Type 2 diabetes mellitus (T2DM) is a common metabolic disease in middle-aged and older adults characterized by chronic hyperglycemia, which leads to long-term macrovascular and microvascular complications of various organ systems. The epidemic of T2DM and its complications raise a global health threat (Zheng et al., 2018). The present literatures have proved that T2DM is a significant risk factor of developing certain mental disorders, including cognitive dysfunction, dementia, and depression (Biessels and Despa, 2018; van Sloten and Schram, 2018; Xue et al., 2019), and older individuals with T2DM progress to dementia at faster rates (Xu et al., 2010; Morris et al., 2014). Although the underlying mechanisms of these disorders are still unaddressed, growing evidences indicate that cerebral microvascular dysfunction is one of the key mechanisms, which may be driven by hyperglycemia, obesity, insulin resistance, and hypertension (van Sloten et al., 2020). Therefore, characterizing the phenotype of cerebral perfusion alterations may advance our understanding of the underlying mechanisms of cognitive aging and mental impairments in T2DM.

As the brain is a highly metabolic organ with limited energy reserves, the metabolically active regions need abundant supply of glucose and oxygen via cerebral perfusion (Coucha et al., 2018). Cerebral blood flow (CBF), commonly defined as the volume of blood delivered to a unit of brain tissue per minute, is responsible for the delivery of nutrients to the brain (Fantini et al., 2016). CBF is also correlated to brain activity, and there is a coupling between metabolically active regions and CBF under normal circumstances (Hoge et al., 1999). Recent studies have observed neurovascular decoupling in T2DM (Hu et al., 2019; Yu et al., 2019; Zhang et al., 2021). Therefore, the cerebral perfusion impairment may cause oxidative metabolism dysfunction of brain and neuronal damage, leading to mental disorders in T2DM.

CBF can be quantitatively measured using single-photon emission computerized tomography (SPECT), positron emission tomography (PET), perfusion computed tomography (PCT), dynamic susceptibility contrast magnetic resonance imaging (DSC-MRI), and arterial spin labeling (ASL). However, SPECT and PET require injection of radiotracers while PCT and DSC-MRI require injection of intravenous contrast agent (Wintermark et al., 2005). Besides, SPECT, PET, and PCT are associated with radiation exposure. Compared with the aforementioned methods, ASL is a non-invasive method to measure CBF by magnetically labeling the inflowing arterial blood water in vivo as an endogenous tracer (Williams et al., 1992). Due to its non-radiation, non-invasiveness, and reliability, ASL is proposed as a promising method to reveal cerebral perfusion biomarkers in various mental disorders (Alsop et al., 2015; Haller et al., 2016; Zhang, 2016).

In the last three decades, growing literatures have attempted to characterize cerebral perfusion patterns in T2DM, but the findings are varied across studies. A recent study systematically reviewed literatures on cerebral perfusion in T2DM and found the reduction of regional cerebral perfusion in multiple locations, including occipital lobe, domains involved in the default mode network and the cerebellum (Wang et al., 2021). However, this study involved various modalities including SPECT, DSC-MRI, and ASL. More importantly, no quantitative synthesizing method was used to conduct meta-analysis of voxel-based studies. As the region of interest (ROI) method has inherent bias and is more liberal in statistical threshold than voxel-based analysis (VBA) method (Radua and Mataix-Cols, 2009; Yao et al., 2021), the quantitative meta-analysis of voxel-based studies can objectively identify regional CBF differences at whole-brain level without any priori hypothesis.

Therefore, we first systematically reviewed literatures on cerebral perfusion in T2DM using ASL and then conducted a quantitative meta-analysis on these voxel-based literatures using Seed-based d Mapping (SDM, formerly Signed Differential Mapping) as primary tool. The SDM is a well-recognized synthesizing method for voxel-based studies and has been used in meta-analysis of cerebral structural and functional alterations in T2DM (Liu J. et al., 2017; Liu et al., 2021; Yao et al., 2021). This study aimed to identify consistent regional CBF alterations in T2DM and explore the potential effects of the clinical characteristics on these perfusion alterations.

To our knowledge, this is the first quantitative meta-analysis to pool the ASL studies to identify the consistent pattern of CBF alterations in T2DM. This systematic review and meta-analysis revealed that the regional CBF was significantly reduced in the patients with T2DM, mainly involving occipital and parietal lobes. These findings indicated the potential neuropathology of visual impairment and cognitive aging in T2DM (Meusel et al., 2014).

The most consistent and significant finding was that the perfusion of occipital lobe was impaired in T2DM. The middle occipital gyrus and calcarine fissure/surrounding cortex in the occipital lobe were important components of visual cortex, which were responsible for vision processing and visual memory (Tootell et al., 1998; Wandell et al., 2007). Previous studies demonstrated that the decreased perfusion in middle occipital gyrus was associated with impaired visuospatial function and visual memory (Xia et al., 2015; Cui et al., 2017). A recent meta-analysis study of functional magnetic resonance imaging (fMRI) also revealed consistent hypoactivity in the middle occipital gyrus and calcarine fissure/surrounding cortex in T2DM (Yao et al., 2021). Beside, recent studies focusing on patients with diabetic retinopathy observed decreased CBF in the bilateral calcarine fissure/surrounding cortex (Huang et al., 2021) and hypoactivity in the middle occipital gyrus (Wang et al., 2017; Qi et al., 2020). These findings indicated that the perfusion and function alterations in occipital lobe, which involving vasculopathy and neuropathy along the visual pathway (Heravian et al., 2012), might be attribute to the potential visual impairment, a common comorbidity of diabetes.

