200 IEEE REVIEWS IN BIOMEDICAL ENGINEERING, VOL. 15, 2022
Moni o ing and Synch oniza ion o Ca diac and
Respi a o y T aces in Magne ic Resonance
Imaging: A Re iew
Ma ina Lad o a , Radek Ma inek , Jan Nedoma , Pa la Hanzliko a , Michael Douglas Nelson ,
Radana Kahanko a , Jind ich B ablik , and Jakub Kola ik
(Me hodological Re iew)
Abs ac —Synch oniza ion o human i al signs, namely
he ca diac cycle and espi a o y excu sions, is necessa y
du ing magne ic esonance imaging o he ca dio ascula
sys em and he abdominal ca i y o achie e op imal image
quali y wi h minimized a i ac s. This e iew summa izes
echniques cu en ly a ailable in clinical p ac ice, as well
as me hods unde de elopmen , ou lines he bene i s and
disad an ages o each app oach, and o e s some unique
solu ions o conside a ion.
Index Te ms—Magne ic esonance imaging (MRI), ca -
diac magne ic esonance imaging (CMRI), MRI igge ing,
ca diac igge ing, espi a o y igge ing.
I. INTRODUCTION
MAGNETIC Resonance Imaging (MRI) is a powe ul
non-in asi e ool o imaging human body s uc u e and
unc ion. The main ad an ages o MRI include he absence o
ionizing adia ion, high con as be ween di e en ypes o so
issues, and i s abili y o image in a bi a y spa ial o ien a ions.
In addi ion o p o iding high- esolu ion images o he body
s uc u e, MRI also p o ides no el pa hophysiologic insigh
in o basic bodily unc ions (e.g., molecula wa e di usion,
Manusc ip ecei ed Augus 28, 2020; e ised Decembe 22, 2020;
accep ed Janua y 25, 2021. Da e o publica ion Janua y 29, 2021;
da e o cu en e sion Janua y 24, 2022. This wo k was suppo ed
in pa by he Eu opean Regional De elopmen Fund in he Re-
sea ch Cen e o Ad anced Mecha onic Sys ems p ojec , unde P ojec
CZ.02.1.01/0.0/0.0/16 019/0000867 wi hin he Ope a ional P og amme
Resea ch, De elopmen and Educa ion, and in pa by he Minis y
o Educa ion o he Czech Republic unde P ojec s SP2020/156 and
SP2021/32. (Co esponding au ho : Radana Kahanko a.)
Ma ina Lad o a, Radek Ma inek, Radana Kahanko a, Jind ich
B ablik, and Jakub Kola ik a e wi h he Depa men o Cybe -
ne ics and Biomedical Enginee ing, Technical Uni e si y o Os-
a a, 70800 Os a a, Czechia (e-mail: ma ina.lad o a@ sb.cz;
adek.ma inek@ sb.cz; adana.kahanko [email p o ec ed]; jind ich.b ablik@
sb.cz; jakub.kola ik@ sb.cz).
Jan Nedoma is wi h he Depa men o Telecommunica ions,
Technical Uni e si y o Os a a, 70800 Os a a, Czechia (e-mail:
jan.nedoma@ sb.cz).
Pa la Hanzliko a is wi h he Os a ska Uni e zi a, 70103 Os a a,
Czechia (e-mail: pa la.hanzliko [email p o ec ed]).
Michael Douglas Nelson is wi h The Uni e si y o Texas, A ling on, TX
76019 USA (e-mail: [email p o ec ed]).
Digi al Objec Iden i ie 10.1109/RBME.2021.3055550
Fig. 1. Examples o MR mo ion a i ac s: (a) blu ing due o mo e
andom espi a o y mo ion and (b) ghos ing a i ac s caused by pe iodic
b ea hing [3].
issuepe usion,o MRspec oscopy).Toge he , hein o ma ion
p o ided can help make mo e accu a e diagnoses, and imp o es
ou abili y o moni o ea men ou comes [1].
This e iew ocuses on MRI applica ions ha equi e i al
sign synch oniza ion, namely wi h he ca diac cycle and/o
espi a o yexcu sions.Indeed, igge ing andga ingde icesand
algo i hms a e necessa y o ensu e high image quali y o egions
wi hin he ches and abdomen (e.g. hea , li e , panc eas). The
ask is o pe o m imaging du ing minimal mo emen o he
hea and ho ax, which causes mo ion a i ac s, appea ing as
shadows o blu ed con ou s on he image ( e e ed o as “ghos -
ing a i ac s” see Figu e 1). Synch oniza ion se es o supp ess
hese a i ac s bu is no wi hou se e al majo challenges. In
pa icula , synch oniza ion o biological signals is limi ed by
inhe en high- equency dis u bances and ex eme magne ic in-
duc ion, which in e e es wi h he measu emen o i al signs,
and ul ima ely comp omises MRI sequence synch oniza ion
[2], [3].
A. Abdominal MRI
Abdominal MRI is widely used due o i s abili y o ex ac
in o ma ion a he le el o issue composi ion and o assess
unc ional s a us, including he me abolic s uc u e o he issue.
I allows clinicians o espond o issue damage o dys unc ion
and o adap he he apy o disease p e en ion s a egy acco d-
ingly. Gi en he e sa ili y o MRI, i can de ec a ple ho a o
abdominal diso de s, such as s ea osis, ib osis, in lamma ion,
This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 License. Fo mo e in o ma ion, see h p://c ea i ecommons.o g/licenses/by/4.0/
LADROVA e al.: MONITORING AND SYNCHRONIZATION OF CARDIAC AND RESPIRATORY TRACES 201
and umo s (malignan and benign). Fo example, when exam-
ining he li e , bile duc s, and panc eas, MR cholangiopanc e-
a og aphy begins o eplace he in asi e endoscopic me hod
in diagnos ic indica ions. Also, MRI can be used o iden i y
in lamma o y o neoplas ic diseases wi h bowel wall abno -
mali ies o classi y some ypes o C ohn’s disease by accu a e
de ec ion o indi idual lesions and e alua ion o disease ac i i y.
Mo eo e , MRI is also an excellen ool o di e en ia e he
benign na u e o asymp oma ic ad enal lesions, he p obabili y
o which inc eases wi h highe age and is becoming a common
p oblem. Clinical e alua ion wi h MRI also plays a c ucial ole
in he diagnosis, planning, and assessmen o ea men o bo h
benign and malignan gynecological condi ions, and has e en
p o ed o be supe io o compu e omog aphy in he diagnosis
o u e ine and ce ical cance [1], [4].
B. Ca dio ascula MRI
Ca dio ascula magne ic esonance imaging (CMRI) is in-
c easingly used o examine hea s uc u e and unc ion. I s
non-in asi e na u e, oge he wi h i s a oidance o ionizing
adia ion, allow o epea able, low- isk cha ac e iza ion o he
myoca dium and i s associa ed componen s, h oughou disease
p og ession/ eg ession. One o he mos equen applica ions
o CMRI in clinical p ac ice is he assessmen o a ial and
en icula mo phology. Due o CMRI’s high spa ial esolu-
ion, h ee-dimensional co e age, and excellen high con as -
o-noise a io (CNR), e alua ion o chambe dimension, size and
shape is highly obus and ep oducible [5]–[7]. As such, CMRI
is ecognized o i s accu a e cha ac e iza ion and de ec ion o
ca diomyopa hies (hype ophic o es ic i e ca diomyopa hy,
e c.), pa aca diac masses (bo h malignan and benign ca diac
umo sincluding seconda y umo s, pseudo umo s andin aca -
i a y h ombi), congeni al hea diseases, and condi ions o he
pe ica dium, sep al de ec s, obs uc i e lesions, pos ope a i e
s a es and mo e. Ca dio ascula MRI also includes s uc u es
beyond he hea , including he ao a and o he majo blood
essels. Fo example, MRI can be used o de ec ao ic aneu ysm
and dissec ion, as well as a he oscle o ic plaque [5], [6], [9].
Synch onizing CMRI wi h he ca diac cycle p o ides ad-
di ional insigh in o ca diac unc ion and myoca dial ia-
bili y, allowing examina ion o myoca dial ischemia and/o
acu e/ch onic myoca dial in a c ion. In eg a ion o CMRI wi h
eloci y/ low measu emen s p o ides insigh in o ca diac hemo-
dynamics and/o al ula unc ion, such as mi al, icuspid
and pulmona y in low e sus egu gi a ion, al ula s enosis,
pulmona y hype ension, o p os he ic al e unc ion [5], [8],
[9].Fu he mo e,asCMRIhase ol ed,pa icula lyo e hepas
10-15 yea s, mo e no el applica ions ha e eme ged, including
MR angiog aphy, myoca dial issue cha ac e iza ion, and my-
oca dial pe usion imaging [10]–[12].
