Neuropsychological testing is a potent tool for measuring cognitive functioning and can capture even the most subtle cognitive changes to uncover evidence of cognitive dysfunction [14]. Cognitive decline (cognitive functions such as memory, thinking, and reasoning may decline over years to decades to a greater extent than would be expected due to age alone) is assessed by longitudinally comparing the degree of cognitive decline in a patient. It does not require the patient to meet definitive criteria for dementia or mild cognitive impairment (MCI). Studies on cognitive decline have helped to identify the danger of preclinical cognitive impairment in advance, providing a simpler technique of testing cognitive function and allowing for a more accurate assessment of the relationship between hypertension and cognitive performance [14, 15]. However, clinical diagnosis of MCI or dementia is equally crucial, as both diagnoses have far-reaching public health consequences [16]. In the case of hypertension, current evidence from epidemiologic studies has found that hypertension is strongly associated with cognitive decline [14], MCI [17], and dementia [2, 18, 19].

Prospective cohort studies may be preferable to cross-sectional studies when examining the causal relationship between hypertension and AD, so we focused primarily on evidence from longitudinal cohort studies. While the correlation between hypertension and cognitive impairment was first established in patients with stroke and Vascular Cognitive Impairment (VCI), subsequent investigations have determined that cognitive decline attributed to hypertension could also occur in the absence of stroke [20]. Multiple studies have found that there may be an age-dependent relationship between hypertension and the risk of AD, and that the association between hypertension and AD at different ages of onset is complex and variable [21, 22].

Several studies have arrived at more consistent conclusions about the relationship between midlife hypertension and the risk of AD. In a meta-analysis of prospective cohort studies, middle-aged hypertension was found to be more strongly associated with AD [19]. In the Honolulu-Asia Aging Study, untreated hypertension levels in midlife were found to be associated with an increased risk of AD 25 years later, and it was suggested that hypertension may be an important mediator or pathologic contributor to AD [23]. Another study with 21 years of follow-up found that elevated systolic blood pressure (SBP) in midlife was an independent risk factor for AD in later life [24]. Moreover, the Atherosclerosis Risk in Communities (ARIC) study revealed that hypertension and prehypertension among middle-aged individuals confer a comparable risk of dementia, with both groups displaying a higher risk relative to normotensive individuals [14]. A national cohort study in Korea also found an increased risk of AD with SBP $\ge$160 mmHg in midlife [25]. In conclusion, all the above studies emphasize that control of hypertension in midlife may benefit cognition in later life.

As for the effect of late-life hypertension on AD there is currently inconsistency. A few studies have found late-life hypertension to be associated with AD. In the Kungsholmen Project, a community-based cohort of 75-year-olds followed for 6 years found that subjects with an SBP $>$180 mmHg had an elevated risk of AD [26]. In Addition, a 15-year follow-up study in 70-year-olds found that participants who developed dementia at age 79–85 years had higher SBP and diastolic blood pressure (DBP) at age 70 years than those who did not, with higher DBP recorded in patients with AD than those who did not develop AD [27]. However, no relative risk assessment for hypertension-related dementia was reported. No other studies have confirmed a strong association between late-life hypertension and AD. In fact, some current studies in older adults ($\ge$70 years) have shown a U-shaped relationship between blood pressure and cognitive performance. The inconsistent results seen in older adults may reflect differences in the cognitive domains assessed, differences in study design, including the way in which follow-up time and rates of cognitive change were modeled, differences in the characteristics of the study populations and different age ranges, and adjustments for co-determinants that may confound the hypertension-cognition association. Reverse causality may also contribute to the observed association between blood pressure (BP) and cognition, particularly in studies with cross-sectional designs. Cognitive impairment is a process that lasts for decades, thus posing a challenge in determining temporality. More research is needed to elucidate the causal relationship between BP and cognition and to better understand the role of medications in the observed associations [28, 29, 30].

Collectively, the above results highlight the potential detrimental impacts of hypertension acquired during midlife. Nevertheless, due to the absence of a gold standard for stratifying dementia subtypes, the subset of studies investigating the correlation between hypertension and specific dementia subtypes might inadvertently complicate the assessment of hypertension’s role in other forms of dementia and cases with mixed underlying causes. Therefore, more rigorous criteria are needed to aid in the diagnosis of dementia subtypes. Furthermore, age-stratified and long-term follow-up studies focused on hypertension could serve as valuable tools for enhancing our comprehension of the intricate connection between hypertension and AD.

