High-Frequency Resting State Connectivity fMRI
University Of New Mexico Health Scis Ctr, Albuquerque NM
Investigators
Abstract
Recent studies using high-speed fMRI techniques have detected resting state connectivity at frequencies up to 5 Hz in the visual and the motor cortices with significantly higher spatial-temporal stability than the corresponding low frequency (< 0.1 Hz) resting state connectivity. This approach has the potential for addressing principal limitations of mapping low frequency resting state connectivity, such as high sensitivity to signal drifts and long time scales necessary for separating major RSNs. However, other studies using lower temporal resolution have been more cautious regarding the possible signal sources or were unable to replicate the findings. None of the published studies have identified a biophysical mechanism. We have recently detected remarkably strong high frequency connectivity in the auditory cortex, both in healthy controls and in patients with brain tumors, with sensitivity and spatial specificity that approaches that of conventional low frequency resting state connectivity, using high-speed multi-slab echo-volumar imaging (MEVI) (136 ms temporal resolution) and a confound-tolerant seed-based sliding window correlation analysis. Our preliminary data also show high frequency connectivity across several other major RSNs, consistent with previous studies. We hypothesize that high frequency connectivity may reflect fast cerebrovascular regulation. This contrast mechanism would enable novel clinical applications that are not feasible with current methodology, such as improved localization of deep sources of inter-ictal epileptogenic activity to guide surgical resection and mapping of disease-related abnormalities in vascular compliance. The specific aims of this study are: (1) Characterize the biophysical mechanisms of high frequency connectivity in healthy controls. We will compare 2D-accelerated MEVI with 68 ms TR and multi-band EPI with 136 ms TR in 12 healthy controls at 3 Tesla. Biophysical modeling based on arterial spin labeling will be used for calibrated fMRI. Filtering of cardiac pulsatility up to the 3rd harmonic will minimize blood vessel contamination. The detection threshold for high frequency connectivity will be determined by simulating correlations in a Rician noise model. (2) Characterize the physiological basis and clinical potential of high-frequency connectivity in patients with brain tumors. We will assess the physiological basis of high-frequency connectivity in 10 patients with brain tumors adjacent to the auditory and sensorimotor cortex by mapping lesion-related displacement of connectivity. We will then compare sensitivity and specificity with task-based fMRI mapping and intra-operative electrocorticography. If successful, this research will enable mapping of neural activity and connectivity at much shorter time scales than currently feasible, thus improving the characterization of the temporal dynamics of functional connectivity, enhancing the spatial-temporal information obtained from combining fMRI with EEG and MEG and informing about the neurophysiological mechanisms that control brain connectivity and neurovascular coupling. The improved tolerance to slowly varying confounding signals and head movement will have considerable clinical impact for investigating difficult to image populations, such as epilepsy, stroke, Parkinson?s disease and vascular disease.
View original record on NIH RePORTER →