Quantification of MRS data in the frequency domain using a wavelet filter, an approximated Voigt lineshape model and prior knowledge.

Quantification of MRS spectra is a challenging problem when a large baseline is present along with a low signal to noise ratio. This work investigates a robust fitting technique that yields accurate peak areas under these conditions. Using simulated long echo time (1)H MRS spectra with low signal to noise ratio and a large baseline component, both the accuracy and reliability of the fit in the frequency domain were greatly improved by reducing the number of fitted parameters and making full use of all the known information concerning the Voigt lineshape. Using an appropriate first order approximation to a popular approximation of the Voigt lineshape, a significant improvement in the estimate of the area of a known spectral peak was obtained with a corresponding reduction in the residual. Furthermore, this improved parameter choice resulted in a large reduction in the number of iterations of the least-squares fitting routine. On the other hand, making use of the known centre frequency differences of the component resonances gave negligible improvement. A wavelet filter was used to remove the baseline component. In addition to performing a Monte Carlo study, these fitting techniques were also applied to a set of 10 spectra acquired from healthy human volunteers. Again, the same reduced parameter model gave the lowest value for chi(2) in each case.

A study of the headgroup motion of sphingomyelin using 31P NMR and an analytically soluble model.

A 31P NMR investigation has been carried out of the headgroup dynamics of sphingomyelin molecules in bilayers for the L alpha and L beta' phases. The resulting line shapes have been analysed in terms of a reduced-parameter model, using van Faassen's method for obtaining an analytic solution to first-order stochastic differential equations to simulate the line shapes of oriented and non-oriented samples. Our treatment results in good fits to measured data but using fewer parameters than traditional methods. Angles and correlation times (tau parallel and tau perpendicular) describing the geometry and dynamics of the headgroup are obtained by optimising the agreement between simulated and experimental data. The results are contrasted with those obtained for the lecithins DMPC and DPPC using a similar analysis. Not only are tau parallel and tau perpendicular now equal in value for the L alpha phase, but this value is also found to be nearly four times larger than the longest correlation time for the lecithins. We interpret this as evidence of inter- and intramolecular hydrogen bonding in the L(alpha) phase of sphingomyelin. In the L beta' phase of sphingomyelin, however, although tau(parallel) and tau(perpendicular) remain equal their value is now 32% smaller than that of the lecithins in the same phase. This indicates less difference between the fluidities in the headgroup region of the two phases of sphingomyelin as compared to that of the lecithins. Another significant difference between the L beta' phases is that the tilt angle for sphingomyelin is found to be nearly twice as large as for the lecithins. We argue that these combined observations point to the existence of hydrogen bonding in this phase also. Again in contrast to our previous work on lecithins, our results provide evidence of a negative diamagnetic anisotropy in sphingomyelin molecules, even in the L beta' phase. This is provided for in our model by the assumption that our unoriented samples consist of prolate ellipsoidal liposomes whose major axes are oriented parallel to the static magnetic field. The apparently different diamagnetic behaviour of sphingomyelin in the present work is due to the higher static field used rather than any intrinsic difference in this respect between sphingomyelin and the lecithins DMPC and DPPC.

The simulation of 31P NMR line shapes of lipid bilayers using an analytically soluble model.

Van Faassen's method for obtaining an explicit solution to a first order stochastic differential equation is applied to the simulation of 31P NMR line shapes of unoriented phospholipid bilayers in the Lalpha phase and of oriented bilayers in both the Lalpha and Lbeta' phases. The effects of the two slowest motions on the density matrix are described by the stochastic Liouville equation (SLE) which is solved analytically using the method of van Faassen. These two slowest motions are assumed to be a rotational re-orientation about the long molecular axis and a uniform wobble of this axis within a conical volume with re-orientation rates characterised by correlation times tau(parallel) and tau(perpendicular) respectively. In the present work the Hamiltonian contains the intramolecular dipole-dipole interaction between the phosphorous nucleus and the four closest methylene protons of the choline headgroup, as well as the anisotropic chemical shielding interaction. Hence the contribution to relaxation from cross correlation between the dipole-dipole and anisotropic chemical shielding interactions is included. The reorientation of the headgroup is assumed to be a rotation sufficiently fast to lead to complete axially symmetric averaging of the Hamiltonian about the rotational axis (the P-O11 bond axis). Evaluation of the line shape in the present work involves only numerical integration and is therefore less computationally demanding than the large matrix inversions involved in the approaches of Campbell, Freed et al. The present theory also uses fewer parameters than that of Dufourc et al. but nevertheless results in good agreement with these authors' measurements on DMPC bilayers, using a fixed value of 10 for the ratio tau(perpendicular)/tau(parallel) in the case of the Lalpha phase. However, in contrast to Dufourc et al., we find that these correlation times are equal for the Lbeta' phase. Finally, we have simulated the decoupled powder line shapes obtained from the Lbeta' phase of DPPC by Campbell and Meirovitch. Again, we get good agreement providing tau(perpendicular)=tau(parallel).

Choice of spectroscopic lineshape model affects metabolite peak areas and area ratios.