C. High-Field MRI
The e is a gene al end in he ield o MRI o inc ease mag-
ne ic ield s eng h. In jus he pas 10 yea s, we ha e wi nessed
an ex ensi e expansion o clinical scanne s om 1.5 T o 3 T,
wi h wo o he majo endo s now o e ing FDA-app o ed
Fig. 2. Compa ison o he quali y o ce eb al angiog aphy images a
(a) 1.5 T, (b) 3 T, and (c) 7 T magne ic ield s eng hs [8].
7 T clinical scanne s. In p inciple, inc easing he magne ic ield
s eng h is di ec ly associa ed wi h inc easing signal in ensi y;
howe e , i b ings o he sou ces o e o . Wi h he highe mag-
ne ic ield s eng h, he inhomogenei y bo h o magne ic and
adio equency ield con ibu e o he deg ada ion o image
quali y, which has o be eclaimed by di e en op imiza ion
echniques [13]–[16]. Thus, be o e he inc eased signal in ensi y
andiso opic esolu ioncanbe ealized,sophis ica edalgo i hms
mus i s be de eloped and in eg a ed. Ano he limi a ion o
imaging in high magne ic ields a e he magne ic e ec s on
he elec ical signals used o moni o ing i al signs o he
pa ien ; challenging synch oniza ion (see mo e in Sec ion II-A).
Toge he , hese inhe en limi a ions o en lead o p olonged
scan imes, and in some cases, imp ac icable long b ea h-hold
imes [17], [18].
Ne e heless, imaging a highe ield s eng hs holds g ea
p omise. The con as be ween blood and issue is much mo e
esol ed, and he e is g ea po en ial o accele a ed imaging
and educing imaging a i ac s; majo ad an ages when imaging
small, as -mo ing s uc u es [8], [19]. Fo example, he in-
c eased signal ene gy has a g ea ad an age in MR angiog aphy,
whe e iny essels, such as ce eb al a e ies, a e mo e isible
a 7 T han a 1.5 T (see Figu e 2). In his con ex , high ield
s eng h MRI p omises o p o ide new insigh in o p e iously
un esol ed ana omical s uc u e and physiological unc ion in
i o [8], [18].
D. MRI Synch oniza ion
Synch oniza ion o MRI (o ga ing/ igge ing) is dependen
upon eal- ime acquisi ion o ca dio espi a o y unc ion. As
such, MRI ga ing can be di ided in o wo main ca ego ies:
espi a o y, which allows educing mo ion a i ac s in
he image esul ing om he ches mo emen du ing a
pa ien ’s ee b ea hing (when scanning o an abdominal
and ho acic a ea), and
ca diac, he goal o which is o elimina e he e ec o
myoca dial mo emen and cap u e a hea scan in he
speci ic phase o i s cycle (when scanning o a ho acic
a ea), usually moni o ed by elec oca diog aphy (ECG)
signal.
202 IEEE REVIEWS IN BIOMEDICAL ENGINEERING, VOL. 15, 2022
Fig. 3. The basic p inciple o he p ospec i e igge ing o (a) mo pho-
logical imaging, (b) cine imaging.
Synch oniza ion o MRI can be di ided in o wo basic p in-
ciples ”p ospec i e and e ospec i e. In he case o p ospec-
i e igge ing, he MRI acquisi ion is igge ed by he desi ed
physiologic e en (e.g. R wa e o ECG signal) wi h a speci ic
delay o by eaching he speci ied le el o inspi a ion. Thus, he
acquisi ion usually coincides wi h he mos mechanically quie
phase o he ca diac/ espi a o y cycle (see Figu e 3) [20].
P ospec i e igge ing in cine imaging allows da a acquisi ion
co e ing mos o he ca diac cycle, de e mined by he numbe o
ca diac phases o segmen s wi hin an R-R in e al. The acqui ed
da aa e so ed in ok-spaces( ep esen ing he indi idual ames)
acco ding o he ca diac phase, and each k-space is illed o e
se e al hea bea s. This unc ional se s a s wi h he R wa e.
Since he du a ion o he R-R in e al a ies sligh ly om
he hea bea o he end, he las 10% o he R-R in e al is
usually no sampled (see Figu e 3(b)). P ospec i e igge ing
has ad an ages in noise il a ion.
The in e al o acquisi ion is o en se such ha he de ec o
in en ionally igno es any high ampli udes o he han R wa es,
p e en ing alse igge ing. This abili y, howe e , is p oblema ic
when moni o ing pa ien s su e ing om apid o i egula ac-
i i y ”a hy hmias. In pa ien s wi h p ema u e hea bea s, he
ea lie R wa e is passed because he g adien pulses igge ed
by he p e ious R wa e a e s ill unning, which esul s in
ex ended acquisi ion ime. This hea a e a iabili y mus be
conside ed when planning he acquisi ion window posi ioning
o a p ospec i ely igge ed examina ion. Fo he de ec ion o
he nex R wa e, he acquisi ion window posi ion should no
be oo close o he nex R wa e. An acquisi ion window placed
oo close may in e e e wi h he nex R wa e and sho en he
co esponding R-R in e al. Du ing he scan, he examine mus
Fig. 4. The basic p inciple o he e ospec i e ga ing.
Fig. 5. The p inciple o ca diac igge ing.
moni o changes in he hea cycle and app op ia ely adap he
igge ing pa ame e s [21], [22].
Re ospec i e ga ing allows con inuous acquisi ion o MR
da a (see Figu e 4). Bo h he image and he hea / espi a o y
signal a e cap u ed o e se e al cycles and used o econs uc
he image by eo de ing, g ouping, o co ela ing wi h phase o
he ca dio espi a o y cycle. This p ocedu e equi es uncon en-
ional “ eal- ime” pulse sequence (e.g., o unc ional imaging)
and acquisi ion so wa e decisions, because he me hod o da a
acquisi ion changes as i uns. In his way, e ospec i e ga ing
uses complex me hods o in e e ence il e ing [20], [21].
The main ad an age o e ospec i e acquisi ion is he abili y
o collec da a om all hea phases. Like p ospec i e igge ing,
e ospec i e me hods may ha e p oblems wi h a hy hmias and
low R-wa e ampli udes. The sys olic and dias olic pe iods a e
une enly al e ed in case o hea a e a iabili y o a hy hmia
e en s, so e ospec i e so wa e is unable o compensa e by
app op ia e da a segmen a ion, esul ing in e o [21], [22].
The ollowing chap e s ou line common p oblems ela ed
o ca diac cycle and espi a o y synch oniza ion, summa ize
a ailable synch oniza ion echniques commonly used in clinical
p ac ice, as well as in oduce se e al no el synch oniza ion
app oaches cu en ly unde de elopmen . We discuss he ad-
an ages and disad an ages o each app oach, especially in he
case o he inc easingly popula high ield s eng h MRI.
II. CARDIAC GATING METHODS
Ca diac ga ing is c i ically impo an o CMRI. In cu en
p ac ice, ca diac ac i i y is mos commonly moni o ed by ECG
signal. The p inciple o ca diac igge ing (see Figu e 5)is
based on he de ec ion o a change in signal o he hea (e.g.,
he R wa e o ECG signal, en icula depola iza ion). Gi en
he p ominence o he R wa e, and i s ou ine use in ca diac
LADROVA e al.: MONITORING AND SYNCHRONIZATION OF CARDIAC AND RESPIRATORY TRACES 203
Fig. 6. Compa ison o ECG- igge ed and non- igge ed images using (a) T2-Weigh ed Fa -Supp essing Sequence, and (b) T1-Weigh ed
Myoca dial-Supp essing Sequence.
physiology o deno e he beginning o he ca diac cycle, ca diac
imaging is o en igge ed by he R wa e [23]. Depending on he
imaging being pe o med, de ec ion o he R wa e may ini ia e
a se ies o e enly imed acquisi ions ac oss he R- o-R in e al
in o de o econs uc a cinema ic image o ca diac sys ole and
dias ole.Al e na i ely, heR-wa emaybeused ode ine he ime
pe iod o a single acquisi ion (e.g. end-sys ole o mid-dias ole).
The mos commonly used hea scan sequences a e cine
T1/T2-weigh ed sequences, a -supp essing T2-weigh ed se-
quences, and myoca dial-supp essing T1-weigh ed sequences.
These examina ions a e usually pe o med unde espi a o y
quiescence, wi h he subjec s holding hei b ea h o he en i e
measu emen pe iod “ om 6 o 25 seconds ” o a oid espi a o y
mo emen a i ac . Figu e 6 shows examples o scans ig-
ge ed by combina ion o ECG signal and a b ea h-holding/ ee-
b ea hing. As can be app ecia ed, combining ECG igge ing
wi h espi a o y quiescence imp o es o e all image sha pness
and elimina es mo ion a i ac s.
