Relationship between changes in resting-state spontaneous brain activity and cognitive impairment in patients with CADASIL

Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) has been recognized as the most common heritable cause of recurrent stroke and vascular dementia. This disease is caused by pathogenic mutations in the NOTCH3 gene on chromosome 19, which encodes a transmembrane receptor primarily expressed in vascular smooth muscle cells and pericytes of the small penetrating cerebral and leptomeningeal arteries [1]. Typical pathological changes of CADASIL include accumulation of a non-amyloid granular osmiophilic material and the NOTCH3 protein in the vessel wall. The majority of patients with CADASIL have a family history in the absence of vascular disease risk factors such as hypertension, diabetes, coronary heart disease, and smoking. The essential clinical manifestations of CADASIL include migraine with aura, subcortical ischemic events, mood disturbances, and cognitive impairment. It has been reported that 80% of patients with this disorder have cognitive deficits, principally in the domains of executive function, information processing speed, attention, and memory [1, 2].

Structural cerebral magnetic resonance imaging (MRI) in patients with CADASIL reveals age-dependent widespread white matter lesions, seen as large symmetrical hyperintensities on T2-weighted and fluid-attenuated inversion recovery (FLAIR) imaging in periventricular areas and in the centrum semiovale [3]. Furthermore, lacunar infarcts and microbleeds are frequently visible in the basal ganglia and brainstem [4, 5]. Studies employing diffusion tensor imaging (DTI), a quantitative measure of white matter integrity, have revealed extensive brain microstructural changes in patients with CADASIL [6], with different cognitive functions being correlated with white matter structural integrity at different sites [7]. Furthermore, other conventional MRI parameters, such as lacunar lesion burden and brain atrophy, have important impacts on cognitive function and have become essential predictors of cognitive decline in patients with CADASIL [8‐13]. One study indicated that reaction time on a simple task may serve as a marker of early cognitive changes in patients with CADASIL [14]. Reaction time is negatively associated with corpus callosum area, determined on three-dimensional T1 MRI sequences using a validated methodology [15]. Collectively, these results demonstrate that MRI measures of brain structures have close associations with cognitive functioning in patients with CADASIL.

Resting-state functional magnetic resonance imaging (fMRI) has been widely used to investigate changes in global functional network connectivity and local spontaneous neuronal activity in the brain at rest. Cognitive functioning, including executive function and processing speed, depend on the integrity of the frontal–subcortical and frontoparietal networks [16‐18]. A preliminary study by Cullen et al. demonstrated that intrinsic frontoparietal functional networks are also detectable in patients with CADASIL and are associated with cognitive measures, particularly measures of executive ability and speed [19]. Although these abnormalities in functional network connectivity can be used to reveal the comprehensive and integrative characteristics of brain areas, that study was unable to assess the abnormalities in regional neuronal activity. The amplitude of low-frequency fluctuation (ALFF) is a method that has been used to evaluate local spontaneous neuronal activity in a resting-state fMRI analysis based on blood oxygenation level dependent (BOLD) signals. It reflects the amplitude of neuronal activity within a single voxel by calculating the square root of the power spectrum in a frequency range (typically 0.01–0.08 Hz) [20]. ALFF is effective for examining neuropsychiatric disease-related neuronal activity within specific brain regions [21]. Previous studies have shown the specific patterns of ALFF in mild cognitive impairment (MCI), Alzheimer’s disease (AD) and subcortical ischemic vascular dementia (SIVD) and these alterations in ALFF have revealed significant correlations with cognitive performance in these patients [22‐24]. Measurement of ALFF may aid in early diagnosis and illness monitoring of AD and may become a sensitive marker of AD pathology such as Aβ/pTau ratio regardless of clinical diagnosis, which can provide novel insights into the pathophysiological mechanisms of AD [25‐27]. Thus, we assumed that brain functional activity is more likely to be altered in these patients in the resting-state based on the aforementioned abnormalities in cognitive behavior and brain structure in patients with CADASIL [1‐5]. However, whether there are changes in ALFF values and whether these aberrant changes have impacts on cognitive impairment in patients with CADASIL remain poorly understood. In this study, we used ALFF analysis to investigate regional spontaneous neuronal activity in patients with CADASIL by resting-state fMRI and examined their correlations with cognitive performance.

The maps of the significant differences in ALFF of the 22 CADASIL and the 44 controls were compared using voxel-wise two-sample t tests with age and gender as covariates. Additionally, the FC maps used voxe-wise two-sample t tests with age and gender as covariates within a brain mask. To address the issue of multiple comparisons, the ALFF and FC statistical maps were assigned thresholds at p < 0.001 (voxel level), and family-wise errors (FWE) were corrected to p < 0.05 at the cluster level. The surviving clusters were reported.

