An fMRI study of the effect of amphetamine on brain activity.

Functional magnetic resonance imaging was used to evaluate the effects of oral d-amphetamine on brain activation elicited by auditory and simple motor tasks in ten normal right-handed subjects. We measured the percent signal change and number of voxels activated by a tone discrimination task and a right hand finger-tapping task after 20 mg of d-amphetamine and after placebo. Compared to placebo, amphetamine significantly increased the number of activated voxels in the left and right primary auditory cortices during the tone discrimination task and increased the number of activated voxels in the ipsilateral primary sensorimotor cortex and right middle frontal area during the motor task. Although highly specific vascular effects of drug cannot be ruled out as an explanation, these results could also mean that amphetamine increases the neuronal activity associated with each of these two tasks.

Creating a library of generalized Fourier sampling patterns for irregular 2D regions of support.

Multiple-region MRI (mrMRI) represents a generalization of the Shannon sampling theorem to permit sparse k-space sampling whenever the scanned object or its high-contrast edges are confined to multiple known regions. Use of an optimal mrMRI sampling pattern produces an image with root-mean-squared (RMS) noise over the supporting regions equal to the RMS noise in a conventional Fourier image with the same total area of support. Analytical solutions for such sampling patterns have been described previously for all arrangements of two or three (noncollinear) supporting regions. This work describes a robust numerical method for creating a library of optimal and near-optimal mrMRI sampling patterns for more complicated geometries. The average noise amplification over all sampling patterns in the demonstration library was only 4%, with 30% of the sampling patterns resulting in no noise amplification whatsoever.

Detecting brain activation in FMRI data without prior knowledge of mental event timing.

Almost all methods of detecting brain activation in fMRI data depend on prior knowledge of mental event timing. For example, the investigator may be required to stipulate the short time intervals during which mental activity occurs. In addition, the hemodynamic response to mental activity is often assumed to be linearly additive, and the shape of that response is frequently estimated or modeled. Analysis methods that do not make these assumptions still require prior knowledge of characteristics of the spatial distribution of neural activity. This paper describes a new method of analyzing fMRI data that does not rely on any of these assumptions. Instead, our approach is based on the following simple premise: the time course of signal in activated voxels will not vary significantly when an entire task protocol is repeated by the same individual. The model-independence of this approach makes it suitable for "screening" fMRI data for brain activation that may have unanticipated timing. Retrospective examination of the time course of the detected signals may be used to understand the nature of the activity. We demonstrate the method by using it to detect brain activation in two subjects who performed hand sensorimotor tasks according to block and single-trial designs.

The mechanical state of intracranial tissues in elderly subjects studied by imaging CSF and brain pulsations.

The biomechanical properties of intracranial tissues influence the mechanical coupling of brain and CSF oscillations to the driving vascular pulsations. Dynamic phase contrast MRI was used to measure the transfer functions that characterize these couplings in normal elderly subjects and patients with Alzheimer's disease. The transfer functions of both groups were significantly different from the previously reported transfer functions of normal young subjects. The data show that vascular pulsations tend to cause greater spinal cord movements and smaller CSF oscillations in the older subjects than in the younger ones. These results are likely to be due to age-related changes in the mechanical state of intracranial tissues.

Reorganization of the hand somatosensory cortex following perinatal unilateral brain injury.

Functional magnetic resonance imaging was used to map the hand somatosensory cortices of nine hemiparetic young adult patients with perinatal unilateral brain injury in the sensorimotor area and five normal subjects. Stimulation of the paretic hand by periodic manual squeezing produced activation in the contralateral hemisphere of three patients and in the ipsilateral hemisphere of three other patients. Paretic hand stimulation produced no activation in either hemisphere of the remaining three patients. Therefore, one-third of the patients demonstrated functional "plasticity" of the brain in the form of inter-hemispheric relocation of the hand somatosensory function. The volume and pattern of activation for both hands was altered for those patients that showed evidence of cortical reorganization to the opposite hemisphere. This differs from the hand motor system, which exhibited inter-hemispheric reorganization in a higher proportion of a related group of hemiparetic subjects.

Multiple region MRI.

