DESCRIPTION NONINVASIVE BIREFRINGENCE COMPENSATED SENSING POLARIMETER Cross Reference To Related Applications This application claims the benefit of U. S. provisional patent application serial no. 60/536,288 filed January 13, 2004.

FIELD OF INVENTION The present invention relates to a birefringence compensated sensing polarimetric system. In one aspect, the poJarimeter system is used to measure and compensate for corneal birefringence when measuring glucose levels in a patient's eye.

BACKGROUND OF THE INVENTION Diabetes Mellitus is a common and serious chronic disease, which afflicts about 177 million people worldwide, 17 million people in the United States and is the fourth leading cause of death. It leads to long-term complications such as coronary artery disease, hypertension, retinopathy, neuropathy and nephropathy. Research indicates that self-monitoring of blood glucose levels prevents or slows down the development of these long term complications.

An optical polarimetric glucose sensor provides a means for the noninvasive measurement of glucose concentration, thereby reducing pain and complications associated with the current invasive methods.

The use of polarimetry in the detection of analyte concentration has existed for several years. Pohjola demonstrated that glucose concentration

in the aqueous humor of the eye is correlated to that of blood. In 1982 March et al, were the first to propose the use of polarimetry to indirectly estimate blood glucose levels via the aqueous humor of the eye. They found in order to measure millidegree sensitive rotations due to glucose at physiological levels a very sensitive and stable polarimeter is required. In the past decade considerable work has been done in the development of such a polarimeter.

Cote et al. reported on the potential of millidegree sensitivity by utilizing a true phase technique. This work was later followed by Cameron et al. who reported on a Faraday based polarimeter using a digital closed-loop feedback technique with sub-millidegree sensitivity. Since then, different polarimetric variations have been illustrated by several groups to measure glucose concentration. Chou et a/. reported on a polarimeter utilizing an optical heterodyne approach with the ability to detect glucose levels below 10 mg/dl ; however, the open loop system lacked stability due to fluctuations in the laser intensity and noise. Most recently, Ansari et a/. proposed a theoretical model using the Brewster's reflection off the eye lens for measuring glucose concentration.

Though aqueous humor of the eye contains glucose, it also has other optically active components that can contribute to the overall optical rotation.

To estimate glucose concentration in the presence of other optically active components, King et al. demonstrated the use of a multi-spectral Pockels cell based system. This work was followed by Cameron et a/. who used a multi-spectral Faraday-based system which also demonstrated the potential to overcome rotations due to the presence of other optically active components. Though glucose concentration in the aqueous humor correlates to that of blood, there is a transport time delay between the diffusion of glucose from the blood into the aqueous humor. If such measurements are to be of benefit to a diabetic person as a reliable predictor of blood glucose concentration, the time delay should be below 10 minutes. In 2001, Cameron et a/. measured the transport time delay in a rabbit model and had shown

this delay to be under the 10 minute threshold. Most recently, Baba et a/. have shown the effects of temperature and pH to be negligible in the normal physiological range.

The main problem currently hindering the development of a viable polarimetric system to indirectly measure blood glucose levels in the aqueous humor of the eye is the birefringence of the cornea associated with motion artifact. Since the birefringence of the cornea is spatially varying, as the cornea moves with respect to the sensing light beam, this motion induced time varying birefringence tends to mask the detected glucose signal.

To date, time varying corneal birefringence due to motion artifact is the main factor limiting in vivo polarimetric glucose measurements in the eye which is not addressed by current glucose sensing polarimeters. U. S. Pat No 5,303, 709 disclosed a system to facilitate diagnosis of retinal eye disease. To minimize effects of corneal birefringence, this system utilized a backscattered beam from the retina coupled to a variable retarderto reduce corneal birefringence contributions on nerve fiber retinal layer measurements. The compensation implementation in the'709 patent incorporated a polarization sensitive confocal system integrated into a scanning laser retinal polarimeter. U. S. Pat. No. 6,704, 106 disclosed a method and system to cancel retardance error in regards to retinal nerve fiber layer measurements. To achieve this, four retardance measurements collected over one complete rotation of a mechanically rotated half-wave retarder are averaged to minimize effects of system birefringence, leaving a mean retardance measurement free of residual polarization bias. In U. S. Pat.

No. 6,356, 036, a system and method for determining birefringence on the anterior segment (i. e. cornea and lens) of a patient's eye was disclosed.

