King’s Research Portal
DOI:
10.1109/TNSRE.2019.2956836
Document Version
Peer reviewed version
Link to publication record in King's Research Portal
Citation for published version (APA):
Dong, J., Geng, B., Khan Niazi, I., Amjad, I., Dosen, S., Jensen, W., & Kamavuako, E. N. (2020). The Variability
of Psychophysical Parameters following Surface and Subdermal Stimulation: A Multiday Study in Amputees.
IEEE transactions on neural systems and rehabilitation engineering , 28(1), 174-180. [8918067].
https://doi.org/10.1109/TNSRE.2019.2956836
Citing this paper
Please note that where the full-text provided on King's Research Portal is the Author Accepted Manuscript or Post-Print version this may
differ from the final Published version. If citing, it is advised that you check and use the publisher's definitive version for pagination,
volume/issue, and date of publication details. And where the final published version is provided on the Research Portal, if citing you are
again advised to check the publisher's website for any subsequent corrections.
General rights
Copyright and moral rights for the publications made accessible in the Research Portal are retained by the authors and/or other copyright
owners and it is a condition of accessing publications that users recognize and abide by the legal requirements associated with these rights.
•Users may download and print one copy of any publication from the Research Portal for the purpose of private study or research.
•You may not further distribute the material or use it for any profit-making activity or commercial gain
•You may freely distribute the URL identifying the publication in the Research Portal
Take down policy
If you believe that this document breaches copyright please contact librarypure@kcl.ac.uk providing details, and we will remove access to
the work immediately and investigate your claim.
Download date: 09. Aug. 2022
> TNSRE-2019-00349 <
1
Abstract—Electrotactile stimulation has been suggested as a
modality for providing sensory feedback in upper limb prostheses.
This study investigates the multiday variability of subdermal and
surface stimulation. Electrical stimulation was delivered using
either surface or fine wire electrodes placed right under the skin
in eight amputees for seven consecutive days. The variability of
psychophysical measurements, including detection threshold
(DT), pain threshold (PT), dynamic range (DR), just noticeable
difference (JND), Weber fraction (WF) and quality of evoked
sensations, was evaluated using the coefficient of variation (CoV).
In addition, the systematic change in the mean of the parameters
across days was assessed in both stimulation modalities. In the case
of DT, PT, DR, and perceived intensity at 100 Hz, the CoV of
surface stimulation was significantly smaller than that of
subdermal stimulation. Only PT showed a significant systematic
change in the mean value across days for both modalities. The
outcome of this study has implications for the choice of modality
in delivering sensory feedback, though the significance of the
quantified variability needs to be evaluated using usability tests
with user feedback.
Index Terms—Prostheses, surface electrotactile stimulation,
subdermal electrical stimulation, sensory feedback, sensation
variability.
I. INTRODUCTION
round 1.6 million people were living with limb amputation
in the year 2005, and it has been estimated that 3.6 million
people will be living with amputation in the United States of
America by the year 2050 [1]. Currently, some of the
functionality of a lost arm can be replaced by a prosthesis. A
prosthesis is defined as an artificial device that replaces a
biological limb both functionally and morphologically. The
human hand has a highly complex structure that comprises
many degrees of freedom; the hand has remarkable capabilities
in performing dexterous and delicate movements. This is
possible due to the sophisticated closed-loop control integrating
This work was supported by the Danish Ministry of Higher Education and
Science – The Danish Council for Independent Research | Technology and
Production Sciences under Grant 1337-00130.
Jian Dong is with the Department of Orthopedics, The Second Hospital of
Jilin University, Changchun 130041, China. He is now with SMI®, the
Department of Health Science and Technology, Aalborg University, 9220
Aalborg, Denmark (e-mail: jd@hst.aau.dk).
Bo Geng, Strahinja Dosen and Winnie Jensen are with SMI®, the
Department of Health Science and Technology, Aalborg University, 9220
Aalborg, Denmark.
efferent motor output and afferent sensory feedback.
Consequently, mimicking the structure and function of the
human hand using an artificial system is a very challenging
task.