Traditional Fourier MR imaging (FT MRI) utilizes the Whittaker-Kotel'nikov-Shannon (WKS) sampling theorem. This theorem specifies the spatial frequency components which need to be measured to reconstruct an image with a known field of view (FOV). In this paper, we generalize this result in order to find the optimal k-space sampling for images that vanish except in multiple, possibly non-adjacent regions within the FOV. This provides the basis for "multiple region MRI" (mrMRI), a method of producing such images from a traction of the k-space samples required by the WKS theorem. Image reconstruction does not suffer from noise amplification and can be performed rapidly with fast Fourier transforms, just as in conventional FT MRI. The mrMRI method can also be used to reconstruct images that have low spatial-frequency components throughout the entire FOV and high spatial frequencies (i.e. edges) confined to multiple small regions. The greater efficiency of mrMRI sampling can be parlayed into increased temporal or spatial resolution whenever the imaged objects have signal or "edge" intensity confined to multiple small portions of the FOV. Possible areas of application include MR angiography (MRA), interventional MRI, functional MRI, and spectroscopic MRI. The technique is demonstrated by using it to acquire Gd-enhanced first-pass 3D MRA images of the carotid arteries without the use of bolus-timing techniques.

Locally focused contrast-enhanced carotid MRA.

With conventional Fourier transform (FT) magnetic resonance imaging (MRI), it is difficult to perform contrast-enhanced three-dimensional (3D) MR angiography (MRA) with the temporal and spatial resolution necessary to depict the carotid arteries. However, locally focused (LF) MRI is a more efficient method that utilizes prior knowledge of the image content to reconstruct images from sparse k-space samples. In this paper, we show how LF MRI can be used to perform high-resolution gadolinium (Gd)-enhanced 3D carotid MRA in less than 10 seconds. First, the accuracy of the technique was demonstrated by comparing LF and conventional (FT) images of a vascular phantom. Then the method was used to perform Gd-enhanced 3D MRA of a patient's carotid arteries. Instead of using bolus timing, the arterial phase was retrospectively identified in a consecutive series of images, just as in X-ray angiography.

Locally focused MRI of interventions.

Certain interventional MR procedures would benefit from T2-weighted imaging because of the sensitivity of T2-weighted images to tissue damage and target lesion contrast. To acquire such images with reasonable temporal resolution, a single-shot acquisition should be used because of the inherently long TR needed for T2 weighting. Unfortunately, most scanners require long readout times (eg, greater than 150 msec) and high bandwidths (eg, greater than 120 kHz) to perform conventional single-shot imaging with high spatial resolution. The resulting images are thus degraded by unacceptable artifacts and noise levels. This study illustrates how to create locally focused MR images that have high spatial resolution in a region of interest and lower spatial resolution elsewhere in the image. Because these images can be created from sparse k-space data, a scanner with modest gradients (eg, 10 mT/m maximal amplitude, 500 microsec minimal rise time) can acquire them after a single excitation with relatively short readout time and low bandwidth. This technique may make it practical to monitor interventions with T2-weighted imaging. The method was illustrated by reconstructing dynamic changes, which were simulated experimentally by moving objects in the vicinity of a normal human head.

Assessment of the biomechanical state of intracranial tissues by dynamic MRI of cerebrospinal fluid pulsations: a phantom study.

We used a cranial phantom to investigate how intracranial mechanical factors [brain compliance and the resistance to the flow of cerebrospinal fluid (CSF)] affect the way in which CSF pulsations are driven by pulsatile transcranial blood flow. Dynamic phase-contrast magnetic resonance imaging (MRI) was used to measure the transfer function between vascular pulsations and pulsatile response of the CSF below the foramen magnum of the phantom. We found that the coupling between the high frequency components of cervical CSF flow and transcranial blood flow was decreased when the phantom was modified to simulate increased brain compliance and increased resistance to CSF flow.

Using prior knowledge of human anatomy to constrain MR image acquisition and reconstruction: half k-space and full k-space techniques.

In this note, we demonstrate how to utilize prior knowledge of human cranial anatomy to constrain full k-space and half k-space acquisition and reconstruction of 128 times 128 images. We used a database of magnetic resonance head images to derive new basis functions which represent the most important features of the head. the "training" images were also used to derive formulas for reconstructing head images from a subset of the usual 128 phase-encoded signals and to determine the optimal k-space locations of those signal measurements. We used this algorithm, called Feature-Recognizing MRI, to reconstruct 128 times 128 head images from 50-60% of the signals filling the full k-space. Furthermore, we combined the algorithm with a conventional half k-space technique to create 128 times 128 images from only 60% of the 80 signals required by the usual unconstrained half k-space imaging. Thus, the prior knowledge represented by the image database, together with a half k-space technique, made it possible to construct accurate magnetic resonance images from only 30-40% of the complete set of 128 signals. In other words, a database of head images was used to devise a 1/3 k-space method for imaging the head.

