Co-reporter:Sanjeewa R. Karunathilaka, Gary W. Small
Analytica Chimica Acta 2017 Volume 987(Volume 987) pp:
Publication Date(Web):22 September 2017
DOI:10.1016/j.aca.2017.08.035
•A nocturnal hypoglycemic alarm based on near-infrared spectroscopy is proposed.•Differential spectra computed relative to an initial reference spectrum are used.•Classification methods place spectra into alarm or non-alarm classes.•The approach is tested with a tissue phantom based on keratin and collagen films.•Correct classifications are 93% for the most demanding test case investigated.An alarm algorithm for detecting episodes of nocturnal hypoglycemia is demonstrated in simulation studies that incorporate the use of a tissue phantom. Based on transmission spectra collected in the near-infrared combination region of 4000–5000 cm−1, pattern recognition methods are used to classify spectra into alarm and non-alarm data classes on the basis of whether or not they signify a glucose excursion below a user-defined hypoglycemic alarm threshold. A reference spectrum and corresponding glucose concentration are acquired at the start of the monitoring period, and absorbance values of subsequent spectra are computed relative to this reference. The resulting differential spectra reflect differential glucose concentrations that correspond to the differences in concentration between each spectrum and the reference. Given the alarm threshold, a database of calibration differential spectra are partitioned into those above and below the threshold. Piecewise linear discriminant analysis is then used to compute a classification model that can be applied to differential spectra collected during the monitoring period in order to identify spectra that signal glucose concentrations in the hypoglycemic range. This alarm algorithm is demonstrated in two multiple-day dynamic studies that incorporate a tissue phantom composed of films of keratin and collagen that approximate the thicknesses of the corresponding proteins found in human tissue.Download high-res image (168KB)Download full-size image
Co-reporter:Sanjeewa R. Karunathilaka and Gary W. Small
Analyst 2015 vol. 140(Issue 12) pp:3981-3988
Publication Date(Web):20 Apr 2015
DOI:10.1039/C5AN00258C
Near-infrared (near-IR) spectroscopy has been investigated for use in direct measurements of human tissue for the purpose of quantifying clinically relevant analytes such as glucose. Spectra collected by transmitting near-IR light through human tissue reveal the presence of both aqueous components and solid structures within the optical path of the measurement. Developing technology for use in making these measurements requires either the availability of human subjects or an in vitro experimental platform that can effectively simulate the spectroscopic properties of tissue. This paper describes the preparation and testing of films of collagen and keratin, the two proteins that comprise the bulk of the solid material in the human epidermis and dermis. By placing these films in the optical path of a near-IR spectrometer together with an aqueous sample cell, a phantom can be constructed that allows experiments to be performed that simulate noninvasive measurements of tissue. In this work, both constant and variable thickness films are prepared and evaluated by use of a regression fit to spectra of human tissue. The stability and spectral reproducibility of the prepared films are also assessed. The regression fits to the human subject spectra yield values of R2 ranging from 0.97 to 0.99 and the films are found to yield spectra that vary by less than a 2% relative standard deviation under three different reproducibility tests.
Co-reporter:Sanjeewa R. Karunathilaka and Gary W. Small
RSC Advances 2014 vol. 4(Issue 67) pp:35405-35414
Publication Date(Web):19 Aug 2014
DOI:10.1039/C4RA07579J
A threshold concentration monitoring procedure based on near-infrared (near-IR) spectroscopy is described for use in continuous process monitoring applications. The method is based on collecting an off-line reference sample and obtaining a near-IR spectrum and corresponding reference concentration at the start of the monitoring period. Subsequently, spectra are collected continuously and ratios are taken with respect to the reference spectrum. The resulting spectra in absorbance units are differential spectra whose effective analyte concentration (termed the differential concentration) is the difference in concentration relative to the concentration in the reference sample. By knowing the reference concentration and a user-specified threshold, a critical concentration can be defined that specifies the threshold in terms of differential concentrations. Determining whether the analyte concentration is within specification can then be addressed as a pattern classification problem and a qualitative classification model can be used to discriminate differential spectra that reflect the two possible states: within or outside of specification. A simulated biological process is used to test the methodology in which a dynamic system of glucose, lactate, urea, and triacetin in the mM range in phosphate buffer is monitored continuously to detect occurrences when the glucose concentration drops below a threshold of 3.0 mM. With the use of three sets of prediction data, one of which was collected 2.5 years after the calibration data, the monitoring algorithm is implemented with 100% successful detections and no false detections.