Research suggests that HTN may have a global impact on cognitive performance and in global cognitive assessments, such as the Mini-Mental State Examination (MMSE) and the Montreal Cognitive Assessment (MoCA), people with vascular risk including hypertension score lower. Individuals with more factors also score lower, which is largely influenced by difficulties related to attention (in the MMSE assessment) and visuospatial executive functioning (in those with lower MoCA scores) [31]. Broader neuropsychological assessments suggest that hypertension most significantly affects executive function [16], motor speed, and attention [32, 33]. These cognitive domains are often associated with subcortical disorders such as common vascular diseases or pure vascular dementia (VaD) [34].

The pattern of cognitive impairment in hypertension-associated dementia is complicated by potential overlapping etiologies and the high prevalence of mixed dementia (MD). Although memory impairment is considered a typical feature of AD compared with VaD [35, 36], it has been found that situational memory is significantly affected in cognitively impaired hypertensive participants and is associated with elevated A$\beta$ Pittsburgh Compound-B positron emission tomography (PiB-PET) binding. This suggests that the pathological mechanisms of AD may play an equally prominent role in hypertensive patients. Conversely, patients with MD (i.e., AD coexisting with VaD) exhibit worse global scores as well as worse attention and visuospatial abilities [37].

Studies focusing only on specific cognitive domains are therefore prone to errors in findings due to misdiagnosis or the effect of difficulties in determining the exact cause throughout an individual’s lifetime. Further incorporation of biomarkers (including serology, cerebrospinal fluid, and imaging) can help to better characterize the disease [38].

The potential mechanisms linking hypertension and AD are complex and multifaceted. Current evidence suggests that a synergistic interplay among various pathological factors may underlie the cognitive impairments associated with hypertension and there are extensive detailed reviews on this topic [39]. Below, we provide a summary of the key points and an overview of the current evidence using neuroimaging markers to elucidate the connection between hypertension and AD (Fig. 1) [40].

Fig. 1.

Common imaging features and potential mechanistic links between hypertension and Alzheimer’s disease (AD). Hypertension exerts a multifaceted influence on the brain, manifesting through alterations in the structure and function of the cerebrovasculature, perturbations in the renin-angiotensin system, the emergence of inflammation, and the induction of oxidative stress. Such pathological processes can lead to the development of cerebral white matter lesions, brain atrophy, microhemorrhages, and the accumulation of pathologic proteins associated with AD. These conditions are discernible through neuroimaging technologies and are recognized for their potential to affect the initiation or advancement of AD. BBB, blood-brain barrier; A$\beta$, amyloid beta; APP, amyloid precursor protein.

The relationship between hypertension and brain structure and functions is not completely understood and current evidence suggests that several pathological factors can interact [41]. Hypertension can produce vascular and functional changes in large and small cerebral vessels in the brain, and small cerebral vessels and arterioles are more vulnerable to the mechanical stress associated with hypertension [42]. It is relevant to highlight the strong relationship between arterial hypertension and cerebral small vessel disease (cSVD); the characteristic findings of cSVD in magnetic resonance imaging (MRI) include lacunar infarcts, white matter hyperintensities (WMH), cerebral microbleeds (CMB), enlarged perivascular spaces, and brain atrophy [42]. Conversely, chronic hypertension induces a rightward shift in the autoregulatory curve and, consequently, an increment of vulnerability to sudden changes in blood pressure, resulting in ischemia or increased risk of brain hemorrhage.

The modified structure of the perivascular space, often observed in conditions such as hypertension, heightens the possibility that the perivascular and paravascular clearance systems, which are vital for the removal of waste products from the brain, may be disrupted, potentially contributing to white matter injury [43, 44]. Specifically, in hypertension, the perivascular spaces become enlarged and distorted [43, 45], a change that could hinder the effective elimination of neurotoxic metabolic byproducts. This is significant as hypertension is also linked to the accumulation of A$\beta$ and tau proteins, indicating a possible reduction in the clearance of these proteins, which are associated with AD pathogenesis. Furthermore, the atherosclerosis of large arteries due to hypertension leads to increased pulsatility in the microvasculature. This increased pulsatility is hypothesized to affect the perivascular space, potentially reducing the clearance of interstitial fluid (ISF) and cerebrospinal fluid (CSF; a colorless, clear fluid found in the ventricles and subarachnoid space of the brain.) [43]. The perivascular space is a critical component of the glymphatic system, which is responsible for the drainage of metabolic waste from the brain. Disruptions to this system can result in the buildup of harmful substances, such as A$\beta$ and tau, which are known to negatively impact brain health.

