senso s
A icle
Capaci i e Sensing o Non-In asi e B ea hing
and Hea Moni o ing in Non-Res ained,
Non-Seda ed Labo a o y Mice
Ca los González-Sánchez 1, Juan-Ca los F aile 2,*, Ja ie Pé ez-Tu iel 2, Ellen Damm 3,
Jochen G. Schneide 3, Heiko Zimme mann 1,4, Daniel Schmi 1and F ank R. Ihmig 1
1F aunho e -Ins i u ue BiomedizinischeTechnik (IBMT), Sulzbach/Saa 66280, Ge many;
[email p o ec ed] (C.G.-S.); [email p o ec ed].de (H.Z.);
[email p o ec ed].de (D.S.); [email p o ec ed].de (F.R.I.)
2ITAP—Uni e sidad de Valladolid, Paseo del Cauce 59, Valladolid 47011, Spain; [email p o ec ed]
3Luxembou g Cen e o Sys ems Biomedicine, Uni e si y o Luxembou g Esch-su -Alze e L-4362,
Luxembou g and In e nal Medicine II, Saa land Uni e si y Medical Cen e , Hombu g 66421, Ge many;
[email p o ec ed] (E.D.); [email p o ec ed] (J.G.S.)
4
Molecula and Cellula Bio echnology/Nano echnology, Saa land Uni e si y, Saa b uecken 66123, Ge many
*Co espondence: [email p o ec ed]; Tel.: +34-983-423-355; Fax: +34-983-423-358
Academic Edi o : Daniel Teichmann
Recei ed: 24 May 2016; Accep ed: 4 July 2016; Published: 7 July 2016
Abs ac :
Animal es ing plays a i al ole in biomedical esea ch. S ess educ ion is impo an
o imp o ing esea ch esul s and inc easing he wel a e and he quali y o li e o labo a o y
animals. To es ima e s ess we belie e i is o g ea impo ance o de elop non-in asi e echniques
o moni o ing physiological signals du ing he anspo o labo a o y animals, he eby allowing
he ga he ing o in o ma ion on he anspo condi ions, and, e en ually, he imp o emen o hese
condi ions. He e, we s udy he sui abili y o comme cially a ailable elec ic po en ial in eg a ed
ci cui (EPIC) senso s, using bo h con ac and con ac less echniques, o moni o ing he hea a e
and b ea hing a e o non- es ained, non-seda ed labo a o y mice. The design has been es ed unde
di e en scena ios wi h he aim o checking he plausibili y o pe o ming con ac less cap u e o
mouse hea ac i i y (ideally wi h an elec oca diog am). Fi s expe imen al esul s a e shown.
Keywo ds: non-in asi e senso ; capaci i e senso s; physiological signals in mice; s ess in mice
1. In oduc ion
Animal es ing has played and plays a i al ole in biomedical esea ch. I s use has e ealed he
he apeu ic e ec i eness and e iciency o new echniques and subs ances, chemicals and medicines,
as well as hei oxici y o sa e y, wi hou pu ing a isk he li es and sa e y o humans.
Despi e he undeniable bene i s ha hese echniques p o ide o mankind, we need a mo al
code ha minimizes he impac o esea ch on animals. This need has been de ended by sec o s bo h
wi hin and ou side he scien i ic communi y since he 19 h cen u y, and has p og essed o cu en
p ac ice, summa ized in he “p oposal o he h ee Rs”: eplacemen , educ ion, and e inemen .
This p oposal was published by Russell and Bu ch in 1959 [
1
], and is based on h ee pilla s ha
scien i ic expe imen a ion mus comply wi h whene e possible:
-
Replacemen : e e s o me hods which a oid o eplace he use o animals in an a ea whe e animals
would o he wise ha e been used. This includes bo h absolu e eplacemen s (i.e., eplacing animals
wi h inanima e sys ems, such as compu e p og ams) and ela i e eplacemen s (i.e., eplacing
mo e sen ien animals, such as e eb a es, wi h animals ha cu en scien i ic e idence indica es
ha e a signi ican ly lowe po en ial o pain pe cep ion, such as some in e eb a es).
Senso s 2016,16, 1052; doi:10.3390/s16071052 www.mdpi.com/jou nal/senso s
Senso s 2016,16, 1052 2 o 16
-
Reduc ion: e e s o any s a egy ha will esul in ewe animals being used o ob ain su icien
da a o answe he esea ch ques ion, o in maximizing he in o ma ion ob ained pe animal.
-
Re inemen : e e s o he modi ica ion o husband y o expe imen al p ocedu es o minimize pain
and dis ess, and o enhance he wel a e o an animal used in science om he ime i is bo n un il
i s dea h.
Despi e hese e o s, millions o animals a e needed each yea in labo a o ies a ound he wo ld.
In he Eu opean Union, nea ly 11.5 million labo a o y animals we e used in 2011, abou 75% o which
we e oden s [2].
In esea ch, i is impo an o minimize all ex e nal in luences ha could undesi ably modi y he
esponse o labo a o y animals, b inging he esul s close o he eali y o he p oblem. By ha ing
be e models we can educe he numbe o subjec s used du ing es s.
One pa ame e ha may in oduce disc epancies is s ess [
3
]. Reducing s ess is impo an o
imp o ing he esul s o esea ch, as well as o inc easing he wel a e and quali y o li e o labo a o y
animals. Any change in he en i onmen al condi ions o an animal can induce s ess, anspo being
o pa icula signi icance, especially o labo a o y animals. In he case o gene ically modi ied animals,
anspo s ess is o e en g ea e impo ance, as specialized companies make shipmen s o hei
specimens on eques o labo a o ies a ound he wo ld, wi h a consequen inc ease in anspo ime
ha exace ba es he possible symp oms.
