Study and Evaluation of Palmar Blood Volume Pulse for

Heart Rate Monitoring in a Multimodal Framework

Hugo Silva

1,2

, Joana Sousa

1,2

and Hugo Gamboa

1,3

PLUX - Wireless Biosignals, S.A., Av. 5 de Outubro, n. 70 - 6, 1050-059 Lisbon, Portugal

IT - Instituto Superior T

ecnico, Av. Rovisco Pais, n. 1, 1049-001 Lisboa, Portugal

CEFITEC - Faculdade de Ci

encias e Tecnologia, Universidade Nova de Lisboa,

2829-516 Caparica, Portugal

Abstract. Within the ﬁeld of biosignal acquisition and processing, there is a

growing need for combining multiple modalities. Clinical psychology is an area

where this is often the case, and one example are the studies where heart rate

and electrodermal activity need to be acquired simultaneously. Both of these pa-

rameters are typically measured in distinct anatomical regions (the former at the

chest, and the later at the hand level), which raises wearability issues as in some

cases two independent devices are used; ﬁnger clip sensors already enable heart

rate measurement at the hand level, however they can be limiting for free living

and quality of life activities. In this paper we perform a study and evaluation of

an experimental blood volume pulse sensor, to assess the feasibility of measuring

the heart rate at the hand palms, and thus enabling the design of more convenient

systems for multimodal data acquisition.

1 Introduction

Blood Volume Pulse (BVP) sensors are a commonly used method for assessing the

cardiovascular activity at the arterial level [1]. Their operating principle is based on

photoplethysmography, that is, by externally applying a light source in the visible or

invisible wavelengths to the tissues, and measuring the amount of light that reaches

a photodetector [2]. The detector can be positioned to measure either by reﬂection or

transmission; as the heart pumps blood through the arteries, and subsequently through

the peripheral vessels, the translucency of the vessels changes due to the increased

blood volume, modifying the way that the emitted light is reﬂected or transmitted to the

photodetector [3]. The typical output of the sensor is then a signal where each cardiac

cycle is expressed as a pulse wave (Figure 1(a)).

In this paper we present a study and evaluation of a palmar BVP sensor for heart rate

measurement, designed to be integrated in multimodal systems for biosignal acquisition

at the hand level. The studied arrangement further expands the current state-of-the-

art in the ﬁeld by improving wearability aspects. Unlike existing systems, taking into

account that the sensor is placed at the hand palm, no additional volume is moved

on to the ﬁngers, enabling the wearer to make normal use of the hand. Experimental

results have shown a good correlation between the measurements taken at the hand

palm and at the ﬁngertip, validating the palmar placement as an adequate alternative to

Silva H., Sousa J. and Gamboa H..

Study and Evaluation of Palmar Blood Volume Pulse for Heart Rate Monitoring in a Multimodal Framework.

DOI: 10.5220/0003884900350040

In Proceedings of the 2nd International Workshop on Computing Paradigms for Mental Health (MindCare-2012), pages 35-40

ISBN: 978-989-8425-92-8

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

(a) Pulse wave signal (b) Common sensor placement

Fig. 1. Example of a Blood Volume Pulse (BVP) signal and typical ﬁngertip sensor.

standard approaches. The remainder of the paper describes the motivation for our work

in Section 2, presents the materials and methods in Section 3, and highlights the results

and conclusions in Sections 4 and 5 respectively.

2 Motivation

The heart rate has became a widespread biosignal measurement for self-management

and assessment, with extensive application in a variety of contexts such as clinical diag-

nosis, sports activities, and affective computing, among others [4–6]. Chest straps have

risen as a standard for unobtrusive data acquisition, and although these form factors are

concealed by the clothing and thus, discrete for regular use, a growing number of ap-

plications require the measurement of multiple parameters simultaneously. This is the

case of clinical psychology [7, 8], where besides the heart rate and derived information,

electrodermal activity is also used as a psychological or physiological arousal indica-

tor. As electrodermal activity is more noticeable at the hand palms and ﬁngers, several

usability difﬁculties arise when both parameters need to be measured simultaneously,

since two independent devices are generally required.