This method involved using a backscattered (i. e. reflected) light beam similar to that disclosed in'709 except the patient's lens reflection intensity through confocal imaging is no longer used as a reference and birefringence of all

segments of the eye that are anterior to the retina are determined using a direct polarization beam. In other words,'036 eliminated the need for a confocal imaging system and the scanning laser polarimeter was now able to use the same path to measure birefringence of the anterior segment of the eye. In regards to the invention disclosed herein, a propagated polarized laser beam, not backscaffered, passes directly through the anterior chamber of the eye and does not interact with the lens or retina. In addition, the compensator is tied to an autonomous controller system to compensate for corneal birefringence effects in real-time.

Accordingly, it is an object of the present invention to provide an improved noninvasive glucose sensing polarimeter that incorporates a new method to overcome the effects of corneal birefringence, therefore, allowing forthe realization of in vivo polarimetric glucose measurements. In addition, such an implementation as described herein wouJd allow for the detection of any optically active molecule in a medium or sample in which birefringence is a problem. Furthermore, the approach to birefringence compensation could be implemented in most types of optical polarimeters, other than the Faraday approach as described herein.

The prior art fails to provide any practical, workable polarimeter system which can consistently provide accurate measurements of the glucose level in human tissue. There is a strong but unmet need for a practical, reliable system which overcomes the problems of the prior art to provide a noninvasive system for measurement of human glucose levels.

Some prior art systems are not noninvasive and certain polarimetric based systems, used for analyte sensing, are not birefringence compensated. The systems do not employ a birefringence compensator, nor do they sense real-time corneal birefringence. The ability to sense and compensate for birefringence effects allows for analyte measurements regardless of light path (i. e. location).

DISCLOSURE OF THE INVENTION The present invention relates to a system and method for compensating for the effects of birefringence and employs an optical birefringence analyzer to sense the real-time corneal birefringence contributions and then provides a feedback signal to a compound electro- optical system that negates the contributions found is a given sample. The birefringence contribution vanishes, thus significantly reducing the main error component for polarimetric measurements.

In one aspect, the present invention relates to a birefringence sensing polarimetric system comprising a means for measuring rotation of a substance in a sample, and a means for computing the value of retardance that need to be applied at a birefringence compensator in order to eliminate any rotation of a polarization vector due to the sample. in certain embodiments, the rotation measuring means comprises at least one Faraday modulator, at least one Faraday compensator, at least one analyzer, at least one detector, at least one amplifier, and at least one controller. Also, in certain embodiments, at least one means for computing the value of retardance that needs to be applied to the birefringence compensator comprises at least one circular analyzer, at least one detector, and at least one controller. The retardance is computed and sent as an input into a compensation portion of the controller wherein the compensation algorithm can be represented by the difference equation: y (n) = x (n) y (n-1) where'y'is the retardance applied to the birefringence compensator and'x'is the computed retardance, and wherein upon completion, there is no circularly polarized component and only linearly polarized light and any birefringence is compensated for.

In another aspect the present invention relates to a method for overcoming corneal birefringence comprising: using a circular Stokes parameter'V'for measuring birefringence compensation, and measuring glucose concentration using a Faraday based glucose sensing polarimeter.

In yet another aspect, the present invention relates to a non-invasive in vivo method for sensing a concentration of an optically active substance in an animal's aqueous humor. The method comprising: aligning a polarizer with a fast axis of the initial retarder to minimize effects of anterior corneal birefringence wherein a polarized laser beam passes through the glucose sample and the posterior corneal surface with a retardance (S), and splitting the laser beam wherein, in order to compensate for the posterior birefringence before determining glucose rotation, output light from the sample and retarder is separated into two paths by the beam splitter such that one beam is passed through an analyzer capable of characterizing at least one of four Stokes parameters (I, Q, U, V), and receiving a second of the split beams in a modulator and modulating the linear polarization vector of the laser.

BRIEF DESCRIPTION OF DRAWINGS Fig. 1 is a prior art schematic illustration of a simple polarimeter.

Fig. 2 is an illustration of linear to elliptical polarization states with varying retardance.

Fig. 3a is a schematic illustration of a digital closed-loop controlled glucose sensing polarimeter.

Fig. 3b is a block diagram of a corneal birefringence compensation module.

Fig. 4a is a block diagram of a glucose sensing polarimetric system.

Fig. 4b is a schematic illustration of a corneal birefringence compensated polarimeter using a single birefringence compensator.

Fig. 4c is a schematic illustration of an expanded birefringence compensator allowing for both anterior and posterior birefringence compensation.

Figs. 5a-5d are graphs showing actual verses predicted glucose concentrations for hyperglycemic glucose-doped water experiments.

Figs. 6a-6d are graphs showing actual verses predicted glucose concentrations for hypoglycemic glucose-doped water experiments.

Fig. 7a is a graph showing detected amplitude versus glucose concentration without corneal birefringence.

Fig. 7b is a graph showing detected amplitude versus corneal birefringence for a fixed glucose concentration without compensation.

