Influence of Arterial Occlusion at Various Cuff Pressures on Systemic

Circulation Measured by rPPG

Leah De Vos

, Gennadi Saiko

, Denis Bragin

3,4

and Alexandre Douplik

2,5

Department of Engineering, Toronto Metropolitan University, Toronto, Canada

Department of Physics, Toronto Metropolitan University, Toronto, Canada

Lovelace Biomedical Research Institute, Albuquerque, U.S.A.

Department of Neurology, University of New Mexico School of Medicine, Albuquerque, U.S.A.

iBest, Keenan Research Centre of the LKS Knowledge Institute, St. Michael’s Hospital, Canada

Keywords: Photoplethysmography, Arterial Occlusion, Microcirculation.

Abstract: Background: Arterial occlusion is a ubiquitous medical procedure, which is used in many clinical scenarios.

However, there is no standard protocol for the selection of the applied pressure. As various pressures may

trigger different physiological responses, it is important to understand these peculiarities. The aim of the

current work is to investigate if there is any difference in the systemic response to the occlusion at various

applied pressures. Methods: Hands of healthy volunteers (10 volunteers) were occluded at the wrist by

inflating the blood cuff to 150 or 200 mmHg. The remote photoplethysmography (rPPG) measurements of

control and experimental hands were taken. To assess systemic response, we have analysed the behaviour of

AC (low frequency, LF at 0.1 Hz rate) components in green and red channels during occlusion and reperfusion.

Results: We have not found a statistically significant difference in the LF spectra between occlusions at 150

and 200 mmHg pressures. Conclusions: We have performed the analysis of low-frequency (0.1 Hz)

components of remote photoplethysmography signals during arterial occlusion at 150 and 200 mmHg. Our

preliminary results show that the systemic response is similar at both levels of occlusions.

1 INTRODUCTION

Arterial occlusion is a ubiquitous medical procedure.

It is used in many clinical scenarios. Probably the

most well-known use of arterial occlusion is blood

pressure measurements using an auscultatory method,

which was originally based on a stethoscope and a

sphygmomanometer. Another important clinical use

of arterial occlusion is to detect endothelial

dysfunction in critically ill patients (Joannides et al.,

2006).

In common clinical scenarios the arterial

occlusion is caused by placing the blood pressure cuff

over the forearm or wrist and inflating it over systolic

blood pressure (Lenders et al., 1991). Common sense

would suggest that inflating the blood cuff just over

systolic pressure would be sufficient to occlude

arteries of the hand. However, it is not the case.

https://orcid.org/0000-0002-5697-7609

https://orcid.org/0000-0003-4894-0061

https://orcid.org/0000-0001-9948-9472

As a rule, an additional pressure of 50 mmHg is

typically considered to be a safe margin

(Kanchanathepsak et al., 2023). However, the exact

relationship between the applied pressure and the

level of occlusion is not well understood and

established in the literature. In particular, it can be

driven by multiple factors such as BMI. As such, the

applied pressure often is a tradeoff between the desire

to occlude vessels and patient’s tolerance to pain. The

latter is particularly important for a bed-side test for

endothelial dysfunction assessment, where the artery

is occluded for 3 min (Saldin, 2019). Thus, it could

be beneficial to get more insights into the differences

at various levels of the occlusion. Such investigation

can be based on observing hemodynamic response,

which can be on a local and systemic level.

Remote photoplethysmography (rPPG) has been

shown its utility in investigation of skin

De Vos, L., Saiko, G., Bragin, D. and Douplik, A.

Inﬂuence of Arterial Occlusion at Various Cuff Pressures on Systemic Circulation Measured by rPPG.

DOI: 10.5220/0012469800003657

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 17th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2024) - Volume 1, pages 313-318

ISBN: 978-989-758-688-0; ISSN: 2184-4305

313

microcirculation (Burton et al., 2023). The

predominant signal source in the PPG is the cardiac

pulsation caused by the ejection of blood from the left

ventricle during cardiac systole, which causes

distensions of blood vessels and changes in the tissue

absorption. These blood waveforms demonstrate

changes in five frequency bands, which are related to

different physiological processes. Heart rate for

normal subjects at rest varies from 60-100 beats per

minute (bpm) (Vital Signs (Body Temperature, Pulse

Rate, Respiration Rate, Blood Pressure), 2022).