Due to the close relation between BVP signals and the cardiac activity, the heart rate

is one of the parameters generally extracted from the collected data. Furthermore, BVP

sensors can be applied to any place in the body irrigated by blood vessels, particularly in

peripheral areas, making it an appealing alternative to the more intrusive cheat mounted

apparatuses. The most common placement is at the ﬁngertip [9] (Figure 1(b)); however,

for long term or continuous use, this anatomic placement limits the regular activities

of it’s wearer as the ﬁngers play a fundamental role in most daily tasks. Furthermore,

for patients with schizophrenia, psychosis, autism, among other, to place biomedical

sensors at the ﬁnger is hard to the medical staff, because these patients have high move-

ments and tendency to remove the ﬁnger clip. Thus, a BVP glove can overcome some

of these issues, since the patient wears a glove, which is hard to remove and it as not

being subject to the ﬁngers movements.

From BVP signal, Heart Rate (HR) and RR intervals can be extracted, allowing the

Heart Rate Variability (HRV) analysis in terms of time and frequency domains. HRV is

widely related to the mental diseases, since it can provide information about the emo-

tional state of the patient, by its correlation with the automatic heart modulation being

a powerful tool for clinical use. For example, the HRV spectral analysis reveals that

the vagal activity is the major contribution to the high frequencies component. Fur-

thermore, it is also argued that the low frequency (LF) and high frequency (HF) ratio

reﬂects the sympathovagal balance or the sympathetic modulation. These parameters

are very important since some authors defend that there is a signiﬁcant increase of LF

band and signiﬁcant increment of the HR for panic disorders patients and signiﬁcant

lower values of R-R intervals and HF peak of spectral analysis in depressive patients

than in the health people.

In addition to the HR and HRV analysis, the BVP sensor can be updated with other

emitter, becomes in an oximetry sensor, which allows to measure both HR and oximetry

parameters. BVP sensor can also be combined with an ECG sensor and through the

Pulse Transient Time (PPT) technique allowing to measure the blood pressure.

Thus, sensors designed for other areas such as the earlobe can also be found, nonethe-

less, the placement at the hand level is particularly advantageous for the design of mul-

timodal measurement units, as it enables the combination of BVP signals with other

parameters (e.g. electrodermal activity).

3 Methodology

Our work targeted the comparison of heart rate measurements taken at the hand palm

and at the ﬁngertip, aiming at the feasibility evaluation of palmar sensors for that pur-

pose. We used a set of two BVP sensors, a standard transmissive sensor placed at the tip

of the ring ﬁnger, and an experimental reﬂective sensor in a side-by-side, dual-emitter,

single detector conﬁguration (Figure 2(a)), placed at the base of the hypothenar emi-

nence, near the point where it meets the wrist (Figure 2(b)). This palmar location was

identiﬁed as the least affected area within hand whenever ﬁnger movements occur. The

palm sensor was compared with a ﬁnger tip since this last one is considered the gold

standard to measure heart rate based on PPG technique for clinical practice. Further-

more, the ﬁnger tip used as standard it was also validated by [10], which compared

the HR and HRV extracted from the ﬁnger tip sensor and ECG sensor. The authors

have shown a strong positive association between all the parameters calculated using

the ﬁnger BVP and ECG signals.

A bioPLUX research wireless biosignal acquisition unit was used for data acquisi-

tion, enabling synchronous sampling and real-time transmission of the collected data

to a base station. For the measurements at the ﬁngertip, a bvpPLUX sensor was used;

for the experimental palmar sensor, the base circuitry of a bvpPLUX was used but the

emitter/detector arrangement was adapted according to what previously described. Raw

sensor data was acquired with 1000Hz sampling rate and 12-bit resolution; through of-

ﬂine signal processing, we computed the heart rate information and performed statisti-

cal analysis of the results. Figure 3 depicts an example of the raw signals obtained with

each of the sensors.

(a) Palmar sensor (b) Hand placement

Fig. 2. Blood Volume Pulse (BVP) sensor and placement at the hand palm.

Fig. 3. Palmar and ﬁngerprint BVP signal.

4 Results

Experimental evaluation was performed on a group of 10 healthy subjects, composed

by 5 males and 5 females, with an average age of 26±3 years, and quietly sat in order to

avoid movement artifacts. We intend to validate the palm position with the ﬁnger and so

no movement artifact must be in the signal. These artifacts mask the BVP signals, not

allowing to visualize the raw BVP signal and extract the cardiac and signal parameters.

Data was acquired at rest for a period of 5 minutes, and analysed with respect to the

accuracy of heart rate calculation (HR), signal-to-noise ratio (SNR), and root mean

square (RMS). Table 1 outlines the main results; which revealed a very strong positive

association between the measurements obtained by both sensors.

