Visualization of altered neurovascular coupling in chronic stroke patients using multimodal functional MRI

Abstract

Evaluation of cortical reorganization in chronic stroke patients requires methods to accurately localize regions of neuronal activity. Blood oxygenation level-dependent (BOLD) functional magnetic resonance imaging (fMRI) is frequently employed; however, BOLD contrast depends on specific coupling relationships between the cerebral metabolic rate of oxygen (CMRO(2)), cerebral blood flow (CBF), and volume (CBV), which may not exist following stroke. The aim of this study was to understand whether CBF-weighted (CBFw) and CBV-weighted (CBVw) fMRI could be used in sequence with BOLD to characterize neurovascular coupling mechanisms poststroke. Chronic stroke patients (n=11) with motor impairment and age-matched controls (n=11) performed four sets of unilateral motor tasks (60 seconds/30 seconds off/on) during CBFw, CBVw, and BOLD fMRI acquisition. While control participants elicited mean BOLD, CBFw, and CBVw responses in motor cortex (P<0.01), patients showed only mean changes in CBF (P<0.01) and CBV (P<0.01), but absent mean BOLD responses (P=0.20). BOLD intersubject variability was consistent with differing coupling indices between CBF, CBV, and CMRO(2). Thus, CBFw and/or CBVw fMRI may provide crucial information not apparent from BOLD in these patients. A table is provided outlining distinct vascular and metabolic uncoupling possibilities that elicit different BOLD responses, and the strengths and limitations of the multimodal protocol are summarized.

Figure 1

(A) The anatomical region of interest (ROI) used for all control and patient volunteers, as well as the separate unthresholded z-stats for the control and patient volunteers. Note that the z-stats are highly asymmetric for the control volunteers, as expected for the unilateral joystick task, however, are smaller and more evenly distributed between cortices for the patients. Average (B) control and (C) patient blood oxygenation level-dependent (BOLD), cerebral blood flow-weighted (CBFw), and cerebral blood volume-weighted (CBVw) reactivity time courses within M1. Gray blocks correspond to stimulus periods and error bars represent standard error over all volunteers in each group. Note the absence of a mean BOLD signal change in patients, yet clear CBFw and CBVw reactivity, thereby suggesting discord in CBF, CBV, and cerebral metabolic rate of oxygen consumption (CMRO₂) coupling relationships following chronic stroke, unclear from BOLD measurements alone.

Figure 2

Representative block-averaged time courses for (A) a control subject and (B–D) three patients. Gray areas depict the stimulus period. For the control, a robust blood oxygenation level-dependent (BOLD) change (P<0.05) is observed with normal cerebral blood flow-weighted (CBFw) (P<0.05) and cerebral blood volume-weighted (CBVw) (P<0.05) reactivity. Alternatively, the patients show a more complex array of variation. (B) A small, positive BOLD change (P<0.05) is explained by a small CBFw increase (P<0.05) and negative CBVw change (P<0.05), implying possible vascular steal in the region. (C) A negative BOLD change (P<0.05) is explained by a large CBVw increase (P<0.05), consistent with vasodilatory autoregulation, and a smaller CBFw change (P<0.05). (D) A patient with effectively no BOLD (P>0.05) and near-negligible CBFw change (P<0.05), yet small vasodilatory CBVw change (P<0.05). Importantly, patients display a more complex array of hemometabolic coupling relationships, which require independent estimations of CBF and CBV to understand fully.

Figure 3

(A) Arterial spin labeling (ASL) kinetic curves for a baseline cerebral blood flow (CBF)=50 mL/100 g per minute and arterial transit time (ATT)=0.6 seconds and different possibilities from changes in these parameters during motor stimulation. Here, the increase in CBF has been taken to be equal to the reduction in ATT, and three scenarios for changes of 15%, 24%, and 33% are shown. (B) The change in ASL difference magnetization (normalized by baseline ASL difference magnetization), for the three activation scenarios in (A). Note that at short inversion time (TI), the reactivity measures depend heavily on choice of TI, yet at longer TI, the contrast is less sensitive to this parameter. These plots provide an exemplar for interpreting the cerebral blood flow-weighted (CBFw) changes in terms of absolute changes in CBF and ATT.