Demonstration of vascular and physiological effects on resting-state fMRI. We used state-of-the-art ultra-fast fMRI acquisition techniques with multivariate physiological monitoring to assess the effect of carbon dioxide (CO2) fluctuations on the resting-state fMRI signal, providing the first detailed assessment of its kind (Golestani et al., NeuroImage 2014). In addition, we demonstrate experimentally the modulation of fMRI-based functional network measurements by non-neural cerebrovascular reactivity (Golestani et al., NeuroImage 2015).
Demonstration of dynamic neurovascular coupling and vascular bias in resting-state functional MRI. The extent of neurovascular coupling is unknown in resting-state fMRI, much less the effect of vascular contributions to resting-state fMRI functional connectivity. Our work, which used a comprehensive set of vascular measures, demonstrated for the first time the spatial variability in resting-state neurovascular coupling as well as the relationship between functional connectivity measures and macrovascular presence (Tak et al., NeuroImage 2014), with critical implications for rs-fMRI data interpretation (Tak et al., Brain Connect 2015).
Demonstration of dissociation between neurovascular and structural variations in healthy brain aging. Structural changes in the brain have long been observed as part of aging and neurodegenerative diseases. While neuronal integrity is irrevocably tied to neurovascular health, the neurovascular mechanism underlying this structural decline has remained unknown. This work clearly demonstrated, for the first time, distinct patterns of vascular and structural changes in normal aging (Chen et al., NeuroImage 2011), and pioneered a new imaging processing methodology (Chen et al., PLoS ONE 2013) for multi-modality imaging in the community of aging.
Elucidation of the dynamic relationship between vascular and metabolic mechanisms of the BOLD (blood-oxygenation level-dependent) fMRI signal. The understanding of neurovascular interactions in the transient BOLD signal is critical to the understanding and interpretation of BOLD fMRI. For the first time, we obtained simultaneous measurement of BOLD-specific blood flow and volume measurements, which experimentally clarified the origins of the BOLD signal transients (Chen and Pike, NeuroImage 2009). 5. Elucidation of the relationship between vascular and metabolic mechanisms of the BOLD signal. We developed MRI techniques to measure venous cerebral blood volume changes (Chen and Pike, NMR Biomed 2009), which led to the quantification of the venous flow-volume relationship in humans (Chen and Pike, NeuroImage 2010). I also developed methodology to quantify the effect of hypercapnic calibration on cerebral metabolism (Chen and Pike, J Cereb Blood Flow 2010). These measurements are critical for the use of techniques such as calibrated BOLD. The methods associated with these publications have been widely discussed, and the results are being adopted by research labs around the world.