Conservatively extending the lower bound to 50 bpm

to consider lower resting heart rates that can occur in

certain people, such as athletes (Doyen et al., 2019),

then the corresponding frequency range is 0.83-1.67

Hz. The normal respiration rate for a healthy subject

is 12 to 20 breaths per minute, corresponding to a

frequency range of 0.20-0.33 Hz (Vital Signs, n.d.).

This PPG signal further contains oscillations in 0.01-

0.02 Hz, 0.02-0.06 Hz, 0.06-0.15 Hz ranges

corresponding to endothelial related metabolic,

neurogenic, and myogenic activities, respectively (Li

et al., 2006). Myogenic range also contains Mayer

waves, which are oscillations in blood pressure that

typically occur at a frequency of 0.1 Hz (Julien, 2006).

The mechanism for Mayer waves is subject to active

debate, but recent findings advocate that the

oscillations are produced by a sympathetic

baroreceptor response to hemodynamic disturbances

(Julien, 2006). The ability to capture Mayer waves by

a smartphone camera was demonstrated in (Burton et

al., 2022). As occlusion and/or reperfusion may trigger

hemodynamic disturbance we hypothesize that this

disturbance can depend on the severity of

occlusion/reperfusion and characterize this difference.

The aim of the current work is to investigate the

differences in systemic physiological response to the

occlusion/reperfusion at various applied pressures.

We used rPPG to analyze skin microcirculation in an

arterial occlusion model with 2 different applied

pressures. We have selected 150 mmHg as a pressure

which is just marginally higher than the systolic

pressure, and 200 mmHg, as a higher pressure with

50 mmHg safety margin. As we aim to investigate the

systemic response, we have analyzed the rPPG signal

in both experimental and control hands.

2 METHODS

2.1 Data Collection

The experimental setup, as seen in Figure 1, includes

the subject sitting with their hands placed side by side

on a raised platform in the prone position. An iPhone

14 camera (Apple, Cupertino, CA, USA) is held by a

tripod above the hands positioned directly above. As

a light source, two rectangular video light panels with

600 LEDs on each panel (NEEWER LED Video

Light, Shenzhen, China) were positioned on either

side of the camera illuminating the hands (light colour

was set to 4600K to maximize the green channel

signal and 100% intensity was set to ensure maximum

signal to noise ratio). A white circle of 1 cm diameter

is also placed in the frame next to the hands for colour

normalization during processing. In each iteration of

data collection, one hand is designated as the

experimental hand, occluded throughout the data

collection, and the other acts as a control. On the arm

of the experimental hand, a pressure cuff is worn

around the wrist to apply pressure during the data

collection.

Figure 1: Experimental setup. (From top to bottom) LED

panels, iPhone 14, pressure cuff and hands placed on

staging table.

Once the setup is established, the data collection

procedure is as follows:

1. The subject's initial blood pressure is

measured from the experimental arm.

2. The subject places both hands on the platform,

in the prone position. At this point, the iPhone

starts recording video at 60 fps.

3. Baseline measurements are recorded for 1

minute (baseline interval).

BIOIMAGING 2024 - 11th International Conference on Bioimaging

314

4. Arterial occlusion is then applied on the

experimental hand for 3 minutes by applying

150 mmHg of pressure (occlusion interval).

5. The pressure is then released and no pressure

is applied to the arm for 4 minutes (reperfusion

interval).

6. At the 8-minute mark data collection

concludes and the iPhone recording video and

the muscle oxygenation monitor are both

turned off.

A rest period of 5 minutes is taken and then the

process is repeated for steps 2-6 with the opposite arm

acting as the experimental hand and with 200 mmHg

of pressure being applied instead of 150 mmHg.

For this work, data from 10 subjects were

acquired (age range: 20-60 years old, 4 males, 6

females) each with no evidence of cardiovascular

disease and no extreme BMI values. The Toronto

Metropolitan Research Ethical Board approved this

study.

2.2 Image Processing

The recorded video footage from the data collection

is then processed and analysed using MATLAB

2023b (Mathworks, Natick, MA, USA). First, to

improve processing time, the video is cut into three

segments each representing sequentially no pressure,

arterial occlusion (150 mmHg or 200 mmHg) and

pressure release. The following processing is applied

to each video segment. A Gaussian filter is applied to

each video frame to reduce the amount of noise in the

video and high-frequency components in the video.