The e is e idence ha anspo p oduces high s ess le els. An assessmen o s ess du ing
handling and anspo o animals is p esen ed in [
4
]. This s ess is no con ined o he anspo ed
animal, bu i can also a ec hei o sp ing [
5
,
6
]. Di e en pa ame e s ha e been used o assess s ess
in animals, such as: co isol, co icos e one and lucose, ood and wa e consump ion and weigh
loss [
7
]. A e high s ess le els, he mice equi e much ime o he hea a e (HR) and b ea hing a e
(BR) physiological signals o e u n o alues wi hin he no mal le els.
Fo example, he need o 24 o 48 h o es o he immune sys em and co icos e one le els
o s abilize a e anspo has been desc ibed [
8
,
9
]. Ano he s udy ha moni o ed s ess indica o s
based on animal beha io and he co icos e one le el a e anspo a ion ound ha mice we e no
comple ely acclima ed a e h ee o ou days [3,10].
Aiming o minimize s ess, some ac o s mus be aken in o accoun :
‚Animal heal h (especially hei abili y o a el).
‚
Design and ma e ials o con aine s, including loading and unloading as well as measu es o
inspec he condi ion o he animal du ing anspo .
‚Numbe o animals in each con aine and he alloca ed space o each animal.
‚En i onmen al condi ions inside he con aine .
‚
Quali y and quan i y o subs a e o ma e ial co e ing he bo om as well as ood and wa e
(o addi ional liquid sou ce).
‚T a el ime.
‚Numbe o ips o changes be ween ehicles.
‚Reco e y o he animal a e he ip.
We belie e i is o g ea impo ance o de elop non-in asi e echniques o moni o ing
physiological signals du ing he anspo o labo a o y animals. This allows he ga he ing o
in o ma ion abou he condi ions o anspo wi h an aim o imp o e hese condi ions acco ding o he
p inciple o he h ee Rs. This will imp o e he quali y o li e o he animal, and will also imp o e he
pe o mance ha his animal p o ides o science, u he dec easing he numbe o animals equi ed
o esea ch.
This pape p esen s he de eloped app oach based on EPIC (elec ic po en ial in eg a ed ci cui )
senso s o measu e elec ic po en ial o cap u ing physiological signals in mice, wi h emphasis on
non-in asi e and con ac less echniques.
Senso s 2016,16, 1052 3 o 16
2. Expe imen al Sec ion
2.1. S essMeasu emen in Labo a o yAnimals
One commonly used me hod o moni o ing he le el o s ess in labo a o y animals du ing
anspo is based on hea a e a iabili y (HRV). This e m is de ined as bea - o-bea changes in
hea a e o a ia ions o he RR in e als ( ime be ween wo R wa es in an elec oca diog am) in
consecu i e ca diac cycles. HRV p o ides a mo e accu a e measu emen o s ess han hose ob ained
simply by obse ing ca diac hy hm, because he ca diac hy hm can a y g ea ly om one animal o
ano he , o depend on physical ac i i y. Measu ing in e -bea s in e als (IBI) and HRV a e p omising
app oaches o e alua ing s ess in animals [
11
]. The impo ance o HRV as a po en ial ma ke o s ess
and heal h is shown in [12].
Howe e , acquisi ion o he hea a e is no a i ial ask in small labo a o y animals. In asi e
echniques a e o en used, based on implan able elec odes placed inside he animal ha migh
ansmi in o ma ion h ough a adio link [
13
,
14
]. These sys ems p o ide a comple e and de ailed
elec oca diog am (ECG) signal, bu equi e ha d wo k and dedica ion by he esea che and p oblems
may a ise necessi a ing pos s ess es o inc easing he isk o dea h o he animal.
While he e a e non-in asi e ECG acquisi ion echniques o oden s, hese o en equi e aining
o he animal and placing i in a ce ain posi ion o de ice. Cap u e o ECG signals wi h we elec odes
is o en associa ed wi h discom o because he elec odes equi e skin p epa a ion and he use o
gels. These gels end o lose mois u e o e ime wi h he consequen loss o ECG signal. Thus,
hese echniques a e no easible o long- e m con inuous moni o ing.
The de elopmen o an ECG moni o ing sys em using EPIC senso s ha do no equi e he
p epa a ion o he skin o di ec con ac wi h he subjec would be o g ea help in imp o ing he
usabili y o hese sys ems.
2.2. Capaci i e Senso s
A pa icula ype o d y elec ode, i s de eloped by Lopez and Richa dson [
15
], is known as
a capaci i e o insula ed elec ode. These elec odes equi e no ohmic con ac wi h he body since
i ac s as a simple capaci o placed in se ies wi h he skin, so ha he signal is capaci i ely coupled.
The ecei ed signal can be connec ed o an ope a ional ampli ie and hen o s anda d ins umen a ion.
The use o a dielec ic ma e ial in good con ac o he skin esul s in a ai ly la ge coupling
capaci ance, anging om 300 pF o se e al nano- a ads. As a esul , a sys em wi h educed
noise and app op ia e equency esponse is eadily achie able using s anda d high-impedance
FET ( ield-e ec ansis o ) ampli ie s.
While we and d y elec odes equi e physical con ac wi h he skin o unc ion, capaci i e
elec odes can be used wi hou con ac , h ough an insula ing laye such as hai , clo hing o ai .