A. Ca diac Ga ing Challenges
A i ac s du ing MRI may o igina e om bo h he pa ien and
he measu emen sys em. Fi s ly, he elec odes and associa ed
connec o sandcablesshouldbe made o non- e omagne ic ma-
e ial o educe magne ic ield dis o ion, whe eas in e e ence
may alsely igge he scanne andcause loss o synch oniza ion
wi h he hea [21], [24]. Blood low in he hea can induce
ol age and gene a e hyd odynamic a i ac s. Since blood is
an elec ically conduc i e liquid whose mo emen p oduces an
elec ical cu en added o he hea signal, a magne ohyd ody-
namic e ec occu s. The elec ical signal gene a ed in his way
usually a ec s he T wa e o ECG signal (see Figu e 7). I he
Fig. 7. Demons a ion o magne ohyd odynamic a e ac a ec ing
ECG wa e o m measu ed: (a) ou side he in luence o MR magne ic
ield, (b) in he 1.5 T magne ic ield, whe e bo h he R wa e and he
T wa e acqui e simila ampli udes.
T wa e occu s a a ime o apid mo emen o blood om he
hea , he T wa e is dis o ed [25]. In addi ion, su ace ecei e
coils a e commonly used in CMRI examina ions o ampli y and
imp o e heacqui ed signal, causing his in e e ence oinc ease
signi ican ly.
O he p oblems, such as diso de s o ca diac conduc ion sys-
em o un ela ed physiological p ocesses (e.g., muscle emo ),
a e addi i e o he in e e ence men ioned abo e. The op imal
image quali y equi es app op ia e elec ode placemen ha
maximizes he ampli ude o he R-wa e while minimizing hese
undamen al a i ac s. Ex eme changes, such as he noise o a
high-ampli ude g adien sys em o he small elec ical cu en s
caused by a change o magne ic ield g adien , a e e y dis up-
i e. Howe e , due o he la ge equency di e ence be ween
204 IEEE REVIEWS IN BIOMEDICAL ENGINEERING, VOL. 15, 2022
Fig. 8. ECG wa e o m wi h i s igge ing signal.
he noise and ECG, high equencies can o en be supp essed by
il e ing [21], [26].
In many applica ions, bo h p ospec i e and e ospec i e e-
jec ion o a hy hmia has been in eg a ed in o acquisi ion and
econs uc ion so wa e. In his case, an accep ance window
ha allows R-R a iabili y mus be speci ied. This s ep usually
inc eases he acquisi ion ime, and he examine has o decide
whe he o educe he scan ime by inc easing he numbe o
segmen s a he expense o image quali y. Howe e , he use o
p ospec i e ga ing ins ead o e ospec i e ga ing can o e come
ex eme de ia ions o he R-R in e al. Typically, he sho e
R-R in e al should be used o schedule he acquisi ion window,
whichmus be sligh lysho e han hesho es R-Rin e al.The
esul ing image will ep esen an incomple e hea cycle, bu a
calcula iono ejec ion ac ion (amongo he ou comemeasu es)
is s ill possible [21], [24].
The a i ac s ela ed o magne ic ield a e pa icula ly p o-
nounceddu ingmeasu ing a highmagne ic ields eng h. When
using ield s eng hs highe han 3 T, he in luence o he mag-
ne ohyd odynamic e ec on he signal inc eases, and ob ain-
ing a clea signal o su icien quali y o aul less igge ing
becomes mo e di icul [27], [28]. As ul a-s ong MRI ields
become mo e widesp ead, solu ions o his sensi i i y issue a e
inc easingly needed. Thus, new me hods o igge ing CMRI,
discussed in mo e de ail below, ha e been de eloped and a e
cu en ly being s udied.
B. Elec oca diog aphy
Sensing o he ECG signal (see Figu e 8) is one o he
bes known and mos commonly used me hods o moni o ing
ca diac ac i i y. The ECG measu emen s consis o placing he
elec odes on he pa ien ’s ches a he speci ic loca ions. I is
measu ed in cu en leads, usually in 12 leads o s anda d non-
MR diagnos ic measu emen s wi h 3 limb elec odes included.
When igge ing CMRI, ewe leads a e su icien , because he
signal does no se e as a ool o a diagnose o he hea unc ion
diso de s bu only o he QRS complexes de ec ion. Thus, he
numbe o elec odes is commonly educed o h ee o ou (see
Figu e 9) [24].
Al hough ECG is a means o quick ca diac unc ion eco ding
om which each phase o he ca diac cycle can be de e mined
o he pu pose o MRI igge ing, many p oblems wi h he
MRI de ice i sel and i s magne ic ield, such as a i ac s o
complica ions when measu ing wi h ce ain ma e ials, a ise.
Fo he bes possible p e en ion o a i ac s, i is necessa y o
Fig. 9. Typical dis ibu ion o ECG elec odes in MR en i onmen s:
(a) – pa allel dis ibu ion ad an ageous o high magne ic ield s eng hs,
(b) – elec ode dis ibu ion no mally used a 1.5 T ield s eng hs.
Fig. 10. Example o a ailable ECG-based senso s: (a) Wi eless ECG
uni o sensing in MR en i onmen [32]. (b) Quad ode ECG elec ode
o he MR ga ing [33].
p epa e he skin well a he loca ions o ECG elec odes (mos
o en loa ing elec odes) placemen , since accu a e de ec ion
o low po en ials equi es minimal skin impedance and op imal
elec odecon ac wi h he skin. Theadded noise ha esul s om
poo elec ode con ac wi h he skin hen leads o e oneous
igge ing du ing he CMRI examina ion. Immedia ely be o e
he examina ion, he hai mus be sha ed, and he skin su ace
sc ubbed o wi h a mild ab asi e soap o gel be o e elec odes
a e applied; complica ing he exam and adding discom o o he
pa ien [20], [21].
Fu he mo e, ECG measu emen s equi e some sa e y p e-
cau ions due o he in e ac ion o he ECG sys em wi h RF and
g adien sys ems, since ECG, as a measu emen wi h elec ically
ac i e componen s, b ings he isk o supe icial hea ing o he
pa ien ’sskino e enbu ns esul ing omhigh ol ageinduc ion
in ECG ha dwa e [29], [30]. To p e en cu en s occu ing in
he ECG leads due o he apid swi ching o g adien ields, he
loca ion o he leads has o comp ise no loops, and he wi es
should be as sho as possible. The possibili y o bu ns due o
in e ac ion wi h he RF ield should be educed by loca ion o
he wi es ou side he esona o s, and he ba e y-powe ed ECG
measu ing uni (see Figu e 10(a)) should be used o ensu e
gal anic sepa a ion o he MR sys em and he pa ien [31].
All o he majo MRI endo s o e ECG ga ing p oduc s,
and se e al comme cially a ailable p oduc s a e also a ail-
able h ough hi d pa y endo s. Mode n p oduc s include
MR-compa ible 3-lead ECG igge ing de ice using a secu e
ixed elec ode layou (see Figu e 10(b)), which es ic s cable
leng h by dic a ing a igh elec ode placemen pa e n and hus,

LADROVA e al.: MONITORING AND SYNCHRONIZATION OF CARDIAC AND RESPIRATORY TRACES 205
Fig. 11. The example o he ECG wa e o m o ca diac igge ing: (a) be o e (I) and a e (II) sequence s a , (b) and (c) du ing he acquisi ion wi h
ailed igge ing, (d) du ing he acquisi ion wi h success ul igge ing, (e) wi h imp ope ly posi ioned ECG elec odes, ( ) wi h mo ion a i ac s.
p o ides op imal signal pe o mance and sa e y due o la ge
elec ode con ac a ea. This new design signi ican ly lowe s
esis ance using he unique gel, and educes a leas pa ly
pa ien p epa a ion ime and discom o hanks o i s o m o
disposable pa ch [33]. Se e al comme cially a ailable sys ems
allow measu emen s o mo e i al signs oge he , such as ECG,
espi a o y unc ion and blood p essu e, so ha p o ide a de-
ailed in o ma ion abou pa ien ’s condi ion du ing examina ion
[34].
As desc ibed abo e, i is also impo an o conside he in-
luence o magne ohyd odynamic e ec s du ing ECG synch o-
niza ion. This e ec causes ECG signal dis o ion and p e en s
a p ope de ec ion o R wa e and subsequen synch oniza ion.
Fo example, K ug e al. [35] and Snyde e al. [36] ha e
shown ha ECG igge ing is no app op ia e o high ields. A
ield s eng hs beyond 3 T, he ECG eco ding is bo h spa ially
and empo ally dis o ed, o he poin ha he R wa e may no
longe be clea ly iden i iable o o e shadowed by he T wa e
dominance, exceeding he R wa e by 20% o i s ampli ude.