Next, three regions of interest (ROI) are manually

selected from the video file, one on the experimental

hand, one on the control hand and one around the

white circle. For each ROI, each video frame’s pixels

are averaged for each colour channel (red, green and

blue (RGB)) resulting in a single red, green and blue

value per frame. This creates signals representing a

time series for each colour channel. To normalize the

time series for the experimental and control hands the

following equations are used,

𝑒

,

 255

𝐸

,

𝑅

,



(1)

𝑐

,

 255

𝐶

,

𝑅

,



(2)

Where i refers to the colour channel (i = red,

green, and blue), j refers to the frame’s number (j = 1

- N, where N is the number of frames), E represents

the experimental hand time series, C represents the

control hand times series and R represents the white

circle time series.

Figure 2: Experiment hand placements, and highlighted

regions of interest in red.

2.3 Data Analysis

The DC component of the signal represents the

average intensity of the signal over time and the AC

component represents changes in the signal

associated with the cardiac cycle. For this work, we

considered the AC component of the signal. As such,

for AC component of the rPPG signal, changes in

amplitude over time due to arterial occlusion events

were analysed, specifically amplitude drop relative to

baseline condition when occlusion is applied and the

amplitude overshoot relative to baseline when the

pressure is released. The calculations performed for

each component are as follows:

2.3.1 AC Component

For the AC component, the low frequency range in

the red and green channels were analysed. To retrieve

the LF spectral range, each channel was run through

a bandpass filter with a frequency range of 0.05-5 Hz.

After the signals are filtered to this frequency range,

the continuous wavelet transforms (CWT) is applied

on the signals calculating the mean power of the

results. After this the power spectral density (PSD) in

the LF range for both the red and green channels can

be calculated. To express the percent drop of the LF

frequency of the experimental hand relative to the

control hand when pressure is applied and the

overshoot when pressure is released, the following

equation is used,

Inﬂuence of Arterial Occlusion at Various Cuff Pressures on Systemic Circulation Measured by rPPG

315

% 𝑐ℎ𝑎𝑛𝑔𝑒 

𝑣𝑎𝑙𝑢𝑒



 𝑏𝑎𝑠𝑒𝑙𝑖𝑛𝑒

𝑏𝑎𝑠𝑒𝑙𝑖𝑛𝑒

𝑥 100

(3)

Where i represents the segment number (i = 2-3),

value

represents the PSD value at 0.1 Hz on

occlusion (i=2) or reperfusion (i=3) segment, and

baseline represents the PSD value at 0.1 Hz on

baseline (i=1) segment. The above equation was

performed on both the red and green channels for

each segment, and for experimental and control

hands.

2.3.2 Statistical Analysis

To assess the statistical significance of the differences

between results, i.e. comparing between AC

components, two-tailed t-tests were performed. In

this case the null hypothesis was that the mean of data

1 is equal to the mean of data 2, essentially that there

is no significant difference between the two datasets.

The alternative hypothesis is that the mean of data 1

is not equal to the mean of data 2 denoting a

significant difference between the two datasets. To

perform the t-test, the following process is followed.

First the t-test statistic is calculated as,

𝑡 

𝑥̅



 𝑥̅

























(4)

Where 𝑥̅



and 𝑥̅



are the means of data 1 and data

2 respectively,𝑠



and 𝑠



are the standard deviations of

data 1 and data 2 respectively, and 𝑛



and 𝑛



and the

sample sizes of data 1 and data 2 respectively. Then

the p-value is determined by first calculating the

degree of freedom by,

𝑑𝑓 



































/













 





/















(5)

Once 𝑑𝑓 is calculated then the p-value can be

found as,

𝑝𝑣𝑎𝑙𝑢𝑒  𝑃𝑇𝑡 | 𝑑𝑓

(6)

Where T is the random variable following a t-

distribution with df and t is the test statistic. From

there if the p-value < α (significance level set as α =

0.05) the null hypothesis is rejected and if the p-value

≥ α the null hypothesis is accepted. For this study the

computations were performed in MATLAB.

3 RESULTS

We have analysed the behaviour of the AC

components at LF spectral range in green and red

channels during occlusion and overshoot. The results

are depicted in Figure 3 and Figure 4, respectively.

The results are also summarized in Table 1 for the

experimental hand and Table 2 for the control hand.

Figure 3: The behaviour of AC components in green (left

panel) and red (right panel) channels during occlusion. The

drop has been calculated as a relative change in AC

component on the 2nd interval compared to the baseline.

Figure 4: The behaviour of AC components in green (left

panel) and red (right panel) channels during reperfusion.

