Multi-biosignals Analysis

The Effect of Peripheral Nerve Stimulation on Skin Conductance and Heart Rate

Variability

Tiago Araújo

, Pedro Dias

, Neuza Nunes

and Hugo Gamboa

Department of Physics, Faculty of Sciences and Technology, New University of Lisbon, Lisbon, Portugal

PLUX - Wireless Biosignals S.A., Lisbon, Portugal

Keywords: Skin Conductance, Blood Volume Pulse, Heart Rate Variability, Peripheral Nerve Stimulation.

Abstract. Objectives: This study aims to evaluate the influence of standard electrical stimulation on human

electrophysiology. Methods: A total of 10 healthy subjects were submitted to the same protocol. The

electrical stimuli were applied on the median nerve of the left wrist. Blood Volume Pulse (BVP) and

Electrodermal Activity (EDA) signals were acquired from the index finger through an oximeter and from

both the abductor pollicis muscle and the 3

palmar interosseous muscle of the right hand, respectively.

Nerve stimulation was performed using increasing intensities current: range from 5 to 30 mA, with 1mA

step and applying 20 stimuli per step. Heart Rate (HR) and Heart Rate Variability (HRV) were computed,

from the analysis of the latency between BVP pulses, in basal state and during stimulation. EDA parameters

response latency, response rise time and readaptation slope were computed for each burst.

Discussion: Electrical stimulation reveals to influence several parameters of the Autonomic Nervous System

(ANS). It was easily detected an EDA rise response for each of the applied bursts and also an increase of the

HRV during stimulation.

1 INTRODUCTION

Recently, the importance of studying biosignals has

been increasing due to its role finding physical and

mental stress. It has already been determined that the

Autonomic Nervous System (ANS) exerts a constant

influence over heart rate (HR) and skin conductance.

Despite all the studies, there has always been a gap

in the understanding of the effects of electrical

stimulation in human physiology (Shetter, A., 1997;

Dimitrijevic, M., 2008).

Skin Conductance (Electrodermal Activity) has

been reported as a potential non-invasive marker for

sympathetic activity, and has been used recently in

psychophysiological research. The Electrodermal

activity (EDA) measurement is based on content of

water and electrolytes in individual parts of the

organism and the spreading of low electrical current

through two electrodes localized on skin surface.

The conductance depends on the amount of

sweat produced by eccrine glands, which are

regulated by the sympathetic nervous system

(Visnovcova, 2013). It’s the sweat level changes that

modify the resistance, and alterations in the EDA

signal are noticed.

The total EDA is composed of a baseline and a

phase level. The baseline is the EDA on a daily

basis. Providing the EDA is not a constant value it

was noted that its oscillations are very small over a

small time interval. Therefore we considered the

EDA baseline to be linear. The phase level is a

variation of a subject’s EDA due to external stimuli

applied on the wrist and reflects the arousal of the

sympathetic nervous system (Dominik, B., 2010).

The EDA parameters usually analysed are the

response latency, response rise time and readaptation

slope.

EDA latency is the time interval between the

application of the first stimulus and the detection of

a response from the subject’s EDA. EDA rise time is

the time taken by an EDA response to rise and peak.

Finally, EDA readaptation slope is defined as the

slope of the line obtain from a linear regression with

all data points past the last stimulus.

The heart rate is very sensitive and readapts

quickly to different stimuli. The sympathetic and

parasympathetic neurons, which are linked to the

sinoatrial node in the heart, have a major

contribution to changes in its beating rate (although

there are more contributions, like the physical and

235

Araújo T., Dias P., Nunes N. and Gamboa H..

Multi-biosignals Analysis - The Effect of Peripheral Nerve Stimulation on Skin Conductance and Heart Rate Variability.

DOI: 10.5220/0005217902350239

In Proceedings of the International Conference on Bio-inspired Systems and Signal Processing (BIOSIGNALS-2015), pages 235-239

ISBN: 978-989-758-069-7

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

mental state). Heart rate variability (HRV) is the

physiological phenomenon in which the heart rate

oscillates around its mean value due to external

influences from the autonomic nervous system

(ANS), or other mental and physical factors (Malik,

M., 1990; Malik, M., 2006). It’s a marker to study

the activity of the regulatory mechanisms and to

analyse the effects (excitatory and inhibitory) they

have upon the heart rate.