Indeed, he ampli ude o he T wa e can e en be augmen ed
a ield s eng hs as low as 1.5T [37], challenging ansla ion
o ECG moni o ing/synch oniza ion a highe ield s eng hs.
While he p e alence o his p oblem emains incomple ely
unde s ood, some epo s es ima e he p oblem o exis in as
many as 30% o cases [27], [38]. This is pa icula ly oubling,
gi en he impo ance o accu a e R wa e egis a ion o image
acquisi ion [27], [28], [39].
Figu e 11 shows examples o he ECG signal wi hin examina-
ion wi h a ious de o ma ions ha p e en p ope synch oniza-
ion. To clea ly demons a e he e ec s o he MR en i onmen
on he ECG signal, Figu e 11(a) shows he ECG signal p io o
ini ia ing a scan sequence (I), and du ing acquisi ion (II). No e
how he ECG signal is dis o ed by he RF pulse du ing acquisi-
ion, p e en ing R wa e delinea ion, simila ly o o he cases in
Figu e 11(b) and Figu e 11(c), when only se e al R wa es we e
de ec ed. The co ec igge signal ob ained om he dis o ed
ECG du ing acquisi ion is only achie ed in Figu e 11(d).Fig-
u e 11(e) shows an example o low ampli ude R-wa es los in he
baseline noise and hus, hei ailed delinea ion. I is caused by
poo elec odes cleaning and s icking, insu icien sha ing, and
imp ope posi ioning o elec odes. Mo ion a i ac s a e deno ed
in Figu e 11( ), when he signal is comple ely dis o ed, and R
wa es a e no de ec able.
Fo all hese easons, mo e ad anced me hods o R wa e
de ec ion and magne ohyd odynamic a i ac supp ession ha e
been de eloped and used, such as noise cancella ion [40], in-
dependen componen analysis [41], [42], nonlinea Bayesian
il e ing [42], wa ele ans o m [44] o hei combina ions. Fo
example, Abi-Abdallah e al. [45] p oposed signal decompo-
si ion using wa ele ans o ma ion and subsequen adap i e
il e ing. Fo his app oach, an o -line wa ele ans o ma ion
p ocess is made o elimina e he delay in signal il e ing as much
as possible, gene a ing a e e ence signal ha can be used wi h
an adap i e il e du ing he eal- ime calcula ion phase. The
algo i hm also calcula es he espi a o y synch oniza ion signal
by ex ac ing a b ea h wa e o m ha modula es he ECG signal.
Then, a combina ion o bo h hea and espi a o y igge ing
signals pe o ms MR synch oniza ion.
Mo emode napp oachincludesa eal- imehighe -o de QRS
de ec o based on adap i e h esholding [46]. The h eshold uses
a ou h cen al momen calcula ion ha ep esen s a signi ican
signal change wi hin he QRS complex compa ed o o he com-
ponen s, such as he T o P wa e. S äb e al. [47] obse ed
only a e occasions o alse nega i e, alse posi i e o misplaced
igge e en so ECG igge inga ields eng h7Tbyincluding
a lea ning phase o he R-wa e de ec ion. The p oposed igge
algo i hm lea ns he shape o he ising edge o he R-wa e
while he subjec is lying ou side he magne bo e, whe e he
magne ohyd odynamic e ec is negligible. A e lea ning phase
is comple ed, he algo i hm compa es di e en de i ed en i ies
o he incoming ECG signal wi h he co esponding en i ies o
he lea ned shape in eal ime.
206 IEEE REVIEWS IN BIOMEDICAL ENGINEERING, VOL. 15, 2022
Fig. 12. Typical placemen o VCG elec odes in MR en i onmen .
Al hough he discussed me hods, imp o ing he R wa e
delinea ion, eached e y high accu acy in ca diac igge ing,
he p ocessing algo i hms ha e unduly high compu a ional e-
qui emen s and ha e no been ye sui able o in oduc ion in o
he medical p ac ice. The insu icien compu ing pe o mance
in he pas was he main limi ing ac o o using hese com-
plex me hods, which emained o be used only in he esea ch
a ea. Howe e , he inc easing pe o mance o mic op ocesso
echnology (mul ico e p ocesso s, p og ammable ga e a ays,
e c.) b ings new ends and possibili ies in a ailable compu e
equipmen , signal p ocessing and analysis o e he pas ew
yea s. On he o he hand, al hough he so wa e imp o emen o
ECG synch oniza ion does no demand he addi ional ha dwa e,
he esea ch o o he sensing me hods (wo king on ano he
physical p inciple) could inc ease he accu acy o ca diac syn-
ch oniza ion igh in he lowes senso y laye . Thus, oge he
wi h ad anced p ocessing algo i hms, hey could p o ide highe
image quali y o diagnos ics.
C. Vec o ca diog aphy
Vec o ca diog aphy (VCG) desc ibes elec ical ac i i y o he
hea by h ee independen loops, each ep esen ing indi idual
phases o he ca diac cycle (P wa e, QRS complex, and T wa e).
The loops can be displayed in a one-dimensional image, simila
o he con en ional ECG, as well as wo- and h ee dimensional
loop econs uc ions, ep esen ed by he ampli ude (mV) o he
elec ical signal be ween leads (i.e. X, Y lead; X, Z lead; Y, Z
lead; o ac oss each o he X, Y, and Z leads). In e ms o diag-
nos ic in o ma ion, VCG is mo e sensi i e o he de ec ion o
hype ophyand ischemichea disease [48]–[51]. In he MR en-
i onmen , VCG is mo e esis an o he magne ohyd odynamic
e ec , since he hea ’s elec ical axis is o ien ed in he opposi e
di ec ion o blood low h ough he hea . Fo his eason, some
s udies ha e begun o in es iga e CMRI synch oniza ion using
he VCG ins ead o he ECG signal, whe e he ec o model is
used o app oxima e he ECG signal om any lead. F om an
applica ion pe spec i e, he main di e ence be ween ECG and
VCG esides in he placemen o he elec odes a ached o he
pa ien ’s ches (see Figu e 12).
As men ioned, VCG has gained in popula i y wi hin he MR
communi y, gi en i s esis ance o he magne ohyd odynamic
Fig. 13. (a) Typical VCG wa e o m and (b) mani es a ion o magne o-
hyd odynamic e ec .
Fig. 14. Illus a ion PPG based sys em p inciples and deploymen :
(a) Example o inge PPG uni o measu emen in MR en i on-
men [32], (b) P inciple o PPG signal sensing om inge , (c) PPG
wa e o m wi h i s igge signal.
e ec (see Figu e 13). Indeed, i s u ili y o ca diac synch o-
niza ion has been ecognized o mo e han wo decades, wi h
a deg ee o success [24], [26], [52]. Like ECG moni o ing,
howe e , he me hod lags in cases o se e e a hy hmia, when
he igge ing sys em is unable o di e en ia e he QRS loop
om he loop gene a ed by he ec opic bea . Likewise, VCG
is simila ly a ec ed by high s eng h and complexi y o he
magne ohyd odynamic signal in 7 T ield s eng h [35].
D. Pulse Wa e
Ca diac synch oniza ion using a pe iphe al pulse measu ed
by pho ople hysmog aphy (PPG) is no as widesp ead as a
s anda d ECG signal measu emen bu is a sui able al e na i e,
especially gi en i s esis ance o MR a i ac s. The PPG signal
(see Figu e 14(c)) is mos o en measu ed by a ligh senso
(see Figu e 14(a)), which wo ks by assessing he blood abso p-
ion o ligh co esponding o changes in blood olume (see
Figu e 14(b)). The ise and all o his measu ed wa e e lec
he sys ole and dias ole p ocess, making i a sui able ool o
igge CMRI [53]. While pulse wa e de ec ion is echnically
simple , in ol ing placemen o a de ice on an easily accessible
inge (as oppose o elec odes placed on he ches ), se e al
majo disad an ages ha e p e en ed i s widesp ead adop ion o
ou ine MR synch oniza ion. Fi s , he ligh senso is suscep ible
o mo emen , especially mo emen o he inge /hand on which
LADROVA e al.: MONITORING AND SYNCHRONIZATION OF CARDIAC AND RESPIRATORY TRACES 207
Fig. 15. Pulse wa es wi h igge ing ma ks (a) signal wi hou a i ac s, (b) signal wi h a i ac s caused by he mode a e mo emen o he inge , (c)
signal dis o ed by s ong and sha p mo emen .
