Spectral fitting routines

FIGURE 58-2 Hydrogen ( H) spectrum from a normal human brain at 4 Tesla field strength. The spectrum is very complicated, comprised of many overlapping peaks which are difficult to resolve from each other. Spectral analysis routines make use of spectral models which are constructed from the individual spectra acquired from each biomolecule from in vitro solutions. This model is then fitted to the raw data and approximate concentrations for each biomolecule are extracted... 

Determine the T, values of the individual carbon nuclei by analysing the data from the C Inversion Recovery experiment using the interactive fit routine of ID WINNMR. For further information consult Modern Spectral Analysis, volume 3 of this series, and the Help tool of ID WIN-NMR. Add the T, values to the C NMR data table. Try to rationalise the T, values with respect to the evaluated structure and the molecular dynamics of the investigated molecule. 

One widely used application of clinical MRS spectral simulation is the creation of prior information for spectral analysis and fitting routines. Well-defined metabolite prior information results in more consistent and complete estimations of the actual data. One example of a parametric model used to fit clinical MRS data is shown below. 

Computations are completed through an iteration process. There is no exact tristimulus match, but a best-fit spectral curve match is calculated. This is not a selective spectral match. Formulations are done using a spectral matching routine. 

Modem pRS systems are accompanied with powerful spectral acquisition and analysis software, which enables the creation of ID (cross section), 2D, and 3D maps of various features from the ID, 2D, or 3D array of spatially resolved Raman spectra. Various features that can be routinely mapped include intensity variations of specific peaks (by plotting the user-defined peak intensity or integrated area under the peak), intensity ratio of two different bands, peak position (by user-defined peak fitting routines such as Gaussian, Lorentzian), and peak widths. The obtained images can be further processed to highlight the spatial variations of the acquired spectra. For example. Boolean maps, which present a binary... 

All impactor and filter samples were analyzed for up to 45 elements by instrumental neutron activation analysis (INAA) as described by Heft ( ). Samples were irradiated simultaneously with standard flux monitors in the 3-MW Livermore pool reactor. The x-ray spectra of the radioactive species were taken with large-volume, high-resolution Ge(Li) spectrometer systems. The spectral data were transferred to a GDC 7600 computer and analyzed with the GAMANAL code (1 ), which incorporates a background-smoothing routine and fits the peaks with Gaussian and exponential functions. 

As approximate fits to spectra, oscillator models often miss essential details in the physics of the material response. Spectra of real samples reveal the consequences of composition, structure, doping, oxidation or reduction, multiplicity of phases, contaminant or introduced charges, etc., on electronic structure. These consequences from sample preparation can qualitatively affect intermolecular forces. To the extent possible, the best procedure is to use the best spectral data collected on the actual materials used in force measurement or materials designed for particular force properties. Given the present progress in spectroscopy, such coupling of spectra and forces may soon become routine. 

When P > 3, exponential curve-fitting procedures for the WSGG spectral model become significantly more difficult for hand computation but are quite routine with the aid of a variety of readily available... 

One method used to isolate a X-ray line from unwanted background and noise, employs equilibrated filters. It consists of linking the concentration of interest to the difference between two measurements. The first is obtained by installing a transmission filter between the sample and the detector to isolate the characteristic radiation of the element wanted and the second by fitting an absorption filter which is opaque to this same radiation. This will enable, for example, to quantify the copper from its main spectral line by using two filters, one made of nickel and the other made of cobalt. The fluorescence originating from the filters themselves is a limiting factor in this method, which is reserved for routine measurements. 

These requirements, along with other recommendations found in the guidance documents available, constitute the fitness-for-purpose requirements for residue methods using mass spectral detection. An excellent discussion of some issues related to the performance of mass spectrometry methods is contained in a 2010 paper, including the misuse of the term sensitivity and the challenges of establishing lower limits for the routine application of mass spectrometric methods. ... 

Havelock Ellis remarked that What we call progress is the exchange of one nuisance for another nuisance. The advent of computerized instruments has certainly made routine a number of difficult spectroscopic measurements. Unfortunately, the sophistication of the procedures now available can result in erroneous results or interpretations if they are not used with caution. A classic example is the use of curve-resolving techniques. There is always a suspicion that a good fit between an observed spectral profile and a number of bands can be obtained providing that a sufficient number of the latter are included in the analysis. Clearly, in such cases a prior knowledge of the number of bands and their frequency would significantly increase our confidence in the results. 

---