Micro-circulation of skin blood: optical monitoring by advanced photoplethysmography techniques

 

Janis Spigulis*, Renars Erts and Uldis Rubins

 

University of Latvia, Physics Department and IAPS, Raina Blvd. 19, Riga, LV-1586, Latvia

 

 

ABSTRACT

 

Blood micro-circulation in upper skin layers has been detected in real time by advanced two-channel photoplethysmography techniques. Design of the sensor device is described, along with results of blood flow measurements and model calculations.

Keywords: Photoplethysmography, optical bio-sensing, cardio-vascular assessment.

 

 

1. INTRODUCTION

 

Micro-circulation of skin blood is a subject of considerable physiological interest due to its role in human metabolism and thermoregulation. There are several techniques used to follow the blood flow in living tissues. Optical methods seem to be the most promising for skin microcirculation studies, since the penetration depth of optical radiation is relatively small, typically 0.1…3 mm dependently on the radiation wavelength. Therefore analysis of the skin back-scattered optical radiation provide valuable selective information on blood flow only in upper skin layers, cutting off the influence of the deeper arteries and veins.

 

The non-invasive reflection photoplethysmography (PPG) method uses the back-scattered optical signals for temporal analysis of the skin blood volume pulsations. The periodical increase of blood volume in micro-vessels due to their dilatation (expansion) during the systolic raise of pressure with the following diastolic contraction (relaxation) over each heartbeat causes corresponding changes in absorption of optical signals traveling within the working volume. Consequently, the  time-dependent parts of the properly filtered PPG signals adequately reflect the skin blood volume pulsations during microcirculation.

 

The PPG techniques have been substantially improved since its origins 1, thanks to the fast development of microelectronics and computer technologies. That opens new horizons for implementation of this method in clinical praxis, self-monitoring and tele-medicine.

 

We have previously studied several options of the PPG signal analysis potential for obtaining information on skin vascular processes and assessment of eventual vascular pathologies:

-          shape analysis of individual PPG heartbeat pulses and time-averaged single-period PPG signals 2,

-          temporal analysis of fluctuations in heartbeat rate and other PPG signal parameters at steady-state conditions 3,

-          analysis of changes in PPG signals under influence of stress, physical exercises or other external factors 4,

-          analysis of the pulse wave propagation in blood vessels by means of the two-channel PPG techniques 5.

 

Two models of the PPG sensor equipment have been developed and clinically tested at University of Latvia – on-wheel equipment with the table-top computer 2, 3 and a portable PPG sensor device with a lap-top computer 4, 5. The recent two-channel version of the portable device will be shortly described in the next paragraph, with further presentation of some clinical data and their analysis in frame of a relatively simple vascular model.

 

*) E-mail janispi@latnet.lv, tel/fax +371 7228249

 

2. THE PORTABLE TWO-CHANNEL PHOTOPLETHYSMOGRAPHY SENSOR DEVICE

 

Recently a portable two channel reflection-type PPG sensor device comprising a set of universal contact probes, bio-signal processing electronics and a lap-top computer has been developed - see Fig. 1. 

 

 

 

 

 

Fig. 2. Application of the PPG contact probe at the neck.
 
               

 

 

 

 

 

 

 

 

                               

 

                a                                                                                                              b

 

Fig. 1. The portable PPG sensor device: a – general view, b – detection of the fingertip PPG signals.

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 2. Application of the contact probe on neck.

 

The basic design of the device is relatively simple. It consists of two optical contact probes, bio-signal amplifying/filtering circuit and a lap-top computer with specially developed software for AD-conversion, storage, processing and display of the PPG signals. All equipment is powered by a rechargeable battery and placed in a hand-held case (dimensions 44x32x11 cm, weight 4.1 kg). The contact probe continuously emits radiation into the skin tissues containing blood micro-vessels, and detects the AC-component of the back-scattered radiation that reflects the blood volume pulsations. The probe comprises a GaAs emitting diode (diameter of the emitting area ~2 mm, power ~10 mW, peak wavelength ~ 940 nm), and a Si photodiode with square detection area ~5x5 mm). Both diodes are closely mounted on a soft plastic pillow and fixed onto the measurement site by means of a sticky band. The band length is adjusted to the fingertip measurements; however, the band easily can be extended by spare bands, if necessary, so providing possibility to take PPG measurements from different locations of the body, e.g., forehead, forearm, knee, neck (Fig. 2).

 

The AC-component of the photodiode output signals is selected, pre-amplified and converted into digital format, then accumulated and processed by the computer. The signal sampling rate can be changed; the time resolution about 5-10 milliseconds appeared to be quite satisfactory for recording of well-resolved heartbeat signals. Parallel real-time detection and recording of two PPG signals from different body sites was ensured by originally developed electronic circuit.