These con ac less elec odes ha e been desc ibed gene ally as simple capaci i e elec odes, bu in
eali y he e is also a small esis i e elemen , since he insula ion also has a non-negligible esis ance.
Signals coupled ia con ac less elec odes can con ain unaccep able le els o noise, equi ing a sui able
design o o he elemen s o inc ease he inpu impedance and neu alize he pa asi ic inpu capaci ance.
One o he majo d awbacks o hese elec odes is hei sensi i i y o mo emen s, which along wi h he
high se ling ime can ende hese sys ems inadequa e in ce ain scena ios.
Capaci i e senso s ha e been used o measu e hea a e in humans ia ei he di ec skin con ac
o h ough one and wo laye s o clo hing wi h no dielec ic gel and no g ounding elec ode [
16
–
18
].
These senso s ha e also been used o moni o human espi a o y a e using a conduc i e ex ile-based
wea able senso [
19
,
20
]. High impedance elec ic po en ial senso s ha e been used in humans o
measu e b ea hing and hea signals [
21
–
23
]. Howe e , we ha e ound no epo o he use o
capaci i e senso s o moni o ing hea a e in mice du ing anspo .
In ou expe imen s wi h mice, we ha e used he PS25251 EPIC capaci i e senso (see Figu e 1) ha
inco po a es a numbe o ools o enhance hei cha ac e is ics, allowing measu emen o he elec ic
Senso s 2016,16, 1052 4 o 16
po en ial. I s main ea u es a e: ul a-high inpu esis ance, ypically 20 G
Ω
, d y-con ac capaci i e
coupling, inpu capaci ance as low as 15 pF, lowe
´
3 dB poin ypically 200 mHz and uppe
´
3 dB
poin ypically 10 kHz. I ope a es wi h bipola powe supply om ˘2.4 V o ˘5.5 V.
Senso s 2016, 16, 1052 4 o 16
capaci i e coupling, inpu capaci ance as low as 15 pF, lowe −3 dB poin ypically 200 mHz and
uppe −3 dB poin ypically 10 kHz. I ope a es wi h bipola powe supply om ±2.4 V o ±5.5 V.
(a) (b)
Figu e 1. PS25251 EPIC senso , on (a) and ea (b) iews.
2.3. Cap u e o Physiological Signals in Roden s
A sys em o simul aneously moni o ing he hea a e and b ea hing a e o anes he ized mice
using a piezoelec ic ansduce is p oposed in [24]. Noncon ac moni o ing o ca dio espi a o y
ac i i y by elec omagne ic coupling has been in es iga ed in humans, by using wo measu emen
modali ies: he capaci i e coupling pa h and he induc i e coupling pa h (magne ic induc ion
moni o ing) [25,26]. In his pape , we p opose o use capaci i e sensing o non-in asi e b ea hing
and hea moni o ing in non- es ained, non-seda ed labo a o y mice.
This esea ch ocuses on he cap u e o physiological signals in labo a o y animals, pa icula ly
C57BL6/J/LDLR
−/−
s ain mice. Animal expe imen s we e app o ed by he Saa land Uni e si y
Medical Cen e animal expe imen a ion o ice (animal p o ocol 30/2012).
The aim o he p oposed expe imen s was he con ac less cap u e o mouse hea ac i i y
(ideally ECG), using he EPIC senso s, and aking in o accoun he ollowing echnical di icul ies:
 Small ampli ude o signal due o he small size o he mouse hea .
 The impossibili y o using di ec me hods o noise educ ion, such as he g ounding o he
mouse o i s connec ion o he DRL (d i en igh leg) signal.
 Low capaci i e coupling wi h he senso when using con ac less echniques.
Because o hese limi a ions, we ha e conduc ed expe imen s o de e mine he easibili y o his
echnique, and also o e alua e o he op ions, such as he use o echniques wi h con ac be ween he
animal and senso s. Di e en con igu a ions ha e been used depending on he ype o expe imen ,
bu he ha dwa e and so wa e was always simila .
2.3.1. Senso Ins umen a ion Ci cui
Usually, he weak ca diac signal appea s as he di e en ial signal be ween wo elec odes. To
emo e common mode noise we used an ins umen a ion ampli ie , INA128 (Texas Ins umen s,
Dallas, TX, USA), which has high common mode ejec ion, low powe consump ion and low p ice.
The senso s equi e a powe supply ol age o ±5.5 V maximum, which is also sui able o he
ins umen a ion ampli ie . Ba e ies we e used as powe supply in o de o educe he
en i onmen al elec ical noise.
F om he p e-amp, he signal was ed o a high-pass i s -o de il e and ampli ied by one
sec ion o an OPA2131 ope a ional ampli ie (Texas Ins umen s), cap u ed by a digi al oscilloscope
(MDO4104B-6, Tek onix, Bea e on, OR, USA) and s o ed on a PC. Figu e 2a shows he
expe imen al se up used o he acquisi ion o physiological signals in oden s. Figu e 2b shows he
ins umen a ion ci cui diag am implemen ed o his expe imen al se up.
Figu e 1. PS25251 EPIC senso , on (a) and ea (b) iews.
2.3. Cap u e o Physiological Signals in Roden s
A sys em o simul aneously moni o ing he hea a e and b ea hing a e o anes he ized mice
using a piezoelec ic ansduce is p oposed in [
24
]. Noncon ac moni o ing o ca dio espi a o y
ac i i y by elec omagne ic coupling has been in es iga ed in humans, by using wo measu emen
modali ies: he capaci i e coupling pa h and he induc i e coupling pa h (magne ic induc ion
moni o ing) [
25
,
26
]. In his pape , we p opose o use capaci i e sensing o non-in asi e b ea hing
and hea moni o ing in non- es ained, non-seda ed labo a o y mice.