Even though advanced prosthetic hands that can partly
replicate the motor dexterity of a natural human hand are
available (e.g. DEKA Hand and iLimb), a continuing challenge
is to restore the sensory function of the hand. For upper limb
prosthetic users, the absence of sensory feedback impedes the
efficient use of their prostheses, which can lead to user
frustration and abandonment of the device [2]. The sensory
awareness, which is available with body-powered prostheses
due to a direct connection between the gripper and the user's
shoulder, does not exist in myo-electrically controlled systems
[3]. In this case, the users must rely primarily on the direct
observation of the device (visual feedback) [4] and secondarily,
on subtle clues such as the sounds of the motor and transmission
(intrinsic feedback) [5]. Therefore, restoring somatosensory
feedback to the prosthesis user can decrease visual attention and
improve control by providing explicit information about the
state of the device.
The somatosensory feedback can be provided using different
stimulation methods to elicit tactile sensations [6]. Current non-
invasive solutions are mostly based on delivering
electrocutaneous stimulation [7], or vibration [8] to the skin on
the residual limb. The residual limb can also be stimulated
mechanically (e.g. pushing the limb by using a force applicator,
squeezing the limb by a cuff, or by stretching the skin) [9]. The
feedback can be restored through invasive methods as well, i.e.
by electrically stimulating peripheral nerves [10]. In this case,
the aim of the stimulation is to activate the same neural
structures that have been used before the amputation, leading to
somatotopic feedback. The same result may be achieved using
non-invasive methods by delivering the stimulus to the
Imran Khan Niazi is with New Zealand College of Chiropractic, 1060
Auckland, New Zealand; SMI®, Department of Health Science and
Technology, 9220 Aalborg University, Denmark; Health and Rehabilitation
Research Institute, AUT University, 1142 Auckland, New Zealand.
Imran Amjad is with Riphah College of Rehabilitation Sciences, Riphah
International University, H-8/2 Islamabad, Pakistan.
Ernest Nlandu Kamavuako is with the Centre for Robotics Research,
Department of Informatics, King’s College London, WC2B 4BG London,
United Kingdom.
The Variability of Psychophysical Parameters
following Surface and Subdermal Stimulation:
A Multiday Study in Amputees
Jian Dong, Bo Geng, Imran Khan Niazi, Imran Amjad, Strahinja Dosen, IEEE Member, Winnie
Jensen and Ernest Nlandu Kamavuako, IEEE Member
A
> TNSRE-2019-00349 <
2
phantom map if it exists on the residual limb [11].
In general, the feedback information is transmitted by
relating a measured prosthesis variable to selected stimulation
parameters. For example, the magnitude of the grasping force
can be communicated using the magnitude or frequency of
stimulation. Therefore, the user needs to learn to relate the
stimulation parameters to the prosthesis state, and this requires
training. The information about grasping force, slippage [12],
hand aperture [13], finger flexion [14], and elbow angle has
been previously encoded and transmitted through sensory
feedback [7], [15].
The electrocutaneous stimulation is an attractive modality to
restore feedback and it has been investigated intensively in the
past [16]. The stimulation can be delivered using simple and
compact circuits and electrodes [17]. Therefore, the
electrotactile interface is convenient for providing multichannel
feedback and integration into a prosthetic socket. Furthermore,
since there are no moving mechanical parts, the stimulation
parameters can be changed fast and independently. This allows
eliciting rich and dynamic tactile sensations. Studies with able-
bodied subjects and amputees have shown that electrotactile
feedback can improve prosthesis control [18], [19].
Nevertheless, a disadvantage of electrocutaneous stimulation
delivered through surface electrodes is that it can produce
uncomfortable and even painful sensations. High voltage is
needed for the stimulation to overcome the skin impedance,
which can also vary depending on the conditions of the
electrode-skin interface. To increase the dynamic range,
between the sensation and pain threshold, larger electrodes are
required. Importantly, these drawbacks may be overcome by
placing the electrodes subdermally, as previously demonstrated
[20], [21]. Subdermal stimulation can lead to substantially more
compact feedback interfaces, since it is based on point
electrodes (wire tip), and it can substantially decrease the
required voltage and current consumption because skin
impedance is bypassed. As a step in this direction,
psychophysical measurements were conducted previously [21]
to evaluate and compare the properties of the surface and
subdermal stimulation.
