Karen J Mullinger^1,2, Matthew Cherukara¹, Susan T Francis¹, and Stephen D Mayhew^2
1SPMIC, School of Physics and Astronomy, University of Nottingham, Nottingham, Nottinghamshire, United Kingdom, ²BUIC, School of Psychology, University of Birmingham, Birmingham, West Midlands, United Kingdom

The fMRI post-stimulus undershoot is widely observed but its origins remain unclear. We recorded EEG-BOLD-ASL responses to 10s-duration flickering and static visual checkerboard stimuli and show that the amplitude of both the BOLD and CBF post-stimulus undershoots in V1 are negatively correlated with the 10-20s post-stimulus power of the occipital EEG alpha (8-13Hz) oscillation. In addition, flicker trials resulted in both significantly larger fMRI undershoots and significantly higher post-stimulus alpha power responses than static trials. These results provide further evidence that the fMRI undershoot arises from a neuronal mechanism and comprises a unique feature of brain activity.

Adam Berrington¹, Andre Gouws², Stuart Clare¹, Peter Jezzard¹, and Uzay Emir^1
1FMRIB Centre, University of Oxford, Oxford, United Kingdom, ²York Neuroimaging Centre, University of York, York, United Kingdom

We investigate neurochemical changes associated with the positive and negative BOLD response. Using a visual stimulus capable of generating PBR and NBR adjacently in visual cortex, along with a two-voxel spectroscopy sequence at 7T, we determine metabolite changes from resting baseline. The relative increase or decrease of certain metabolites e.g. glucose and glutamate, between NBR and PBR, hints at alternate metabolic processes behind either response. Baseline concentrations of GABA and glutamate are also found to correlate to negative BOLD strength suggesting a link between inhibition-excitation balance and BOLD response.

Aaron Benjamin Simon¹, Zachary Smith², Richard Buxton², and David Dubowitz^2
1Bioengineering, University of California San Diego, La Jolla, CA, United States, ²Radiology, University of California San Diego, La Jolla, CA, United States

How does sustained hypoxia affect the brain oxygen metabolism (CMRO₂) response to a stimulus? We applied a novel multimodal Bayesian approach for estimating the CMRO₂ response to stimuli in human subjects after 6 hrs, 2 days and 7 days of sustained hypoxia (3800m elevation, 12.5% O₂). The acute response to a CO₂ challenge was a significant reduction in CMRO₂, and this did not change with up to a week of acclimatization. In contrast, the CMRO₂response to a visual stimulus was significantly reduced for up to 2 days of sustained hypoxia, but recovered to the normoxic response by 7 days.

Harshan Ravi^1,2, Peiying Liu¹, Shin-Lei Peng¹, and Hanzhang Lu^1
1Advanced Imaging Research Center, University of Texas at South Western Medical Center, Dallas, Tx, United States, ²Department of Bioengineering, University of Texas at Arlington, Arlington, TX, United States

Cerebrovascular reactivity (CVR) reflects the ability of the brain vasculature to dilate in response to a vasoactive stimulus. CVR is usually measured with CO2 inhalation while continuously acquiring BOLD MRI images. The biggest drawback of CVR is its poor sensitivity and reliability. Multiband EPI is a fast-imaging technology that allows the acquisition of multiple 2D slices simultaneously. In fMRI and DTI applications, it has been shown that multiband EPI provides an SNR advantage over conventional EPI. In this work, we examined the SNR benefit of multiband acquisition in CVR mapping, by comparing the data collected using multiband with those using conventional EPI.

Carlos C. Faraco¹, Jeroen C.W. Siero², Megan K. Strother¹, Daniel F. Arteaga¹, and Manus J. Donahue^1
1Radiology and Radiological Sciences, Vanderbilt University School of Medicine, Nashville, TN, United States, ²Department of Radiology, University Medical Center Utrecht, Utrecht, Netherlands

Identification of cerebrovascular compromise in the clinic is frequently performed using catheter angiography, which is insensitive to tissue-level hemodynamics and is sub-optimal for longitudinal monitoring. Alternatively, hypercarbic-normoxic BOLD fMRI may be used to assess cerebrovascular reactivity (CVR). As a significant proportion of patients for whom CVR assessment is desirable are hypoxic, and may be operating near reserve capacity, hypercarbic-hyperoxic (HC-HO) BOLD fMRI is advisable. However, HC-HO stimuli introduce several experimental confounds, including non-specific venous O2 rebinding to dHb. Here we demonstrate that administration of baseline HO, before HC-HO, resolves or reduces the drawbacks associated with HC-HO CVR-weighted BOLD imaging.

