Kensuke Sekihara

Generalized Wiener estimation of three-dimensional current distribution from biomagnetic measurements

Proposes a method for estimating three-dimensional (3-D) biocurrent distribution from spatio-temporal biomagnetic data. This method is based on the principle of generalized Wiener estimation, and it is formulated based on the assumption that current sources are uncorrelated. Computer simulation demonstrates that the proposed method can reconstruct a 3-D current distribution where the conventional least-squares minimum-norm method fails. The influence of noise is also simulated, and the results indicate that a signal-to-noise ratio of more than 20 for the uncorrelated sensor noise is needed to implement the proposed method. The calculated point spread function shows that the proposed method has very high spatial resolution compared to the conventional minimum norm method. The results of computer simulation of the distributed current sources are also presented, including cases where current sources are correlated. These results suggest that no serious errors arise if the source correlation is weak.

Image restoration from non-uniform magnetic field influence for direct Fourier NMR imaging

A new technique is proposed for NMR image restoration from the influence of main magnetic field non-uniformities. This technique is applicable to direct Fourier NMR imaging. The mathematical basis and details of this technique are fully described. Modification to include image restoration from non-linear field gradient influence is also presented. Computer simulation demonstrates the effectiveness of this technique for both Fourier zeugmatography and spin-warp imaging.

Spatially resolved nmr spectroscopy using phase-modulated spin-echo trains

Abstract A method for spatially resolved NMR spectroscopy is described. A spin-echo train induced by periodically inverting a field gradient is further phase modulated by prior applications of other field gradients. Each one-dimensional spin-echo train is converted to two-dimensional (spectroscopic and spatial) data by echo rearrangement to give two sets of echo arrays. Multidimensional Fourier transformation of phase-modulated echo arrays provides spatially resolved high-resolution spectra or spectroscopically resolved spin images, reducing measurement dimensions by one compared with the usual multidimensional Fourier transform method, and allowing high-speed spectroscopic studies of spatially inhomogeneous objects such as biological systems. Two techniques for expanding the spectral bandwidth, which is limited by the reciprocal of the field-gradient inversion period, are described. A proper combination of the two techniques can provide the widest bandwidth without unnecessarily increasing the total measurement time. Experimental demonstrations have been made using spatially one- and two-dimensional proton test samples.

High-speed spatially resolved NMR spectroscopy using phase-modulated spin-echo trains. Expansion of the spectral bandwidth by combined use of delayed spin-echo trains

Ces trains dechos de spin sont induits par la combinaison dinversions periodiques dun gradient de champ et des techniques de codage de phase par applications impulsionnelles dautres gradients de champ

Maximum-likelihood estimation of current-dipole parameters for data obtained using multichannel magnetometer

A method of reducing the influence of external noise magnetic field on the accuracy of estimating current dipole parameters is proposed. It utilizes the spatial correlation of external noises, and is applied to data measured using a multichannel magnetometer. Computer simulation demonstrates the effectiveness of the proposed method.<<ETX>>

Average-intensity reconstruction and Wiener reconstruction of bioelectric current distribution based on its estimated covariance matrix

Proposes two methods for reconstructing current distributions from biomagnetic measurements. Both of these methods are based on estimating the source-current covariance matrix from the measured-data covariance matrix. One method is the reconstruction of average current intensity distributions. This method first estimates the source-current covariance matrix and, using its diagonal terms, it reconstructs current intensity distributions averaged over a certain time. Although the method does not reconstruct the orientation of each current element at each time instant, it can retrieve information regarding the current time-averaged intensity at each voxel location using extremely low SNR data. The second method is Wiener reconstruction using the estimated source-current covariance matrix. Unlike the first method, this Wiener reconstruction can provide a current distribution with its orientation at each time instant. Computer simulation shows that the Wiener method is less affected by the choice of the regularization parameter, resulting in a method that is more effective than the conventional minimum-norm method when the SNR of the measurement is low.<<ETX>>

A method of measuring field-gradient modulation shapes. Application to high-speed NMR spectroscopic imaging

A method is proposed for measuring waveforms of a time-varying field gradient used for high-speed NMR imaging. It utilises a self-encoding gradient which calibrates the gradient modulation to be measured. Results are obtained in the form of a time integral of the gradient waveform, which can be used directly for Troppers image reconstruction. Spectroscopic one-dimensional imaging experiments have been performed using rectangular and cosine gradient modulationsl. It has been demonstrated that, in both cases, significant improvement can be achieved by making use of actually measured gradient modulations for the reconstruction. It has also been shown that in addition to Troppers reconstruction, appropriate deconvolution is necessary to obtain images with less distortion in the case of cosine modulation.

A new method of measuring static field distribution using modified Fourier NMR imaging

A new method for mapping the static magnetic field of NMR imagers is presented. The method uses a conventional Fourier imaging pulse-sequence with the exception that the time interval between the 90 degrees and 180 degrees pulses does not equal that between the 180 degrees pulse and spin-echo formation. In images obtained from the pulse sequence, the difference between those two time intervals causes phase shifts that depend on the static magnetic field distribution. Thus, the information on the field distribution can be extracted from these phase shifts. This method is quicker and more easily implemented than the method proposed by Maudsley et al. (1979, 1983, 1984).

Details of simulated annealing algorithm to estimate parameters of multiple current dipoles using biomagnetic data

The details of the simulated annealing algorithm proposed to estimate the parameters of multiple current dipoles using biomagnetic data are described. The effects of the choices of such numerical conditions as the amount of estimate transitions, the equilibrium criterion, and the temperature decrement on the algorithms performance are discussed. Incorrect results from the computer simulation obtained with inappropriate choices of such conditions are presented. The verification of a near-optimum convergence by reheating and reannealing is demonstrated. A modification of this algorithm is proposed for implementation by parallel computer, and the validity of this parallel algorithm is demonstrated by computer simulation.

Reduction of brain noise influence in evoked neuromagnetic source localization using noise spatial correlation

In magnetoencephalographic measurements, magnetic fields caused by spontaneous brain activities not related to the neural activities under study are often referred to as brain noise. This is because the accuracy in neural source localization is considerably degraded by such spontaneous neuromagnetic fields. This paper reports the experimental results of applying the previously proposed noise covariance method to reducing the degradation caused by brain noise and to improving the accuracy in localizing auditory-evoked neural sources. Firstly we present the results of our experiments using measured brain noise and computer-generated signal fields. These results confirm that the covariance method can, in principle, improve the accuracy of evoked neural source localization. Next, the method was applied to source localization for actual neuromagnetic fields evoked by speech sounds. The results obtained strongly suggest that the method is effective in processing actual evoked neuromagnetic data.