2D locally focused MRI: applications to dynamic and spectroscopic imaging.

Conventional magnetic resonance images have uniform spatial resolution across the entire field of view. A method of creating MR images with user-specified spatial resolution along one dimension of the field of view was described recently by the authors. This paper presents the 2D generalization of this technique, which allows the user to specify arbitrary spatial resolution in arbitrary 2D regions. These images are reconstructed from signals that sparsely sample the k-space representation of the image. Therefore, locally focused images can be acquired in less time than that required by Fourier imaging with uniformaly high resolution. In this paper the authors show how to increase the temporal resolution of dynamic imaging (e.g., interventional imaging) by using high resolution in areas of expected change and lower resolution elsewhere. Alternatively, by matching the local spatial resolution to the expected edge content of the image, it is possible to avoid the localized truncation artifacts that mark Fourier images reconstructed from the same number of signals. For example, the authors show how proton spectroscopic images of the head may be improved by using high resolution in the neighborhood of scalp lipids that might otherwise cause truncation artifacts.

Frameless stereotaxy with real-time tracking of patient head movement and retrospective patient-image registration.

The accuracy of a novel frameless stereotactic system was determined during 10 surgeries performed to resect brain tumors. An array of three charge-coupled device cameras tracked the locations of infrared light-emitting diodes on a hand-held stylus and on a reference frame attached to the patient's skull with a single bone screw. Patient-image registration was achieved retrospectively by digitizing randomly chosen scalp points with the system and fitting them to a scalp surface model derived from magnetic resonance (MR) images. The reference frame enabled continual correction for patient head movements so that registration was maintained even when the patient's head was not immobilized in a surgical clamp. The location of the stylus was displayed in real-time on cross-sectional and three-dimensional MR images of the head; this information was used to predict the locations of small intracranial lesions. The average distance (and standard deviation) between the actual position of the mass and its stereotactically predicted location was 4.8 +/- 3.5 mm. The authors conclude that frameless stereotaxy can be used for accurate localization of intracranial masses without resorting to using fiducial markers during presurgical imaging and without immobilizing the patient's head during surgery.

Hemodynamically independent analysis of cerebrospinal fluid and brain motion observed with dynamic phase contrast MRI.

Brain and cerebrospinal fluid (CSF) movements are influenced by the anatomy and mechanical properties of intracranial tissues, as well as by the waveforms of driving vascular pulsations. The authors analyze these movements so that the purely hemodynamic factors are removed and the underlying mechanical couplings between brain, CSF, and the vasculature are characterized in global fashion. These measurements were used to calculate a set of impulse response functions or modulation transfer functions, characterizing global aspects of the vasculature's mechanical coupling to the intracranial tissues, the cervical CSF, and the cervical spinal cord. These functions showed that a sudden influx of blood into the head was rapidly accommodated by some type of intracranial reserve or capacity. After this initial response, an equal volume of CSF was driven through the foramen magnum over the next 200-300 ms as the intracranial reserve relaxed to its base-line state.

Locally focused MRI.

Conventional magnetic resonance images are reconstructed by Fourier transformation and have uniform spatial resolution across the entire field of view (FOV). This paper describes a way of creating MR images that have higher spatial resolution in some areas than in others. High resolution imaging can be confined to just those areas where it is necessary to resolve strong edges without truncation artifact. Such locally focused images can be acquired in less scan time than that required to image the entire FOV with uniformly high resolution. Images are reconstructed from a subset of the usual phase-encoded signals required to create a uniformly well-resolved image. The measured signals are usually nonuniformly scattered in k-space. Functional and interventional imaging may benefit from this technique, which makes it possible to acquire a rapid series of dynamical images that have high resolution in areas of expected change and lower resolution elsewhere. Spectroscopic images may be improved by using high resolution in the neighborhood of sharp edges (e.g., scalp lipids) that might otherwise cause truncation artifacts.