Co-reporter:Chamathca P. S. Kuda-Malwathumullage
Journal of Applied Polymer Science 2014 Volume 131( Issue 13) pp:
Publication Date(Web):
DOI:10.1002/app.40476
ABSTRACT
The commercial importance of polyamides (PAs) motivates the development of chemical analysis tools for use in characterizing their structure and properties. Near-infrared (IR) spectroscopy offers advantages in this regard because of its simplicity of sample preparation and compatibility with sample thicknesses on the order of millimeters. For applications in which the measurement of sample temperature is difficult with a conventional probe, the work presented here demonstrates the ability to determine the temperature of PA 66 directly from its near-IR spectrum. Temperature-induced changes in spectral shape in the 4000–5000 cm−1 region are extracted through application of the discrete wavelet transform, and the resulting preprocessed spectra are submitted to partial least-squares regression to construct predictive models for temperature. These models are tested across different samples of PA 66 and over a time span of 7 weeks. Errors in predicted temperatures averaged 1.50°C over the range of 21–105°C. © 2014 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2014, 131, 40476.
Co-reporter:Chamathca P. S. Kuda-Malwathumullage
Journal of Applied Polymer Science 2014 Volume 131( Issue 16) pp:
Publication Date(Web):
DOI:10.1002/app.40645
ABSTRACT
The hygroscopic nature of polyamide (PA) polymers motivates the development of analysis tools for use in assessing their moisture content. Among possible analysis techniques, near-infrared (near-IR) spectroscopy is non-destructive, requires little or no sample preparation, and is compatible with sample thicknesses on the order of mm. The work reported here makes use of transmission near-IR spectroscopy in the combination region (5000–4000 cm−1) to develop a protocol for assessing the moisture content of PA 66 samples directly from their spectral intensities after preprocessing with the standard normal variate transform and partial least-squares. The method is compatible with online or continuous monitoring applications and can be calibrated without the use of destructive reference measurements such as thermogravimetric analysis. The long-term calibration performance of the technique is evaluated, and on a scale of 0–100% moisture uptake, the standard error of prediction is found to average 1.4% over 6 months. © 2014 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2014, 131, 40645.
Co-reporter:Boyong Wan and Gary W. Small
Analyst 2011 vol. 136(Issue 2) pp:309-316
Publication Date(Web):18 Oct 2010
DOI:10.1039/C0AN00103A
A novel synthetic data generation methodology is described for use in the development of pattern recognition classifiers that are employed for the automated detection of volatile organic compounds (VOCs) during infrared remote sensing measurements. The approach used is passive Fourier transform infrared spectrometry implemented in a downward-looking mode on an aircraft platform. A key issue in developing this methodology in practice is the need for example data that can be used to train the classifiers. To replace the time-consuming and costly collection of training data in the field, this work implements a strategy for taking laboratory analyte spectra and superimposing them on background spectra collected from the air. The resulting synthetic spectra can be used to train the classifiers. This methodology is tested by developing classifiers for ethanol and methanol, two prevalent VOCs in wide industrial use. The classifiers are successfully tested with data collected from the aircraft during controlled releases of ethanol and during a methanol release from an industrial facility. For both ethanol and methanol, missed detections in the aircraft data are in the range of 4 to 5%, with false positive detections ranging from 0.1 to 0.3%.