 Special software was developed for the PPG bio-signal acquisition,

 processing and data storage, offering various options, e.g.

·         Filling the first window for patient data – name, age,          

gender, complains, doctor’s comments, etc.;                                                         

·         Pre-setting of the measurement time schedule;

·         The PPG signal registration and display in real time;

·         Signal clean-up (special filtering algorithm) and calculation of the mean single-period PPG (SPPPG) signal shape;

·         Calculation of specific cardio-vascular parameters for the registered;

·         Display of the corresponding PPG parameter set with subsequent cardio-vascular assessment results.

 

 

3. THE MEASUREMENT RESULTS: SOME EXAMPLES

 

The newly developed bio-sensor device had undergone several tests, and some interesting clinical results will be presented and discussed below.

Fig. 3. Comparison of the mean SPPPG signals taken at different body locations – middle fingertips (both arms), forehead and carotid artery – for two persons.

               

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 4. The observed changes in shapes of the skin SPPPG signals at various contact site distances from the arteries:

            a – Carotid artery, b – Temporal artery.

 

The mean single-period PPG (SPPPG) signals taken at four locations of the body for two persons are presented on Fig. 3. Differences in shapes of the signals detected from the middle fingertips of both arms, forehead and the carotid artery (on the neck) for the same person were clearly observed. One can also note also person-to-person differences in the mean SPPPG signals taken at the same locations of the body. Each of the mean SPPPG signals obviously contains specific features related to the person’s cardio-vascular condition; the assessment criteria are still elaborated in discussions with medical doctors.

As mentioned above, the limited penetration depth of the optical radiation (~ 2-3 mm under the skin surface for the wavelength range used in this study) is the reason why the PPG signals selectively reflect mainly the blood micro-circulation in the upper skin layers. Additionally, in the vicinity of arteries the arterial blood flow also influences the PPG signal shape. For instance, Fig. 4 illustrates the recorded mean SPPPG signal changes when the contact probe was placed at different distances from the artery projection to skin surface; two arteries (carotid and temporal) were regarded. The SPPPG signal recorded right over the artery obviously comprises two components of the blood volume pulsations, originated by subcutaneous micro-circulation and arterial macro-circulation. Most probably, each of them can be separated by means of proper signal analysis algorithms. Possibility to extract the arterial pulsations from the integral PPG signals would give additional advantages to the PPG methodology. In particular, it would help to increase accuracy of the pulse wave velocity measurements using the two-channel method.

 

Many PPG studies have been performed using fingertip as the measurement site. The fingertip PPG signals are relatively strong thanks to the anatomical feature - large vascular bed and, consequently, total volume of the pulsating blood. This is a typical peripheral blood supply situation when the artery pulse waves can be substantially re-shaped during their propagation via the highly resistant micro-vessels. The fingertip signals sometimes are pretty complicated, e. g. including two secondary peaks (notches) – see Fig. 5. The question - weather or not the fingertip PPG signals are suitable for physiological assessment - is still broadly discussed.

 

 

Fig.5. The observed fingertip PPG signal with two notches.        Fig.6. The time-shifted PPG signals recorded

                       simultaneously at the fingertip and toe.

 

 

Advantage of the newly developed two-channel approach is the availability of additional diagnostic information on the vascular blood flow resistance – it can be obtained by measuring the heartbeat wave propagation time between two body sites as the time shift between the two corresponding PPG pulses. For example, Fig. 6 illustrates distinct time delay between the PPG signals recorded simultaneously at the left fingertip and left toe. Time resolution of the device is ~ 0.01 s (it can be improved for an order of magnitude, if necessary), so reliable blood pulse wave propagation velocity estimations for diagnostic needs can be provided in this way.

 

 

4. THE VASCULAR MODELING AND SIMULATION

 

There are many open questions regarding the skin blood micro-circulation, therefore modeling and computer simulation of the vascular blood flow dynamics could be very helpful. In order to interpret the recorded shapes of PPG signals, a simple vascular model is regarded below, with assumption that the blood vessel elasticity assures linear proportionality between the blood volume and the blood pressure pulsations.

 

The blood pressure changes induced by heart contractions are propagating along the blood vessels as pulse waves, with subsequently pulsating blood flow. The human vascular system is elastic and multi-branched, and each branching partially reflects back the pressure wave. The whole vascular network from aorta till capillaries consists of numerous Y-branches, therefore modeling of a single Y-branching situation (Fig. 7) is very important. The relationship between incoming, outgoing and reflected pressures in this case can be represented by the following equations  6 :

                                                                  ( 1 ),

and

                                                           ( 2 ),

where is the peripheral reflectance,  – the peripheral permeability at the vessel’s branch, – incoming pressure,  – reflected pressure,  and – pressures in respective outgoing branches,  – blood flow impedance of the base vessel, , – impedances of  the branched vessels. Impedance Z is a physiologically significant vascular parameter:

                                                                                           ( 3 ),

where is the density of blood,  - pulse wave propagation velocity and  - the blood vessel cross-section.

