The aim of our research is to characterize the dynamic organization of functional network activity in the brain, with the ultimate goal of understanding the alterations that occur in pathological states and using targeted interventions to restore normal activity.  Several ongoing projects that advance these objectives are described briefly below.  More details can be found in the papers listed on the Publications page.

Neural basis of functional connectivity

Functional connectivity is typically measured as the correlation between the blood oxygenation level dependent (BOLD) MRI signal from different areas of the brain over a series of minutes.  While the BOLD signal reflects some aspects of neural activity, it can be contaminated by physiological or system noise.  To determine the extent to which the BOLD correlations correspond to coordinated neural activity, we have developed a method for simultaneous multisite microelectrode recording and MRI in the anesthetized rat, which is technically very challenging due to mutual interference between the two modalities.  Preliminary results have been reported in JoVE and Brain Connectivity, showing that the relationship between functional connectivity and neural activity is highly dependent on the state of the animal.  We are currently attempting to modify the method for use in unanesthetized animals to mitigate the potentially confounding effects of anesthesia on both neural activity and vascular tone; to identify anesthetic-specific relationships between BOLD and electrophysiology; and to establish a neural basis for variations in BOLD connectivity observed using shorter time windows than traditional cross correlation analysis.

Sliding window correlation

Several groups, including our lab, have recently shown that changes in functional connectivity occur on much shorter time scales than typically used for traditional analysis.  However, randomly-matched or modeled signals can show similar variability, making it difficult to determine whether the changes have a neural basis.  We have shown that the correlation between large scale networks in short time windows (~10 s) before a vigilance task predicts the reaction time on the task, strongly suggesting a neural contribution.  We are currently exploring the relationship between sliding window correlation of different frequencies of electrical activity and the simultaneously-recorded MRI in the anesthetized rat to further support a neural basis for the variability.

Quasiperiodic BOLD patterns

In addition to variability in sliding window correlation, the BOLD signal also exhibits reproducible, quasiperiodic spatiotemporal dynamics.  These were first reported by our group in the anesthetized rat, but further work showed that they can also be detected in humans, where they involve areas of the well-known default mode and task positive networks.  The patterns can be characterized with an autoregressive pattern finding algorithm that produces a template of the pattern and a time course showing the strength of the pattern.  Our work has shown that the patterns can also be found when cerebral blood volume contrast is used, and that they are not related to cardiac or respiratory cycles.  The involvement of the large, anticorrelated networks observed in the human studies suggest that these patterns may be involved in fluctuations in attention and performance.

DC recording and MRI

A few groups have used MRI and electrophysiology to examine the neural basis of the BOLD signal, but because the frequencies of the electrical recordings are much higher than the BOLD frequencies of interest (1-100 Hz vs 0.1 Hz), the recorded electrical activity is usually converted to bandlimited power and downsampled before comparison with the BOLD signal.  However, with special amplifiers, it is possible to record very low frequency electrical activity that matches the frequencies observed in the BOLD signal.  We are currently pursuing this approach in the rat with our simultaneous MRI and microelectrode recording protocol and in the human using DC-EEG and MRI.