Vascular risk factors (e.g., obesity and hypertension) are associated with cerebral small vessel disease, Alzheimer’s disease (AD) pathology, and dementia. Reduced perfusion may reflect the impaired ability of blood vessels to regulate blood flow in reaction to varying circumstances such as hypercapnia (increased end-tidal partial pressures of CO2). It has been shown that cerebrovascular reactivity (CVR) measured with blood-oxygen-level-dependent (BOLD) MRI is correlated with cognitive performance and alterations of CVR may be an indicator of vascular disfunction leading to cognitive decline. However, the underlying mechanism of CVR alterations in BOLD signal may not be straight-forward because BOLD signal is affected by multiple physiological parameters, such as cerebral blood flow (CBF), cerebral blood volume, and oxygen metabolism. Arterial spin labeling (ASL) MRI quantitatively measures blood flow in the brain providing images of local CBF. Therefore, in this study, we measured CBF and its changes using a dynamic ASL technique during a hypercapnia challenge and tested if CBF or CVR was related to cognitive performance using the Mini-mental state examination (MMSE) score. Seventy-eight participants underwent cognitive testing and MRI including ASL during a hypercapnia challenge with a RespirAct computer-controlled gas blender, targeting 10 mmHg higher end-tidal CO2 level than the baseline while end-tidal O2 level was maintained. Pseudo-continuous ASL (PCASL) was collected during a 2-min baseline and a 2-min hypercapnic period. CVR was obtained by calculating a percent change of CBF per the end-tidal CO2 elevation in mmHg between the baseline and the hypercapnic challenge. Multivariate regression analyses demonstrated that baseline resting CBF has no significant relationship with MMSE, while lower CVR in the whole brain gray matter (β = 0.689, p = 0.005) and white matter (β = 0.578, p = 0.016) are related to lower MMSE score. In addition, region of interest (ROI) based analysis showed positive relationships between MMSE score and CVR in 26 out of 122 gray matter ROIs.

To reduce the total scan time of multiple postlabeling delay (multi-PLD) pseudo-continuous arterial spin labeling (pCASL) by developing a hierarchically structured 3D convolutional neural network (H-CNN) that estimates the arterial transit time (ATT) and cerebral blow flow (CBF) maps from the reduced number of PLDs as well as averages.Methods: A total of 48 subjects (38 females and 10 males), aged 56-80 years, compromising a training group (n = 45) and a validation group (n = 3) underwent MRI including multi-PLD pCASL. We proposed an H-CNN to estimate the ATT and CBF maps using a reduced number of PLDs and a separately reduced number of averages. The proposed method was compared with a conventional nonlinear model fitting method using the mean absolute error (MAE). Results:The H-CNN provided the MAEs of 32.69 ms for ATT and 3.32 mL/100 g/min for CBF estimations using a full data set that contains six PLDs and six averages in the 3 test subjects. The H-CNN also showed that the smaller number of PLDs can be used to estimate both ATT and CBF without significant discrepancy from the reference (MAEs of 231.45 ms for ATT and 9.80 mL/100 g/min for CBF using three of six PLDs). Conclusion:The proposed machine learning-based ATT and CBF mapping offers substantially reduced scan time of multi-PLD pCASL.

A CNN algorithm was proposed to estimate ATT and CBF simultaneously, using multi-TI ASL, acquiring ASL images at multiple PLDs. Hierarchical structure of CNN was used to reduce the estimation error of ATT and CBF. The proposed method successfully estimated ATT and CBF maps using reduced numbers of TIs or averages with a higher accuracy than a conventional non-linear model fitting. The successful estimation of ATT and CBF using our H-CNN may allow a total scan time reduction with a smaller number of TIs or averages and improvements of image quality when a part of acquisition is corrupted.