Fig. 16. Sel -Ga ing wa e o m wi h i s igge signal.
he senso is loca ed. Figu e 15 illus a es ypical mo emen
a i ac , and he associa ed igge ing e o ha accompanies
such mo emen . Spiche e al. [53], [54] ha e a emp ed o
o e come his limi a ion by in oducing a con ac less ideo-
based PPG sys em, bu ha e no ye succeeded in widesp ead
clinical in eg a ion o his app oach. Second, he delay be ween
ca diac ac i i y (i.e. mechanical ejec ion o blood) and de ec ion
o ha ac i i y in he pe iphe y c ea es a sizable ime delay (i.e.
se e al hund ed millisecond delay needed o pulse o a el o
he pe iphe y), he ex en o which canno be accu a ely, no
ep oducibly co ec ed o [27]. The e o e, in mos cases, his
igge ing me hod is only used when o he igge ing me hods
ail.Themos no ableexcep ionis lowimaging o ce eb ospinal
luid h ough ce eb al en icles and openings, o which PPG
igge ing is supe io o ECG ga ing [55].
E. Sel -Ga ing Me hods
Sel -ga ing (S-G) echniques, see [97]–[100], elimina e he
need o ex a ha dwa e o CMRI synch oniza ion by ecei ing
he igge ingin o ma iondi ec ly omMRsignals.Theme hod
eliesei he onacqui ing adial(o spi al) k-spaceda a ha co e
he k-space cen e in each eadou o on acqui ing addi ional
(non-phase-encoded) na iga o eadou s h ough he k-space
cen e . This majo ad an age is no wi hou i s own d awbacks
howe e ,assuchanapp oachsigni ican lyex ends hescan ime.
The p inciple o S-G CMRI is based on changes o mo emen
o olume in he image. Tha means a se ies o consecu i e
echoes demons a es peak al e a ions co esponding o p opo -
ional changes in o al ans e se magne iza ion due o o gan
mo emen and changes in blood olume. This da a is hen
p ocessed and segmen ed o c ea e an ECG compensa ing signal
(see Figu e 16), acco ding o which he image is subsequen ly
econs uc ed acco ding o he associa ed ca diac cycle, ep e-
sen ed by he ampli ude and phase o he echo [97]. Due o
i s na u e, his signal is esis an o a i ac s ha may a ec
igge ing signals o o he me hods (ECG, VCG, e c.). The
S-G me hod is based on a i s -di e ence de ec ion algo i hm;
howe e , many s udies deal wi h mo e complex ypes o inc ease
he accu acy o ca diac synch oniza ion [99].
Nijm e al. [99] compa ed he classical me hod, based on
he i s di e ence calcula ion (including he signal il e ing by
i s de i a i e), wi h mo e complex me hods, such as he median
empla e ma ching echnique (wi h calcula ion o he co ela ion
be ween he o iginal signal and he empla e, de e mined by hal
he RR median calcula ed by he i s di e ence me hod) and
polynomial ma ching echnique (a polynomial wi h leng h o
an RR in e al median is used as a empla e he e). The cubic
polynomial me hod achie ed he lowes e o and he highes
SNR and was deemed “compa able” o ECG synch oniza ion.
I is wo h no ing howe e , ha none o he S-G me hods
ou pe o med he ECG measu emen echnique o he o iginal
i s di e ence me hod.
In o de o elimina e e iciency de ici s o he p e iously
in oduced S-G me hods, he o he s include using modula ion
o MR echo ha occu s h ough: (1) issue mo emen i sel
(echo-peak me hod), (2) kymog am (1D ga ing signal om
he empo ally e ol ing 2D cen e o mass), and (3) 2D low-
esolu ion image co ela ion [101]–[105]. In each case, he se
o iews a each ca diac phase was de i ed using ime-s amp
in o ma ionandcomplexlinea in e pola ionalgo i hm.A linea
eg essionalgo i hmwasused ocompu esignalcomponen s o
indi idual sampled k-space posi ions (i.e., spa ial- equency do-
mains) ac oss ca diac phases. Then, he con olu ional algo i hm
o image econs uc ion c ea ed an image o each ca diac cycle.
The delay o all ob ained S-G signals is 2.5 ms. The echo-peak
me hodseems obe he mos p ac icalo he h eeme hods es ed
conside ing he image quali y, alues o he ime a iabili y o
he igge ing signal, and low compu a ional cos . Mo eo e ,
he co ela ion echnique equi es an ope a o ’s in e ac ion. All
h ee me hods achie ed e y simila esul s o image quali y
compa ed o ECG synch oniza ion [97].
Mos cen al line S-G me hods esul s in a doubling o he
acquisi ion ime, while adial s eak a i ac s a e encoun e ed
wi h he p ojec ion econs uc ion me hod and can in e e e
wi h image in e p e a ion. To o e come hese limi a ions, C owe
e al. [100] u ilized S-G igge ing on a double-echo sequence
208 IEEE REVIEWS IN BIOMEDICAL ENGINEERING, VOL. 15, 2022
whe e he second g adien echo wi hou phase coding was
used o gene a e a igge signal. A sampling o he second
echo p o ided subjec i e image quali y sco e be ween good
and excellen , which means a sa is ac o y esolu ion and con-
as o image in e p e a ion and de ini ion o ine ana omic
s uc u es. Also, no s a is ically signi ican di e ences we e ob-
ainedbe weenS-Gand ECG-ga edimagesandmass pa ame e s
measu emen s in all olun ee s a clinically p ac ical acquisi ion
imes, ex ended negligibly. This echnique is no suscep ible o
he adials eaka i ac s,bu he ewasnos a is icallysigni ican
di e ence in image quali y be ween S-G and ECG synch oniza-
ion.
Mos S-G me hods ha e been success ully used o e ospec-
i e ga ing o CMRI, which is no as sensi i e o changes in hea
a e (as in he case o p ospec i e igge ing) and p o ides an
image h oughou hehea cycle,whichispa icula lyimpo an
o de e mining compulsi e sys olic o dias olic p ocesses [99].
The S-G echnique is unsui able o p ospec i e igge ing,
which equi es accu a e de e mina ion o igge delay in o de
o place an acquisi ion window in he a ge phase o he ca diac
cycle [28], [97], [99]. Fu he mo e, hese me hods a e lagging
in e y high hea a e and some hea diseases exp essing
hemsel es by e y mild myoca dial con ac ion and elaxa ion
h oughou he hea cycle [100]. Hiba e al. [98] o e came his
challenge o scanning a a e y as hea a e (up o 600 bpm)
when es ing he S-G echo-peak me hod du ing imaging a a ’s
hea . The S-G igge ing p o ided be e imaging o papilla y
muscles han ECG ga ing and achie ed highe SNR alues.
F. Op ical Me hods
The use o ibe op ic senso s b ings an inno a i e app oach
o biological signals measu emen and is inc easingly being
handled by cu en esea ch. Op ical senso s enable sensing a
wide ange o signals, such as sound mani es a ions, p essu e
changes, o empe a u e a ia ions. All hese signals a e mea-
su ed non-in asi ely simply by a aching he senso o he sens-
ingloca ion,e.g.,apa ien ’sches o back.Theop icalmeasu ing
o i al signs in he MRI en i onmen has he basic ad an ages
o i s ha mlessness since he senso is made only o op ical ibe
ma e ial and p o ec i e elemen s (cases). Mo eo e , he senso s
a e highly immune o he a i ac s a ising om magne ic and
high- equency elec omagne ic ields, so no a i ac s occu in
he image. O he bene i s include he e y small dimensions o
he senso s, hei low weigh , and he minimiza ion o wi es.
Howe e , he signi ican disad an age o op ical me hods is he
high cos and size o he in e oga ion and measu ing uni , which
allows he con e sion o signals in o digi al o ma . On he o he
hand, his uni can be used o e alua e he esul s o mul iple
senso s. Thus, he p ice o he sys em can be dec eased when
comp ised o he depa men wi h mul iple MR scanne s, which
is usual in he clinical p ac ice.
Today, he mos commonly used ypes o op ical senso s
include senso s based on ligh in e e ence, so-called in e e -
ome ic senso s, and senso s wi h Fibe B agg G a ing (FBG),
whiche alua echangeso ligh e lec edon heg a ings uc u e.
The mino bu cu en ly e ol ing pa consis s o mic o- and
Fig. 17. P inciple o mic o-bending op ical senso . The de ice de ec s
changes in he inpu ligh due o he mechanical ac ion ( o ce) on he
op ical ibe .
Fig. 18. Illus a ion o PCG based sys em p inciples and deploymen :
Typical in e e ome ic senso placemen o he ca diac ac i i y mea-
su emen in posi ion: (a) “ s anding, (b) ” supine; (c) Example o an
In e e ome ic senso (d). PCG wa e o m wi h i s igge signal.
mac o-bending ibe op ic senso s, which wo k on he p inciple
o e alua ing changes in ligh ou pu caused by op ical ibe
bends c ea ed h ough a “sandwich” mic o-bende s uc u e (see
Figu e 17) [56]–[58].
