Convolution spectral imaging

• Published in: Journal of magnetic resonance (1969) Vol. 79; no. 2; pp. 236 - 254
• Main Authors: Cockman, Michael D, Mareci, Thomas H
• Format: Journal Article
• Language: English
• Published: Orlando, FL Elsevier Inc 01.09.1988; Academic Press
• Subjects: Exact sciences and technology; Infrared, submillimeter wave, microwave and radiowave instruments, equipment and techniques; Instruments, apparatus, components and techniques common to several branches of physics and astronomy; Physics; Submillimeter wave, microwave and radiowave spectrometers; magnetic resonance spectrometers, auxiliary equipments and techniques; Convolution; Nuclear magnetic resonance; Nuclear magnetic resonance imaging; Spectrum
• ISSN: 0022-2364; 1557-8968
• DOI: 10.1016/0022-2364(88)90217-X

A new method of spectral imaging is introduced. It is based on a well-known threedimensional chemical-shift imaging technique in which spatial and chemical-shift data are collected separately. The new method, however, mixes spectral information into each spatial dimension. This is accomplished in part by requiring that the magnitude of the controlled inhomogeneity introduced by the applied gradients be smaller than the magnitude of the inherent inhomogeneity which gives rise to the chemical-shift phenomenon. This has the benefit of producing a greater signal-to-noise ratio than techniques in which the spectral information is obliterated by large applied gradients. A second requirement of the new method is that the length of a time delay, during which chemical-shift encoding occurs, and the amplitude of a gradient, during which spatial encoding occurs, are changed in concert with each other. This simultaneous stepping in the encoding dimension and reduced applied gradient amplitudes in each dimension have the effect of creating modulation functions whose Fourier transforms are the convolutions of functions of spatial position and spectral frequency. The new technique is therefore called “convolution spectral imaging.” Because the spectral information does not comprise a separate dimension of a data set, a reduction in the number of dimensions required for acquisition and display occurs. This reduces both measurement time and processing requirements. Convolution spectral imaging is most effective when the spectral resonances of the sample are well-separated and when the sample dimensions are small. Thus the technique may be best suited for microscopic NMR imaging at high magnetic field strengths.