Fig. 7c is a graph showing detected amplitude versus corneal birefringence for a fixed glucose concentration and birefringence compensation.

Fig. 8 is a flowchart glucose measurement controller.

Figs. 9a and 9b are graphs showing simulation results for uncompensated and compensated hypergJycemic glucose data.

Figs. 10a, 10b and 10c are FFT graphs that show the effect of glucose rotation and birefringence: Fig. 10a-without any glucose solution; Fig. 1 Ob-with a glucose concentration of 200 mg/dl ; and Fig. 10c-with a glucose concentration of 200 mg/dl and retardance of 5 degrees.

Figs. 11a, 11b, and 11c are calibration graphs for the uncompensated glucose-doped water experiments (a)- (c).

Figs. 12a, 12b, and 12c are validation graphs for uncompensated glucose-doped water experiments (a)- (c).

Figs. 13a, 13b, and 13c are calibration graphs for the compensated glucose-doped water experiments (a)- (c).

Figs. 14a, 14b, and 14c are validation graphs for compensated glucose-doped water experiments (a)- (c).

BEST MODE OF CARRYING OUT THE INVENTION In one aspect of the present invention, optical polarimetry is applied to the development of a noninvasive glucose sensor. Optical polarimetry relies on

the optical activity of glucose to rotate the linear polarization of light that is proportional to concentration. The glucose concentration in the aqueous humor contained within the anterior chamber of the eye provides an indirect measure of blood glucose concentration. The glucose concentration in the aqueous humor of the eye is correlated with that of blood. In order to measure millidegree sensitive rotations due to glucose at physiological levels, a very sensitive and stable polarimeter is required.

One problem with using the aqueous humor of the eye as the sensing site is that it has other optically active components that could contribute to the overall optical rotation. Other optically active components that may be included in the aqueous humor of the eye include the latic acid, albumin and the like. For a given substance, wavelength, ORD characteristics and molecular make-up of these samples need to be calculated. These relationships are described in the following pages of this specification. The other problem with using the aqueous humor as the sensing medium is the transport time delay between the diffusion of glucose from the blood into the aqueous humor. If such measurements are to be of benefit to a diabetic person as a reliable predictor of blood glucose concentration, the time delay should be below 10 minutes. Still another problem currently hindering in vivo polarimetric glucose detection with the eye as the sensing site is the birefringence of the cornea associated with motion artifact. The cornea is a birefringent material and the birefringence varies as a function of the corneal position. This corneal birefringence associated with motion artifact alters the state of polarization of the input beam, thus, masking the glucose signature.

The present invention relates to a system and a method which overcome the variations in the birefringence of the cornea. In addition, the inventive system described herein compensates for corneal birefringence and the azimutal angle of the fast and slow corneal optical axes.

Specific Rotation A beam of light is composed of electro-magnetic waves oscillating perpendicular to direction of light propagation. Normally, light exists in an unpolarized state. Unpolarized light has electro-magnetic oscillations that occur in an infinite number of planes, A device known as a linear polarizer only transmits light in a single plane while eliminating or blocking out light that exists in other planes. The light exiting the polarizer is known as plane polarized light.

Chiral organic molecules are molecules that do not contain a structural plane of symmetry. They rotate the polarization plane of light as it propagates through the sample. These molecules are collectively known as optically active. Depending on the molecules'confirmation, the plane of polarization may either be rotated clockwise or counter-clockwise. Molecules possessing the ability to rotate light to the left or counter-clockwise are denoted as levorotatory (L-) and those that rotate light to the right or clockwise are referred to as dextrorotatory (D+). Glucose is a dextrorotatory optically active molecule. The specific rotation of glucose dissolved in water is +52.6 °/ (dm g/ml).

The equation which relates optical rotation to a molecules specific rotation is given by equation (1) Where M is the specific rotation of an optically active compound, a is the observed rotation in degrees, L is the path length in dm, and C is the sample concentration in grams of mass per ml of solution.

For a given chiral substance, the wavelength dependence of specific rotation provides the Optical Rotatory Dispersion (ORD) characteristics of the constituent molecule. Every optically active molecule possesses its known unique ORD curve based on its molecular makeup.

The relationship between wavelength and specific rotation is given by Drude's equation. Equation (2) is an approximation of Drude's equation and is valid only outside the absorption region for the molecule of interest. If the specific rotation of a chiral molecule is known at two different wavelengths, equation (2) can be solved for ko and 20 and the specific rotation can be calculated for any wavelength within the region. Other tissue of the body may be used as sensing sites. Tissues of ear, nose and the thin skin areas between fingers and toes may be sensed. These are non-bony tissues. If used, wavelength and molecular makeup will have to calculate for these sensing sites.

