Determination of concentrations of biological substances using raman spectroscopy and artificial neural network discriminator

• US 5,553,616 A
• Filed: 11/30/1993
• Issued: 09/10/1996
• Est. Priority Date: 11/30/1993
• Status: Expired due to Term

• Current Assignee: Florida Institute of Technology Incorporated
• Original Assignee: Florida Institute of Technology Incorporated
• Inventors: Ham, Fredric M., Cohen, Glenn M.
• Primary Examiner(s): Smith, Ruth S.
• Application Number:
  US08/160,033
• Time in Patent Office: 1,015 Days
• Field of Search: 128/633, 128/634, 128/664, 128/665, 250/339, 250/340, 250/341, 250/339.01, 250/339.07, 250/340, 250/339.09, 250/341.1, 250/341.5, 395/21, 395/23, 395/24, 395/25, 395/924, 356/39, 356/301
• US Class Current: 600/316
• CPC Class Codes: A61B 5/14558 by polarisation
• CPC Class Codes: A61B 5/418 lymph vessels, ducts or nodes
• CPC Class Codes: A61B 5/6826 Finger
• CPC Class Codes: A61B 5/6838 Clamps or clips
• CPC Class Codes: A61B 5/7267 involving training the classification device
• CPC Class Codes: G01N 21/65 Raman scattering
• CPC Class Codes: G16H 10/40 for data related to laboratory analysis, e.g. patient specimen analysis
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: Y10S 128/925 Neural network
• CPC Class Codes: Y10S 706/924 Medical

Abstract

The concentration of a substance, such as glucose, in a biological sample, such as human tissue (e.g. the skin of an index finger) is non-invasively determined by directing the output beam of a laser diode onto and into the skin so as to cause Raman scattering. The output of a charge coupled device, upon which the scattered light is spatially dispersed according to frequency is digitized and applied to a processor. The processor compares the Raman scattering intensity characteristics of the sample with a comparative model, in particular, an artificial neural network discriminator (ANND). The ANND is trained with a plurality of Raman spectral characteristics from biological fluids or tissue, possessing known Raman scattered light intensities versus wavelength characteristics at known concentrations. A preferred implementation of the ANND employs fuzzy adaptive resonance theory-mapping (ARTMAP), which has robust noise rejection capabilities and can readily handle nonlinear phenomena.

Claims

1. A method of non-invasively deriving a measure of the concentration of glucose in a human body fluid comprising the steps of:

(a) irradiating a human body tissue containing said human body fluid with monochromatic light in the near infrared spectral region emitted by a monochromatic light source, thereby causing said glucose within the irradiated human body tissue to produce Raman scattering of said monochromatic light into multiple, spatially separated, frequency shifted wavelength components associated with molecular characteristics of said glucose;

(b) tracking variations in light emitted by said monochromatic light source;

(c) directing said multiple, spatially separated, frequency shifted wavelength components resulting from said Raman scattering through a holographic notch filter, which removes the Rayleigh component of scattered light, onto a photosensitive detector, which produces electrical output signals representative of said multiple, spatially separated, frequency shifted wavelength components associated with molecular characteristics of said glucose;

(d) generating ratio signals representative of the ratios of said electrical output signals representative of said multiple, spatially separated, frequency shifted wavelength components resulting from said Raman scattering produced in step (c) to variations in monochromatic light emitted by said monochromatic light source tracked in step (b), so as to eliminate effects of variations of said monochromatic light source; and

(e) processing said ratio signals by means of a fuzzy adaptive resonance theory artificial neural network discriminator that has been trained with a plurality of training exemplars, corresponding to scattered light components representative of different concentrations of glucose, and which is robust to a human body fluid sample not previously encountered, and quantitatively deriving therefrom concentrations of said glucose in human body fluid based upon the spectral intensities of detected wavelengths of said spatially separated, frequency shifted wavelength components.

2. An apparatus for non-invasively deriving a measure of the concentration of glucose in a human body fluid comprising:

a monochromatic light source which emits monochromatic light in the near infrared spectral region, and irradiates

human body tissue containing said human body fluid with said monochromatic light, thereby causing glucose within the irradiated human body tissue to produce Raman scattering of said monochromatic light into multiple, spatially separated, frequency shifted wavelength components representative of molecular characteristics of said glucose;

a holographic notch filter;

optical elements which direct said multiple, spatially separated, frequency shifted wavelength components resulting from said Raman scattering through said holographic notch filter, which removes the Rayleigh component of scattered light incident thereon;

a first photosensitive detector device, which is disposed to receive light passing through said holographic notch filter and produces electrical output signals representative of said multiple, spatially separated, frequency shifted wavelength components representative of molecular characteristics of said glucose;

a second photosensitive detector device, which is disposed to receive monochromatic light emitted by said monochromatic light source and produces signals representative of variations in light emitted by said monochromatic light source;

a ratio signal processor, which is coupled to receive said signals representative of variations in light emitted by said monochromatic light source and said electrical output signals, and which generates ratio signals representative of the ratios of said electrical output signals representative of said multiple, spatially separated, frequency shifted wavelength components resulting from said Raman scattering to variations in monochromatic light emitted by said monochromatic light source, so as to eliminate effects of variations of said monochromatic light source; and

a fuzzy adaptive resonance theory artificial neural network discriminator that has been trained with a plurality of training exemplars, corresponding to scattered light components representative of different concentrations of glucose, and which is robust to a human body fluid sample not previously encountered, said fuzzy adaptive resonance theory artificial neural network processing said ratio signals and quantitatively deriving therefrom concentrations of said glucose in human body fluid based upon spectral intensities of the detected wavelengths of said spatially separated, frequency shifted wavelength components.