In conclusion, vascular damage caused by hypertension results in brain dysfunction due to hypoxia-ischemia and increased production of A$\beta$ due to the increased processing of amyloid precursor protein (APP) by secreted enzymes and decreased clearance of AD-related proteins. In turn, A$\beta$ and tau proteins alter vascular function and amplify the deleterious effects of hypertension on the vasculature.

Hypertension impairs the structural and functional integrity of the cerebral microcirculation, promoting microvascular thinning, cerebral microvascular endothelial dysfunction, and neurovascular uncoupling, as well as impairing cerebral blood supply [39]. This insufficient blood supply, in turn, triggers metabolic irregularities within the brain parenchyma. Consequently, there is an increased activity of beta-secretase 1, an enzyme responsible for cleaving APP into beta-amyloid, thus promoting the formation of amyloid plaques [46]. HTN also increases the pressure in the lumen of cerebral vessels, leading to increased formation of reactive oxygen species and weakening of the vessel wall, resulting in disruption of the blood-brain barrier (BBB). This allows the contents of the plasma to enter the central nervous system (CNS) thereby triggering neuroinflammation. Mouse model studies have effectively demonstrated that neuroinflammation is a pivotal link connecting hypertension and AD. Additionally, BBB disruption impedes the efficient clearance of beta-amyloid, which further promotes its deposition. Taken together, these series of events significantly contribute to both the onset and progression of AD [20].

While the heightened risk of AD associated with hypertensive patients is evident and the impact of hypertension on cognition is relatively well-established, a degree of uncertainty remains regarding whether HTN directly influences AD neuropathology or simply serves as a concurrent factor contributing to cognitive impairment and dementia [47].

An autopsy study reported that midlife hypertension was associated with more senile plaques and NFT pathology at death, showing that the apolipoprotein E (APOE) $\epsilon$4 allele (especially $\epsilon$4 homozygosity) was associated with hypertension through the elevation of CSF tau and phosphorylated tau levels. However, in that study, hypertension did not interact similarly with the effect of carrying the $\epsilon$4 allele on CSF A$\beta$ levels [48]. Other studies have reported that late-life hypertension (BP) is associated only with NFT, if not with both NFT and senile plaques [49, 50]. In vivo AD biomarker studies based on PET imaging or CSF analysis have produced similarly contradictory results. Although several studies have shown a significant correlation between hypertension and A$\beta$ deposition in the brains of older adults [51, 52], many other studies have not found such a direct relationship [53, 54, 55, 56, 57, 58, 59, 60, 61]. However, some of these findings suggest that current hypertension (rather than a history of hypertension) synergistically modulates the relationship between A$\beta$ and tau deposition in the brain in later life [61, 62].

Several epidemiologic and clinicopathologic studies have reported an association between hypertension and AD. However, a challenge to the interpretation of the relationship between hypertension and AD is the substantial lag in time from the start of hypertension studies to the pathologic diagnosis of AD. The increasing popularity of amyloid PET imaging and structural MRI provides an opportunity to better understand the role of hypertension and brain pathology in vivo. In population-based or special (enriched) cohort studies, large sample sizes are needed to distinguish the separate roles of SBP, DBP, pulse pressure, mean arterial pressure, and carotid-femoral pulse wave velocity (PWV) for each imaging feature. Longitudinal studies that prioritize the assessment of perfusion deficit, white matter changes, cortical volume changes, microbleeds, and amyloid accumulation are needed to understand their interrelationships. Future imaging techniques that can detect the involvement of specific targets will be needed to understand which cells or cellular components are the sites of initial pathogenic injury in hypertension. The cell types and endpoints most suitable for treatment remain undetermined.