Polarimetry The optical instrument used to measure rotation due to an optically active sample is a polarimeter. The main components of a polarimeter are a light source 10, a polarizer 12, a sample cell container 14, a second polarizer 16 known as the analyzer, and a detector 18, as shown in Fig 1.

As the beam passes through the sample, the plane of polarization will rotate according to the concentration of the sample and path length of the container. The amount of rotation due to the sample can be determined using the analyzer. If the analyzer is oriented perpendicular to the initial polarizer, theoretically no light will be transmitted if a sample is present. If an optically active sample is then introduced into the system, the intensity of transmitted light will be proportional to the amount of rotation in polarization due to the sample. Thus, the detected light intensity is related to the sample's concentration assuming a constant path length.

Polarimetry and the eye as the sensing site For in vivo polarimetric glucose detection, a suitable sensing site is

required. Several tissues in the body, such as the skin, are extremely scattering in nature. This scattering effect tends to significantly depolarize the light making it difficult to measure the small rotations due to physiological glucose levels. The eye is unique in that the cornea provides a low scattering window into the body. The diffusion or secretion of glucose into the aqueous humor of the eye correlates with blood glucose levels with a time delay.

These reasons make the eye a preferred sensing site.

Although the eye is virtually void of scatter and has a glucose concentration correlated to blood, it also has its own drawbacks as a sensing site. The main drawback is the spatially varying birefringence of the cornea associated with motion artifact.

Corneal Birefringence If the optical properties of a substance are same in all the directions regardless of its orientation, the substance is said to be isotropic. In many crystalline structures and some organic substances the optical properties are not the same in all directions and they have more than one index of refraction and these materials are known as anisotropic.

Birefringence is a property of anisotropic substances in which two orthogonally oriented different refractive indices of light exist, the ordinary refractive index, rio (along the slow axis) and extraordinary refractive index rie (along the fast axis). Light that is polarized along the x-axis experiences a different index of refraction, and therefore travels at a different speed than does light polarized along the y-axis. This difference in the speed of propagation between the x and y polarized components induces a phase difference. Depending on the magnitude of the components and the relative phase retardance (b), different states varying from linear to elliptical are shown in Fig 2.

The cornea is a birefringent anisotropic substance. Corneal birefringence is due to stroma that are composed of sheets of lamellae which are further

composed of coliagen fibers aligned parallel to each other. Each successive sheet of lamella is oriented differently with respect to the previous layer.

Each of these layers contains its own inherent birefringence and the degree of the arrangement of the lamella determines the overall birefringence. In many studies it has been shown that magnitude of retardation increases along the radius towards the periphery of the cornea. Birefringence of the cornea associated with the motion artifact is the major problem currently hindering in vivo polarimetric glucose measurements.

Digital closed loop controlled glucose sensing polarimeter The block diagram of the system used for in vitro glucose detection is shown in Fig 3a (prior art). A diode laser 20 emitting 1mW of power at a wavelength of 670 nm (red) is used as a light source. The laser beam is polaried by an initial polarizer 22 present in the optical system. Modulation of the polarization vector is provided by a Faraday modulator 24 driven by a sinusoidal function generator 26 at a frequency of 804Hz and modulation depth of 10. In certain embodiments, a power amplifier 28 can be used between the function generator 26 and the Faraday modulator 24. This modulated signal propagates through a sample cell container 30 constructed of optical grade glass with a path length of 1 cm. A Faraday compensator 32 is then used to provide feedback compensation within the system. The purpose of this compensator 32 is to nullify or eliminate any rotation due to glucose. Following the Faraday compensator 32 is another polarizer 34 which functions as an analyzer. The analyzer 34 transforms the modulated polarization vector into intensity modulation according to Malus'law. The intensity is detected by a silicon photo diode detector 36 and amplified by a wide bandwidth amplifier (not shown) which outputs a voltage proportional to the detected light intensity. The amplified output and modulation signal are sent as inputs to a lock-in amplifier and controller program through a data acquisition (DAQ) board. In certain embodiments, the controller program is

connected to a PC 41, which, in turn, is operatively connected to the power supply 42. The lock-in amplifier measures the signal component present at the modulation frequency, and based on this signal, the controller decides the course of action to be taken in order to compensate for any rotation due to the optically active sample. The output of the controller is applied to the Faraday compensator 32 through a GPIB controlled DC power supply 42.

This output is used to drive the Faraday compensator 32 to nullify the system.

The intensity of light which is detected by the system is given by equation (3) Where cl,,, is the modulation frequency, 0 is the modulation depth of the faraday modulator. is the difference in the rotation due to glucose and Faraday compensator. The detected signal consists of a dc term, a frequency-doubled term, and the signal of interest at the modulation frequency o,,, which is used as the input into the control system.

