Koichi Oshio

Single-shot 3D imaging techniques improve arterial spin labeling perfusion measurements

Arterial spin labeling (ASL) can be used to measure perfusion without the use of contrast agents. Due to the small volume fraction of blood vessels compared to tissue in the human brain (typ. 3–5%) ASL techniques have an intrinsically low signal‐to‐noise ratio (SNR). In this publication, evidence is presented that the SNR can be improved by using arterial spin labeling in combination with single‐shot 3D readout techniques. Specifically, a single‐shot 3D‐GRASE sequence is presented, which yields a 2.8‐fold increase in SNR compared to 2D EPI at the same nominal resolution. Up to 18 slices can be acquired in 2 min with an SNR of 10 or more for gray matter perfusion. A method is proposed to increase the reliability of perfusion quantification using QUIPSS II derivates by acquiring low‐resolution maps of the bolus arrival time, which allows differentiation between lack of perfusion and delayed arrival of the labeled blood. For arterial spin labeling, single‐shot 3D imaging techniques are optimal in terms of efficiency and might prove beneficial to improve reliability of perfusion quantitation in a clinical setup. Magn Reson Med 54:491–498, 2005.

The ASYST Software for Scientific Computing

The acceptance of ASYST by the scientific community could dramatically change the way scientific data are handled and reduce the need for extensive in-house software development for many applications. However, there is no substitute for a well-conceived use of any software. We feel that the full acceptance of the ideas and concepts pioneered by Adaptable Laboratory Software and other software houses will depend on the ability of the scientific community to fully test and verify the procedures used by such products. Only then can the results produced by these software packages be subjected to confirmation which is crucial to rigorous scientific endeavor.

Iterative estimation of T2 to correct echo planar magnetic resonance images

An iterative algorithm has been developed to correct for distortions in echo planar images caused by short T2 components. The values of T2 are initially estimated from a set of images produced by the inverse Fourier-transform of the geometric mean of Hermitian symmetric points. The estimated T2 values are then used to compute k-space data, which when compared with the true data, provide error data sets and corresponding images to iteratively refine the estimates of T2. Images corrected for T2 decay are thereby generated at specified echo times. Computer simulation studies of several phantoms show good convergence under a variety of conditions. This procedure should enable wider data acquisition windows to be utilized in echo planar or spin echo images, leading to better resolution or better signal-to-noise ratio. >

Correction of T2 distortion in multi-excitation RARE sequence

Correction schemes have been implemented to correct for T2 distortions in a multiexcitation RARE (rapid acquisition with relaxation enhancement) sequence where data from multiple echoes and multiple excitations are combined. Computer simulation studies and human imaging studies have been conducted to develop and test the correction procedures. A direct method and an iterative technique have been investigated. The direct technique utilizes Hermitian symmetry of the T2 weighted data and is shown to reduce distortions in T2 weighted images. The iterative scheme begins with an estimation of T2, wherefrom k-space data are computed and compared to the true data to provide error images. The error images are then used to refine iteratively the reconstructed images at a specified echo time. The iterative procedure has been used to improve T1 weighted images acquired through a sequence based on acquisition of two half-plane Fourier samples. These correction techniques should enable a practical implementation of RARE for producing T1 and T2 weighted images comparable to standard spin echo images.

A new generation of SPECT and PET systems based on position sensitive photomultipliers

The design and image reconstruction aspects of a head SPECT (single photon-emission computerized tomography), a body SPECT, and a body PET (photon emission tomography) system, all based on 13-cm-diameter position-sensitive photomultiplier tubes (PSPMTs) have been investigated. Wobbling is introduced to improve the resolution of the head SPECT from 1.2 cm to 0.9 cm at 10 cm. By combining data from adjacent tubes, the resolution of the body ring SPECT is improved from 2.1 cm to 1.7 cm at 20 cm. Sampling gaps are reduced in the ring PET after one rotational increment. Results of computer simulation studies to reconstruct tomographic images from these systems are presented. The PSPMT-based imaging systems greatly facilitate dynamic imaging and should lead to significant reductions in electronic components over existing systems. >

Neural network approach to segmentation of magnetic resonance head images

A novel image segmentation scheme based on a neural network has been implemented to segment magnetic resonance head images. A three‐layer perceptron‐type neural network, trained with backward error propagation algorithm was used. The scheme utilizes first‐echo intensity and computed T2 values to construct a two‐parameter space for classification. After training on a selected slice, the method successfully segments all slices for a given subject without any further human interaction.

Simulation of phase effects in magnetic resonance flow imaging

Using a phantom containing stationary and flowing regions, the phase-related effects of flow on spin echo and echo planar images have been investigated. The role of parameters such as flow direction, mean velocity, velocity modulation, gradient pulses, echo time and data sampling time in producing artifacts and the reduction of these artifacts by a flow compensating pulse have been examined. It is shown that, due to the finite data sampling window, the flow compensating pulse cannot correct fully for phase shifts even for first-order motion. >

SQUID neuromagnetometric reconstruction of brain activity

Abstract not available.