Background

Moyamoya disease (MMD) is a chronic cerebrovascular occlusion disease, of which the underlying pathogenesis has not yet been fully elucidated [1]. Chronic hypoperfusion in patients with MMD will inevitably damage white matter fibers [2]. The low perfusion or microstructurally altered tissue may appear normal on conventional magnetic resonance images, and are called normal-appearing white matter (NAWM) [3]. Aberrant neural functioning develops because of damage to white matter fibers or the neural pathway, which is linked to worse cognitive outcomes [4].

Resting-state functional magnetic resonance imaging (rs-fMRI) is widely used to study brain functional organization in patients with stroke and neurodegenerative diseases [5]. Functional connectivity is defined as the temporal correlation that equals a Pearson’s r, in the high amplitude, low-frequency spontaneously generated blood oxygen level-dependent (BOLD) signal between voxels or brain regions [6]. Exploring the characteristics of brain functional connectivity changes is beneficial because it can promote our knowledge regarding cognitive impairment secondary to chronic ischemia in MMD and is crucial for predicting the reversibility of neurocognitive dysfunctions after revascularization surgery. Thus, functional connectivity changes may become early features of cognitive impairment and a prognostic hallmark of MMD. Some researchers even suggest that rs-fMRI should be performed as part of routine clinical investigations that enable assessment of multiple brain systems in MMD [7].

Diffusion tensor imaging (DTI) has been used to quantify microstructural alterations in white matter in vivo, and is especially useful for patients with neurological disorders [8]. The main mechanism of DTI involves measuring the magnitude and directionality of random water movement to generate information regarding tissue or fiber integrity of the direct connections between different brain regions [9]. Fractional anisotropy (FA) and mean diffusivity (MD) are the primary tensor metrics that reflect the overall physiological and pathological alterations of white matter [10]. In addition, axial diffusivity (AD) and radial diffusivity (RD), which reflect axon integrity and myelin sheath integrity, respectively, can be useful for understanding the underlying physiology in many aspects [9, 7, 18, 19], the current study performed ROI-based analyses via functional connectivity for the first time and further combined tensor metrics to provide additional evidence for our findings.

There is growing evidence that brain functional changes may occur prior to the structural changes and imply a dissociation between functional impairment and structural damage [20, 21]. While the BOLD signal contrast depends on the interplay among hemodynamic factors [22], this may explain the physiology underlying reduced interhemispheric or intrahemispheric functional connectivity after stroke; thus, changes in brain function are likely to be more sensitive than that in structural pathways.

Our study showed that the functional connectivity of several paired regions decreased in MMD patients. There is increasing evidence that intrahemispheric or interhemispheric connectivity changes that occur in ischemic disease [23] are linked to cognitive decline [24]. Here, we also confirmed that changes in functional connectivity were correlated with cognitive performances in MMD, although significant correlation was only demonstrated in the functional connectivity between SMA-IFGorb in left hemisphere. The functional pathway of the left SMA-IFGorb had a significant positive correlation with the MMSE and MoCA, and a negative correlation with the TMT-B. Since the larger TMT-B scores indicate worse performance, it means that this pathway is positively correlated with general cognitive function and processing speed or set-shifting abilities.

The SMA complex is located in the posterior part of the superior frontal gyrus, which is bounded by the cingulate gyrus, anterior central gyrus, and superior frontal gyrus [25]. Neuroimaging studies have confirmed that the SMA is involved in the cognitive process of working memory, executive control and processing speed [26,27,28]. Through fiber tracking imaging, researchers found that the most important subcortical connection with the SMA included the primary motor area, the cingulate, the corpus callosum, the inferior frontal gyrus (IFG), and the striatal complex-internal capsule [29]. It is important to mention that the frontal aslant tract has been identified as a white matter tract connecting the SMA to the IFG, and it was shown to have a role in language function, working memory, social community tasks, attention, and processing [30]. Our results were consistent with the abovementioned anatomical and imaging studies of the SMA.

MMD is an ischemic disease that affects patients for a long period of time, and chronic insufficient blood supply leads to demyelination of white matter fibers and destruction of axon integrity. However, such patterns of subtle damage referred to as NAWM are usually not demonstrated in the structural images; thus, preventing us from detecting the brain abnormalities that are caused by MMD on conventional MRI sequences. Here, in this study, we excluded patients with cerebral lesions larger than 3 mm. In ischemic MMD with symptoms including TIA, dizziness, or headache, there is always no apparent cerebral lesion. In hemorrhagic MMD, cerebral lesions could be the residual minor destruction of brain structures after the absorption of hematoma. And especially in patients with primary IVH, there is always no evidence of lesions in the parenchyma. Fortunately, with the development of MR technology, we can use emerging methods to elucidate brain damage by capitalizing on many different aspects other than structure.