Ideally, a feedback interface needs to produce stable and
repeatable sensations. This is even more important when using
subdermal stimulation since the electrodes are meant to stay
within the tissue for its lifetime, contrary to surface stimulation
where they will be reapplied with each donning and doffing of
the prosthesis. The short-term stability of subdermal
stimulation has been tested in our previous study for up to eight
hours [22]. However, long-term stability plays an important
role in achieving the long-lasting functional sensory feedback
and verifying its usability in clinical applications for amputees
[23], [24]. In general, the multiday variability of the commonly
used psychophysical measurements has received less attention
in the literature.
Therefore, the aim of this study was to investigate the
variability of psychophysical measurements over the course of
seven days when using subdermal versus surface stimulation in
upper-limb amputees. The psychophysical measurements that
were investigated systematically in the present study were
detection threshold (DT), pain threshold (PT), dynamic range
(DR), just noticeable difference (JND), Weber fraction (WF)
and the subjective quality of evoked sensations.
II. METHODS
A. Subjects
Nine male upper-limb amputees (33.6 ± 12.9 years old, 13.7
± 11.1 years after amputation) were recruited from Railway
General Hospital, Rawalpindi, Pakistan (Table I). Subjects
provided written informed consent and the study adhered to the
Helsinki Declaration. The ethical committee of Riphah
International University (N-ref# Riphah/
RCRS/REC/000121/20012016) approved the study protocol.
All subjects had undergone traumatic amputation of their
dominant hand/arm. None of the subjects abused cannabis,
opioids or other drugs. They had no record of previous
neurological, musculoskeletal or mental illnesses, lack of
ability to cooperate, fear of injections; and they were all
phantom pain-free. One subject was excluded from the study
because a pain threshold could not be reached even with a
current amplitude of 40 mA (the highest possible current to
deliver) and at that high stimulation level, strong muscle
twitches were evoked.
B. Experiment procedure
A single session was performed each day for seven
consecutive days. The psychophysical measurements were
collected in the order of DT, PT, JND and sensation evaluation
in each session. All seven sessions were scheduled at the same
time of the day. The subdermal electrode was disconnected
after each session and it remained under the skin for the
duration of the experiment. The insertion site and the wire were
wrapped with a medical bandage between different sessions, to
minimize displacement of the electrode during daily activities.
The surface electrodes were disposed following each session.
This protocol was selected to mimic the real-life application in
which the surface electrodes are reapplied with each donning
and doffing, while the subdermal electrodes will be placed
permanently. The common ground electrode was also removed
at the end of the session. The same common ground electrode
was reused for three or four sessions depending on the
stickiness and conductivity.
TABLE I
D
EMOGRAPHIC
D
ATA
OF
THE
S
UBJECTS
Subject
Age
Years after
amputation
Amputation level
Cause of
amputation
1
59
31
Transradial
Trauma
2
29
15
Transradial
Trauma
3
19
14
Wrist disarticulation
Trauma
4
30
3
Partial hand
Trauma
5
52
2
Partial hand
Trauma
6
36
31
Wrist disarticulation
Trauma
7
32
7
Wrist disarticulation
Trauma
8
19
14
Partial hand
Trauma
9
26
14
Partial hand
Trauma
All subjects were undergoing dominant side amputation.
> TNSRE-2019-00349 <
3
C. Stimulation
A programmable stimulator (ISIS Neurostimulator, Inomed,
Germany) was used to generate biphasic, rectangular,
symmetric pulses with a pulse width of 200 μs. The stimulator
was controlled by a custom-made program implemented in
LabVIEW version 2015 running on a laptop.
Commercially available surface electrodes (Ambu Neuroline
700, 20 mm 15 mm) and subdermal fine wire electrodes were
used to deliver the electrical stimulation. The subdermal wire
electrodes were made of Teflon-coated stainless steel (A-M
Systems, Carlsborg WA, diameter 50 µm), with 5-mm tip
exposed [21]. Each subject was checked to see if the stump of
his forearm had enough normal skin (the skin without any
visible scar and any abnormal sensation) for electrode
placement. The two stimulation electrodes were positioned on
the dorsal side of the proximal end of the stump. The subject
was seated on a chair with their stump exposed, and the skin of
the dorsal stump was shaved in the area of approximately 2 cm
3 cm. The skin location was cleaned with a 70% alcohol swab
and the wire was inserted subdermally using a 25-gauge
hypodermic needle. The rest of the wire electrode was fixed to
the skin by Fixomull® stretch tape to avoid any displacement.