Stokes/Mueller model for the designed corneal birefringence compensation system A generalized block diagram of the corneal birefringence compensation module of the glucose sensing polarimetric system is illustrated in Figure 3b.

The main components present in the birefringence compensator optical system are a laser, polarizer, retarder (anterior birefringence), sample, retarder (posterior birefringence), and the birefringence compensator. The use of Stokes vector and Mueller matrix theory provides a way to model the system for computing the birefringence. The Stokes/Mueller model for this optical system is given by the following matrix system (eqn 3.1) Birefringence compensator 0 0 0 Q 0 cos2 (2y,) +sinZ (2y,) *cos (8,) sin (2y,) cos (2y,) (1-cos (8,))-sin (2y,) sin (8,) U 0 sin (2y,) cos (2ri) (1-cos (bu)) sin2 (2y,) +cos2 (2y,) *cos (8,) cos (2, y,) sin (61) V O sin (2Y) sin (6l) cos (2y,) sin (8.) cos (8,) Posterior retarder 1 0 0 0 (3. 1) * O cos2 (2Y) +sin2 (2Y) *cos (8) sin (2y) cos (2y) (l-cos (8))-sin (2y) sin (8) 0 sin (2y) cos (2y) (1-cos (8)) sin2 (2y) +cos2 (2y) *cos (8) cos (2y) sin (8) 0 sin (2y) sin (8) cos (2y) sin (8) cos (8) Sample Polarizer Laser 1 0 0 0 0. 5-0. 5 0 0 1 0 cos (2y.) sin (2 (pg) 0 * 1 O 5 0 5 0 0 * O 0-sin (2zpg) cos (2cps) 0 2 0 0 0 0 0 0 0 0 100000

where the system of matrices are presented in opposite order to the direction of light propagation. y'is the azimutal angle and'J'is the retardance due to the posterior corneal retarder,'y,'is the azimutal angle and '#1' is the retardance due to the birefringence compensator, 0, is the rotation due to glucose. This model does not account for the initial retarder as its optical axis is aligned with the initial polarization state. Initially, the birefringence compensator does not compensate for the retardance, therefore, Yl and 6, would be zero and the Mueller matrix of the compensator would result in an identity matrix. Since, elliptically polarized light is represented only by the Stokes parameter'V', simplifying the equation yields : Rearranging terms in equation [3.2], and assuming g to be negligible for physiological glucose levels, retardance Jis given by: , 5p 2V (3. 3) predicted-Siri giri2'Y

In the above equation, the Stokes parameter'V'is detected using a circular analyzer, as described herein.

Stokes/Mueller model for the glucose sensing polarimetric system The generalized block diagram for the glucose sensing polarimetric system is illustrated in Figure 4a. The main components present in the optical system of the designed polarimeter are a laser, polarizer, sample, birefringence compensator, Faraday modulator, Faraday compensator, analyzer, and detector. The use of Stokes vector and Mueller matrix theory provides a way to model the polarized light beam throughout the optical system. For glucose sensing, the birefringence compensator cancels out any retardance due to the sample. Therefore, there is end effect is the contributions of the birefringence compensator and sample cancel out, therefore the Mueller matrix combination of these two, result is an identity Mueller matrix. For simplification, eliminating the birefringence compensator and the sample retardance from the optical system, the matrix representation of the system is given by: Analyzer Faraday Compensator I 1 0 0 u v v Q 1 1 1 0 0 * 0 cos (-2of) sin (-2of 0 U 200 0 0 0-sin (-2of) cos (-2ouf 0 V 0 0 0 0 0 0 0 1 Faraday Modulator Sample 0 0 0 1 0 0 0 * 0 cos (-2sin ()) sin (-2sin (o)) 0 0 cos (20g) sin (20g) 0 0-sin (-26'sin (<B)) cos (-2sm (o)) 0 0-sm (2) cos (2#g) 0 0 0 0 10001 Polarizer Laser 1-1 0 0 1 *1-1 1 0 0 * O 20 0 0 0 0 00000

where Of is the rotation in polarization due to the compensation Faraday rotator, Og is the rotation due to the optically active sample, #m is the modulation depth, co,, is the modulation frequency and t is the time. This system of matrices when multiplied through can be simplified to the equation below Applying the substitution # = #g - #f and further simplifying the equation (4) for intensity is given by the equation 5 If the intensity equation (5) is further simplified and the assumption is made that sin (x) # x for x « 1, the equation above (5) can be reduced to:

Applying the identity sin2x = I-1 cos2x, to the equation above (6) yields, The equation above (7) describes the intensity of light, which is detected by the system at any instance of time. As can be seen, the detected signal consists of a DC term, a frequency-doubled term, and the signal of interest at the modulation frequency ?,, which is used as the input into the control system.

