substrate. Alterations in L-type Ca2 + current, rectifier background K + current, and expression of gap junction connexin have been observed, contributing to re-entry. Also, structural remodeling such as atrial enlargement and fibrosis can occur, contributing to the persistence of AF re-entry and can be arrhythmogenic for cardiomyocytes. Moreover, changes in the autonomic nervous system can contribute to the initiation and maintenance of AF. Beta-adrenergic receptor activation can alter L-type Ca2 + current, RyR2 open probability, and SR CA2 + leak through CaMKII and protein kinase phosphorylation. These changes can increase the risk of delayed afterdepolarizations and the development of ectopic firing [95].

Autonomic changes with AF
Altered sympathetic nerve activity in AF patients is well documented. Earlier experimental work using rapid atrial pacing to induce AF saw a significant increase in MSNA during rapid atrial pacing [150]. In addition, Ikeda and his colleagues have found augmented single-unit recordings of MSNA in heart failure patients with chronic AF but not with multi-unit recordings of MSNA [54]. The limitation of this study is that it did not assess AF patients without other comorbidities, such as heart failure, to avoid confounding factors. However, the study supports the notion that having chronic AF is associated with altered sympathetic nerve activity even in pre-existing cardiovascular conditions such as heart failure.
Altered autonomic activity in AF is also evident from catheter ablation studies. Cui and colleagues found altered MSNA during radiofrequency catheter ablation in areas close to pulmonary veins [17]. They reported a decreased MSNA during the procedure and increased cardiac parasympathetic tone, as measured by heart rate variability. This finding demonstrates the augmented sympathetic nerve activity in individuals with AF and the potential for reversibility of altered sympathetic outflow with effective treatment such as ablation. It is believed that radiofrequency catheter ablation influenced parasympathetic afferents, which decreased MSNA during the procedure. More recently, Mukai and colleagues observed that after 12 weeks of catheter ablation in AF patients, MSNA decreased significantly (P < 0.01) compared to baseline levels obtained prior to the procedure [92]. Moreover, it was found that percent changes in MSNA were significantly associated with left atrial volume [92], highlighting the potential contribution of altered autonomic nerve activity on cardiac modification, such as left atrial volume in AF patients. Overall, these findings indicate chronically altered autonomic activity in AF patients and further studies are needed to reveal the central mechanism underlying these disturbances.