The p inciple o measu ing hea a e using in e e ome ic
senso s (see Figu e 18(c)) is gene ally based on phonoca dio-
g aphy (PCG) signal (see Figu e 18(d)), gene ally consis ing
o se e al hea sounds (S1 ”S4). S ill, in mos cases, only he
i s wo sounds a e e iden , o which he i s (S1) e lec s he
en icula sys ole. The mechanical-acous icac i i yo he hea
and he mechanical ac i i y o he lungs cause changes in he
e ac i eindexo heco eandin heleng ho hemeasu inga m
o he senso placed on hebody (seeFigu e 18(a),(b)).The in e -
e ome ic measu ing sys em hen e alua es hese changes. This
me hod o ca diac ac i i y moni o ing has indeed been desc ibed
p e iously [59]–[67]. The p inciple o hea a e measu emen
using an FBG senso is based on ballis oca diog aphy (BCG,
see Figu e 19(d)) and has been desc ibed p e iously [69]–[76].
Ca diac ac i i y is mani es ed physiologically in he ches a ea
by a sligh p essu e ac ion occu ing due o he mechanical
LADROVA e al.: MONITORING AND SYNCHRONIZATION OF CARDIAC AND RESPIRATORY TRACES 215
Fig. 29. The p inciple o na iga o igge ing in CMRI.
image da a. Howe e , da a needs o be acqui ed h oughou he
hea cyclein CMRI, so igge ingby one na iga ionecho wi hin
one cycle is insu icien o his pu pose [141].
Pe e s e al. [139] used wo na iga ion echoes wi hin one
hea cycle o CMRI espi a o y igge ing ” jus be o e he
QRS complex occu ed and 500 ms a e ha . Da a is ecei ed
when bo h na iga o posi ions all in o he accep ance window.
The e was no signi ican di e ence be ween he image quali y o
he alues o he obse ed pa ame e s (le en icula unc ion
and olume) by igge ing and b ea h-holding, bu he CNR o
hese images showed di e ences on behal o he b ea h-hold
me hod (47 ±14 s. 21 ±10). Howe e , he na iga ion echoes
p o ided su icien image quali y o diagnosis and inc eased
he e iciency and speed o he examina ion ( o al ime o
b ea h-holding was 10 minu es, including pauses, o igge ing
3.7 minu es), which b ings bene i s o pa ien s unde s ess o
comple ely incapable o coope a ion in b ea hing.
D. Sel -Ga ing Me hods
Like ca diac synch oniza ion, S-G echniques (see
Sec ion II-E) also deal wi h espi a o y igge ing, because
he a o e-men ioned me hods o moni o ing espi a o y ac i i y
may lag i he posi ion o he ches wall changes wi h espec
o he hea ’s posi ion du ing se e al espi a o y cycles. The
S-G echnique o en allows igge ing by bo h i al unc ions
in CMRI examina ions, whe e ob aining he indi idual signals
by p ope il a ion (each signal has a di e en equency
esponse), see [98], [141]–[143].
The me hod is based on calcula ing he co ela ion be ween
he igge image and he a ge image ob ained in he same
hea phase and he desi ed posi ion o he b ea h cycle. La son
e al. [141] il e he b ea h cu e ob ained by S-G signals and
co ela ion using low-pass FIR il e and speci y a espi a o y
h eshold (see Figu e 30), which de ines he da a u he used
o econs uc he image a each phase o he hea cycle. This
p ocedu e inc eased he sha pness and con as o he image
compa ed o he ee-b ea hing image wi hou igge ing, bu
he di e ence in image quali y compa ed o he b ea h-hold
Fig. 30. Respi a o y Sel -Ga ing wa e o m wi h illus a ion o ga ing
p inciple.
Fig. 31. Example o he mic o-bend senso implemen a ion in o
ma [162].
echnique was no signi ican , simila ly o [142], [143]. Ne -
e heless, ca diac and espi a o y S-G signals co ela e s ongly
wi h ex e nally measu ed signals using ECG (R=1.00 ±0)
and espi a o y bel s (R=0.82 ±0.1) [143].
O he success ul applica ions o he e ospec i e espi a o y
S-G me hods we e ca ied ou in lung UTE (ul a-sho echo
ime) imaging in [165]–[168]]. The lung in es iga ion is pa ic-
ula ly challenging no only because o espi a o y and ca diac
mo ion limi ing he image quali y, bu also in insically low MR
signal caused by he low wa e concen a ion, mul iple ai ” issue
in e aces, o low p o on densi y o he lung pa enchyma. Since
na iga o sequence causes a concomi an sa u a ion o lung
pa s, S-G me hod ep esen s less complica ed way o ee-
b ea hing lung imaging [166], [169]
Real- ime S-G igge ing has been demons a ed in [144],
[145]. This app oach p omises o signi ican ly educe acqui-
si ion ime while main aining image quali y simila o he e -
ospec i e me hod bu equi es mo e sophis ica ed so wa e o
signal de ec ion and p ocessing.
E. Op ical Me hods
The ield o op ical senso s and ibe s also in es iga es he
espi a o y synch oniza ion o MR, simila o he ca diac (see
Sec ion II-F), when bo h i al signs a e o en combined o he
measu emen o he pa ien ’s s a us du ing MRI. Fo measu e-
men o b ea hing ac i i y by op ical senso s, esea che s use
in e e ome ic senso s [61]–[66], [146]–[149], FBG senso s
placed unde he pa ien ’s back [70]–[73], [75], as well as
mic o/mac o-bending ibe s [90], [93], [150]–[154], which a e
o en pa o so-called sma ex iles [155]–[161]. Example o
he mic o-bend senso implemen a ion in o ma is shown in Fig-
u e 31. The di e ence in measu emen o espi a ion compa ed
o ca diac ac i i y is he signal il e ing in a di e en equency

216 IEEE REVIEWS IN BIOMEDICAL ENGINEERING, VOL. 15, 2022
band o main ain only he desi ed b ea h cu e, i.e., wi hin he
ange o 0.05–2 Hz.
Yoo e al. [163] p esen ed an op ion o measu ing only
espi a o y ac i i y, which esponds o changes in ai empe -
a u e du ing inhala ion and exhala ion, acqui ed by a special
senso . The au ho s designed wo di e en ypes o op ical
senso s o moni o ing he espi a o y a e ” nasal and ab-
dominal. A senso a ached o he nasal ca i y measu es he
ai low using a he moch omic pigmen ha changes colo de-
pending on he empe a u e a ia ions wi hin he b ea h cycle
(inhala ion/exhala ion). The second ype o senso is loca ed
on he pa ien ’s abdomen and measu es b ea hing ac i i y as
he ci cum e ence a iance. The au ho s e i ied he use o he
p oposed senso s wi hou de e io a ing he MR image, which
p omises a sui abili y o he ibe -op ic espi a ion senso s o
espi a o y moni o ing du ing su gical p ocedu es pe o med
inside an MRI sys em.
Measu emen o espi a o y a e by FBG senso s in MR en i-
onmen waspe o medin[68],[74],[77],[78],[80],[81],[164].
All hese s udies eached he Bland-Al man de i ed accu acy o
he espi a o y a e g ea e han 95% ela ed o con en ional
espi a o y senso s, so he op ical me hod appea s o be a good
al e na i e o MRI espi a o y igge ing. In connec ion wi h
hea a e moni o ing, his echnique becomes e y in e es ing
o MRI examina ionssince onesenso wouldbeable omeasu e
he o e all i al signs o he pa ien , mo eo e con inuously, as
desc ibed in [83].
Fajkus e al. [132] es ed espi a o y igge ing a a 3 T
ield s eng h using an FBG senso applied o he nasal oxygen
capsule. They il e ed he ob ained signal in he band o 0.1 ”
0.5 Hz and compa ed he calcula ed espi a o y ac i i y wi h
he con en ional measu emen by espi a o y bel s, whe e he
measu emen by op ical senso eached an accu acy based on
he Bland-Al man analysis o o e 95%. Acco ding o expe s,
he images om bo h igge ing me hods we e diagnos ically
bene icial, bu highe con as and sha pness o he con ou s
we e achie ed when using he op ical senso . Also, acco ding o
quan i a i e image analysis, images igge ed by FBG achie ed
be e esul s han espi a o y e e ence.
Respi a o y igge ing using a mic o-bend senso was es ed
a a ield s eng h o 1.5 T in [162]. The op ical signal eached
a e y high co ela ion R =0.971 wi h measu emen s by s an-
da d espi a o y senso s. No mo emen a i ac s o blu y pa s
occu ed in he image. The di e ence be ween he op ical and
na iga o me hods in assessing he SNR o he kidney and
he CNR kidney-spleen was no signi ican , bu he na iga o
me hod gained a signi ican ly be e diagnos ic quali y o he
images.