Corneal Birefringence compensated glucose sensing polarimeter Fig. 4b is a schematic illustration of a corneal birefringence compensated polarimeter using a single birefringence compensator. The block diagram of a corneal birefringence compensated glucose sensing polarimeter 50 is shown in Figure 4b. A light source such as a red diode laser module 52 is used as the light source. The laser 52 is initially polarized by a polarizer 54 and is oriented such that the maximum transmission is obtained. The polarizer 54 is aligned with the fast axis of the initial retarder, which minimizes the effect of birefringence due to the anterior side of the cornea. The polarized laser beam then traverses through a sample 56. The cornea birefringence induces a phase retardance (b) in the polarized laser beam resulting in a change in the state of polarization from linearly polarized light to elliptically polarized light and masking the signature of glucose. In order to compensate for the birefringence, a birefringence compensator 60, which is another electro-optical retarder, applies a retardance that cancels out any effect due to the sample birefringence.

The value of the retardance that needs to be applied at the birefringence compensator is computed by determining the circularly

polarized Stokes parameter'V'. For calculating this Stokes parameter, the elliptically polarized light is split into two routes by a non-polarizing laser line beam splitter 62. One route passes through a circular analyzer 64, which is a quarter wave plate followed by a 45° linear polarizer, capable of characterizing the circularly polarized Stokes parameter'V'. This beam is then directed towards a detector 66. The detector output is digitized using a data acquisition board (DAQ) 70. In certain embodiments, this is the input into a feedback controller program implemented using a PC 71. This controller outputs a voltage proportional to the birefringence compensation retardance. The compensation algorithm can be represented by the difference equation ) =) +-1) (8) where 'y' is the retardance applied to the birefringence compensator and'x' is the computed retardance. Upon completion, there is no circularly polarized component and only linearly polarized light ; therefore, the birefringence due to the posterior side of the cornea is compensated for.

In the other route, the linearly polarized laser beam is used for measuring glucose rotation. A Faraday modulator 80 is then used to modulate the linear polarization vector of the laser. The Faraday modulator 80 is driven by a sinusoidal source at 1058 Hz. A Faraday compensator 82 provides feedback compensation within the system. The purpose of this compensator 82 is to nullify or eliminate any rotation of the polarization vector due to the glucose sample. Following the Faraday compensator 82 is another polarizer 84, known as the analyzer, with its transmission axis oriented perpendicular to that of the initial polarizer 54. The analyzer 84 transforms the modulated polarization vector into intensity modulation according to Malus'law. The beam is then directed towards a photo-detector 86 whose output is a voltage proportional to the detected light intensity. The output of the detector 86 is then amplified by a wide bandwidth amplifier (not shown).

The amplified output and modulation signal are sent as inputs to a

lock-in amplifier and controller program through a data acquisition board 70.

In certain embodiments, the controller program is operatively connected to the PC 71. The lock-in amplifier measures the signal component present at the modulation frequency, while rejecting low and high frequency electro- magnetic noise. The intensity that impinges on the detector is given by equation (8); it is modeled using Stokes vector and Mueller matrix theory, which is described in detail below.

In equation 9, fJm is the modulation depth, co is the modulation frequency, t is time, and 0 = 0,-Of where 0, and Of are the rotations in polarization due to the glucose sample and Faraday compensator, respectively. As can be seen from equation (9), the relative amplitude of the sinusoidal term at the modulation frequency is proportional to the rotation due to the glucose sample assuming no compensation (ouf =0). This is used as the input into the controller, which forces the net rotation in polarization to zero. The output of the controller is applied to the Faraday compensator through a GPIB controlled DC power supply 92. Upon completion, the output voltage of the controller is proportional to the glucose concentration of the sample.

Fig. 4c is a schematic illustration of a more robust implementation of corneal birefringence compensator. In Fig. 4c, instead of using a single birefringence compensator, as illustrated in Fig 4b, two compensators are implemented to compensate for both anterior and posterior birefringence effects of the sample. Regardless of the compensation approach, the preferred compensation retarder is a liquid crystal retarder, but others may be employed such as a mechanical retarder or a photoelastic modulator (PEM). The path of the polarized laser beam may vary. While a parallel to the segments of the eye are preferred, a path perpendicular or at an angle to retina or lens may be employed.

Real time glucose controller program The flow chart of the glucose controller program is illustrated in Fig. 8.

Lock-in amplifiers are used to detect very small AC signals, and accurate measurements can be made even when the signal is obscured with noise many times larger. They use a technique known as phase sensitive detection to single out the component of the signal at a specific reference frequency and phase. Noise signals at frequencies other than the reference frequency are rejected and do not affect the measurement.