HRV can be analysed on time or frequency

domains. Through the analysis of the root mean

square of the

successive RR intervals we can obtain a

measure of the sympathetic and parasympathetic vagal

activity (Clifford, G., 2002). One way to measure the HR

is through the Blood Volume Pulse (BVP) sensor. With

the cardiac pulse, erythrocytes will change the spatial

alignment and this will affect the blood’s opacity. The

BVP sensor makes use of opacity variations from the tip

of the finger to gather the cardiac pulse on a certain

instant.This study

aims to analyse the effects of

electrical stimulation on human physiology, based

on skin’s galvanic response and on Blood Volume

Pulse (BVP) signals.

2 METHODS

2.1 Subjects and Room Conditions

In this study a total of 10 healthy subjects, composed

of 5 males and 5 females, with a mean age of 24.10

years (standard deviation of 2.38 years), were

submitted to a previously developed heart rate and

skin conductance acquisition protocol. All of the

subjects were healthy, with no physical or

neurophysiological disorders registered.

The subjects were seated on a comfortable chair

with both their arms relaxed. It was required of them

to feel comfortable for the acquisition to begin.

Earmuffs were placed on the subjects’ ears to

prevent any noise distraction. Moreover the room

was kept silent during the acquisition and the

subjects were asked to remain motionless and

relaxed.

The stimuli were applied through two disposable

electrodes on the median nerve of the left wrist and

two electrodermal activity acquisition electrodes

were placed on the abductor pollicis muscle and on

the 3

palmar interosseous muscle of the right hand.

An oximeter was also placed on the index finger of

the right hand to measure the heart rate of the

subjects. The sensor makes use of colour variations

from the tip of the index finger (due to blood’s

opacity) to gather the blood volume on a certain

instance in time.

Although the majority of the subjects didn’t

know about the acquisition protocol, the acquisition

had to be repeated on two of the subjects (1 and 2)

due to low quality signal reading.

2.2 Acquisition Protocol

The first part of the acquisition took 4 minutes and

no stimuli were applied to the subjects. This segment

of the acquisition protocol had the purpose of

acquiring the basal BVP and EDA.

The second segment featured the application of

the electrical stimuli on the medial nerve of the left

wrist. Square current modulated waveform was

used. The electrical stimuli consisted of 6 bursts of

20 repetitive stimuli each and with increasing

intensities (ranging from 5 to 30 mA). In each burst

the interval between stimuli was 0.9 seconds and the

time interval between bursts was of 10 seconds.

The final segment of the acquisition was similar

to the first one, a 4-minute acquisition to record the

subjects’ recovery after the application of stimuli.

Through all of the phases, these segments were

recorded at a 2000Hz acquisition rate. A combined

wireless, miniaturized and synchronized unit was

specifically developed for multi-biosignal

acquisition (Plux, 2014) and nerve stimulation

(Araújo, T., 2012).

2.3 Processing

After the acquisition from all subjects the data was

processed using a script in Python. The data was

stored in a .txt file (for each subject) and organized

in columns, each column belonging to a certain

parameter and each line to a recorded frame.

The data recorded was processed according to

steps bellow:

1. Conversion of frames to seconds, since 1000

frames corresponded to a single second.

2. Search for all HR peaks. HR data was plotted

in a time/HR graph to ensure all peaks found by the

script matched the plotted peaks.

3. Determination of mean heart rate for all 3

segments.

4. Determination of HR amplitude and HRV for

all 3 segments.

5. Analysis of EDA time latency, response rise

time and readaption slope via plotting and linear

regression.

These steps were repeated for all 10 subjects.

BIOSIGNALS2015-InternationalConferenceonBio-inspiredSystemsandSignalProcessing

236

3 RESULTS

Although it might be a difficult task to study the

influence of electrical stimulation due to external

factors, which are hard to control, the parameters

chosen provide a subjective analysis of the subjects’

reactions to stimulation. Parameters chosen for EDA

include EDA latency time, EDA rise time and EDA

readaptation slope.

Table 1 shows the HR and HRV for each subject

by determining standard deviation of the successive

hearth pulses intervals.

Table 1: Mean values and respective standard deviation

error for the heart rate (in beats per minute) obtained

during segments 1, 2 and 3.