The heart rate exhibits a coefﬁcient of determination R

> 98%, and a p− value >

0.05, showing that the outcomes are not statistically different. The measurement di-

vergence between both sensors is 1bpm at most, as shown by the average heart rate

difference of 0.25bpm ± 0.45. Another interesting ﬁnding is a pulse latency between

Table 1. Fingertip and hand palm BVP sensors comparison.

Fingertip Hand Palm

M ± SD M ± SD

HR (bpm) 70.13 ± 10.51 70.38 ± 10.07

SNR (dB) 132.70 ± 61.43 173.30 ± 60.21

RMS (V) 0.22 ± 0.12 0.07 ± 0.05

the hand palm and the ﬁngertip; from the collected data, an average 0.14 ± 0.38s lag

was found, with the maximum being 0.73± 0.08s for one of the tested subjects, and the

minimum being 0s for two others. At the palmar level, the experimental BVP sensor

used in the tests also presented a higher SNR, and a lower RMS.

5 Conclusions

The heart rate is one of the most widely adopted biomedical indicators in use to date,

and state-of-the-art research has focused on ﬁnding wearable and easy to use acquisi-

tion methods applicable in free living and quality of life activities. While chest straps

currently provide a convenient acquisition possibilities, with a very high level of dis-

cretion, a problem arises in applications that require the combination of heart rate with

other indicators that need to be measured in body areas other than the chest.

Clinical psychology is one such area, where measurements as the electrodermal

activity (typically measured at the hand palms or ﬁngers), also provide important psy-

chophysiological information. Blood Volume Pulse (BVP) sensors, stand as alternative

method for heart rate assessment at the hand level, however current measurement ap-

proaches greatly limit the subjects normal activities, as they require the sensor to be

clipped to a ﬁnger. In our work, we further extend the stat-of-the-art in the ﬁeld, by

evaluating the feasibility of using the BVP for heart rate assessment at the palmar level,

thus enabling the design of more integrated and practical multimodal measurement sys-

tems.

Results obtained from real-world data have revealed a high correlation between

measurements taken at the hand palm and at the standard ﬁnger tip location (R

98%), as shown also by the average divergence of 0.25bpm. Another interesting ﬁnd-

ing was a latency between the palmar and ﬁnger BVP pulses; this, together with further

validation on a larger set of test subjects will be the focus of future work within our

group and in daily-use condition. This will allow to lead ﬁlter studies in order to under-

stand how to remove the movement artifacts into the BVP signal and, consequently, to

improve the feasibility of the measures. Furthermore, usability and hygiene conditions

will be also evaluated, which will enable to make adjustment in the glove and become

it more comfortable and usable for clinical practice.

Acknowledgements

This work was partially funded by the Seventh Framework Programme (FP7) under the

”ICT4Depression” project (ref. 248778), and by the Fundac¸

ao para a Ci

encia e Tecnolo-

gia (FCT) under the grant SFRH/BD/65248/2009, whose support the authors gratefully

acknowledge.

References

1. Duun, S., Haahr, R., Birkelund, K., Thomsen, E.: A ring-shaped photodiode designed for use

in a reﬂectance pulse oximetry sensor in wireless health monitoring applications. Sensors

Journal, IEEE 10 (2010) 261–268

2. Webster, J.: Design of Pulse Oximeters. Taylor & Francis (1997)

3. Reisner, A., Shaltis, P., McCombie, D., Asada, H.: Utility of the photoplethysmogram in

circulatory monitoring. Anesthesiology 108 (2008) 950

4. Picard, R. W.: Affective Computing. 1st edn. The MIT Press (2000)

5. Chung, E. K.: Pocketguide to ECG Diagnosis. Blackwell Publishing Professional (2000)

6. Haahr, R., Duun, S., Thomsen, E., Hoppe, K., Branebjerg, J.: A wearable electronic patch

for wireless continuous monitoring of chronically diseased patients. In: Medical Devices and

Biosensors, 2008. ISSS-MDBS 2008. 5th International Summer School and Symposium on,

IEEE (2008) 66–70

7. Boucsein, W.: Electrodermal Activity. 1st edn. Springer (1992)

8. Trull, T. J.: Clinical Psychology (with InfoTrac). 7 edn. Wadsworth Publishing (2004)

9. Nonin: Nonin - nonin medical inc. [online] avalilable at: http://www.nonin.com/ [accessed

5 september 2011] (2011)

10. Medeiros, J., Martins, R., Palma, S., Gamboa, H., Reis, M.: Development of a blood volume

pulse sensor to measure heart rate variability. In: Proceedings of the IBEROSENSOR 2010.

(2010)