Changes in the brain with AF
Several reports indicate a connection between AF and cognitive decline, with most studies showing an accelerated cognitive decline in patients with AF. The Framingham Heart Study investigated the association between AF and cognitive performance and found that AF is associated with poor attention and a fall in executive function, particularly in men [98]. There are also reports of an association between AF and cognition, with AF associated with an increased risk of developing mild cognitive impairment and dementia [2,40]. The odds are increased with infarctions in the brain [40]. Of 5150 participants followed up over 7 years, approximately 10.7% developed AF, and cognitive performance (measured using a Modified Mini-Mental State examination) declined faster in individuals who developed AF [139]. These findings are further supported by a meta-analysis of 77,668 patients that examined the association between AF and dementia [118]. The study's mean follow-up time was 7.7 years and showed that 15% (11,700 participants) had AF, and 6.5% (4773 participants) of the participants developed dementia. In pooled adjusted hazard ratio comparing incident dementia in those with AF and without AF, the hazard ratio was 1.42 (P < 0.001), indicating a higher risk of developing dementia with AF. Overall, evidence suggests an association between AF and cognitive decline. However, there has been a lack of research investigating the changes in functional activation in AF patients without dementia or mild cognitive impairment [129].
These impairments may result from changes in brain function, and a recent pioneering study investigated the association between AF and functional brain changes [128]. They found decreased connectivity in the default mode network in AF patients but no change in executive, visuospatial, and salience networks. Interestingly, default mode network connectivity was also reduced in the hypertensives, which shows similarities between these two disease states. However, the functional activation of other networks, such as the salience network, differed in the two groups, as hypertensives showed a reduced activation, and no change was detected in the AF patients. These alterations could reflect the disease-specific changes in functional activation and supports the need for a separate investigation of functional connectivity in different disease states.
As mentioned above, the default mode network consists of the precuneus, medial prefrontal, and posterior cingulate cortices, brain areas shown to be involved in the regulation of MSNA [56]. However, no studies have investigated the functional changes in brain regions modulating the spontaneous generation of MSNA in AF. Given that the autonomic nervous system plays a significant role in cardiovascular changes seen in AF, future studies using the MSNA-coupled fMRI approach will be beneficial for revealing the underlying mechanism of cognitive decline and cardiovascular changes in AF.
Structural abnormalities are also reported in AF patients. Earlier studies looking at the association between AF and structural changes in the brain reported changes in the brain volume in AF. A cross-sectional MRI study in stroke-free AF patients reported a significant reduction in hippocampus volume [65]. The study also measured these individuals' cognitive capacity and showed decreased attention, executive function, and memory-related tasks in people with AF. Given that the hippocampus has a significant role in memory and executive function, this study demonstrates a possible contribution of structural abnormalities to the underlying altered cognitive functions in AF patients [111]. Similar reports on the consequences of AF in the brain were reported in the Framingham Heart Study Offspring cohort. They found a significant reduction in total brain volume and specific brain regions, including frontal areas [106]. In contrast to the previous study, they did not report any changes in hippocampal volume. Since the Framingham Heart Study was a longitudinal study that took the average of multiple measurements of MRI on AF patients, the result may reflect the possibility that alterations in the hippocampal volume could be under continuous remodeling or are variable in different participants. Nonetheless, with all results combined, the presence of AF in patients could result in volumetric changes in the brain.
AF can result in other structural abnormalities in the brain. In a longitudinal analysis of 963 participants, Berman and his colleagues observed that 3% developed AF. Those with AF had higher odds of developing cerebral infarction and worsened sulcal and ventricular grade in the brain [10]. It is suggested that the presence of cerebral infarctions may lead to the development of white matter hyperintensity lesions [149]. However, there are mixed reports regarding the association of white matter hyperintensity with AF. In a retrospective study of patients presenting with embolic stroke, AF was associated with the presence of white matter hyperintensity patches in anterior subcortical regions [84]. However, early volumetric studies consistently report no changes in white matter hyperintensity lesions due to AF [106,133]. One possible explanation for this discrepancy could be that Mayasi's study on patients with embolic stroke is likely to be confounded by increased infarctions due to stroke. This explanation could further support the association between the development of white matter hyperintensity and cerebral infarction. However, more studies on the functional and structural changes in patients with AF will be needed to validate the discrepancies in the association between AF and white matter hyperintensity.

Conclusions
Aging and risk conditions for cardiovascular disease associated with aging-hypertension and atrial fibrillation-are independently associated with cardiovascular, autonomic, and structural and functional changes in the brain. While cardiovascular changes in each condition differ, increased MSNA is commonly observed in all these states. Significant changes in the brain occur in areas known to mediate sympathetic outflow, particularly in areas associated with the default mode network, which comprises the medial prefrontal cortex, posterior cingulate cortex, and precuneus. Hence, the structural and functional change in the specific brain areas may associate with the changes in MSNA.
Furthermore, these underlying changes may predispose risk groups towards developing cardiovascular diseases. Future studies are needed to investigate the neural pathway in which these changes occur by measuring autonomic nerve activity together with structural and functional magnetic resonance imaging [8, 68,69,80,82,126,127] as well as other advanced techniques such as magnetoencephalography (MEG) [112]. By correlating how structural and functional changes in the brain may influence cardiovascular outcomes through the autonomic nervous system, we hope to guide future therapeutics in rectifying dysregulation of autonomic and cardiovascular control by the brain.