Cerebral small vessel diseases (SVD) are a group of neuropathological processes caused by dysfunctions of small blood vessels in the brain. Three pathological subforms of subcortical small vessel disease are distinguished: arteriosclerosis/microatheroma in 200–800 μm diameter arterioles, fibrinoid necrosis/lipohyalinosis of 40–300 μm diameter arterioles and concentric hyaline thickening/ateriolosclerosis of 40–150 μm diameter arterioles [1]. These three often occur in the same brains and often, but not invariably, in the setting of longstanding arterial hypertension or diabetes [2,3]. Disease of these vessels is responsible for primary intracerebral hemorrhages and discrete small cerebral infarcts (lacunar infarcts) in the deep gray nuclei and white matter (WM) [1]. Together these make up one-third of all strokes [2,4]. In addition, SVD causes the more diffuse widespread WM hyperintensities (WHM) in T2 weighted images, which are characterized by demyelination, axonal loss and gliosis without frank infarction [1]. Earlier studies found WM lesions in 56% and 96% of participants with no history of stroke and with ages ≥40 years [5] and ≥ 65 years [6], respectively. Silent brain infracts were found in 14.2% and 21% of subjects with ages ≥40 years [5] and ≥ 60 years [7], respectively. More severe SVD is associated with increased cognitive [6,8], psychiatric [9], and physical disabilities [6,10,11].

The prevalence, distribution and degree of SVD is inferred from the prevalence and degree of the end organ damage manifested by intracerebral hemorrhage, lacunar infarcts and WMH [12]. While these brain lesions attributed to SVD can be characterized due to their larger sizes and clear visualization by conventional MRI, it remains challenging to noninvasively measure the early pathological changes of the underlying small vessels. Post-mortem data demonstrate that the penetrating arteries (PAs) in WM are important foci for SVD [13,14]. An imaging method for measuring the early structural and functional changes of PAs in WM in vivo for both cross-sectional and longitudinal studies would help to illuminate the etiopathogenesis of SVD and help develop effective treatment strategies.

Structural changes can be inferred by measuring the size of PAs, while functional changes can be inferred by measuring their flow velocity and pulsatility. In this study, we evaluated the feasibility of detecting differences in the size and flow characteristics of WM PAs between patients with diabetes mellitus (DM) and healthy controls (HC) with ultra-high field 7 T MRI. PA diameter and flow velocity can be obtained after correcting for partial volume effects by model based analyses of complex difference images (MBAC) [15]. Although age [16,17], blood pressure [18], and SVD [19] related pulsatility alterations have been reported, pulsatility measurement was not pursued here as we only focused on quantitative measures of PAs, which requires sufficient signal to noise ratio (SNR) at individual PA level to perform MBAC from phase contrast (PC) MRI for correcting partial volume effects [15]. Separation of the data into multiple cardiac phases will make the SNR insufficient.