The overshoot has been calculated as a relative change in

AC component on the 3rd interval compared to the baseline.

Table 1: Comparison of different metrics at 150 and 200

mmHg on the experimental hand. Specifically mean,

standard deviation (in brackets), and p-value (with

significance noted by (ns), no statistical significance, when

p > 0.05).

Metric

Mean @ 150

mmHg

Mean @ 200

mmHg

p-value

Reperfusion

Red Channel

155.26

(268.91)

84.69

(122.72)

0.46

(ns)

Reperfusion

Green Channel

92.18

(158.02)

38.60

(72.34)

0.34

(ns)

Occlusion Red

Channel

-15.91

(82.94)

-10.51

(43.26)

0.86

(ns)

Occlusion

Green Channel

-32.80

(47.77)

-34.30

(27.79)

0.93

(ns)

BIOIMAGING 2024 - 11th International Conference on Bioimaging

316

Table 2: Comparison of different metrics at 150 and 200

mmHg on the control hand. Specifically mean, standard

deviation (in brackets), and p-value (with significance

noted to (ns), no statistical significance, when p > 0.05).

Metric

Mean @ 150

mmHg

Mean @ 200

mmHg

p-value

Reperfusion

Red Channel

3.70

(67.42)

68.15

(155.25)

0.244

(ns)

Reperfusion

Green Channel

35.05

(99.15)

118.91

(129.85)

0.12

(ns)

Occlusion Red

Channel

-6.18

(72.96)

24.19

(111.01)

0.48

(ns)

Occlusion

Green Channel

20.57

(80.03)

93.23

(102.29)

0.09

(ns)

4 DISCUSSION

Here, we present an initial pilot investigation of the

microvasculature hemodynamic during

occlusion/reperfusion captured by a smartphone

camera. Data was collected as described from ten

healthy subjects. We have performed the analysis of

remote photoplethysmography signals in the low

frequency (0.1 Hz) range for two levels of occlusion:

mild occlusion (150 mmHg) and occlusion with a

safety margin (200 mmHg).

As we aimed to investigate the systemic response,

we analyzed the rPPG signal in both the experimental

and control hands.

The behaviour of the LF component is very

similar for both levels of occlusion. In particular, we

have not found any statistically significant

differences between distributions caused by 150 and

200 mmHg either during occlusion or reperfusion.

During occlusion, AC amplitudes in the

experimental hand in the 0.1 Hz range drop for both

150 and 200 mmHg pressures (see Figure 3). The

drop distributions are characterized by very similar

means and standard deviations. It holds true for both

the green and red channels.

During occlusion, AC amplitudes in the control

hand in the 0.1 Hz range increase for both 150 and

200 mmHg pressures in the green channel (see Figure

3). The increase distributions are characterized by

very similar means and standard deviations. It holds

true for both green and red channels.

During reperfusion, the behaviour of the 0.1 Hz

range component in the experimental hand in 150 and

200 mmHg cases are also characterized by very

similar means. However, standard deviations for 150

and 200 mmHg pressures are quite different in both

red and green channels (see Figure 4).

Moreover, during reperfusion, one can see that

there is a substantial (by 50-70% in red channel)

increase in 0.1 Hz oscillations in the experimental

hand.

We hypothesize that these low-frequency

oscillations can be attributed to Mayer waves. While

Mayer waves share the same frequency range as

myogenic activities (0.06-0.15Hz), their origins are

different. Mayer waves are the sympathetic activity

with baroreflex activation. Myogenic oscillations are

local and independent of the sympathetic nervous

vasoconstriction.

Our conclusion regarding the origin of the 0.1 Hz

amplitude increase is based on two observations.

Firstly, we see similar results during reperfusion in

both experimental and control hands. Thus, it speaks

in favour of systemic response. Secondly, the results

are quite similar for red and green channels, which

have different sampling depth. Thus, similarity in

these responses also points in favour of Mayer waves,

as local regulation should demonstrate some

differences between capillary network (sampled by

the green channel) and deep vascular plexus (sampled

by the red channel)

Thus, we can conclude that both lower and higher

pressures are probably triggering a similar systemic

response in the form of a sympathetic baroreceptor

response to hemodynamic disturbances.

