Introduction

Patients with chronic steno-occlusive disease of the cerebrovascular system are at risk of ischemic stroke as well as cognitive decline. Functional connectivity mapping has advanced the understanding of neural mechanisms underlying cognitive impairment. While functional network reorganization has been found in stroke and cerebrovascular disease^1-3, little is known about remote effects of hemodynamic compromise on network topology of distant brain areas, such as the hemisphere contralateral to vessel occlusion/stenosis. To address this gap, we aimed to evaluate the efficiency of functional connectivity in the contralateral hemisphere among patients with unilateral chronic steno-occlusive disease, and to investigate the relationship between network efficiency and CVR, a measure of cerebral hemodynamic reserve.

Methods

Twenty-one patients with unilateral chronic steno-occlusive disease of the anterior circulation (age, 48.43 ± 14.46 years; 16 females) and twenty age-matched healthy controls (age, 40.75 ± 16.21 years; 8 females) were included. All patients underwent a 20-minute blood oxygen level–dependent (BOLD) acquisition using a gradient-echo echo-planar imaging (EPI) sequence (TR/TE = 2000/30 ms, FA = 78°, voxel size = 3.44 × 3.44 × 4 mm³, 30 slices). After 5 minutes of baseline BOLD acquisition, acetazolamide (ACZ) was slowly infused over 3-5 minutes. CVR was calculated as percentage augmentation of BOLD signals between pre- and post-ACZ infusion. Similarly, BOLD images were acquired from the healthy controls for 13.33 minutes, but without the administration of ACZ. The first 5 minutes of data from both patients and healthy participants were used for network analysis. To calculate graph-theory metrics, forty-five regions of interest (ROIs) were defined according to the Automated Anatomical Labeling atlas⁴ for each hemisphere. Network efficiency was quantified in terms of the local efficiency E_local and the nodal efficiency E_nodal. In the healthy controls, the mean and standard deviation of network efficiency were calculated for both left and right hemispheres separately as a reference standard for the patients. To quantify abnormality in the hemisphere contralateral to cerebrovascular disease in each patient, network efficiency was scored in terms of z-value relative to the mean and standard deviation of the respective hemisphere of the healthy controls calculated above. The correlation between network efficiency and ROI-specific CVR was evaluated using a mixed-effects model. Subject-specific slope and intercept were modeled as random effects to assess the linear relationship.

Results

In general, both E_local and E_nodal were lower in the patients as compared with the healthy controls (Fig 1A and B). Fig 2 shows the z-maps of E_local and E_nodal of the patients relative to the controls. Negative z-scores were observed in most of the ROIs, suggesting compromised network efficiency in the patient cohort. One-sample t-test showed that z-scores of network efficiency were significantly lower than zero at different sparsity levels (20%, 30% and 40%) (Table 1). Visual inspection suggested an association between network efficiency and CVR (Fig 1). Mixed-effects model showed significant correlation between E_nodal and CVR (P = 0.009), while no significant correlation was observed between E_local and CVR (P = 0.976) (Fig 3).

Discussion and Conclusion

Our study provides evidence for altered functional connectivity distant from injured sites⁵. We found reduced network efficiency in the contralateral hemisphere among patients with chronic cerebrovascular disease. Although we focused on the apparently normal hemisphere in this study, its hemodynamic response function might still be altered, contributing to the compromised topological properties. Further studies using simultaneous EEG-fMRI may shed further light on the physiological basis of brain functional networks.

Acknowledgements

No acknowledgement found.