Although most of the current research has concluded that a direct association between hypertension and tau is lacking [60], a recent study from China has introduced a novel perspective. The study reported a noteworthy link between a history of hypertension and elevated systolic blood pressure with tau levels in the CSF, specifically in individuals under the age of 65 years [53]. Furthermore, other studies have suggested that hypertension may influence tau levels under certain conditions. For instance, individuals in the hypertension group who are carriers of the APOE $\epsilon$4 allele demonstrated a pathological correlation with increased levels of mid-tau (including total tau and phosphorylated tau) [55]. Additionally, a significant correlation between tau concentration and the hypertensive group was observed following a decrease in mean pulse pressure [63]. Moreover, a recent study found that although hypertension had no direct relationship with A$\beta$ and tau deposition, it could synergistically modulate the interaction between cerebral A$\beta$ load and tau deposition in later stages of life [60, 61]. These emerging insights suggest that the development of strategies for blood pressure control in mid- to later life could provide benefits for both the prevention and treatment of AD.

Hydrogen sulfide (H${}_2$S) is an important gaseous transmitter that regulates neuronal and vascular homeostasis [64]. It has been shown that the cerebral microvasculature produces H${}_2$S and, in the vascular system, it has a wide range of inflammatory and immunomodulatory functions [65]. Sulfides protect against vascular disease [66, 67], and free sulfides promote cerebral vasodilation [68], protect against systemic and focal cerebral ischemia injury [69, 70], and contribute to BBB integrity [71, 72]. Lower sulfide levels predict cardiovascular disease risk [72]. In the nervous system, free sulfide acts as a calcium-based neuromodulator [72]. In addition, H${}_2$S preserves cognitive performance in experimental dementia models, possibly by improving antioxidant status, attenuating oxidative stress [73, 74], and inflammation in vivo and in vitro [74]. In AD-related dementia (ADRD), recent studies have found that increased plasma total sulfide toxicity is associated with elevated acidic soluble sulfide and bound sulfide [75]. Cognitive and microvascular disease indicators correlated with H${}_2$S levels. Total plasma sulfide is the strongest indicator of ADRD [75]. It plays a partial role in the relationship between cognitive dysfunction and the volume of cerebral white matter lesions, the latter being an indicator of microvascular disease. Thus, plasma H${}_2$S and sulfide may be potential biomarkers for the link between hypertension and AD [76]. Carrying out further large prospective experiments to explore the causal role of sulfide in microvascular disease and ADRD could help further explore the mechanism of action of AD and HTN.

The recent availability of imaging and biochemical biomarkers of AD has provided the opportunity to investigate the relationship between AD pathology and hypertension [77, 78, 79, 80]. From the beginning of the 20th century to the present, functional MRI (fMRI) has increased in popularity [81], as it has been shown that brain functional connectivity has been altered in hypertensive patients who show no signs of macroscopic damage on structural MRI [82, 83]. At the same time, fMRI signals are generated by neurovascular coupling, making it sensitive to the effects of hypertension and more favorable for observing the effects of hypertension on brain function.

Several studies have highlighted the structural brain alterations that are characteristic of AD, which include diminished hippocampal volume, reduced cortical thickness, and disruptions in the microstructure of white matter fiber tracts [2, 82]. These suggest the presence of alterations in functional brain activity as well as in brain network connectivity, particularly within the default network (DMN) [84, 85]. Shah et al. [86] previously observed distinctive cerebral white matter lesions in hypertensive patients. They further uncovered a significant correlation between cerebral white matter load and DMN regions by studying the relationship between cerebral white matter load and functional cerebral networks in hypertensive patients, revealing that greater white matter lesion volumes correlated with lower functional connectivity (reflects the functional relationship between brain regions, FC) in the DMN [87]. Moreover, this was accompanied by poorer performance on cognitive tests, suggesting that WMH could be a potential mechanism through which hypertension impacts the progression of cognitive decline in AD.

Another study delved into the interaction between hypertension and individuals carrying the APOE $\epsilon$4 allele. The study demonstrated that the presence of this allele exacerbated brain function impairment, particularly in the frontal and parietal regions of hypertensive patients. While fewer studies directly examined the effect of hypertension on brain function, those that did found disruptions in functional connectivity within the frontal-parietal network, dorsal attention network (DAN), and DMN networks in hypertensive patients. These findings align with the observed decline in executive function and attention in individuals with hypertension [88].

Collectively, these studies underscore the structural and functional transformations that hypertension induces in the brain, although there is currently a lack of convincing large cohort prospective studies exploring the causal relationship between hypertension and cognitive function.