This esea ch ocuses on he cap u e o physiological signals in labo a o y animals, pa icula ly
C57BL6/J/LDLR
´/´
s ain mice. Animal expe imen s we e app o ed by he Saa land Uni e si y
Medical Cen e animal expe imen a ion o ice (animal p o ocol 30/2012).
The aim o he p oposed expe imen s was he con ac less cap u e o mouse hea ac i i y
(ideally ECG), using he EPIC senso s, and aking in o accoun he ollowing echnical di icul ies:
‚Small ampli ude o signal due o he small size o he mouse hea .
‚
The impossibili y o using di ec me hods o noise educ ion, such as he g ounding o he mouse
o i s connec ion o he DRL (d i en igh leg) signal.
‚Low capaci i e coupling wi h he senso when using con ac less echniques.
Because o hese limi a ions, we ha e conduc ed expe imen s o de e mine he easibili y o his
echnique, and also o e alua e o he op ions, such as he use o echniques wi h con ac be ween he
animal and senso s. Di e en con igu a ions ha e been used depending on he ype o expe imen ,
bu he ha dwa e and so wa e was always simila .
2.3.1. Senso Ins umen a ion Ci cui
Usually, he weak ca diac signal appea s as he di e en ial signal be ween wo elec odes.
To emo e common mode noise we used an ins umen a ion ampli ie , INA128 (Texas Ins umen s,
Dallas, TX, USA), which has high common mode ejec ion, low powe consump ion and low p ice.
The senso s equi e a powe supply ol age o
˘
5.5 V maximum, which is also sui able o he
ins umen a ion ampli ie . Ba e ies we e used as powe supply in o de o educe he en i onmen al
elec ical noise.
F om he p e-amp, he signal was ed o a high-pass i s -o de il e and ampli ied by one
sec ion o an OPA2131 ope a ional ampli ie (Texas Ins umen s), cap u ed by a digi al oscilloscope
(MDO4104B-6, Tek onix, Bea e on, OR, USA) and s o ed on a PC. Figu e 2a shows he expe imen al
se up used o he acquisi ion o physiological signals in oden s. Figu e 2b shows he ins umen a ion
ci cui diag am implemen ed o his expe imen al se up.
Senso s 2016,16, 1052 5 o 16
Senso s 2016, 16, 1052 5 o 16
(a)
(b)
Figu e 2. Expe imen al se up used o cap u e physiological signals in oden s (a).F om le o igh :
Lap op o signal s o age and p ocessing, powe supply, oscilloscope o cap u e signals, he
ins umen a ion ci cui , and a cage. Ins umen a ion ci cui diag am implemen ed o his
expe imen al se up (b).
Figu e 2b shows he i s sec ion (on he le ) wi h EPIC senso s and powe ing (Vdd = 5 V and
Vss = −5 V). The second sec ion (in he middle) includes he ins umen a ion ampli ie ,
common-mode noise ejec ion and i s ampli ica ion. The inal sec ion includes a second
ampli ica ion and analog/digi al con e e (ADC).
2.3.2. Signal P ocessing
We used MATLAB so wa e ( e sion R2014b) o implemen a se ies o algo i hms o imp o e
he quali y o he s o ed signal. We ou inely used a no ch il e o emo e 50Hz noise (and i s
mul iples) and a low-pass il e . This was ollowed by a polynomial spline i and sub ac ion om
he o iginal signal. I he signal was s ill noisy, we used a 5 h o de polynomial Sa i zky-Golay il e ,
o gene alized mo ing a e age o smoo h i . Time and equency domain plo s we e made a e each
one o hese il e s.
2.4. Pla o ms De eloped o Cap u ing Physiological Signals in Roden s
Ini ial expe imen s equi ed minimizing he mo emen o he mouse du ing he cap u e o
physiological signals. Thus, we de eloped h ee pla o m sys ems o ca y ou es s. In Table 1, we
summa ize key aspec s o he h ee sys ems.
We ini ially es ed plas ic ubes (sys em A) in which holes we e made o ensu e adequa e
ai low o allow o no mal b ea hing. Two senso s, one on each side, we e placed on he ou side,
wi h he animal in be ween. Finally, he plas ic ube was su ounded wi h g ounded aluminum oil
in an a emp o educe elec ical noise picked up by he senso s.
Figu e 2.
Expe imen al se up used o cap u e physiological signals in oden s (
a
).F om le o
igh : Lap op o signal s o age and p ocessing, powe supply, oscilloscope o cap u e signals,
he ins umen a ion ci cui , and a cage. Ins umen a ion ci cui diag am implemen ed o his
expe imen al se up (b).
Figu e 2b shows he i s sec ion (on he le ) wi h EPIC senso s and powe ing (Vdd = 5 V and
Vss =
´
5 V). The second sec ion (in he middle) includes he ins umen a ion ampli ie , common-mode
noise ejec ion and i s ampli ica ion. The inal sec ion includes a second ampli ica ion and
analog/digi al con e e (ADC).
2.3.2. Signal P ocessing
We used MATLAB so wa e ( e sion R2014b) o implemen a se ies o algo i hms o imp o e
he quali y o he s o ed signal. We ou inely used a no ch il e o emo e 50 Hz noise (and i s
mul iples) and a low-pass il e . This was ollowed by a polynomial spline i and sub ac ion om
he o iginal signal. I he signal was s ill noisy, we used a 5 h o de polynomial Sa i zky-Golay il e ,
o gene alized mo ing a e age o smoo h i . Time and equency domain plo s we e made a e each
one o hese il e s.