 

Fig. 7. Model of the Y-branched blood vessel.

 

The dynamic pressure (and volume) changes represent superposition of the incoming and reflected pressure waves. To simulate the temporal behavior of pressure changes at the system output, the incoming pressure waveform is considered to be the Gaussian function:

                                                                (4),

where  is amplitude of the pulsating pressure,  – Gaussian rise-time and  – Gaussian bandwidth. The branching back-reflected pressure wave then will reach the base vessel entrance after traveling the distance 2L:

 

                             (5).

 

Generally, multiple reflections may be expected within the base segment and branches, if the input and output are reflective, too. The basic expression for temporal behavior of the outgoing blood pressure via the branch 1 in frame of such multi-reflection approximation can be represented as follows:

   (6),

where  - the output pressure,  - number of reflections in the base,  - number of reflections in the branch, , R0 – reflectivity of the base input, R1 – reflectivity of the branch output, L1 - length of the branch.

 

Equation (6) was used to simulate the pressure waveforms for the case with two equal output branches of diameters twice smaller than that of the base vessel. Calculations were performed for situations without reflections and with one, two and three reflections, using the following model parameters: t0 = 0.25 s, a = 100 s-2, Z1 = Z2 = 4Z0, R0 = 0.5, R1 = R2 = 0.33, T = 1.33, L1 = L2 = 0.75 L, L/c = 0.2 s, L1/c = L2/c = 0.15 s, n = m = 3.

 

a                                              b                                              c                                              d

Fig.8. The blood pressure waveforms in the branch output: a - without reflections (dashed line – the input Gaussian waveform), b - with one reflection, c - with two reflections, d - with three reflections.

 

The simulation results are presented at Fig. 8. The time shift between initial pressure pulse (dotted curve) and output pressure signal is well pronounced in all cases. Each reflection causes one additional peak in the output signal; however, the amplitudes of reflected signals are falling very fast, so the influence of the third reflection to the integral signal shape practically can be neglected (Fig. 8, d). One can note that the simulated two-reflection signal shape (Fig. 8, c) contains two “notches”, similarly to the observed mean SPPPG signals taken from fingertip (Fig. 5).

 

 

5. CONCLUSIONS

 

We tried to find answers to the following questions:

1.        Can the newly developed PPG sensor device be useful for studies of skin blood micro-circulation dynamics?

2.        Can we detect/separate the micro-circulation and arterial macro-circulation influences in the recorded PPG signals?

3.        Can we determine directly the pulse wave propagation velocity in arteries by the two-channel PPG techniques?

4.        Is there a way to set-up a simple model of the skin blood micro-circulation dynamics being in general agreement with the experimentally obtained data?

 

This pilot study gave evidence that all the above questions can be answered positively. Of course, additional studies are necessary to characterize the numerous specific situations in details. Further studies on the mentioned topics are underway.

 

 

ACKNOWLEDGMENTS

 

The authors are deeply grateful to Prof. Juris Aivars and MD Indulis Kukulis for valuable discussions. The financial support from Latvian Council of Science (grant # 01.0067) and Ministry of Education and Science (grant # TOP 02-13) is highly appreciated.

 

REFERENCES

 

1.        A. B. Hertzman, “Photoelectric plethysmograph of the finger and toes in man”, Proc. Soc. Exp. Biol. Med. 37, pp. 1633-1637, 1937.

2.        J. Spigulis, G. Venckus, M. Ozols, “Optical sensing for early cardiovascular diagnostics”, Proc. SPIE 3911, pp. 27-31, 2000.

3.        K. Prieditis, “Detection and computer processing of the photoplethysmography bio-signals”, BSc Thesis, University of Latvia, Riga, 2002, 43 p.

4.        J. Spigulis, I. Kukulis, E. Fridenberga, G. Venckus, “Potential of advanced photoplethysmography sensing for non-invasive vascular diagnostics and early screening”, Proc. SPIE 4625, pp. 38-43, 2002

5.        J. Spigulis, M. Ozols, R. Erts, K. Prieditis, “A portable device for optical assessment of the cardiovascular condition”, Proc. SPIE 5123, pp. 313-319, 2003.

6.        C. Caro, T. Pedley, R. Schroter, W.Seed, The mechanics of the circulation, Oxford University Press, NY-Toronto, 1978.