Pearson’s correlation analysis was used to detect correlations with clinical scores in patients. Demographic data between two groups were compared using Pearson’s chi-squared test for gender and independent-sample t tests for age, MMSE, MoCA, HAMD and HAMA. Significant differences were determined based on p < 0.05.

CADASIL highly resembles sporadic SIVD. However, an ideal experimental model of sporadic SIVD has been absent until now. CADASIL could serve as a model for studying the relationship between cerebrovascular pathology and cognitive impairment in patients with sporadic SIVD and thereby become a model of pure vascular dementia [1, 2]. Therefore, research about the mechanisms underlying CADASIL will help in early recognition and prevention of CADASIL and sporadic SIVD.

This is the first study to investigate the aberrant changes in intrinsic, spontaneous brain activity in patients with CADASIL by measuring ALFF values from resting-state fMRI signals. The ALFF values revealed abnormal brain activity in the right Pcu/CU, bilateral SFG and left cerebellar anterior and posterior lobes, weak FC between the areas with abnormal ALFF and other brain areas, and close associations between changes in ALFF values and cognitive performance.

In this study, a pronounced decease in ALFF in the right Pcu/CU was observed in patients with CADASIL. The Pcu and CU are located in the medial parietal and inferior occipital lobes, respectively [30, 31]. Three distinct FC patterns have been demonstrated in the Pcu based on resting- state fMRI data. The anterior Pcu is functionally connected with the sensorimotor region, the central Pcu with the cognitive/associative region, and the posterior Pcu with adjacent visual cortical regions [30]. Therefore, the Pcu has been considered the hub of the frontoparietal central executive network, and it is thus a vulnerable brain region [32]. The Pcu is often activated during retrieval of episodic memory, and this activation is related to visual imagery during memory processes [33]. Furthermore, some findings have suggested that Pcu volume is associated with impaired anterograde memory due to posterior cortical atrophy, which is driven by well-recognized deficits in visual imagery processing [34]. The Pcu and CU are also involved in several distinct domains, including cognitive reserve, self-awareness of memory, and strategic learning [31]. In this study, the decreased ALFF values in the right Pcu/CU suggest that patients with CADASIL might have the impaired Pcu/CU cognitive functions described above and that Pcu/CU dysfunction might be involved in the pathogenic mechanisms underlying CADASIL. Here, the patterns of functional change in the Pcu/CU of patients with CADASIL were partially consistent with patients with SIVD, MCI or AD [22, 23, 35, 36]. Studies using ¹⁸F-fluorodeoxyglucose positron emission tomography in patients with CADASIL have shown hypometabolism in the Pcu/CU [37]. In addition, Craggs et al. used post-mortem brains from patients with CADASIL to observe severe vessel sclerosis in these two brain regions [38]. Similar results, such as cortical thinning, hypometabolism, hypoperfusion, and loss of synapses in the Pcu/CU, have also been reported in patients with AD by structural MRI, fMRI, and electron microscopy [39‐41]. These findings may account for the low resting-state activation in the Pcu/CU of patients with CADASIL in the present study.

We also observed an increase in ALFF values in the bilateral SFG and left cerebellar anterior and posterior lobes in the patients with CADASIL. The SFG brain region belongs to Brodmann’s area 10 (BA 10), which is the largest frontal brain region and is known as the anterior prefrontal cortex (PFC), frontopolar prefrontal cortex, or rostral prefrontal cortex [42]. BA 10 has been implicated in multiple integrative roles, including working memory, episodic memory retrieval, abstract reasoning, decision making, behavior control, monitoring of internal states, odor evaluation, nociceptive processing, and perceptual metacognition [42‐44]. The cytoarchitecture of BA10 is characterized by more dendritic spines, a higher spine density, and a lower cell body density compared with other areas of the cortex, suggesting that BA 10 is more likely to accomplish these tasks of information integration [45]. In contrast, the cerebellum includes the cerebellar anterior and posterior lobes. Anatomical, neuroimaging, clinical, and behavioral studies have demonstrated that the cerebellum is engaged in the motor domain as well as higher-level cognitive and affective functions. It has been established that sensorimotor tasks activate the anterior lobe, whereas the posterior lobe is involved in higher-level tasks, including language and verbal working memory, executive functions, spatial tasks, and emotional processing. Therefore, a “sensorimotor” and a “cognitive” cerebellum has been proposed [46‐48]. In particular, language and executive tasks activate the posterior lobe and are involved in prefrontal–cerebellar loops [46]. Taken together, the present data show an increase of ALFF activities in the bilateral SFG (BA 10) and left cerebellar anterior and posterior lobes in patients with CADASIL. Therefore, we inferred that the higher spontaneous activity in BA 10 and the cerebellum might be compensatory mechanisms for cognitive deficits in these patients, acting to enhance their functional connectivity. Similarly, increased brain activities in the SFG and posterior lobe of the cerebellum have also been reported in patients with AD and SIVD [22, 49].