The input signal to the controller program is a signal proportional to the detected optical polarization rotation. The feedback controller applies a voltage proportional to the glucose rotation (i. e. lock-in voltage) to the Faraday compensator, to attempt to negate for the polarization rotation due to glucose. The controller program continues until the lock-in output is sufficiently near zero, which means there is no desired signal component at the modulation frequency. Upon completion, all net optical polarization rotation due to glucose is eliminated ; therefore, the output voltage of the controller which is applied to the Faraday compensator is proportional to the glucose concentration of the sample.

Example-Experimental polarimetric glucose detection The Faraday based glucose sensing system was evaluated for zero birefringence contribution for two sets of hyperglycemic 0-600 mg/dl and hypoglycemic 0-100 mg/dl concentration ranges. Least squares linear regression was used to compute a calibration model for glucose prediction.

Validation of the calibration model was performed using an independent data set. The actual versus predicted glucose concentrations for both calibration and validation are pfofted in Figs. 5a-d and Figs. 6a-d, respectively. Each data set possesses a high degree of linearity with all correlation coefficients exceeding r = 0.9880. The mean standard error of calibration (SEC) and

standard error of prediction (SEP) are 5.4047 mg/dl and 5. 4388 mg/dl, respectively.

The birefringence compensated system was simulated for different glucose concentrations varying in steps of 50 from 0-500 mg/dl without a birefringent sample in the optical system. As expected the glucose concentration versus amplitude graph is a linear plot Fig 7 (a). Now, the system was evaluated with the presence of a variable retarder as the sample (e. g. to simulate corneal birefringence); however, without compensating for the birefringence. For this case, the given glucose concentration is fixed at 500 mg/dl and the retardance was varied from 0-2Tr in steps of 0.01. As can be seen in Fig 7 (b), the detected signal varies with the birefringence which masks the glucose signature. In the ideal case for no corneal birefringence, the curve should be constant for a fixed glucose concentration.

If the retardance due to the cornea is computed from equation (3.3) and is compensated using the birefringence compensator, it can see that the amplitude no longer varies with retardance and is a constant for a given glucose concentration as shown in Fig 7 (c). Therefore, the sample birefringence no longer affects the optical polarization rotation measurement.

Example-Simulated noninvasive corneal birefringence compensated glucose sensing polarimeter To better understand the technique employed for corneal birefringence compensation and glucose sensing, the designed system, as depicted in Fig 4b, was simulated in all aspects. In addition, frequency analysis on the detected signal was performed using the fast Fourier transform (FFT). The simulation program, allowed for the user to arbitrarily change the glucose concentration as well as the amount for corneal birefringence. Furthermore, the simulation has the option of enabling/disabling both the glucose and birefringence controllers separately.

The simulation was initially run for different glucose concentrations and retardances. Two sets of experiments were performed using a glucose concentration range of 0-600 mg/dl in increments of 50 mg/dl, one without birefringence compensation and one after birefringence compensation. The retardance values were randomly chosen for different glucose concentrations. The calibration graphs for the uncompensated and compensated data are shown in Figs 9a and 9b. As expected, the uncompensated data set possesses a low degree of linearity with the correlation coefficient r=-0.1536. The SEC for the uncompensated and compensated data is 1252 mg/dl and 0 mg/dl respectively. The 1000 fold decrease in the SEC after compensation demonstrates the operation of the presented corneal birefringence compensation method.

Example-Corneal birefringence masking of glucose polarization rotation The FFT plots in Figs 10a-c illustrate how corneal birefringence masks the signature of glucose. In the Figs 10a and 10b, it is seen that a change in glucose concentration changes the magnitude of the detected signal at the modulation frequency (i. e. signal component at 1058 Hz). Fig 1 Oa corresponds to 0 mg/dl glucose concentration and Fig 10 b corresponds to 200 mg/dl glucose concentration. It should be noted, however, the magnitude at twice the modulation frequency (i. e. 2116 Hz) remains constant which is consistent with eqn 9. As illustrated in Fig 1 Oc, with maintaining the glucose concentration at 200 mg/dl, however, inducing 5 degrees of corneal retardance, it is seen that when birefringence is induced, the magnitude at the modulation frequency (i. e. 1058 Hz) in comparison with Fig 10b changes, therefore in essence masking the signature of glucose.

Example-Experimental results with the designed corneal birefringence compensation system

Using the designed corneal birefringence compensated glucose sensing system and the method described herein, three sets of experiments each, one without birefringence compensation and one after birefringence compensation were performed with a glucose concentration range of 0- 5000mg/dl in steps of 500mg/dl. Sample birefringence values between 3A/4 and A were chosen in small increments.