The surface electrode was placed just next to the wire (Fig. 1).
The pre-gelled common ground electrode (PALS Platinum, 40
mm 64 mm, oval) was applied to the dorsal side of the upper
arm next to the elbow.
D. Psychophysical measurements
1) Detection threshold and pain threshold
The smallest stimulus that can be detected by the subject is
called DT. The DT was measured using a staircase method by
delivering single pulses with an inter-pulse interval of 2 s [25].
To initialize the staircase, an approximate DT was first
determined using the method of limits [26]. The subjects
received a series of pulses gradually increasing in steps of 0.3-
0.5 mA for surface and 0.1-0.3 mA for subdermal stimulation.
The steps were chosen randomly within the indicated range to
avoid any anticipation bias by the subjects [22]. The amputee
reported verbally when he first felt the stimulation. This
amplitude was then used as the initial amplitude in the staircase
procedure. During the staircase testing, a series of stimuli were
delivered to the subjects with the amplitude that was adjusted
adaptively based on subject responses. After each pulse, the
subject reported if he felt the stimulation. If the subject detected
the stimulus, the amplitude was increased, otherwise decreased
(in steps of 0.03-0.05 mA for surface and 0.01-0.03 mA for
subdermal stimulation). The amplitude changes from ‘increase’
to ‘decrease’ or vice versa was defined as a ‘reversal’. The
staircase procedure stopped after 10 reversals or after 30 stimuli
were delivered. The DT was computed as the average of the last
seven reversals.
The stimulus amplitude at which the subject starts to feel pain
is referred to as the PT. The PT was measured using the method
of limits [26] by delivering a single pulse with increasing
amplitude in steps of 0.3-0.5 mA for surface and 0.1-0.3 mA for
subdermal stimulation [22] and with an inter-pulse interval of 2
s. Three measurements were performed, and the PT was
determined as the average of those measurements.
The DR was calculated by dividing PT by DT. A larger DR
indicates that a wider range of electrical stimulation amplitudes
is tolerated by the subject and may also generate a wider range
of sensations (more room to operate).
2) Just noticeable difference
The smallest change in the stimulus amplitude that can be
detected by a subject is called the JND. The JND was
determined using the method of limits [26]. The amplitude of
the baseline stimulus was set at 3×DT in both surface and
subdermal stimulation. If this intensity was higher than the PT,
a lower amplitude of 2×DT or 1×DT was used. Two stimuli
were delivered sequentially and there was a 2-s break between
the pulses. The first pulse was always set to the baseline
amplitude whereas the amplitude of the second pulse was
increased in steps of 0.01-0.11 mA for surface and 0.02-0.07
mA for subdermal. The pairs of pulses were delivered until the
subject reported that he could feel the difference in the intensity.
The difference was recorded as the JND. Finally, the procedure
was repeated three times and the average of the three JNDs was
used for data analysis.
The ratio between the JND and the baseline amplitude is
called WF [18]. The WF was calculated using equation as
follows:
𝑊𝐹 = ΔI/Δ
where ΔI is JND and Δ is the baseline amplitude. According to
Weber’s law, the WF should be approximately constant, and
therefore, it can be used to estimate the JND for different
baselines. Hence, the WF characterizes the resolution of the
perceptual system.
3) Sensation evaluation
A computerized questionnaire was designed (Fig. 2) to
collect the subjective experience of the stimulation [13]. The
subjects were asked to report on the sensation quality, intensity,
comfort, and location. The questionnaire included 12 pre-
defined words that could be selected by the subject to describe
the quality. For the stimulus intensity, the subjects were asked
to indicate a number from a numerical rating scale (NRS),
where 0 represented no stimulation and 10 represented the
Fig. 1. Diagram of electrodes placement.
> TNSRE-2019-00349 <
4
maximum intensity. The stimulus comfort was reported using a
Likert-type scale, where one indicated very comfortable, four
neutral and seven very uncomfortable sensations. To describe
the location of the perception, the subjects could select one of
the three options, namely, 'local', 'radiation' and 'referred'.