Analyses of structural brain studies conducted on individuals with both AD and hypertension have highlighted distinct patterns of atrophy when compared with AD patients without hypertension. Notably, in the hypertensive group, significant atrophy was observed primarily in the temporal lobe, frontal lobe, and cingulate gyrus [88, 89, 90]. These findings imply that hypertension might play a role in the initiation and progression of AD by influencing critical brain regions. While research exploring brain function in individuals with combined AD and hypertension is more limited and currently exhibits some variation due to diverse analytical approaches, it remains crucial to acknowledge the impact of hypertension on resting-state brain function in AD patients, especially the disruption of the DMN region [82, 84]. However, some literature has reported enhanced the posterior cingulate cortex (PCC) functional connectivity in AD/MCI patients, with areas such as the ventral medial prefrontal, bilateral dorsolateral frontal, and left middle cingulate gyrus. Moreover, heightened connectivity has been noted between the left inferior parietal lobule and the PCC in AD patients with hypertension [91, 92]. These findings are indicative of a potential compensatory mechanism, suggesting that these regions might enhance their function in response to early-stage AD/MCI-associated damage to other brain areas. However, other studies did not identify this compensatory phenomenon. Variations in sample sizes and analytical methodologies are factors that might contribute to these discrepancies. This area of research currently features substantial gaps, and further studies are essential to provide deeper insights and more conclusive findings.

The introduction of multimodal imaging techniques has helped us to better understand the pathogenesis shared in the imaging features of hypertension and AD. Notably, these shared features include WMH, cerebral microhemorrhages, and atrophy. Meanwhile, diffusion tensor imaging is another useful imaging technique developed for assessing the anatomic integrity of white matter [93]. Some parameters of the microstructural abnormalities derived from diffusion tensor imaging (DTI) have been reported to help differentiate between dementias. Some studies have found that compared with conventional MRI, DTI may provide a more objective method for the differential diagnosis of VaD and AD disease patients who have only mild white matter alterations on T2-weighted imaging. Compared with VaD patients, AD patients had lower fractional anisotropy (FA) values in the anterior frontal lobe, temporal lobe, hippocampus, inferior-fronto-occipital fascicles (IFOF), genu of the corpus callosum (GCC), and the cingulate fasciculus (CF); and higher apparent diffusion coefficient (ADC) values in the temporal lobe and hippocampus [94]. Parahippocampal tracts were found to be mainly affected in AD, while VaD showed more spread white matter damages associated with thalamic radiation involvement. The genu of the corpus callosum was predominantly affected in VaD, while the splenium was predominantly affected in AD, revealing the existence of specific patterns of alteration useful in distinguishing between VaD and AD [95]. Based on the above evidence, cerebrovascular dysfunction, including HTN, is associated with the accumulation of AD pathology. The “vascular dysregulation hypothesis” proposes that an imbalance between blood flow-based substrate supply and brain energy demand caused by vascular risk factors such as hypertension exacerbates AD and ADRD [96, 97]. The current view is that cerebrovascular dysfunction occurs early in ADRD and therefore may be a potential diagnostic marker and target for intervention. Cerebrovascular dysfunction induces pathophysiology that may contribute to the common symptoms of AD, including cerebral atrophy, cortical thinning, and white matter deterioration. These can be observed by multimodal magnetic resonance, facilitating exploration of the potential link between hypertension and AD [76].

DTI, fMRI, and PET are three imaging techniques widely used in neuroscientific research and clinical diagnostics, and despite their unique advantages, they have some limitations. DTI is primarily used to depict white matter fiber bundles in the brain, but its ability to image grey matter and subcortical structures is limited and it is not sensitive enough for certain types of white matter lesions. fMRI can non-invasively measure brain activity, but its signal can be affected by a variety of physiological and operational factors, such as vascular saturation, the complexity of the task design, and the level of participant cooperation, which can lead to challenges in data interpretation. PET assesses metabolic and neurochemical activity in the brain using radiotracers, and although it provides detailed information about brain function, it carries radiation risks, is costly, and often requires special radiopharmaceuticals, which limits its use in some situations. The introduction of multimodal imaging techniques has helped us to better understand the pathogenesis shared in the imaging features of hypertension and AD.