The uncompensated and compensated data obtained is illustrated in Table 1 below.

Table 1 : Glucose Uncompensated data Compensated data Concentration (Voltage) (Voltage) (mgldl) Set 1 Set 2 Set 3 Set 1 Set 2 Set 3 0 2. 075 2.025 1.997 3.54 3.5 3.466 500 0. 952 0.93 0.942 3.394 3.35 3.365 1000 1. 67 1.598 1. 59 3.226 3. 187 3. 182 1500 0. 42 0. 5 0.35 2.956 2.941 2.935 2000 0. 557 0.52 0.535 2.862 2.835 2.847 2500 0. 405 0.35 0.343 2.754 2.731 2.725 3000 0. 862 0.851 0.876 2.65 2.648 2.667 3500 0. 571 0.55 0.54 2.397 2.37 2. 412 4000 0. 157 0.14 0.125 2.369 2.35 2.33 4500 0. 671 0.652 0.665 2.288 2.284 2.283 5000 0. 27 0.33 0.302 2.168 2.132 2.14

The first three sets of experiments were performed without compensation to show the effect of corneal birefringence. The calibration and validation graphs for the uncompensated data sets are shown in Figures 11a-c and Figures 12a-c, respectively. Validation of the calibration models

for the uncompensated data was performed using the other data sets as independent data sets. By using the least squares calibration model, the computed slope is-41. 48 and the intercept is 5617. As can be seen from Table 2 below, the uncompensated data set possesses a very low degree of linearity with the mean correlation coefficient r = 0.7074. The mean SEC and SEP are 1656 mg/dl and 1647 mg/di, respectively.

Using the method described herein, the second three sets of experiments were performed with birefringence compensation. Validation of the calibration models for the uncompensated data was performed using the other data sets as independent data sets. The calibration and validation graphs for the compensated data are shown in Figs. 13a-c and Figs. 14a-c, respectively. For the designed system and the least squares calibration model, the computed slope is-3731 and the intercept is 12796. As can be seen in Table 2, each data set possesses a high degree of linearity with all correlation coefficients exceeding 0.9894. The mean SEC and SEP for the compensated data are 228 mg/dl and 230 mg/dl, respectively.

Table 2: Summary statistics for the collected data sets Data Model Correlation Standard Standard Coefficient error of error of (r) calibration validation (mg/dl) (mg/dl) Uncompensated a 0. 7147 1623 1690 data b 0. 7148 1622 1560 c 0. 6929 1725 1691 Compensated a 0. 9895 242 247 data b 0. 9901 234 232 c 0. 9917 215 213

The compensated data unlike the uncompensated data possesses a high degree of linearity with the correlation coefficient exceeding r=0.9890 for both calibration and validation. Also, there is a 7-fold drop in the SEC and SEP after compensation. These data and results demonstrate the described birefringence compensation technique and the benefits of using such a method in polarimetric glucose measurements.

Other Embodiments-The device described in the attached documents depict only one form in which the applied theoretical approach to corneal birefringence compensation may be realized. It is to be understood that other embodiments of the present invention include the use of other approaches, such as, but not limited to the following, and that such approaches are within the contemplated scope of the present invention: 1) Other optical mechanisms and approaches for the control and handling of polarized light (e. g. instead of using liquid crystal variable retarders, other similar methods to control of polarized light could be through the mechanical movement of optical elements such as fixed retarders, photoelastic modulation, the Pockels effect, etc...) 2) The control algorithms for the corneal birefringence compensator could be implemented in a variety of forms, such as, through the use of a proportional integral differential (PID) controller or by other similar methods.

3) The birefringence compensator could be extended through other methods as described in (1) to provide enhanced and more robust control to achieve birefringence compensation (e. g. extension from single axis, as described, to three-axis birefringence compensation).

4) The birefringence analyzer could be extended to more fully characterize the state of polarized light to provide feedback to other implementations of the birefringence compensator (e. g. a three-axis variable birefringence compensator).

The above descriptions of the preferred and alternative embodiments of the present invention are intended to be illustrative and are not intended to be limiting upon the scope and content of the following claims.

All of the compositions and methods disclosed and claimed herein can be made and executed without undue experimentation in light of the present disclosure. While the compositions and methods of this invention have been described in terms of the foregoing illustrative embodiments, it will be apparent to those skilled in the art that variations, changes, modifications, and alterations may be applied to the compositions and/or methods described herein, without departing from the true concept, spirit, and scope of the invention. More specifically, it will be apparent that certain agents that are both chemically and physiologically related may be substituted for the agents described herein while the same or similar results would be achieved. All such similar substitutes and modifications apparent to those skilled in the art are deemed to be within the spirit, scope, and concept of the invention as defined by the appended claims.