The work has certain limitations. Firstly, the

measurements were performed on just 10

participants. Thus, larger studies are required to

generalize the results. Secondly, the short time

interval (5 min) between the measurements on left

(150 mmHg occlusion) and right (200 mmHg

occlusion) hands was taken. While quite a significant

amount of time was allowed for baseline and

reperfusion measurements (1 and 4 min,

respectively), it potentially still may impact the blood

flow in the control (left) hand during 200 mmHg

occlusion of the right hand. To mitigate this risk,

more time (e.g. 10 min) needs to be allowed between

experiments in future.

In future work we plan to compare rPPG with

contact PPG, which also measures microcirculation in

skin, and investigate other frequency ranges.

5 CONCLUSIONS

We have performed the analysis of low-frequency (0.1

Hz) components of remote photoplethysmography

signals during arterial occlusion and reperfusion at 150

and 200 mmHg. Our preliminary results show that the

systemic response is similar at both levels of

occlusions.

Inﬂuence of Arterial Occlusion at Various Cuff Pressures on Systemic Circulation Measured by rPPG

317

ACKNOWLEDGEMENTS

We would like to thank the volunteers who

participated in our study, without whom our work

would not have been possible. The authors

acknowledge funding from NSERC Alliance (A.D),

NSERC Personal Discovery (A.D. and G. S.), and

Toronto Metropolitan University Faculty of Science

Discovery Accelerator program (G.S.).

REFERENCES

Burton, T., Saiko, G., Cao, M., & Douplik, A. (2023).

Remote photoplethysmography with consumer

smartphone reveals temporal differences between

glabrous and nonglabrous skin: Pilot in vivo study.

Journal of Biophotonics, 16(1), e202200187.

https://doi.org/10.1002/jbio.202200187

Burton, T., Saiko, G., &Douplik, A. (2022). Remote PPG

Imaging by a Consumer-grade Camera under Rest and

Elevation-invoked Physiological Stress Reveals Mayer

Waves and Venous Outflow: Proceedings of the 15th

International Joint Conference on Biomedical

Engineering Systems and Technologies, 153–159.

https://doi.org/10.5220/0010883100003123

Doyen, B., Matelot, D., & Carré, F. (2019). Asymptomatic

bradycardia amongst endurance athletes. The Physician

and Sportsmedicine, 47(3), 249–252.

https://doi.org/10.1080/00913847.2019.1568769

Joannides, R., Bellien, J., &Thuillez, C. (2006). Clinical

methods for the evaluation of endothelial function – a

focus on resistance arteries. Fundamental & Clinical

Pharmacology, 20(3), 311–320.

https://doi.org/10.1111/j.1472-8206.2006.00406.x

Julien, C. (2006). The enigma of Mayer waves: Facts and

models. Cardiovascular Research, 70(1), 12–21.

https://doi.org/10.1016/j.cardiores.2005.11.008

Kanchanathepsak, T., Pukrittayakamee, N. C.,

Woratanarat, P., Tawonsawatruk, T., &

Angsanuntsukh, C. (2023). Limb occlusion pressure

versus standard tourniquet inflation pressure in minor

hand surgery: A randomized controlled trial. Journal of

Orthopaedic Surgery and Research, 18(1), 539.

https://doi.org/10.1186/s13018-023-04000-3

Lenders, J., Janssen, G. J., Smits, P., &Thien, T. (1991).

Role of the wrist cuff in forearm plethysmography.

Clinical Science (London, England: 1979), 80(5), 413–

417. https://doi.org/10.1042/cs0800413

Li, Z., Leung, J. Y., Tam, E. W., &Mak, A. F. (2006).

Wavelet analysis of skin blood oscillations in persons

with spinal cord injury and able-bodied subjects.

Archives of Physical Medicine and Rehabilitation,

87(9), 1207–1212; quiz 1287.

https://doi.org/10.1016/j.apmr.2006.05.025

Saldin, T. (2019). Assessing Endothelial Dysfunction

Estimating the Differences Between 3 Minute and 5

Minute Reactive Hyperemia. Master’s Theses.

https://doi.org/10.15368/theses.2019.4

Vital Signs (Body Temperature, Pulse Rate, Respiration

Rate, Blood Pressure). (2022, June 14).

https://www.hopkinsmedicine.org/health/conditions-

and-diseases/vital-signs-body-temperature-pulse-rate-

respiration-rate-blood-pressure

Vital Signs: How to Check My Vitals at Home. (n.d.).

Cleveland Clinic. Retrieved November 21, 2023, from

https://my.clevelandclinic.org/health/articles/10881-

vital-signs

BIOIMAGING 2024 - 11th International Conference on Bioimaging

318