Fast Eddy Current Compensation by Feedback Linearization Neural Networks: Applications in Diffusion-Weighted Echo Planar Imaging
Date Issued
2005
Date
2005
Author(s)
Hwang, San-Chao
DOI
en-US
Abstract
Diffusion-weighted magnetic resonance imaging (DWI) sensitizes the magnetic resonance images to the diffusive mobility of water and maps water diffusion in tissue. It can highlight the microstructural characteristics of biological tissues and serve as a useful imaging tool for both clinical diagnosis and basic medical research.
A large number of images with different magnitudes and directions of the diffusion sensitizing gradients need be acquired in order to estimate the diffusion properties. For efficiency, these images are usually acquired using diffusion-weighted echo-planar imaging sequence (DW-EPI). The rapid switching of the gradient pulses of DW-EPI can generate eddy currents in conducting surfaces surrounding the gradient coils. Although generation of eddy currents is greatly decreased in an active shielded gradient system, this can still occur especially when using large and rapidly rising and falling diffusion sensitization gradient pulses.
This study describes the application of the feedback linearization neural networks, known from neural network computing, to the problem of gradient preemphasis. This approach of preemphasis adjustment doesn’t require an iterative procedure between measurement and adjustment, therefore is essentially instantaneous in its execution. Based on our study, gradient compensation determined by our procedure effectively suppressed eddy current induced geometric distortion and spatial shift of diffusion-weighted EPI images. Comparing the manual preemphasis adjustments, this approach not only is reliable and accurate but also can reduce the spent time from several hours to several minutes. We have successfully applied this technique to the pig heart fiber tracking with diffusion tensor echo planar imaging (DT-EPI). In the future, the human brain white matter connectivity will be also studied.
Subjects
擴散權重迴訊平面影像
渦電流效應
線性迴歸類神經網路
DW-EPI
Eddy current
Feedback linearization neural networks
Preemphasis adjustment.
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-94-D83503049-1.pdf
Size
23.31 KB
Format
Adobe PDF
Checksum
(MD5):1c413e89da092f66479de482c2ffe679