Co-reporter:Boyong Wan, Gary W. Small
Analytica Chimica Acta 2010 Volume 681(1–2) pp:63-70
Publication Date(Web):29 November 2010
DOI:10.1016/j.aca.2010.09.022
Wavelet analysis is developed as a preprocessing tool for use in removing background information from near-infrared (near-IR) single-beam spectra before the construction of multivariate calibration models. Three data sets collected with three different near-IR spectrometers are investigated that involve the determination of physiological levels of glucose (1–30 mM) in a simulated biological matrix containing alanine, ascorbate, lactate, triacetin, and urea in phosphate buffer. A factorial design is employed to optimize the specific wavelet function used and the level of decomposition applied, in addition to the spectral range and number of latent variables associated with a partial least-squares calibration model. The prediction performance of the computed models is studied with separate data acquired after the collection of the calibration spectra. This evaluation includes one data set collected over a period of more than 6 months. Preprocessing with wavelet analysis is also compared to the calculation of second-derivative spectra. Over the three data sets evaluated, wavelet analysis is observed to produce better-performing calibration models, with improvements in concentration predictions on the order of 30% being realized relative to models based on either second-derivative spectra or spectra preprocessed with simple additive and multiplicative scaling correction. This methodology allows the construction of stable calibrations directly with single-beam spectra, thereby eliminating the need for the collection of a separate background or reference spectrum.
Co-reporter:Yusuf Sulub and Gary W. Small
Analytical Chemistry 2009 Volume 81(Issue 3) pp:1208
Publication Date(Web):January 8, 2009
DOI:10.1021/ac801746n
Multivariate calibration models based on synthetic single-beam near-infrared spectra are used to demonstrate the ability to maintain viable calibrations over extended time periods. Glucose is studied over the physiological concentration range of 1−30 mM in a buffered aqueous matrix containing varying levels of alanine, ascorbate, lactate, urea, and triacetin. By employing a set of 25 test samples measured 23 times over a period of 325 days, partial least-squares (PLS) calibrations based on synthetic spectra are observed to outperform conventional calibrations that use a set of 64 measured calibration samples. The key to the success of this approach is the use of a set of spectra of phosphate buffer collected on each prediction day to construct synthetic calibration spectra that are specific to that day. This allows the incorporation into the calibration model of nonanalyte spectral variance that is unique to a particular day. In this way, the adverse effects of instrumental drift or other sources of spectral variance on prediction performance can be minimized. Through the application of this methodology, values of the standard error of prediction (SEP) for glucose concentration are maintained to a range of 0.50−0.95 mM and an average of 0.68 mM over the 325 days of the experiment. These results are significantly better than those obtained with conventional models based on measured calibration samples. Over the same time period, a PLS model based on measured calibration spectra in absorbance units produced values of SEP that ranged from 0.41 to 2.02 mM and an average of 1.23 mM.
Co-reporter:Toshiyasu Tarumi, Yuping Wu and Gary W. Small
Analytical Chemistry 2009 Volume 81(Issue 6) pp:2199
Publication Date(Web):February 10, 2009
DOI:10.1021/ac802023w
Multivariate calibration models are constructed through the use of Gaussian basis functions to extract relevant information from single-beam spectral data. These basis functions are related by analogy to optical filters and offer a pathway to the direct implementation of the calibration model in the spectrometer hardware. The basis functions are determined by use of a numerical optimization procedure employing genetic algorithms. This calibration methodology is demonstrated through the development of quantitative models in near-infrared spectroscopy. Calibrations are developed for the determination of physiological levels of glucose in two synthetic biological matrixes, and the resulting models are tested by application to external prediction data collected as much as 4 months outside the time frame of the calibration data used to compute the models. The calibrations developed with the Gaussian basis functions are compared to conventional calibration models computed with partial least-squares (PLS) regression. For both data sets, the models based on the Gaussian functions are observed to outperform the PLS models, particularly with respect to calibration stability over time.
