We report that visual stimulation produces an easily detectable (5-20%) transient increase in the intensity of water proton magnetic resonance signals in human primary visual cortex in gradient echo images at 4-T magnetic-field strength. The observed changes predominantly occur in areas containing gray matter and can be used to produce highspatial-resolution functional brain maps in humans. Reducing the image-acquisition echo time from 40 msec to 8 msec reduces the amplitude of the fractional signal change, suggesting that it is produced by a change in apparent transverse relaxation time T2. The amplitude, sign, and echo-time dependence of these intrinsic signal changes are consistent with the idea that neural activation increases regional cerebral blood flow and concomitantly increases venous-blood oxygenation.Magnetic-resonance imaging (MRI) of rodent brains at high (7-T) magnetic-field strength shows proton signal-intensity alterations related to blood oxygenation in regions close to local blood vessels (1-3). We have termed this phenomenon blood oxygenation-level-dependent (BOLD) contrast and have demonstrated that the underlying mechanism is a magnetic-susceptibility variation caused by deoxyhemoglobin, an endogenous paramagnetic contrast agent. It was further demonstrated that this magnetic-susceptibility effect could be used to measure in vivo changes in hemodynamics. For example, pharmacologically induced changes in cerebral blood flow and oxygen utilization produce measurable changes in BOLD contrast in the rat cerebral cortex. Similar results have recently been demonstrated in cat brain (4).There is increased evidence that a local elevation in human-brain venous-blood oxygenation accompanies an increase in neuronal activity (5-8). For example, positron emission tomography imaging experiments demonstrate stimulation-produced increases in regional cerebral blood flow without significantly changing local oxygen use, thus predicting an elevation in venous-blood oxygenation (6, 7). This result suggested that BOLD contrast imaging could be used to map human mental operations. To examine whether detectable intrinsic magnetic-susceptibility changes are produced in the human brain in response to neuronal activation, we studied the effect of visual stimulation on gradient echo images of human visual cortex acquired at high-magneticfield strength. In general, high-field strength increases the magnitude of susceptibility contrast effects, accentuating BOLD contrast. MATERIALS AND METHODSMRI experiments were done with a 4-T whole-body imaging system with actively shielded gradient coils [Sisco (Sunnyvale, CA)/Siemens (Erlangen, F.R.G.)]. Approval for these human experiments was obtained from the institutional review board of the University of Minnesota Medical School. Radiofrequency power deposition was kept two orders of magnitude below Food and Drug Administration specificabsorption rate guidelines. A snugly fitted head holder with a curved-surface radiofrequency coil (14 cm in diameter) was used to limit...

It recently has been demonstrated that magnetic resonance imaging can be used to map changes in brain hemodynamics produced by human mental operations. One method under development relies on blood oxygenation level-dependent (BOLD) contrast: a change in the signal strength of brain water protons produced by the paramagnetic effects of venous blood deoxyhemoglobin. Here we discuss the basic quantitative features of the observed BOLD-based signal changes, including the signal amplitude and its magnetic field dependence and dynamic effects such as a pronounced oscillatory pattern that is induced in the signal from primary visual cortex during photic stimulation experiments. The observed features are compared with the results of Monte Carlo simulations of water proton intravoxel phase dispersion produced by local field gradients generated by paramagnetic deoxyhemoglobin in nearby venous blood vessels. The simulations suggest that the effect of water molecule diffusion is strong for the case of blood capillaries, but, for larger venous blood vessels, water diffusion is not an important determinant of deoxyhemoglobin-induced signal dephasing. We provide an expression for the apparent in-plane relaxation rate constant (R2*) in terms of the main magnetic field strength, the degree of the oxygenation of the venous blood, the venous blood volume fraction in the tissue, and the size of the blood vessel.

A hemispheric asymmetry in the functional activation of the human motor cortex during contralateral (C) and ipsilateral (I) finger movements, especially in right-handed subjects, was documented with nuclear magnetic resonance imaging at high field strength (4 tesla). Whereas the right motor cortex was activated mostly during contralateral finger movements in both right-handed (C/I mean area of activation = 36.8) and left-handed (C/I = 29.9) subjects, the left motor cortex was activated substantially during ipsilateral movements in left-handed subjects (C/I = 5.4) and even more so in right-handed subjects (C/I = 1.3).

Purpose To develop a strategy for training a physics‐guided MRI reconstruction neural network without a database of fully sampled data sets. Methods Self‐supervised learning via data undersampling (SSDU) for physics‐guided deep learning reconstruction partitions available measurements into two disjoint sets, one of which is used in the data consistency (DC) units in the unrolled network and the other is used to define the loss for training. The proposed training without fully sampled data is compared with fully supervised training with ground‐truth data, as well as conventional compressed‐sensing and parallel imaging methods using the publicly available fastMRI knee database. The same physics‐guided neural network is used for both proposed SSDU and supervised training. The SSDU training is also applied to prospectively two‐fold accelerated high‐resolution brain data sets at different acceleration rates, and compared with parallel imaging. Results Results on five different knee sequences at an acceleration rate of 4 shows that the proposed self‐supervised approach performs closely with supervised learning, while significantly outperforming conventional compressed‐sensing and parallel imaging, as characterized by quantitative metrics and a clinical reader study. The results on prospectively subsampled brain data sets, in which supervised learning cannot be used due to lack of ground‐truth reference, show that the proposed self‐supervised approach successfully performs reconstruction at high acceleration rates (4, 6, and 8). Image readings indicate improved visual reconstruction quality with the proposed approach compared with parallel imaging at acquisition acceleration. Conclusion The proposed SSDU approach allows training of physics‐guided deep learning MRI reconstruction without fully sampled data, while achieving comparable results with supervised deep learning MRI trained on fully sampled data.