Currently, there is intense research being conducted on white matter integrity and brain functional connectivity in neurological diseases. Studies have shown that white matter abnormalities are correlated with disrupted functional connectivity, suggesting that common mechanisms may underlie fibers and functional connectivity [31]. Our study revealed that the fiber integrity and functional connectivity between the SMA and the IFGorb in the left hemisphere were disrupted before the development of a significant brain lesion in MMD patients. Such early detection of brain abnormalities at the preclinical stage is conducive for implementing preventive interventions to abate cognitive decline in MMD.

FA, MD, RD, and AD are the main tensor metrics that are derived from DTI. When the white matter fibers are damaged, the diffusion range of water molecules in the fibers increase, and the values of FA, MD, and RD will also increase accordingly. Our study showed that the FA of the fibers between the SMA and the IFGorb were not significantly different between MMD patients and HCs. This might be the result of the existence of cross fibers between the two brain regions, or it is highly likely that it occurred because the disruption of the fibers happened in all directions.

The interpretations of the changes in the AD are controversial. Theoretically, the AD values decrease when axons are injured, but other studies report that the AD increases in damaged white matter regions [32, 33]. The potential mechanism consists of increased extracellular water content that is secondary to atrophy of white matter fibers, and alterations in the water flow in the axons occur due to the breakdown of certain components of the cytoskeleton, which allows water molecules to move more rapidly parallel to the axons [32, 34]. To summarize, in the case of intensive axonal loss, the increased isotropic diffusion is likely to enhance both RD and AD [33].

Though it seems a bit of redundancy across the cognitive assessment package that we used, there still lies some differences between MMSE, MoCA, and TMT. For example, the MMSE is likely a better test for more severe cognitive impairment and MoCA for mild conditions [35]. TMT is also different from the block that is included in MoCA. Research suggested that TMT had a better performance than MoCA in some cases [36]. On the other hand, the TMT is a more precise indicator of reduced visual processing speed [37]. While they have different applications and results, combining these measurements may provide us with a more comprehensive understanding of cognitive impairment and recovery after revascularization surgery in MMD patients.

CONN is a Matlab-based software that has been widely used for the calculation, demonstration, and investigation of brain network connectivity. Connectivity assessment techniques include seed-to-voxel network maps, ROI-to-ROI connectivity matrixes, graph characteristics of network systems, brain interconnection, inherent connectivity, generalized psychophysiological interaction models, and other voxel-to-voxel measures. Unlike seed-based correlation analysis, which is a hypothesis-driven approach based on a chosen region, ROI-to-ROI analysis is a data-driven method that can produce information regarding the entire brain connectivity pattern. Therefore, we used ROI-based analyses in this study, which we considered to be more powerful for investigating the abnormal brain functional connectivity at the regional level.

White matter integrity impairment and cognitive disruption in MMD have been demonstrated in several studies, aberrant functional connectivity is another form of brain damage caused by MMD [2, 7, 18, 38]. The impaired brain functional and structural connectivity between the SMA and IFG in the left hemisphere provides us a novel perception of the pathophysiology of MMD and might be applied as the hallmark in the evaluation of disease progression and prognosis of MMD in the future.

There are some limitations to our study that must be mentioned. First, like other MMD imaging studies, a small number of patients were investigated due to the rare incidence of the disease, which limits the generalizability of our results. It is necessary to conduct a multi-center, large-sample study to confirm our results. Due to the impact of COVID-19, the collection of control cases has also been affected, though a matched study is more appropriate. Second, vascular abnormalities of MMD mainly involve the anterior circulation of the brain, and therefore, our study did not include the cerebellum as a ROI. It is worth noting that some studies indicate the existence of a functional connection between the cerebellum and the brain [5]. Third, anatomical and imaging studies have shown rich fiber connections between the SMA and the IFG that are associated with cognitive functioning, whereas our results show only that the breakdown of the connection between the SMA and the IFGorb in the left hemisphere is significantly correlated with the cognitive performance of MMD patients. Fourth, further studies are required to determine whether functional or structural connectivity is initially damaged before the onset of cognitive dysfunction in patients with MMD.

Conclusion

In Moyamoya disease, there is damaged brain functional connectivity between the supplementary motor area and the inferior frontal gyrus in the left hemisphere, which is correlated with incomplete integrity of white matter fibers, and may contribute to impaired cognitive performance. These findings could be useful in the evaluation of disease progression and prognosis of moyamoya disease.