2.4. Pla o ms De eloped o Cap u ing Physiological Signals in Roden s
Ini ial expe imen s equi ed minimizing he mo emen o he mouse du ing he cap u e o
physiological signals. Thus, we de eloped h ee pla o m sys ems o ca y ou es s. In Table 1,
we summa ize key aspec s o he h ee sys ems.
We ini ially es ed plas ic ubes (sys em A) in which holes we e made o ensu e adequa e ai low
o allow o no mal b ea hing. Two senso s, one on each side, we e placed on he ou side, wi h he
animal in be ween. Finally, he plas ic ube was su ounded wi h g ounded aluminum oil in an
a emp o educe elec ical noise picked up by he senso s.

Senso s 2016,16, 1052 6 o 16
Table 1. Summa y o he sys ems es ed o cap u e o physiological signals in oden s.
Sys em Name Sho Desc ip ion
A Plas ic ube The e a e holes o main aining an adequa e ai low. Senso s can be
placed inside o ou side and so can be g ounded.
B Pla o m High ca dboa d pla o ms whe e he animal s ays. Con ac o
con ac less echniques can be used. Con ac o con ac less g ounding
can be achie ed wi h aluminum oil.
C Ma ix o senso s A ay o 16 EPIC senso s a anged in a 4-by-4 ma ix wi h minimal
dis ance be ween hem. Coppe ape was used o g ounding. The
ma ix was connec ed o a sel -designed p in ed ci cui boa d.
Se e al expe imen s we e pe o med, wi h a di e en app oach, using he small pla o m
(sys em B) shown in Figu e 3.The e a e wo EPIC senso s on he pla o m and he mouse is placed
o e i . I did no mo e om he a ea whe e he senso s a e posi ioned.
Senso s 2016, 16, 1052 6 o 16
Table 1. Summa y o he sys ems es ed o cap u e o physiological signals in oden s.
Sys em Name Sho Desc ip ion
A Plas ic ube
The e a e holes o main aining an adequa e ai low. Senso s can be placed
inside o ou side and so can be g ounded.
B Pla o m
High ca dboa d pla o ms whe e he animal s ays. Con ac o con ac less
echniques can be used. Con ac o con ac less g ounding can be achie ed
wi h aluminum oil.
C Ma ix o
senso s
A ay o 16 EPIC senso s a anged in a 4-by-4 ma ix wi h minimal dis ance
be ween hem. Coppe ape was used o g ounding. The ma ix was
connec ed o a sel -designed p in ed ci cui boa d.
Se e al expe imen s we e pe o med, wi h a di e en app oach, using he small pla o m
(sys em B) shown in Figu e 3.The e a e wo EPIC senso s on he pla o m and he mouse is placed
o e i . I did no mo e om he a ea whe e he senso s a e posi ioned.
Figu e 3. Pla o m (sys em B) on which he mouse is placed. Two EPIC senso s and he g ounding
can be dis inguished.
Due o he di icul y o ge ing wo o he mouse legs placed exac ly o e he wo senso s, we
p oceeded o design he ma ix o senso s (sys em C) shown in Figu e 4a, ha can accommoda e 16
EPIC senso s a anged in a 4-by-4 ma ix. A e placing he senso s and making he necessa y
connec ions, he op was co e ed wi h an adhesi e coppe ape excep , ob iously, he a eas whe e
he senso s a e loca ed.
(a) (b)
Figu e 4. (a) Ma ix o EPIC senso s (sys em C) used o physiological signals cap u ing in mice;
(b) Sel -designed PCB used o con ol he ma ix o EPIC senso s.
Wi h he aim o imp o ing he connec ion and con ol o his ma ix o senso s, a p in ed ci cui
boa d (PCB) was designed and buil (see Figu e 4b). I con ains an in e ace o connec ion and
powe managemen o up o 16 EPIC senso s, capable o educing powe consump ion by
modi ying he numbe o senso s powe ed a a ime, also ensu ing adequa e lexibili y o u u e
de elopmen s. This boa d also allows he use o selec he pai o senso s o be used o he ou pu
signal and o calcula e he DRL signal.
Figu e 3.
Pla o m (sys em B) on which he mouse is placed. Two EPIC senso s and he g ounding can
be dis inguished.
Due o he di icul y o ge ing wo o he mouse legs placed exac ly o e he wo senso s,
we p oceeded o design he ma ix o senso s (sys em C) shown in Figu e 4a, ha can accommoda e
16 EPIC senso s a anged in a 4-by-4 ma ix. A e placing he senso s and making he necessa y
connec ions, he op was co e ed wi h an adhesi e coppe ape excep , ob iously, he a eas whe e he
senso s a e loca ed.
Senso s 2016, 16, 1052 6 o 16
Table 1. Summa y o he sys ems es ed o cap u e o physiological signals in oden s.
Sys em Name Sho Desc ip ion
A Plas ic ube
The e a e holes o main aining an adequa e ai low. Senso s can be placed
inside o ou side and so can be g ounded.
B Pla o m
High ca dboa d pla o ms whe e he animal s ays. Con ac o con ac less
echniques can be used. Con ac o con ac less g ounding can be achie ed
wi h aluminum oil.
C Ma ix o
senso s
A ay o 16 EPIC senso s a anged in a 4-by-4 ma ix wi h minimal dis ance
be ween hem. Coppe ape was used o g ounding. The ma ix was
connec ed o a sel -designed p in ed ci cui boa d.