In the present study, negative associations were found between ALFF values for the bilateral SFG (BA 10) and cognitive performance, including MMSE and MoCA scores, in patients with CADASIL. Conversely, there was a positive association between the ALFF value for the right Pcu/CU and MMSE scores in these patients. These results indicate that patients with greater global cognitive impairment may exhibit more obvious compensation for ALFF activity in the bilateral BA 10, while similar protective effects were not found in the Pcu/CU. Compensatory reactions in spontaneous brain activity have also been detected in patients with SIVD [22, 49]. According to the different cognitive domains in MMSE and MoCA, we also found negative associations between ALFF values for the right BA 10 and left cerebellar anterior and posterior lobes and orientation and working memory scores, and between ALFF values for the left cerebellar anterior and posterior lobes and attention and executive function scores. These results indicated that resting-state ALFF analysis may be predictive of cognitive deficits, particularly in the domains of orientation, working memory, attention and executive function in CADASIL. In our study, we also found positive associations of the ALFF values of the left BA 10 and left cerebellar anterior and posterior lobes with Fazekas scores and of ALFF values of the bilateral BA 10 with mRs scores, indicating that increased BA 10 and cerebellum brain activity could aid in detecting illness severity in patients with CADASIL.

It has been widely accepted that CADASIL is a subcortical disease with infarcts and leukoencephalopathy. However, postmortem neuropathological examinations and transcranial MRI have shown that patients with CADASIL have widespread neuronal apoptosis in the cerebral cortex, predominantly in layers 3 and 5, and cholinergic dysfunction in the motor cortex [50, 51]. These cortical deficits are related to subcortical infarcts and white matter lesions underlying the primary motor cortex. Therefore, CADASIL has been considered a disease of disconnection due to disruption of either the cortico–cortical or cortical–subcortical networks in the white matter of the frontal lobe, which may explain the executive dysfunction and motor deficits in patients with CADASIL [38, 52]. Recent evidence from resting-state FC highlights the connections between BA 10 and multiple cognitive networks, including the default mode network (DMN), the executive control network, the salience network, and the sensorimotor network. The DMN includes the medial PFC (mainly BA 10), the Pcu, the posterior cingulate cortex (PCC), and the PHG, and plays a vital role in internal process monitoring and memory retrieval [42]. Furthermore, DTI data have offered novel anatomical evidence for large BA 10 fibers contributing to the human cortico–ponto–cerebellar system in the pontine nuclei, indicating that BA 10 may play a vital role in increasing the connections between the cortico–cortical or cortical–subcortical networks by participating in the motor system [53]. Although weaker FC was observed between BA 10 and the major sensorimotor cortex, including the precentral gyrus and postcentral gyrus, and between BA 10 and the subcortical pons in the present study, we considered them as insufficient compensation. Therefore, we proposed the protective mechanism of BA 10 and the sensorimotor areas counteracting the disconnection between the cortical and subcortical networks in patients with CADASIL. Here, decreased FC between the Pcu/CU and CU and between the Pcu/CU and PHG was also detected, consistent with the aberrant connectivity in the occipital–parietal lobe in patients with MCI and AD [54]. Furthermore, we inferred that the dysfunction in the DMN, including BA 10, Pcu, PCC, and PHG, might be implicated in the pathological mechanism underlying CADASIL.

Some limitations of this study should be discussed. First, the relatively small sample of patients with CADASIL reduced the power to detect significant effects. Further study will be necessary with a larger sample size. Second, we did not classify the patients according to the degree of illness. In the future, we will observe functional changes in early- and late-stage patients. Third, a brief screening of cognitive function, including the MMSE and MoCA, was used in this study. Future studies will focus on more detailed cognitive domains, such as executive function, processing speed, attention, and working memory.

This study revealed spontaneous aberrant ALFF values and correlations of these values with cognitive functions in patients with CADASIL. It is well known that resting-state fMRI is a quick and repeatable method without radiation injury, and it has therefore been used to measure features of brain function for early detection and illness monitoring. Therefore, spontaneous brain activity may have the potential to become a predictor of cognitive status in patients with CADASIL. These findings will help to elucidate the causal mechanisms underlying CADASIL and to develop therapeutic strategies.