'Local' represents that the sensation was just beneath the
electrodes. 'Radiation' indicates that the sensation radiated out
from the electrodes. 'Referred' represents that the sensation
appeared at the other part of the body. All the answers were
recorded on a computer via LabVIEW. The NRS and Likert
scale were implemented using sliders and therefore the subject
could indicate any number within the allowed range.
To evaluate the sensation quality, intensity, comfort, and
location (questionnaire contents), the stimulation was delivered
to the subjects in the form of 1-s pulse trains. The amplitude of
the pulse trains was set to 3×DT and the frequency at 20 Hz and
100 Hz. These frequencies were selected as a) they elicit
different sensations, 20 Hz elicits vibration and 100 Hz fused
tingling, and b) they are within the range typically used for
sensory feedback [27]. The amplitude of 1× DT or 2×DT was
used if 3×DT was over PT. Each frequency was delivered to the
subjects three times, and after each delivery, the subject was
asked to fill in the aforementioned questionnaire. The stimuli at
50 Hz and 80 Hz were used as oddballs and delivered 2 times
each. Twenty pulse trains (10 surface stimuli and 10 subdermal
stimuli) were delivered to the subjects and the order of
application of different frequencies was randomized. With
sensation quality and location, the score was recorded as 1 if the
specific word was selected; otherwise, the score was recorded
as 0. Finally, the selection ratios (average scores of the 3
stimulation sequences) were used for data analysis for the
sensation quality and sensation location of each item. With
intensity and comfort, the average value of the three stimulation
sequences was used for data analysis.
E. Data analysis
The coefficient of variation (CoV) was computed to evaluate
the variability of DT, PT, WF, DR, intensity, and comfort
across seven days. CoV was calculated as follows:
𝐶𝑜𝑉 =
(
𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛/𝑀𝑒𝑎𝑛
)
× 100
Finally, 8 CoVs (8 subjects) were obtained as a measure of
within-subject variability across seven days for surface and
subdermal stimulation. Then, the parametric paired sample t-
test was used to compare CoVs between the surface and
subdermal stimulation if the data were normally distributed,
otherwise, the non-parametric Wilcoxon signed-rank test was
used. The CoV values were expressed in percent.
One-way repeated-measures ANOVA was applied to detect
if the psychometric parameters (DT, PT, DR, WF) changed
significantly across the seven days when the data were normally
distributed (Shapiro-Wilk test), otherwise, the Friedman test
was used. In both cases, post hoc pairwise tests (Tukey’s HSD
criterion) were performed if a significant difference was
detected across days.
Skilling-Mack test was used for difference detection across
the seven days in sensation quality, intensity, comfort, and
location data since this test can be used in any block design and
in the presence of missing data. Results are reported as mean ±
standard deviation (M ± SD). Statistics were performed using
IBM SPSS version 25 except the Skilling-Mack test, which was
performed using Statext v3.0. The statistical significance
threshold was set at p < 0.05.
III. RESULTS
A. PT, DT, and DR
The average CoVs of the DT and PT for surface and
subdermal stimulation are shown in Fig. 3. There was a
significant difference (p ˂ 0.05) between the surface and
subdermal stimulation in the CoVs of both DT and PT. The
CoVs for DT were 13.41 ± 5.11 % vs 22.30 ± 5.06 % and for
PT were 16.25 ± 6.93 % vs 20.00 ± 3.60 % in surface and
subdermal stimulation, respectively. Therefore, both DT and
PT were more variable in the case of subdermal stimulation.
There was no significant difference across seven days in the
mean DT of both surface and subdermal stimulation (Fig. 4 A
and B). However, the PT in surface and subdermal stimulation
increased across the seven-day period as shown in Fig. 4 C and
D. The regression fit lines were significantly different from zero
(p ˂ 0.05). The mean PT of surface stimulation changed
significantly across seven days, as detected by the Friedman test.
The post hoc tests revealed that the PT of the second (19.76 mA
± 4.59 mA) and third day (20.60 mA ± 4.43 mA) was
Fig. 3. The average (mean ± standard deviation) coefficient of variation (CoV
)
for detection threshold (DT) and pain threshold (PT)
across seven days for
subdermal (gray) and surface (black) stimulation. ⁎ p ˂ 0.05.
Fig. 2. Questionnaire for sensation evaluation.