Co-reporter:Boyong Wan and Gary W. Small
Analyst 2008 vol. 133(Issue 12) pp:1776-1784
Publication Date(Web):11 Sep 2008
DOI:10.1039/B802557F
Passive Fourier transform infrared (FT-IR) remote sensing measurements are applied to the detection of methanol vapor plumes released from a chemical manufacturing facility. With the spectrometer mounted in a downward-looking mode on a fixed-wing aircraft, overflights of the facility are made during the methanol release. Signal processing and pattern recognition methods are applied to the acquired data for the purpose of constructing an automated classification algorithm for the methanol detection. The analysis is based on the use of short, digitally filtered segments of the raw interferogram data collected by the spectrometer. The classifiers are trained with data collected on the ground by use of an experimental protocol designed to simulate background conditions observed from the air. Optimization of the digital filtering and interferogram segment parameters leads to successful classifiers based on 100 or 120 interferogram points. The optimal interferogram segment location is found to be 95-points displaced from the centerburst, and the best performing digital filters are centered on the methanol C–O stretching band at 1036 cm−1 and have a passband full-width at half-maximum of 100 to 160 cm−1. The best classifiers achieve classification errors of less than 1% and are observed to be resistant to possible interference effects from species such as ethanol and ozone. This work demonstrates the utility of airborne passive FT-IR remote sensing measurements of volatile organic compounds under complex background conditions such as those encountered while monitoring an operating industrial facility.
Co-reporter:Yusuf Sulub and Gary W. Small
Analyst 2007 vol. 132(Issue 4) pp:330-337
Publication Date(Web):14 Feb 2007
DOI:10.1039/B615279A
Quantitative calibration models are developed for passive Fourier transform infrared (FT-IR) remote sensing measurements of open-air-generated vapors of ethanol. These experiments serve as a feasibility study for the use of passive FT-IR measurements in quantitative determinations of industrial stack emissions. A controlled-temperature plume generator is used to produce plumes of known concentrations of pure ethanol and mixtures of ethanol and methanol. Analyte plumes are generated over the path-averaged concentration range of 20–300 ppm-m and stack temperatures of 125, 150, 175, and 200 °C. A novel experimental setup is employed in which an ambient temperature polyvinyl chloride backdrop is placed behind the emission stack and used as a target for the passive IR measurements. An emission FT-IR spectrometer with telescope entrance optics is then employed to view the generated plumes against the backdrop. Signal processing techniques based on signal averaging and bandpass digital filtering are applied to both interferogram and single-beam spectral data obtained from these measurements, and the resulting filtered signals are used as inputs into the generation of multivariate partial least-squares (PLS) calibration models. Successful calibration models are obtained with both interferogram and spectral data, and neither analysis requires the collection of separate IR background data. For a set of validation data collected on a different day from the calibration measurements, standard errors of prediction of 30.6 and 32.2 ppm-m ethanol are obtained for the PLS models based on interferogram and spectral data, respectively.
Co-reporter:Kirsten E. Kramer, Gary W. Small
Vibrational Spectroscopy 2007 Volume 43(Issue 2) pp:440-446
Publication Date(Web):11 March 2007
DOI:10.1016/j.vibspec.2006.05.025
Procedures for data acquisition and data processing are evaluated for the optimal computation of absorbance values based on Fourier transform near-infrared transmission spectra. Samples consisting of physiological levels (1–20 mM) of glucose in an aqueous matrix of variable levels of bovine serum albumin and triacetin are studied in the combination spectral region (5000–4000 cm−1). The weak glucose signals in this region define a challenging analysis that is extremely sensitive to the effects of instrumental drift. The impact of different procedures for obtaining absorbance estimates is evaluated in the context of multivariate calibration models based on partial least-squares (PLS) regression. Replicate calibration and prediction data acquired over 6 months are used to study the robustness of PLS models with respect to time. The recommended protocol for the absorbance calculations is based on the collection of a large group of individual background spectra during the instrumental warm-up period. Seven procedures are tested for obtaining optimal backgrounds for use with either the calibration or prediction data sets. When the developed methodology is employed, standard errors of prediction are maintained in the range of 1.0 mM for spectra acquired up to 6 months after the collection of the calibration data. This level of performance compares favorably to daily internal cross-validation errors of 0.5–0.9 mM.
.jpg)