Heart Rate (bpm)

Segment 1 Segment 2 Segment 3

Mean

(SD) Mean (SD) Mean (SD)

Subject

80.32 (8.79) 81.57 (8.24) 79.00 (8.26)

Subject

66.57 (7.99) 64.87 (5.91) 66.99 (8.24)

Subject

69.52 (2.89) 71.96 (3.81) 69.32 (3.25)

Subject

64.28 (5.57) 61.05 (4.15) 64.49 (5.44)

Subject

70.09 (5.62) 64.40 (4.88) 69.84 (5.38)

Subject

72.43 (5.68) 69.80 (3.72) 72.07 (4.66)

Subject

67.14 (4.40) 71.01 (6.43) 65.47 (4.85)

Subject

83.23 (5.53) 83.85 (5.86) 83.19 (5.97)

Subject

61.94 (6.32) 61.74 (5.01) 60.63 (6.36)

Subject

63.17 (4.22) 62.37 (3.22) 63.18 (3.68)

Figure 1 shows the subjects’ EDA during segment 1

and segment 2. Regarding the subject’s EDA from

segment 1 (Figure 1A), analysis of the results

showed that prior to the application of the stimuli

small oscillations were detected but deemed of low

importance. For segment 3, the behaviour was the

same as presented in Figure 1A. The EDA signals

from all subjects during segment 2 of the acquisition

protocol are presented in Figure 1B.

Table 2 presents EDA latency for all subjects and

for each burst of 20 stimuli. Negative values are

result of a response that was detected before the

burst was applied. There was no measurable

response from subjects 8 and 10 and subject 5

presented a signal with very low activity, which was

particularly hard to analyse.

Table 3 presents the EDA rise time for each of

the detected subjects’ EDA pulses to rise and peak.

4 DISCUSSION

The EDA signal in its absolute value has high inter-

subject variability, which makes it difficult to

establish a measurement range capable of

comprising significant population. To overcome this

effect, it is mandatory to establish amplitude / gain

independent objective parameters. In this study, only

7/10 subjects revealed significant responses of

sympathetic nervous system to the electrical

stimulation. For segment 1, EDA results showed, in

general, a continuous smooth descent tendency for

all subjects (Fig. 1A). For segment 2, EDA results

showed, for most of the subjects, that the EDA

signal has some events in response to the electric

stimulus (Fig. 1B). This allows the analysis of the

time latency between the reception of the stimulus

by the median nerve and the ANS reaction to it.

The stimulus/response latency is a parameter

inherent to the personal physiology and

consequently has high inter-subject variability. In

half of the subjects analysed, a slight tendency to

decrease the latency with the increase of the

stimulation intensity is noticed. This was expected

given the ANS constant re-adaptations.

One subject reveals a notorious response arousal

even before the stimulus application. Curiously, due

to poor quality of the acquisition for this subject, the

protocol needed to be repeated and the results here

exposed are from the second acquisition. This was

the only subject who had been exposed to the

protocol before.

When analysing the EDA rise time, a lower

standard deviation both intra and inter subject is

observed, when compared with Segment 1. The

response rise time seems to be a parameter which

does not correlate with the increase of stimulus

intensity, presenting very stable results within the

same subject.

Another pointer for the ANS management used

in this work was the analysis of BVP signal.

The BVP signal constitutes per se a direct signal

from the vascularization physiology. With the

cardiac pulse, erythrocytes will change the spatial

alignment and this will affect the blood’s opacity.

Multi-biosignalsAnalysis-TheEffectofPeripheralNerveStimulationonSkinConductanceandHeartRateVariability

237

Figure 1: EDA signals obtained for all subjects. A) Example window obtained from segment 1 of the acquisition protocol

(basal activity); B) Signal obtained from segment 2 of the acquisition protocol (six bursts of electrical stimulation).

Table 2: Stimulus response latency of EDA signal for each stimulation burst.

Latency (s)

Burst 1

5mA

Burst 2

10mA

Burst 3

15mA

Burst 4

20mA

Burst 5

25mA

Burst 6

30mA

Mean

(SD)

Subject 1

2,05 1,60 1,30 0,92 2,34 1,19

1,57 (0,49)

Subject 2

0,00 0,00 0,95 1,00 0,88 0,89

0,63 (0,43)

Subject 3

-- -- -- -- -- --

Subject 4

0,92 0,83 0,80 1,14 1,16 0,73

0,93 (0,17)

Subject 5

-- -- -- -- -- --

Subject 6

2,42 2,16 1,55 0,59 1,02 0,97

1,45 (0,66)

Subject 7

-- * 1,20 4,40 -- 1,92 -- *

2,51 (1,37)

Subject 8

2,94 2,00 2,00 1,66 1,70 1,13

1,91 (0,55)

Subject 9

2,81 2,80 3,00 -- -- 0,99

2,40 (0,82)

Subject 10

-- -- -- -- -- -- --

*subject with event previous to stimulation

Table 3: Stimulus response rise time of EDA signal for each stimulation burst.