Se e al expe imen s we e pe o med, wi h a di e en app oach, using he small pla o m
(sys em B) shown in Figu e 3.The e a e wo EPIC senso s on he pla o m and he mouse is placed
o e i . I did no mo e om he a ea whe e he senso s a e posi ioned.
Figu e 3. Pla o m (sys em B) on which he mouse is placed. Two EPIC senso s and he g ounding
can be dis inguished.
Due o he di icul y o ge ing wo o he mouse legs placed exac ly o e he wo senso s, we
p oceeded o design he ma ix o senso s (sys em C) shown in Figu e 4a, ha can accommoda e 16
EPIC senso s a anged in a 4-by-4 ma ix. A e placing he senso s and making he necessa y
connec ions, he op was co e ed wi h an adhesi e coppe ape excep , ob iously, he a eas whe e
he senso s a e loca ed.
(a) (b)
Figu e 4. (a) Ma ix o EPIC senso s (sys em C) used o physiological signals cap u ing in mice;
(b) Sel -designed PCB used o con ol he ma ix o EPIC senso s.
Wi h he aim o imp o ing he connec ion and con ol o his ma ix o senso s, a p in ed ci cui
boa d (PCB) was designed and buil (see Figu e 4b). I con ains an in e ace o connec ion and
powe managemen o up o 16 EPIC senso s, capable o educing powe consump ion by
modi ying he numbe o senso s powe ed a a ime, also ensu ing adequa e lexibili y o u u e
de elopmen s. This boa d also allows he use o selec he pai o senso s o be used o he ou pu
signal and o calcula e he DRL signal.
Figu e 4.
(
a
) Ma ix o EPIC senso s (sys em C) used o physiological signals cap u ing in mice;
(b) Sel -designed PCB used o con ol he ma ix o EPIC senso s.
Wi h he aim o imp o ing he connec ion and con ol o his ma ix o senso s, a p in ed ci cui
boa d (PCB) was designed and buil (see Figu e 4b). I con ains an in e ace o connec ion and powe
managemen o up o 16 EPIC senso s, capable o educing powe consump ion by modi ying he
numbe o senso s powe ed a a ime, also ensu ing adequa e lexibili y o u u e de elopmen s.
This boa d also allows he use o selec he pai o senso s o be used o he ou pu signal and o
calcula e he DRL signal.
Senso s 2016,16, 1052 7 o 16
3. Resul s and Discussion
3.1. Tes T ial 1: Plas ic Tube Pla o m (Sys em A)
The mouse was placed inside a plas ic ube wi h holes o ensu e adequa e ai low o no mal
b ea hing. EPIC senso s we e placed on he sides o he animal and he ube was su ounded wi h
aluminum oil connec ed o a g ound (GND). The signal cap u ed using his sys em is shown in
Figu e 5a. A e applying he no ch il e o he o iginal signal, he signal shown in Figu e 5b is ob ained.
Senso s 2016, 16, 1052 7 o 16
3. Resul s and Discussion
3.1. Tes T ial 1: Plas ic Tube Pla o m (Sys em A)
The mouse was placed inside a plas ic ube wi h holes o ensu e adequa e ai low o no mal
b ea hing. EPIC senso s we e placed on he sides o he animal and he ube was su ounded wi h
aluminum oil connec ed o a g ound (GND). The signal cap u ed using his sys em is shown in
Figu e 5a. A e applying he no ch il e o he o iginal signal, he signal shown in Figu e 5b is
ob ained.
The signals o Figu e 5 show a clea pe iodic pa e n. To s udy i s cha ac e is ics we analyzed
his signal in he equency domain (Figu e 6).
(a)
(b)
Figu e 5. Tes ial 1: signal cap u ed wi h mouse placed inside a plas ic ube (a); Signal a e no ch
il e ing a 50Hz (b).
Figu e 6. Tes ial 1: ans o med signal in he equency domain.
Figu e 5.
Tes ial 1: signal cap u ed wi h mouse placed inside a plas ic ube (
a
); Signal a e no ch
il e ing a 50 Hz (b).
The signals o Figu e 5show a clea pe iodic pa e n. To s udy i s cha ac e is ics we analyzed his
signal in he equency domain (Figu e 6).
Senso s 2016, 16, 1052 7 o 16
3. Resul s and Discussion
3.1. Tes T ial 1: Plas ic Tube Pla o m (Sys em A)
The mouse was placed inside a plas ic ube wi h holes o ensu e adequa e ai low o no mal
b ea hing. EPIC senso s we e placed on he sides o he animal and he ube was su ounded wi h
aluminum oil connec ed o a g ound (GND). The signal cap u ed using his sys em is shown in
Figu e 5a. A e applying he no ch il e o he o iginal signal, he signal shown in Figu e 5b is
ob ained.
The signals o Figu e 5 show a clea pe iodic pa e n. To s udy i s cha ac e is ics we analyzed
his signal in he equency domain (Figu e 6).
(a)
(b)
Figu e 5. Tes ial 1: signal cap u ed wi h mouse placed inside a plas ic ube (a); Signal a e no ch
il e ing a 50Hz (b).
Figu e 6. Tes ial 1: ans o med signal in he equency domain.
Figu e 6. Tes ial 1: ans o med signal in he equency domain.
Senso s 2016,16, 1052 8 o 16
Figu e 6shows a signi ican componen a 50 Hz, co esponding o elec ical noise. Howe e ,
he e is an in e es ing egula a angemen a 6, 12, 18, 24, 30 Hz and so on (6 Hz mul iples) o
dec easing powe . This was ini ially assumed o be elec ical noise o unknown o igin and i s ha monics
ha he cap u ed signal shows, bu u he es s disca ded his possibili y, since his signal was cap u ed
only when he mouse was p esen .