Shin-Lei Peng¹, Harshan Ravi¹, Min Sheng¹, Binu Thomas¹, and Hanzhang Lu^1
1Advanced Imaging Research Center, University of Texas Southwestern Medical Center, Dallas, Texas, United States

Hypercapnia challenge has been proposed as a method for the calibration experiment because CO2 does not alter brain metabolism. However, the assumption that hypercapnia challenge is iso-CMRO2 has not been validated. This study is to investigate a gas challenge is truly ¡§iso-metabolic¡¨ by adding a hypoxic component to the hypercapnic challenge (hypercapnic-hypoxia). We used TRUST MRI to monitor the subject¡¦s CMRO2 during hypercapnic-hypoxia challenge. Our data showed the neural suppression effect of hypercapnia appears to be nullified by the metabolic enhancement effect of hypoxia. We therefore propose hypercapnic-hypoxia is an iso-metabolic challenge that may be used for calibrated fMRI studies.

Shruti Agarwal¹, Raag Airan¹, Sachin K. Gujar¹, Haris I. Sair¹, and Jay J. Pillai^1
1Division of Neuroradiology, Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University School of Medicine, Baltimore, Maryland, United States

Reduced or impaired cerebrovascular reactivity (CVR) in patients with brain tumors and other cerebral diseases can generate false negative or spuriously decreased BOLD fMRI activation during performance of presurgical mapping of motor and language cortex in patients who are potentially candidates for lesion resection. This phenomenon, referred to as neurovascular uncoupling (NVU) is an under-recognized but nevertheless critical limitation of clinical BOLD fMRI. In this study we demonstrate the feasibility of minimization of false negative motor task-based activation through use of a novel breath-hold CVR-based calibration method to effectively compensate for NVU in patients with perirolandic low grade gliomas.

Molly Gallogly Bright¹, Joseph Whittaker¹, Ian Driver¹, and Kevin Murphy^1
1CUBRIC, School of Psychology, Cardiff University, Cardiff, Wales, United Kingdom

We previously identified coupling of vascular-neural networks in BOLD fMRI data using simultaneous neural stimuli and hypercapnia paradigms. We repeat this experiment without a hypercapnic stimulus, observing robust changes in physiology time-locked to the neural tasks. These task-correlated physiologic effects are significantly more associated with the vascular networks identified previously, and identify an additional vascular network in bilateral motor cortices. Task-correlated physiology significantly affects multiple vascular networks and may mimic, spatially and temporally, neural network activations.

Audrey P. Fan¹, Andreas Schaefer², Laurentius Huber², Steffen N. Krieger², Harald E. Moeller², Arno Villringer², and Claudine J. Gauthier^2,3
1Richard M. Lucas Center for Imaging, Stanford University, Stanford, CA, United States, ²Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, ³Concordia University, Montreal, Quebec, Canada

This study investigated whether BOLD signal changes during visual stimulus and gas challenge depend on baseline oxygen extraction fraction (OEF₀), as measured by quantitative susceptibility mapping (QSM) in cerebral veins. We also directly compared absolute OEF₀ values in the visual cortex by QSM and by a novel respiratory-calibration method (QUO2). In eight healthy volunteers, BOLD-ASL scans were acquired during gas breathing and visual stimulus, and gradient echo scans were acquired at rest for QSM reconstruction. Good fidelity was observed between BOLD-ASL signal changes and baseline OEF₀ by QSM, as predicted by biophysical models. Our findings also reveal encouraging concordance between absolute OEF₀ by QSM (30.6±2%) and by QUO2 (31.5±12%) that warrants examination in a larger cohort.

Alan J Stone¹ and Nicholas P Blockley^1
1FMRIB, Nuffield Department of Clinical Neurosciences, Oxford, United Kingdom

There is a need for an easy-to-implement MR method capable of providing quantitative parametric maps of OEF and DBV across the brain in a clinically feasible timescale. In this study a FLAIR GESEPI ASE sequence is used to provide a refined measure of R₂′ through optimisation of the data acquisition. The components of the sequence are selected to minimise confounds associated with CSF partial volume effects (FLAIR), macroscopic field inhomogeneities (GESEPI) and R₂-weighting (ASE) of the R₂′ signal. This affords a simplified application of the qBOLD signal model, allowing parametric maps of OEF and DBV to be produced in a streamlined process.