On the relationship between feature-recognizing MRI and MRI encoded by singular value decomposition.

This paper describes the similarity between two methods of non-Fourier MRI: feature-recognizing MRI (FR MRI) and MRI with encoding by singular value decomposition (SVD MRI). Both methods represented images as truncated expansions of non-Fourier basis functions; these basis images were derived from prior image data by using closely-related mathematical techniques: the Karhunen-Loeve decomposition (or principal components analysis) and singular value decomposition, respectively. We demonstrate that FR and SVD MRI are equivalent in the following sense: given the same prior image data, they lead to exactly the same basis functions. FR MRI utilized prior images of the same body part in many "training" subjects, thought to be similar to the "unknown" subject to be imaged. SVD MRI utilized a single prior image of one subject in order to perform dynamic imaging of that subject. We demonstrate that the basis function expansion derived from a single prior image may not be capable of representing new features (features not found in the prior image). Therefore, the SVD basis functions may be inappropriate for dynamic imaging.

Using an image database to constrain the acquisition and reconstruction of MR images of the human head.

A training set of MR images of normal and abnormal heads was used to derive a complete set of orthonormal basis functions which converged to head-like images more rapidly than Fourier basis functions. The new image representation was used to reconstruct MR images of other heads from a relatively small number of phase-encoded signal measurements. The training images also determined exactly which phase-encoded signals should be measured to minimize image reconstruction error. These signals were nonuniformly scattered throughout k-space. Experiments showed that head images reconstructed with the new method had less serious truncation artifacts than conventional Fourier images reconstructed from the same number of signals. The resulting images were characterized by spatially variable spatial resolution and were particularly well-resolved in regions where the training images had structural detail.

Functional magnetic resonance studies of the reorganization of the human hand sensorimotor area after unilateral brain injury in the perinatal period.

Functional magnetic resonance imaging was used to map the hand sensorimotor area of hemiparetic adolescents and young adults who had suffered unilateral brain damage in the perinatal period. Unlike normal subjects, who exhibit cortical activation primarily contralateral to voluntary finger movements, the hemiparetic patients' intact hemispheres were equally activated by contralateral and ipsilateral finger movements. Our findings are consistent with previous clinical observations and animal experiments which suggest that the immature brain is able to reorganize in response to focal injury.

Retrospective registration of X-ray angiograms with MR images by using vessels as intrinsic landmarks.

Conventional x-ray angiography (XRA) images are projections of the vasculature with high spatial and temporal resolution, while magnetic resonance (MR) angiography (MRA) and MR imaging data show the three-dimensional locations of vessels relative to brain parenchyma. The authors have developed a retrospective method of registering these studies, which makes it practical to produce multimodality displays of this complementary information. Registration was performed by matching vessels seen on both XRA and MRA images. First, the authors determined the coordinates of the center lines of a few "landmark" vessels on the XRA image and the three-dimensional locations of the corresponding intraluminal voxels in the MRA volume. Registration was performed by rotating and translating the MRA-MR imaging volume until the perspective projection of the MRA landmark vessels matched the corresponding vessel center lines on the XRA image. Experiments with phantoms and patients indicated that the two studies were registered with an average error of less than 2 mm. A linked-cursor display was developed to show correspondence between points on the registered XRA and MRA-MR images.

Functional mapping of human motor cortical activation with conventional MR imaging at 1.5 T.

A conventional 1.5-T magnetic resonance (MR) imager was used to detect signal intensity changes on T2*-weighted images of human motor and sensory cortices during performance of hand and tongue movements. Narrow receiver bandwidths were used to improve the signal-to-noise ratio. Protocols consisting of baseline, motor task, rest, and second motor task periods were performed by nine volunteers. Two-dimensional cross correlation was applied to correct in-plane translation and rotation of the head during the imaging session before the control images were subtracted from the task images. Measurements obtained during finger movement tasks indicated a 3%-8% increase in signal intensity near the contralateral central sulcus and smaller ipsilateral signal intensity increases. Bilateral signal intensity increases were also observed during tongue movement studies. A retrospective image registration technique was used to map the signal changes onto conventional anatomic images, which were used to create integrated three-dimensional models of brain structure and function. These integrated images showed that the highest signal intensity due to hand movement was near the putative central sulcus.