An algo i hm was used o loca e he peaks (in Figu e 5) and calcula e he a e age a e, ob aining
a alue o 364 peaks pe minu e. This alue is oo low o co espond o a hea bea , because in
mice hea bea anges om 310 o 840 bea s pe minu e (bpm). Mo eo e , unde s ess condi ions
(mouse inside a plas ic ube in a limi ed space and ecen ly handled by a human), i is expec ed ha
hese alues would be e en highe .
The no mal b ea hing a e in elaxed mice, howe e , usually anges om 80 o 230 b ea hs pe
minu e. The e o e, ou i s hypo hesis was o associa e he signal o Figu e 5(bo om) o a mouse
b ea hing signal, p obably p oduced by he dis ance a ia ion be ween he senso and he mouse’s
body due o i s ches mo emen s du ing inspi a ion and expi a ion. Acco dingly, we pe o med mo e
es s using con ac echniques o see i , in his way, i was possible o cap u e any kind o ca diac signal,
using GND and DRL g ounding echniques.
3.2. Tes T ial 2: Pla o m (Sys em B) wi h Senso s and GND G ounding
In his es , he mouse was placed on a small aised pla o m, which ca ied wo EPIC senso s.
The signal shown in Figu e 7a has been ob ained om he mouse wi h one paw in con ac wi h one o
he EPIC senso s, and he o he in con ac wi h he g ounding (GND).
Senso s 2016, 16, 1052 8 o 16
Figu e 6 shows a signi ican componen a 50 Hz, co esponding o elec ical noise. Howe e ,
he e is an in e es ing egula a angemen a 6, 12, 18, 24, 30 Hz and so on (6 Hz mul iples) o
dec easing powe . This was ini ially assumed o be elec ical noise o unknown o igin and i s
ha monics ha he cap u ed signal shows, bu u he es s disca ded his possibili y, since his
signal was cap u ed only when he mouse was p esen .
An algo i hm was used o loca e he peaks (in Figu e 5) and calcula e he a e age a e,
ob aining a alue o 364 peaks pe minu e. This alue is oo low o co espond o a hea bea ,
because in mice hea bea anges om 310 o 840 bea s pe minu e (bpm). Mo eo e , unde s ess
condi ions (mouse inside a plas ic ube in a limi ed space and ecen ly handled by a human), i is
expec ed ha hese alues would be e en highe .
The no mal b ea hing a e in elaxed mice, howe e , usually anges om 80 o 230 b ea hs pe
minu e. The e o e, ou i s hypo hesis was o associa e he signal o Figu e 5 (bo om) o a mouse
b ea hing signal, p obably p oduced by he dis ance a ia ion be ween he senso and he mouse’s
body due o i s ches mo emen s du ing inspi a ion and expi a ion. Acco dingly, we pe o med
mo e es s using con ac echniques o see i , in his way, i was possible o cap u e any kind o
ca diac signal, using GND and DRL g ounding echniques.
3.2. Tes T ial 2: Pla o m (Sys em B) wi h Senso s and GND G ounding
In his es , he mouse was placed on a small aised pla o m, which ca ied wo EPIC senso s.
The signal shown in Figu e 7a has been ob ained om he mouse wi h one paw in con ac wi h one
o he EPIC senso s, and he o he in con ac wi h he g ounding (GND).
Since he (mainly 50 Hz) noise in Figu e 7a masks he de ailed signal cha ac e is ics, we used a
no ch il e o ob ain he signal shown in Figu e 7b.
(a)
(b)
Figu e 7. Tes ial 2: ime signal wi hou any il e ing (a); Signal a e no ch il e ing a 50 Hz (b).
The signal o Figu e 7b shows clea pe iodic a ia ions. Figu e 8 shows he ans o ma ion o
he equency domain.
Figu e 7. Tes ial 2: ime signal wi hou any il e ing (a); Signal a e no ch il e ing a 50 Hz (b).
Since he (mainly 50 Hz) noise in Figu e 7a masks he de ailed signal cha ac e is ics, we used a
no ch il e o ob ain he signal shown in Figu e 7b.
The signal o Figu e 7b shows clea pe iodic a ia ions. Figu e 8shows he ans o ma ion o he
equency domain.
Senso s 2016,16, 1052 9 o 16
Senso s 2016, 16, 1052 9 o 16
Figu e 8. Tes ial 2: ans o med signal in he equency domain.
The signal shown in Figu e 7b clea ly shows se e al la ge peaks, simila o hose shown in
Figu e 5b, p oduced by he b ea hing o he mouse. I s hy hm has been calcula ed using peak
de ec ion in he signal, esul ing in 196 b ea hs pe minu e, which i s pe ec ly wi h one o he
dominan equencies o he spec um (see Figu e 8), loca ed a 3.25 Hz (195 min
−1
). The e is a second
peak in Figu e 8 a 11.5 Hz (690 bpm). Al hough high, his is wi hin he no mal hea a e ange o
mice.
To check he co espondence o he s onges signal (lowe equency) wi h he espi a o y
hy hm o he mouse, a ideo came a was used o eco d bo h he mo emen o he mouse and he
cap u ed signal (Figu e 9).
Figu e 9. S ill ame om one o he cap u ed ideos o e i y he co ela ion be ween he cap u ed
signal and he b ea hing a e.
