This paper extends the use of principal-component analysis in fully automatic fashion using PCA, each spectrum must be spectral quantification to the estimation of frequency and phase individually phased, often by hand, and shifted in frequency. shifts in a single resonant peak across a series of spectra. The This requires significant time and introduces the possibility estimated parameters can be used to correct the spectra accordof observer bias especially in the adjustments of small phase ingly, resulting in more accurate peak-area estimation. Further, angles.
the removal of the variations in phase and frequency caused by
In this paper, we extend the use of PCA as an automatic instrumental and experimental fluctuations makes it possible to method of spectral quantification to determine the frequency determine more accurately the remaining variations, which bear misalignment and phase misadjustment in a single resonant biological significance. The procedure is demonstrated on simupeak across a series of spectra. This information can be used lated data, a 3D chemical-shift-imaging dataset acquired from a either to remove the variations due to frequency or phase, cylinder of inorganic phosphate (P i ), and a set of 736 31 P NMR in vivo spectra taken from a kinetic study of rat muscle energetics. i.e., in a reference peak, or to quantitate significant frequency In all cases, the procedure rapidly and automatically identifies the (or phase) shifts related to the problem being investigated. frequency and phase shifts present in the individual spectra. In In this paper, we will deal solely with frequency and phase the kinetic study, the procedure is used twice, first to adjust the shifts, since in our experience they are the major source of phase and frequency of a reference peak (phosphocreatine) and variations in the data. Moreover, the problem of correcting then to determine the individual frequencies of the P i peak in each for the linewidth variations requires an alternative mathematof the spectra which further can be used for estimation of pH ical treatment and will be addressed in a future publication.
changes during the experiment.