Another consistent finding was the reduced perfusion in parietal lobe in T2DM. Our quantitative meta-analysis identified decreased CBF in superior parietal gyrus. Previous neuroimaging studies also demonstrated gray matter volume loss (Roy et al., 2020) and functional dysconnectivity (Cui et al., 2016; Liu L. et al., 2017) in superior parietal gyrus in patients with T2DM. The superior parietal gyrus was involved in aspects of attention and visuospatial orientation, including the manipulation of information in working memory (Koenigs et al., 2009), which was impaired in patients with T2DM (Chen et al., 2014; Huang et al., 2016). Working memory is a fundamental cognitive process in the brain and it is crucially important for most higher-order cognitive functions (Baddeley, 2003). T2DM has been consistently associated with an increased risk of dementia and mild cognitive impairment (Reijmer et al., 2010; Beeri and Bendlin, 2020), and the structural and functional abnormalities in the brain are thought to underlie these cognitive deficits (Yao et al., 2021). Previous studies indicated that increased activation strength in parietal lobe was positively associated with memory improvement in patients with mild cognitive impairment (Belleville et al., 2011; Corriveau-Lecavalier et al., 2019). Therefore, it suggests that the decreased perfusion in superior parietal area may underlie the neuropathology of cognitive deficits in T2DM.

Our meta-analysis results also showed decreased CBF in the left caudate nucleus and increased CBF in the right supplementary motor area in T2DM, which were not commonly reported in ROI studies. Besides, it should be noted that the right supplementary motor area showed significant publication bias (Egger test p = 0.001). The caudate nucleus, a component of the dorsal striatum, has an important role in cognitive function and spatial working memory (Postle and D’Esposito, 2003; Grahn et al., 2008). The functional abnormalities of the caudate nucleus may also lead to motor dysfunctions (McColgan et al., 2015; Ji et al., 2018), which have been observed in patients with T2DM (Gorniak et al., 2014; Ochoa et al., 2016). Meanwhile, the supplementary motor area play a role in the direct control of movement, especially in finger movement (Shibasaki et al., 1993; Tanji and Shima, 1994), and the diabetic peripheral neuropathy may lead to sensory impairments in the motor system (Allen et al., 2016). Thus the deficits of corticostriatal circuit between the head of caudate nucleus and supplementary motor area may be the neuropathology for motor dysfunction in T2DM. The increased perfusion in supplementary motor area might suggest a compensation for the functional deficits of corticostriatal circuit in T2DM.

Although ASL has been the widely used neuroimaging approach in brain perfusion, the acquisition parameters and analytic methods varies among ASL studies, bringing potential bias. For example, quantitative assessment of perfusion with ASL is hampered by the transport time from the labeling position to the tissue, known as arterial transit time (ATT) (Alsop et al., 2015). PASL and PCASL are both labeling approaches using single PLD/inversion time (TI) but differ fundamentally in spatial extent and time of labeling and labeling delay (As shown in Table 3). Besides, one of the PASL study used multiple TI approach (Shen et al., 2017), which estimated both CBF and ATT via fitting data. Our subgroup meta-analysis found no overlap of regional CBF alteration between PASL and PCASL, suggesting the labeling approach might have a great impact on cerebral perfusion. As for the spatial resolution and PVE, our subgroup meta-analysis showed that the cerebral perfusion alteration in left caudate nucleus were not reproducible. One possible reason might be that the caudate nucleus was close to lateral ventricle and more likely to contain a mixture of gray matter and cerebrospinal fluid (Jezzard et al., 2018). Besides, as gray matter atrophy was observed in T2DM (Yao et al., 2021), there might be potential overestimation of decreased perfusion in regions where both perfusion and gray matter volume were reduced (Chappell et al., 2021). Future studies should attempt to conduct analysis with and without PVE correction to investigate its influence. In summary, even though some regional perfusion alterations could be affected by the heterogeneity of acquisition parameters and analytic methods, the increased CBF in right supplementary motor area and decreased regional CBF in right middle occipital gyrus and superior parietal gyrus were robust in 3 of 4 subgroup analyses.

There are several limitations in this study. First, the sample size of patients with T2DM included in some studies was relatively small. Second, near half of ASL studies in T2DM were not included in quantitative meta-analysis because of the use of ROI approach without available coordinates and corresponding effect sizes. Third, there were heterogeneity between the included studies. The confounding factors such as age, illness duration, blood glucose control, and comorbidities might affect CBF. Although we sought to identify the potential effects of some confounding factors, the results were negative, which also should be taken caution as only few data were available in the metaregression analysis. It is also difficult to avoid false-negative results even though voxel-based meta-analytical methods have good control for false-positive results (Radua et al., 2012). Fourth, although this review reveals the association between neuropathology and visual impairment and cognitive aging in T2DM, whether the vascular mechanism underlying these disorders remains inconclusive. Further research would be required to determine causation.

In conclusion, this systematic review and meta-analysis revealed consistent cerebral perfusion alterations in T2DM, characterized by decreased CBF in occipital and parietal lobes. These findings suggested the neuropathology of visual impairment and cognitive aging in T2DM.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

JL and PZ conceived and designed the study. JL, XY, YL, and HX collected the data. JL and XY analyzed the data and drafted the manuscript. PZ revised the final manuscript. JL, JR, and PZ provided funding for the study. All authors reviewed the manuscript, contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.