Rise Time (s)

Burst 1

5mA

Burst 2

10mA

Burst 3

15mA

Burst 4

20mA

Burst 5

25mA

Burst 6

30mA

Mean (SD)

Subject 1

2,74 4,00 3,70 3,00 2,36 3,94

3,29 (0,63)

Subject 2

4,74 4,19 4,55 4,92 3,47 4,54

4,40 (0,47)

Subject 3

-- -- -- -- -- --

Subject 4

7,63 3,13 6,40 3,28 2,93 5,60

4,83 (1,82)

Subject 5

-- -- -- -- -- --

Subject 6

3,14 3,15 3,93 5,03 2,50 2,83

3,43 (0,84)

Subject 7

-- * 3,10 2,59 -- 3,00 -- *

2,89 (0,22)

Subject 8

3,84 4,00 4,09 3,75 4,20 2,90

3,79 (0,43)

Subject 9

5,68 4,56 4,30 -- -- 3,93

4,62 (0,65)

Subject 10

-- -- -- -- -- -- --

*subject with event previous to stimulation

BIOSIGNALS2015-InternationalConferenceonBio-inspiredSystemsandSignalProcessing

238

The BVP signal of 5/10 subjects showed an

amplitude decreasing with the beginning of the

electrical stimulation. An example of this pattern can

be seen on the subject of Fig. 4. This is justified by a

vasoconstriction effect caused by adaptations of the

ANS to the electrical stimulation. Vasoconstriction

of the blood vessels increases the spatial density and

consequent alignment of erythrocytes. This fact will

increase the opacity of the finger, leading to lower

signal detection by the sensor used for the detection

of the BVP signal.

The BVP peaks also enable the analysis of the

HR and HRV through the computation of the

standard deviation (SD) of subsequent peaks latency.

Regarding the HR analysis it was observed that the

majority of the subjects (5/10) showed a decrease in

the HR and HRV during stimulation. In 2/10

subjects the opposite effect is verified: the HR and

HRV increase during the stimulation period. The

remainder subjects do not show a direct influence of

the stimulation stage.

For the majority of the subjects, it is

understandable that the electrical stimulation has an

influence on the ANS, which affects the vagus nerve

and thus the sinus node. For all subjects, on Segment

3, HR and HRV return to the Segment 1 basal

values. We can assume that after the stimulation the

subject’s ANS recovered easily. Therefore,

peripheral nerve stimulation did not have any short-

term consequences on the subjects’ HRV.

Electrical stimulation revealed to influence

several parameters of ANS. This influence can now

be taken into account on standard uses of peripheral

nerve stimulation.

REFERENCES

Visnovcova Z., Calvoska A. and Tonhajzerova I. (2013)

Heart rate variability and electrodermal activity as

noninvasive indices of sympathovagal balance in

response to stress, Comenius University, Slovak

Republic.

Thong, T., Li, K, Mcnames, J., Aboy, M., Goldstein, B.

(2003) Accuracy of Ultra-Short Heart Rate Variability

Measures, Oregon, United States of America.

Malik, M. (2006) Heart rate variability – Standards of

measurement, physiological interpretation and clinical

use, American Society for Pacing and

Electrophysiology.

Clifford, G., Tarassenko, L. (2002) Signal Processing

methods for Heart Rate Variability, Department of

Engineering Science, University of Oxford.

Dominik B., Guillaume F., Friston, K., Dolan, R. (2010),

Modulating event-related skin responses, University

College London, United Kingdom.

Malik, M., Camm, A. (1990) Heart rate variability –

Electrophysiology, pacing and arrhythmia, St.

George’s Hospital Medical School, London, England.

Araujo, T., Nunes, N., Gamboa, H. (2012) Miniaturized

wireless controlled electrostimulator. In proceedings

of biodevices – international conference on biomedical

electronics and devices. Biostec 2012, Vilamoura,

Portugal.

PLUX – Wireless Biosignals, SA (2014) [Online]

Available from http://www.plux.info [Accessed July

, 2014]

Multi-biosignalsAnalysis-TheEffectofPeripheralNerveStimulationonSkinConductanceandHeartRateVariability

239