IV. DISCUSSION
The e iew o CMRI synch oniza ion me hods shows ha
oday’s esea ch ocuses mo e on he no el ones han imp o ing
he echniques mos widely used in clinical p ac ice. I is due o
ela i ely complica ed pa ien p epa a ion in case o ECG/VCG
and especially due o he inc easing complica ions o high ield
s eng h examina ion. Also, ela i ely pa ien - iendly PPG is
no sui able because o mo ion a i ac suscep ibili y and long
physiological delay. The con empo a y esea ch ocuses on wo
majo miles ones: (1) ad ancemen o so wa e ha can achie e
adequa e accu acy in he de ec ion o synch oniza ion poin s in
he signal o supp ess magne ic ield a i ac s, and (2) de elop-
men o new me hods o measu ing i al signs ha could eplace
s anda d ECG measu emen s in u u e.
Thanks o he ex eme inc ease in compu ing pe o mance
in he ecen yea s, i is possible o signi ican ly imp o e he
signal/image quali y by using ad anced p ocessing echniques.
Howe e , elying only on so wa e algo i hms can esul in a
numbe o compu e e o s, which can lead o he dis o ion o
some image segmen s and hus, supp ess impo an diagnos ic
in o ma ion. The e o e, i is undoub edly impo an o pay a en-
ion o he basic en i y o he en i e measu ing chain (senso s),
which could p o ide a desi able accu acy al eady on he lowes
laye and hus, oge he wi h using so wa e algo i hms, e en
be e o e all esul s.
Cu en ly, acous ic and accele ome ic igge ing me hods a e
mos sui able o in oduce in o p ac ice, p o iding sa e y and
com o o he pa ien s and high esis ance o a i ac s caused by
magne ic ields; e iciency o acous ic senso s has al eady been
success ully es ed a 7 T. The only d awback is suscep ibili y
o he PCG and SCG o in e e ence caused by g adien coil
swi ching, and acco dingly, he delay in signal ansmission o
he MR uni due o il a ion. O he in es iga ions, which a e s ill
in hei in ancy, such as op ical o DUS measu emen s, show
high-quali y hea a e calcula ions, bu some limi a ions ha e
no ye p o en hei usabili y in medical p ac ice: he op ical
me hod equi es a e y expensi e e alua ion uni , which could
become mo e accessible o e ime, and inaccu acy in DUS
measu emen s o pa ien s wi h ca diac diso de s means a majo
complica ion in ca diac examina ions.
The esea ch me hods ha e se e al common disad an ages:
Non-s anda dized senso placemen in con as wi h ECG
o VCG. The senso loca ion highly depends on he spe-
ci ic pa ien and his body s uc u e; some signals show
e y di e en wa e o ms when placed a di e se poin s o
he ches (e.g. SCG). The e o e, esea ch should handle
he possibili ies o measu emen wi hou he complica ed
a achmen o he senso o he pa ien ’s body, such as
senso s implemen ed in ex iles o ma s.
Signal delay compa ed o ECG signal which is gi en
by se e al a ibu es: (1) a physiological delay be ween
he R wa e and he al e na i e igge poin ( i s hea
sound, J-wa e e c., see Figu e 32), and (2) o he delays
supe imposed caused by he ansmission medium, elec-
onics ensu ing signal s a e and con e sion o a digi al
signal, o il a ion used o emo e undesi able compo-
nen s. Such a delayed igge signal may lag behind he
ECG igge o hund eds o milliseconds, collec ing da a
om he w ong phase o he ca diac cycle. The e o e,
he image may lose i s diagnos ic alue, and he acqui-
si ion needs o be epea ed, ex ending he examina ion
ime. Thus, supp ession o he signal delay in some way,
such as so wa e cus omiza ion o igge ing, should be
used.
LADROVA e al.: MONITORING AND SYNCHRONIZATION OF CARDIAC AND RESPIRATORY TRACES 217
TABLE I
THE CLASSIFICATION OF THE PRESENTED SYNCHRONIZATION METHODS
Fig. 32. Physiological delay o he indi idual hea signals compa ed
o he ECG wi h ma ked peaks ( igge ).
The p esen ed S-G me hods could sol e he p oblems o sen-
so placemen and signal delay e sus ECG. Ano he ad an age
o S-G me hods is he possibili y o espi a o y and ca diac
ga ing al oge he . Al hough his double igge ing p olongs he
equi ed scan ime signi ican ly, he o al examina ion ime is
sho ened in coope a ing pa ien s. In case o pa ien s unable o
assis in b ea hing, espi a o y synch oniza ion is a necessi y
o a success ul examina ion. Because espi a o y compensa ion
using na iga ion echoes lags due o he e y small acquisi ion
window igge ed by only one na iga ion echo pe hea cycle,
a single b ea h-hold examina ion s ill p e ails, which can cause
p oblems o many pa ien s.
Howe e , sho comings o S-G echniques lie in he ealiza-
ion o synch oniza ion i sel . In gene al, hese me hods ha e
low ha dwa e equi emen s a he expense o he complexi y o
so wa e, which p ocesses MR da a o gene a e a igge signal.
The e o e, me hods o measu ing i al unc ions, which enable
o dis inguish a espi a o y unc ion om he hea signal by
app op ia e il a ion (BCG, SCG), become p e e ed. Consid-
e ing he una ailabili y o op ical me hods, acous ic and SCG
me hodsa e cu en lyconside ed asmos p omisingme hods o
combining ca diac and espi a o y synch oniza ion and should
emain o be he subjec o u he esea ch in his a ea.
Table I subjec i ely summa izes me hods o ca diac and
espi a o y MRI synch oniza ion wi h he classi ica ion o hei
accu acy and a i ac esis ance:
Accu acy e e s o he success o co ec de ec ion o
ca diac unc ion wi h sequence ini ia ion in he desi ed
phaseo hehea / espi a o ycycle,whichenhancesimage
quali y. I is e alua ed wi h espec o cases o some
abno mali ies (e.g., disease, a hy hmia) o echnical lim-
i a ions, no o he o igin o MR a i ac s. Accu acy is
classi ied as ollows:
High ”me hodp o ides e yaccu a ehea / espi a o y
phasede ec ionandisminimallya ec edbyabno mal-
i ies, pa ien ’s physiology o mo emen s.
Medium ” me hod p o ides accu acy sui able o MR
ga ing bu is limi ed by some o he abo e-men ioned
p oblems.
Low ” me hod is no ecommended o MR ga ing due
o la ge ange o limi a ions.
Resis ance o MR a i ac s indica es he esis ance o e -
ec s o he MR en i onmen on a gi en signal (in luence
o elec omagne ic ield, magne ohyd odynamic e ec o
acous ic dis u bance) and deg ee o independence om
magne ic ield s eng h, classi ied as:
High ” me hod is comple ely esis an o MR a i ac s
and independen om magne ic ield s eng h.
Medium ” me hod is a ec ed only by one o he abo e-
men ioned MR in luences.
Low ” me hod is suscep ible o all e ec s o he MR
en i onmen .
Complexi y includes compu a ional equi emen s, a ail-
abili y, and he p ice o senso s o measu ing uni . Com-
pu a ional cos is a c ucial ac o in implemen a ion o he
measu ing sys em since i de e mines a unc ion o he
sys em in eal- ime. P ice is also a key poin o me hod
218 IEEE REVIEWS IN BIOMEDICAL ENGINEERING, VOL. 15, 2022
easibili y o clinical p ac ice. A me hod eligible o
in oducingin op ac iceshould accomplishacomp omise
be ween i s complexi y and pe o mance/accu acy.
Speci ic limi a ions closely ela e o he p e ious pa ame-
e sbu includede ailedknowledgeo hedisad an ageso
each me hod, which mus be elimina ed o educed as bes
as possible in o de o in oduce he me hod in o p ac ice
success ully.
V. CONCLUSION
Synch oniza ion o he ca diac cycle and espi a o y excu -
sions is necessa y o ca dio ascula and abdominal MRI o
op imal ep esen a ionand imagequali y.Thepape summa izes
echniquescu en ly a ailable and/o unde de elopmen ha o-
cus on inc easing p oblems wi h measu emen in high magne ic
ields, nowadays eaching up o 7 T. Wi h an ex eme inc ease in
compu ing pe o mance in ecen yea s, so wa e-based me h-
ods, such as S-G and na iga o s, a e he cen e o a en ion, and
hey unde wen a huge expansion. Howe e , hey s ill ha e some
disad an ages (compu e e o s, dis u bing s eady-s a e), which
could be sol ed using he ha dwa e-assis ed me hods based on
he non-elec ical measu emen s. F om such hese echniques,
acous ic and accele ome ic ones a e he mos p omising no
only due o he high accu acy bu also hanks o hei low p ice
and easy implemen a ion. Mo eo e , hey would p o ide he
possibili yo moni o ing he espi a o ysignal oge he wi hca -
diac unc ion and hus, would enable a “double” MR igge ing
app oach, impo an in ca dio ascula examina ions, especially
in non-coope a ing pa ien s. Wi h his ounda ion in place, he
u u e o MR synch oniza ion is encou aging; howe e , mo e
wo k is needed in o de o maximize he echnology a ailable
and de elop new app oaches no ye imagined.