Figu e 10 shows he signal cap u ed. This signal is ela ed o he espi a ion o mouse, since i
e lec s he mo emen s o he mouse’s ches . I can be seen ha he equency o his signal inc eases
o a sho ime when he mouse s a s sni ing. This signal shows he sudden change in he
b ea hing pa e n, jus as he sni ing occu s. As shown in Figu e 10, he mouse’s espi a o y a e
changes e y ab up ly, om 143 o abou 465 b ea hs pe minu e. The signal becomes sa u a ed
b ie ly, bu he equency is s ill ecognizable.
Figu e 8. Tes ial 2: ans o med signal in he equency domain.
The signal shown in Figu e 7b clea ly shows se e al la ge peaks, simila o hose shown in
Figu e 5b, p oduced by he b ea hing o he mouse. I s hy hm has been calcula ed using peak
de ec ion in he signal, esul ing in 196 b ea hs pe minu e, which i s pe ec ly wi h one o he
dominan equencies o he spec um (see Figu e 8), loca ed a 3.25 Hz (195 min
´1
). The e is a second
peak in Figu e 8a 11.5 Hz (690 bpm). Al hough high, his is wi hin he no mal hea a e ange
o mice.
To check he co espondence o he s onges signal (lowe equency) wi h he espi a o y hy hm
o he mouse, a ideo came a was used o eco d bo h he mo emen o he mouse and he cap u ed
signal (Figu e 9).
Senso s 2016, 16, 1052 9 o 16
Figu e 8. Tes ial 2: ans o med signal in he equency domain.
The signal shown in Figu e 7b clea ly shows se e al la ge peaks, simila o hose shown in
Figu e 5b, p oduced by he b ea hing o he mouse. I s hy hm has been calcula ed using peak
de ec ion in he signal, esul ing in 196 b ea hs pe minu e, which i s pe ec ly wi h one o he
dominan equencies o he spec um (see Figu e 8), loca ed a 3.25 Hz (195 min
−1
). The e is a second
peak in Figu e 8 a 11.5 Hz (690 bpm). Al hough high, his is wi hin he no mal hea a e ange o
mice.
To check he co espondence o he s onges signal (lowe equency) wi h he espi a o y
hy hm o he mouse, a ideo came a was used o eco d bo h he mo emen o he mouse and he
cap u ed signal (Figu e 9).
Figu e 9. S ill ame om one o he cap u ed ideos o e i y he co ela ion be ween he cap u ed
signal and he b ea hing a e.
Figu e 10 shows he signal cap u ed. This signal is ela ed o he espi a ion o mouse, since i
e lec s he mo emen s o he mouse’s ches . I can be seen ha he equency o his signal inc eases
o a sho ime when he mouse s a s sni ing. This signal shows he sudden change in he
b ea hing pa e n, jus as he sni ing occu s. As shown in Figu e 10, he mouse’s espi a o y a e
changes e y ab up ly, om 143 o abou 465 b ea hs pe minu e. The signal becomes sa u a ed
b ie ly, bu he equency is s ill ecognizable.
Figu e 9.
S ill ame om one o he cap u ed ideos o e i y he co ela ion be ween he cap u ed
signal and he b ea hing a e.
Figu e 10 shows he signal cap u ed. This signal is ela ed o he espi a ion o mouse, since i
e lec s he mo emen s o he mouse’s ches . I can be seen ha he equency o his signal inc eases
o a sho ime when he mouse s a s sni ing. This signal shows he sudden change in he b ea hing
pa e n, jus as he sni ing occu s. As shown in Figu e 10, he mouse’s espi a o y a e changes e y
ab up ly, om 143 o abou 465 b ea hs pe minu e. The signal becomes sa u a ed b ie ly, bu he
equency is s ill ecognizable.
Senso s 2016,16, 1052 16 o 16
21.
Ku onen, E. Epic Senso s in Elec oca diog am Measu emen . Mas e ’s Thesis, Oulu Uni e si y o Applied
Sciences, Oulu, Finland, 2013.
22.
P ance, R.J.; Bea dsmo e-Rus , S.T.; Wa son, P.; Ha land, C.J.; P ance, H. Remo e de ec ion o human
elec ophysiological signals using elec ic po en ial senso s. Appl. Phys. Le .
2008
,93, 033906-1–033906-3.
[C ossRe ]
23.
Ha land, C.; Cla k, T.; P ance, R. High esolu ion ambula o y elec oca diog aphic moni o ing using
w is -moun ed elec ic po en ial senso s. Meas. Sci. Technol. 2003,14, 923–928. [C ossRe ]
24.
Sa o, S.; Yamada, K.; Inagaki, N. Sys em o simul aneously moni o ing hea and b ea hing a e in mice
using a piezoelec ic ansduce . Med. Biol. Eng. Compu . 2006,44, 353–362. [C ossRe ] [PubMed]
25.
Teichmann, D.; Foussie , J.; Jia, J.; Leonha d , S.; Wal e , M. Noncon ac moni o ing o ca dio espi a o y
ac i i y by elec omagne ic coupling. IEEE T ans. Biomed. Eng. 2013,60, 2142–2152. [C ossRe ] [PubMed]
26.
Teichmann, D.; De Ma eis, D.; Ba el , T.; Wal e , M.; Leonha d , S. A bendable and wea able
ca dio espi a o y moni o ing de ice using wo noncon ac senso p inciples. IEEE J. Biomed. Heal h In .
2015,19, 784–793. [C ossRe ] [PubMed]
©
2016 by he au ho s; licensee MDPI, Basel, Swi ze land. This a icle is an open access
a icle dis ibu ed unde he e ms and condi ions o he C ea i e Commons A ibu ion
(CC-BY) license (h p://c ea i ecommons.o g/licenses/by/4.0/).