Negative Blood-Oxygen-Level-Dependent (BOLD) Response
Fig. 1:
Left) Distance-dependent variations and spatial distributions of the nHRF in the visual cortex.
Middle) Depth-dependent variations of the nHRF and its parameters in the visual cortex.
Right) Distance-dependent variations and spatial distributions of the nHRF in the somatosensory cortex.
To the best of our knowledge, distance/depth-dependent trends of the nHRF have not been thoroughly investigated. Using high spatiotemporal resolution, we will characterize the nHRF and its parameters in both visual and sensorimotor cortices on both hemispheres, as its distance/depth-dependent variations along the cortical surface and within the gray matter (Fig. 1). In order to investigate if the NBR has complex dynamics, we will also examine the temporal linearity of the NBR with various stimuli durations.
The BOLD response in functional magnetic resonance imaging (fMRI) is driven by changes in cerebral blood flow (CBF) and oxygen metabolism (CMRO2), which have been shown to change in parallel with neural activity. In order to properly interpret fMRI data, an understanding of these physiological processes and how they change is critical to possibly reveal information about local neural activity and possible sources of brain dysfunction.
The negative BOLD response (NBR), the decrease in BOLD signal in regions adjacent to the positive BOLD response (PBR) and/or ipsilateral to a stimulus, has been frequently observed in BOLD fMRI experiments. The NBR and its origin has been explored extensively, but typically with poor spatiotemporal resolution and with stimuli that have a long duration. This has led to a dearth of understanding regarding the spatial and temporal properties of the NBR. Not only that, but the sign and timing of the corresponding CBF/CMRO2 dynamics for are uncertain due to inconsistent results and interpretations, meaning the underlying vascular and metabolic mechanisms remain controversial.
To quantitatively evaluate abnormality of local neurovascular and neurometabolic activities caused by pathological disruption, such as in epilepsy, stroke, and Alzheimer’s disease, this proposed research for characterization of the NBR dynamics and its underlying physiology in healthy population must be investigated first. Therefore, we will investigate the properties of the NBR evoked by a brief stimulus, the negative hemodynamic response function (nHRF), as well as its corresponding neurovascular and neurometabolic dynamics in the human brain.
Fig. 2:
Left) nHRF (orange) and CBF (green) measurements for two subjects. Dashed lines show 68% confidence intervals.
Right) Schematic diagram for CMRO2 estimation.
In addition, we will obtain arterial spin labeling (ASL) measurements with high spatiotemporal resolution in order to quantify CBF dynamics and compare the results with our BOLD measurements (Fig. 2). Then, using a general transport computational model for oxygen diffusion that was recently developed, we will quantify CMRO2 dynamics using the CBF/BOLD measurements. This will allow us to gain insight into the neurovascular and neurometabolic activities associated with the NBR, which can provide huge utility in research for normal brain functions.
-
High Resolution Functional Magnetic Resonance Imaging
Fig. 3:
Left) Segmentation of brainstem/thalamus used to create a surface.
Right) Strong HRFs evoked by multisensory stimulus in multiple subcortical regions for two subjects, as well as in early visual cortex (V1).
Our major goal is to provide the first detailed dynamic characterization of subcortical neurovascular coupling. To accomplish this, high spatiotemporal HRF measurements will be collected using a multi-sensory stimulus and compared across a variety of subcortical nuclei and as a function of depth in laminar structures such as the Superior Colliculus (SC) (Fig. 1). Additionally, we will implement a correction of dynamics off-resonance in k-space (DORK) to significantly reduce physiological noise and motional blurring, increasing our SNR in subcortical regions.
Subcortical brain regions are essential for normal human cognitive function, and play many critical roles in human behaviors such as the orientation of attention, arousal, and the modulation of sensory signals to cerebral cortex. Traumatic brain injury (TBI) in the brainstem can cause severe visual impairment, spasticity, coordination deficits, and organic psychosis, as well as deficits in multisensory integration. Despite this, clinical subcortical research studies are limited and mostly oriented around structural MRI and white-matter (WM) damage using MR diffusion tensor imaging (DTI).
The use of BOLD fMRI is crucial to understand the subcortical neurovascular response, especially in cases where there is no structural abnormality, such as mild TBI. High spatiotemporal resolution is required for collecting subcortical HRF measurements due to much lower a signal-to-noise (SNR) ratio and having to accurately resolve the temporal dynamics of the HRF. One study which characterized the subcortical HRF found differences compared to those evoked in the cortex (despite limited spatial resolution), warranting further investigation of the HRF in human subcortical regions.
Fig. 4:
Left) Proton Density Weighted Imaging (PWDI) for more accurate segmentation: Boundary for superior colliculus (SC) in blue, periaqueductal gray (PAG) in red.
Right) Diffusion Tensor Imaging (DTI) for tract-based analysis: Color-coded FA, absolute FA, and tractography from SC.
Following this, Proton Density Weighted Imaging (PWDI) and Diffusion Tensor Imaging (DTI) scans will be collected in order to more precisely segment subcortical nuclei as well as examine global and local WM abnormalities, respectively (Fig. 4). This will allow us to test for significant correlations between these metrics and HRF-based parameters for heathy control and patient populations.
Fig. 5:
Left) Schematic illustration of proximal integration hypothesis as a basis for prompt arterial dilation model.
Right) Examples of model fits to measured BOLD responses as well as CBF/CMRO2 responses corresponding to the model predictions.
Finally, we will also collect ASL measurements with high spatiotemporal resolution in order to extend our oxygen transport model, based on the proximal-integration hypothesis (Fig. 5), to human subcortical regions. Using both CBF and BOLD measurements, we will validate the accuracy of the model and use it to predict dynamics of subcortical neurovascular coupling as well as identify differences between healthy control and patient populations.
-
Clinical Application of the Hemodynamic Response Function
TBD