< Back to previous page

Project

Multispectral Speckle Plethysmography investigating its Origin in Tissue by comparing Temporal and Morphological Features with Clinical References

Background

Monitoring of hemodynamic variables is a cornerstone of modern medicine which has facilitated the diagnosis and analysis of many diseases. The main methods for different types of hemodynamic monitoring used in medical settings include electrocardiogram (ECG), continuous finger arterial pressure monitors (fiAP), photoplethysmography (PPG), and continuous wave doppler flowmetry. However, lack of continuity and specificity of tissue location, as well as in some cases obtrusiveness, remain limitations for continuous monitoring of patients. Furthermore, these systems often fail to detect microcirculatory critical changes. Timely detecting disparity between microcirculation and macrocirculation can further improve the monitoring and diagnosis of highly prevalent diseases including diabetes and peripheral artery disease, as well as lethal conditions such as sepsis. Finding a solution to overcome these limitations can have a significant impact in the outcome of patients.

Speckle Plethysmography (SPG) is a form of laser speckle contrast imaging (LSCI) applied to human tissue to extract time-domain signals highly related to hemodynamics. SPG technology was originally introduced as a method for monitoring microvascular blood flow. However, in previous research the morphology of the signals produced by SPG highly resembled macrovascular blood pressure, which indicated possible misconceptions about the capabilities of the technology. Considering the potential of SPG to provide with a different perspective of human hemodynamic variables in real-time, this technology was selected as the central subject of this research.

Thesis objectives

The purpose of this thesis is to investigate the manner SPG technology can help overcome the limitations of current hemodynamic monitoring technologies. To explore the potential of SPG signals for hemodynamic measurements, SPG is compared to different current hemodynamic monitoring methods. SPG is also compared to conventional PPG throughout this thesis because both biophotonic signals have common limitations and common morphological features.

The first objective of this thesis aims to compare the time-domain accuracy of SPG and PPG beat-to-beat heart rate to the R-R peak intervals of ECG. ECG is chosen as reference because it is the most accurate tool to measure the electrical activation of the heart. While previously published literature on SPG research was performed with devices obtaining signals using a fingerclip in transmissive illumination mode, this investigation of time-domain accuracy applies a remote camera in reflective mode.

 

The second objective aims to compare the performance of SPG and PPG under different optical configurations and wavelengths. The effect of transmissive and reflective illumination mode on SPG and PPG signals is compared at two wavelengths with different types of illumination i.e., coherent, and incoherent. While the results of the first two objectives provide information on the potential application and optimal configuration of SPG, the relation of SPG signals to other physiological variables (e.g., blood pressure) has yet to be explored.

The third objective aims to compare the morphology and Pulse Arrival Time (PAT) of continuous fiAP signal with SPG and PPG signals. For this objective, fiAP signal serves as reference for continuous blood pressure measurement providing information about the similarities and differences between SPG and PPG signals. A limitation of the investigations performed for the three first objectives is that the quality of reflective PPG (R-PPG) signals derived from the camera video stream is not usable for morphological analysis purposes. Hence, a clinical reference device is used to provide transmissive PPG (T-PPG) as a reference.

The fourth objective aims to compare simultaneous SPG and PPG signals obtained in real time-from the same video stream, while a cold pressor test (CPT) is used to temporarily affect the peripheral circulation systemically. The effects of this test are explored by analysing and comparing the signal-to-noise ratio (SNR), harmonic ratio, and alternating component (AC) amplitude of PPG, SPG, and fiAP signals.

Materials and methods

In order to investigate these objectives, various systems were designed, built, and evaluated. Opposite to previous SPG systems which are transmissive and monochrome, these are the first designs that offer the possibility for the simultaneous recording of dual-wavelength reflective SPG (R-SPG). The first system allows for the remote measurement of SPG and PPG at 639 nm and 850 nm with two different optical configurations i.e., transmission and reflection. The remote camera signals recorded by this system are post-processed using two parallel pipelines to derive both SPG and PPG from the same recorded video data. The quality of the camera-derived R-PPG signal is, however, by far inferior to the camera-derived R-SPG signal. Hence, PPG signals from a clinical device (using a transmission mode fingerclip) are used for the comparison. The second system allowed for real-time contact measurement of simultaneous R-SPG and reflective PPG (R-PPG). The second system produced real-time SPG and PPG signals from the camera data also allowing for recording of raw data, obtained from the same video stream at 639 nm and 850 nm. In addition, both systems were synchronized with PPG, ECG and fiAP reference systems.

Results

The investigation on the first objective reveals that remote SPG signal timing is at least equally (likely even more) accurate than contact PPG for beat-to-beat interval calculation when referenced against ECG R-R peak intervals. Furthermore, SPG calculated without dividing by the mean frame intensity was found significantly more robust against ambient light changes than PPG.

The investigation on the second objective reveals that in contrast to PPG, which performs better in transmissive mode, SPG performs better in reflective mode, at least with laser diodes of low power and low coherence length. Furthermore, this investigation revealed by means of the intensity threshold area (ITA) method that SPG measured in transmissive mode suffers from PPG contamination. The signal quality analysis shows that SPG does not significantly profit from an increase in integration time, whereas PPG does.

The investigation on the third objective reveals that waveform morphology of SPG has a higher correlation with fiAP than PPG. In addition, PAT of remote R-SPG shows comparable results to the contact fingerclip transmissive PPG (T-PPG) reference.

The investigation on the fourth objective reveals that SPG shows consistently higher SNR than PPG and that SPG showed less SNR variability, even during CPT. At 850 nm, the AC-amplitudes of SPG and PPG signals both show a significant reduction during CPT, but this is not seen for 639 nm. In addition, the harmonic ratio analysis demonstrates significant differences between SPG and PPG, with the SPG harmonic ratio revealing a higher correlation to fiAP than PPG.

Conclusions and future perspectives

The work developed in thesis shows that SPG is a more accurate and robust method for real-life situations, with regards to SNR and ambient light rejection. Also, remote SPG measurements are significantly easier to achieve than remote PPG measurements, while remote heart rate SPG measurements have comparable accuracy to contact PPG.  This implies that SPG has potential for monitoring situations where contact must be avoided. Furthermore, the ITA method revealed that current transmissive SPG systems in principle are contaminated by PPG, which may deserve attention.

This investigation shows that SPG and PPG contain different information about the vascular system, even when both are derived from the same photons. While SPG is more related to blood pressure and flow pulsations than PPG, it is less related to blood volume pulsations. In addition, SPG shares common features with continuous blood pressure measurements. This is an indication that SPG could have potential for non-invasive blood pressure assessment.

The two wavelengths explored show differences which might be caused by penetration depth variations, but more wavelengths should be explored to reach further conclusions. Hypothetically, wavelength multiplexing could be a method to measure SPG signals from different tissue depths, which could enable the simultaneous analysis of micro- and macrocirculation. Apart from the addition of other wavelengths, also other technologies could be added to the system to provide information that simultaneous multiwavelength SPG and PPG fail to detect.

In fact, the research in this thesis has demonstrated that a compact R-SPG system is also capable of delivering R-PPG signals, wherefore future work could focus on combining both parameters and, hence, the best of both worlds.

Date:1 Nov 2018 →  12 Oct 2023
Keywords:photonics, biomedical
Disciplines:Nanotechnology, Design theories and methods, Sensors, biosensors and smart sensors, Other electrical and electronic engineering
Project type:PhD project