The use of Lorentzian model lineshapes leads to systematic errors in the quantification of in vivo (1)H NMR spectra. Experimental lineshapes are better modeled by the Voigt (mixed Lorentzian-Gaussian) function, leading to more accurate fits (reduced chi(2)). In this work, results from a group of 41 subjects are presented. It is shown that not only are the estimated metabolite peak areas affected by the choice of lineshape model, but so too are the metabolite ratios. For example, the NAA/choline ratio was 1.92 +/- 0.06 (mean +/- standard error) using the Lorentzian lineshape model and 1.85 +/- 0.05 using the Voigt lineshape model. The corresponding figures for NAA/creatine were 2.32 +/- 0.06 and 2. 10 +/- 0.05 respectively, which are significantly different for the two lineshape models. An explanation of this previously unreported effect is given. This finding clearly has serious implications for the methodology and reporting of spectroscopic studies.

An analytical derivation of a popular approximation of the Voigt function for quantification of NMR spectra.

The approximation of the Voigt line shape by the linear summation of Lorentzian and Gaussian line shapes of equal width is well documented and has proved to be a useful function for modeling in vivo (1)H NMR spectra. We show that the error in determining peak areas is less than 0.72% over a range of simulated Voigt line shapes. Previous work has concentrated on empirical analysis of the Voigt function, yielding accurate expressions for recovering the intrinsic Lorentzian component of simulated line shapes. In this work, an analytical approach to the approximation is presented which is valid for the range of Voigt line shapes in which either the Lorentzian or Gaussian component is dominant. With an empirical analysis of the approximation, the direct recovery of T(2) values from simulated line shapes is also discussed.

Chemical differences between long and short amosite asbestos: differences in oxidation state and coordination sites of iron, detected by infrared spectroscopy.

OBJECTIVES: Short fibres of amosite asbestos (SFA), obtained by ball milling of long fibres (LFA), have been shown to be less pathogenic than long fibres. Accumulating evidence suggests an important role for differences in surface chemistry between fibres. Iron has been implicated in the pathogenesis of asbestos fibres. In this study infrared (IR) spectroscopy was used to compare LFA and SFA in terms of the coordination and oxidation state of iron at the three cation sites (M1, M3, M1). METHODS: Infrared was used to examine LFA ad SFA, when dry and when hydrated in the presence and absence of the chelators desferroxamine and ferrozine. With appropriate software the proportions of iron and its oxidation states in the overlapping peaks were resolved and assigned, and the three coordination sites were identified. Data were obtained from 10 samples of both lengths of fibre for each of the four treatments. Iron release was also monitored. RESULTS: Iron was significantly more oxidised in LFA than SFA. Further oxidation of the dry fibres with water, ferrozine, or desferroxamine tended to abolish these differences. There were also significant differences between the proportions of iron held in the different coordination sites of the fibres. For LFA, a higher proportion of its iron was held in the cation sites coordinating less with iron and more with Mg. Interestingly, the sites coordinating single irons were significantly more oxidised than multiple sites. The single iron sites were more oxidised in LFA than SFA and were more readily oxidised by the treatments. CONCLUSIONS: Important chemical differences between LFA and SFA were found. There seemed to be some mobility of iron near the surface. Based on these data it is speculated that the 1 iron surface site may be important in pathogenesis.

Use of Voigt lineshape for quantification of in vivo 1H spectra.

Quantification of NMR visible metabolites by spectral modeling usually assumes a Lorentzian or Gaussian lineshape, despite the fact that experimental lineshapes are neither. To minimize systematic fitting errors, a mixed Lorentzian-Gaussian (Voigt) lineshape model was developed. When tested with synthetic FIDs, the Voigt lineshape model gave more accurate results (maximum error 2%) than either Lorentzian (maximum error 20%) or Gaussian models (maximum error 12%). The three lineshape models gave substantially different peak areas in an in vitro experiment, with the Voigt model having a much lower chi2 (2.1 compared with 5.2 for the Lorentzian model and 6.2 for the Gaussian model). In a group of 10 healthy volunteers, fitting of 1H spectra from cerebral white matter gave significantly different peak areas between the methods. Even when area ratios were taken, the Lorentzian model gave higher values (+5% for NAA/choline and +2% for NAA/creatine) than the Voigt lineshape model, whereas the Gaussian model gave lower values (-2% and -1%, respectively).

13C-NMR determination of the molecular dynamics of cholesterol in dimyristoylphosphatidylcholine vesicles.

Using 13C-NMR measurements of T1, T2 and the nuclear Overhauser enhancement factor at 50.32, 90.56 and 150.87 MHz, we have measured the dynamics of cholesterol in dimyristoylphosphatidylcholine (DMPC) vesicles from 28 to 50 degrees C. Using the model-free approach of Lipari and Szabo, we have found that at 37 degrees C the motion of the rigid steroid ring can be described by an equal contribution from two effective motions with correlation times of 63 and 0.85 ns. The C26 and C27 carbon atoms of cholesterol were found to have an effective correlation time of 8 +/- 2 ps and a value for the square of the generalised order parameter of 0.03 +/- 0.01. The corresponding values for the C25 carbon atom were 17 +/- 4 ps and 0.09 +/- 0.02, showing slower motion and greater order for this carbon atom, which is nearer to the rigid steroid ring. Apart from the effect of vesicle size on T2, no concentration dependence of the dynamics of cholesterol was detected over the cholesterol concentration range 2-30 mol%. The order parameters and correlation times from the present 13C-NMR experiments are shown to be compatible with those from 2H-NMR experiments. This establishes the validity of the present approach, which we are currently extending to low concentrations of cholesteryl oleate in DMPC vesicles.