The pathology of AD is indeed intricate, encompassing not only the accumulation of A$\beta$ and tau proteins but also involving a multitude of markers related to vascular pathology and neuroinflammation. For instance, in addition to A$\beta$ and tau, other biomarkers such as WMH (de Leeuw F-E et al., 2004 [98]) and neurofilament light (NFL) [99] have gained recognition as contributors to AD pathology. While WMH were previously considered a natural outcome of aging and often associated with cerebrovascular diseases (especially hypertension) [100, 101], recent advancements have highlighted their association with multiple pathological pathways. Recent evidence suggests that WMH could be linked to tau levels, thus linking them to the incidence of AD [102, 103].

NFL, a biomarker associated with significant axonal injury, has gained attention due to its increased levels in the CSF of individuals with AD [91]. Recent studies have facilitated the analysis of NFL not only in the CSF but also in the blood (serum and plasma), offering a more accessible means of monitoring [104]. An intriguing link has been suggested between NFL levels and white matter alterations in AD [87, 105, 106]. A recent study by Walsh et al. [99] found a significant positive correlation between NFL and WMH volume in a population with different cognitive impairments and a large proportion of age-related, independent vascular risk factors (including hypertension). However, further analysis by the same research group found that the combination of hypertension and the APOE $\epsilon$4 allele was associated with increased WMH [98, 107]. The authors further hypothesized that the mechanisms behind the WMH observed in AD patients with comorbid hypertension may be more complex. The study did not find an interaction between hypertension and AD pathophysiologic processes in response to WMH loading because the two have a superimposed effect in response to increased WMH loading, which may lead to earlier and more pronounced cognitive decline [98].

Cerebral microhemorrhages have traditionally been attributed to two main causes: cerebral amyloid vascular degeneration and hypertension [98, 108]. Hypertension-related microhemorrhages are typically found deep within the brain parenchyma, often localized in regions such as the internal capsule. Notably, microhemorrhages in cortical sites are more often associated with the presence of the APOE $\epsilon$4 allele and elevated levels of A$\beta$ [108]. A significant correlation between hypertension and increased microhemorrhagic foci has been observed in patients with AD [109, 110]. However, the interaction between A$\beta$ and AD in the formation of microhemorrhages lacks conclusive evidence and requires further investigation for a more comprehensive understanding.

The report found that patients with vascular or MD presented with a more severe load of white (periventricular and deep) and grey matter lesions compared with AD. The high number of juxtacortical microhemorrhages observed in the group with MD suggests that cerebral amyloid angiopathy (CAA) may be a relevant pathological finding within this group [111]. Microhemorrhages may cause cognitive impairment by affecting executive function, information processing, and memory [112]. However, the mechanisms by which microhemorrhages alter cognitive function remain to be determined, and it is currently believed that the direct cognitive effects of microhemorrhages are related to localized damage or dysfunction of adjacent brain tissue. Alternatively, microhemorrhages may be a more general marker of the severity of small-vessel pathology associated with CAA, modulating altered vascular supply and reduced cortical function [113].

Indeed, neuropathological studies have recognized CAA as a major cause of MD in combination with Alzheimer’s or Lewy body pathology [114, 115]. Furthermore, CAA has been significantly associated with subcortical white matter damage, particularly in APOE $\epsilon$4 carriers [116]. This suggests that CAA can be fully recognized as a prototype of MD, in which AD and cerebrovascular lesions coexist. Therefore, special attention should be paid to microhemorrhages during diagnostic tests for dementia.

Hypertension is currently thought to cause brain atrophy [117, 118], but the exact mechanism remains elusive [118]. Part of the reason is that hypertension-mediated WMH leads to a reduction in brain volume due to white matter thinning. Other potential factors include reduced brain volume due to brain tissue destruction caused by cerebral microinfarcts, cerebral atherosclerosis, and impaired clearance of abnormal proteins in AD patients [119].

A greater number of periventricular lesions, deep white matter lesions, deep grey matter lesions, and enlarged perivascular spaces was observed in vascular dementia compared with AD, while MD showed a significantly greater number of periventricular lesions, deep white matter lesions, deep gray matter lesions, and deep and juxtacortical microhemorrhages. Comparing VaD and MD, VD showed a higher number of perivascular spaces in the basal ganglia and centrum semiovale, while MD showed more deep and juxtacortical microhemorrhages. Gray and white matter lesions predominate in vascular and MD, while deep and juxtacortical microhemorrhages predominate in MD, suggesting that cerebral amyloid angiopathy may be the main underlying pathology in cases of mixed-based cognitive decline.