REFERENCES
[1] N. C. Gou soyiannis, Ed., Clinical MRI o he Abdomen. Be lin, Heidel-
be g: Sp inge Be lin Heidelbe g, 2011.
[2] A.Va gheseandD.J.Pennell,Eds.,Ca dio ascula Magne ic Resonance
Made Easy. Edinbu gh, U.K.: Chu chill Li ings one, 2008.
[3] A. S adle , W. Schima, A. Ba-Ssalamah, J. Ke enbach, and E. Eisen-
hube , “A i ac s in body MR imaging: Thei appea ance and how o
elimina e hem,” Eu . Radiol., ol. 17, no. 5, pp. 1242–1255, Ap . 2007.
[4] R. C. Semelka, M. A. B own, E. Al un, and R. C. Semelka, Abdominal-
Pel ic MRI, 4 h ed. Hoboken, NJ, USA: Wiley, 2016.
[5] G. Cons an ine, K. Shan, S. D. Flamm, and M. U. Si anan han, “Role o
MRI in clinical ca diology,” Lance , ol. 363, no. 9427, pp. 2162–2171,
2004.
[6] T. Sa o e al., “Valida ion s udy on he accu acy o echoca diog aphic
measu emen s o igh en icula sys olic unc ion in pulmona y hype -
ension,” J. Ame . Soc. Echoca diog ., ol. 25, no. 3, pp. 280–286, 2012.
[7] P.C.Yange al., “New eal- ime in e ac i e ca diac magne ic esonance
imaging sys em complemen s echoca diog aphy,” J. Ame . College Ca -
diol., ol. 32, no. 7, pp. 2049–2056, 1998.
[8] T. Niendo , D. K. Sodickson, G. A. K ombach, and J. Schulz-Menge ,
“Towa d ca dio ascula MRI a 7 T: Clinical needs, echnical solu ions
and esea ch p omises,” Eu . Radiol., ol. 20, no. 12, pp. 2806–2816,
2010.
[9] G. Pons-Llado and F. Ca e as, A las o P ac ical Applica ions o Ca dio-
ascula Magne ic Resonance. New Yo k, NY, USA: Sp inge Science
Business Media, 2005.
[10] R. R. Edelman, H. P. Ma le, D. J. A kinson, and H. M. Hoogewoud, “MR
angiog aphy,” Ame . J. Roen genol., ol. 154, no. 5, pp. 937–946, 1990.
[11] V. M. Fe ei a, S. K. Piechnik, M. D. Robson, S. Neubaue , and T. D.
Ka ami sos, “Myoca dial issue cha ac e iza ion by magne ic esonance
imaging,” J. Tho acic Imag., ol. 29, no. 3, pp. 147–154, 2014.
[12] M. Je osch-He old, O. Muehling, and N. Wilke, “MRI o myoca dial
pe usion,” Semina s Ul asound, CT MRI, ol. 27, no. 1, pp. 2–10, 2006.
[13] M. Schä , S. Koze ke, S. E. Fische , and P. Boesige , “Ca diac SSFP
imaging a 3 esla,” Magn. Reson. Med., ol. 51, no. 4, pp. 799–806,
2004.
[14] H. J. Michaely e al., “Analysis o ca diac unc ion compa ison be ween
1.5 Tesla and 3.0 Tesla ca diac cine magne ic esonance imaging,” In es .
Radiol., ol. 41, no. 2, pp. 133–140, 2006.
[15] R. Neza a , M. S ube , R. Ouwe ke k, A. M. Gha ib, M. Y. Desai, and
R. I. Pe ig ew, “B1-insensi i eT2 p epa a ion o imp o ed co ona y
magne ic esonance angiog aphy a 3 T,” Magn. Reson. Med., ol. 55,
no. 4, pp. 858–864, 2006.
[16] A. N. P ies , P. M. Bansmann, M. G. Kaul, A. S o k, and G. Adam,
“Magne ic esonance imaging o he co ona y essel wall a 3 T. using an
obliquely o ien ed ein e sion slab wi h adiaba ic pulses,” Magn. Reson.
Med., ol. 54, no. 5, pp. 1115–1122, 2005.
[17] M. E. Ladd e al., “P os and cons o ul a-high- ield MRI/MRS o human
applica ion,” P og. Nucl. Magn. Reson. Spec osc., ol. 109, no. 14,
pp. 1–50, 2018.
[18] A. M. Gha ib, A. Elagha, and R. I. Pe ig ew, “Ca diac magne ic es-
onance a high ield: P omises and p oblems,” Cu . P oblems Diagn.
Radiol., ol. 37, no. 2, pp. 49–56, 2008.
[19] F. on Knobelsdo -B enkenho e al., “Ca diac chambe quan i ica ion
using magne ic esonance imaging a 7 esla-a pilo s udy,” Eu . Radiol.,
ol. 20, no. 12, pp. 2844–2852, 2010.
[20] A. D. Els e , “Ques ions and answe s in MRI,” 2021. [Online]. A ailable:
h ps://m iques ions.com/mo ion- ela ed-a i ac s.h ml
[21] M. S. Naci , A. Za odni, N. Kawel, E.-Y. Choi, J. A. C. Lima, and D. A.
Bluemke, “Ca diac magne ic esonance imaging and i s elec oca dio-
g aphs (ECG): Tips and icks,” In . J. Ca dio asc. Imag., ol. 28, no. 6,
pp. 1465–1475, 2012.
[22] R. Y. Kwong, M. Je osch-He old, and B. Heyda i, Eds., Ca dio ascula
Magne ic Resonance Imaging, 2nd ed., New Yo k, NY, USA: Sp inge
New Yo k, 2019.
[23] R. B. an Heeswijk, G. Bonanno, S. Coppo, A. Co is ine, T. Kobe ,
and M. S ube , “Mo ion compensa ion s a egies in magne ic esonance
imaging,” C i . Re . Biomed. Eng., ol. 40, no. 2, pp. 99–119, 2012.
[24] R. Geo ge, J. Dela C uz, R. Singh, and R. Ilango an, “MRI Mas e ” 2021.
[Online]. A ailable: h ps://m imas e .com/index-3.h ml
[25] G. M. Pohos and K. S. Nayak, Handbook o Ca dio ascula Magne ic
Resonance Imaging, 2nd ed., New Yo k, NY, USA: In o ma Heal hca e,
c2007.
[26] S. E. Fische , S. A. Wickline, and C. H. Lo enz, “No el eal- ime -wa e
de ec ion algo i hm based on he ec o ca diog am o accu a e ga ed
magne ic esonance acquisi ions,” Magn. Reson. Med.: O icial J. In .
Soc. Magn. Reson. Med., ol. 42, no. 2, pp. 361–370, 1999.
[27] T. F auen a h e al., “Acous ic ca diac igge ing: A p ac ical solu ion
o synch oniza ion and ga ing o ca dio ascula magne ic esonance a
7 esla,” J. Ca dio asc. Magn. Reson., ol. 12, no. 1, pp. 1–14, 2010.
[28] M. Becke e al., “Compa ison o le en icula unc ion assess-
men using phonoca diog am- and elec oca diog am- igge ed 2D SSFP
CINE MR imaging a 1.5 T. and 3.0 T,” Eu . Radiol., ol. 20, no. 6,
pp. 1344–1355, 2010.
[29] H. Kugel e al., “Haza dous si ua ion in he MR bo e: Induc ion in ECG
leads causes i e,” Eu . Radiol., ol. 13, no. 4, pp. 690–694, 2003.
[30] S. A. Lange and Q. N. Nguyen, “Cables and elec odes can bu n pa ien s
du ing MRI,” Nu sing 2006, ol. 36, no. 11, pp. 11–18, 2006.
[31] T. F auen a h, T. Niendo , and M. Kob, “Acous ic me hod o synch o-
niza ion o magne ic esonance imaging (MRI),” Ac a Acus ica Uni ed
Acus ica, ol. 94, no. 1, pp. 148–155, 2008.
[32] The Mammendo e Ins i u Fü Physik Und Medizin, “MIPM: MRI-
compa ible p oduc s,” 2021. [Online]. A ailable: h ps://www.mipm.
com/en/ou -p oduc s/accesso ies
[33] “MiRTLE Medical, LLC,” MiRTLE,” 2016-2017. [Online]. A ailable:
h p://www.mi lemed.com/p oduc .h ml
[34] BIOPAC Sys ems Inc., “BIOPAC Sys ems Inc.,” 2021. [Online]. A ail-
able: h ps://www.biopac.com/p oduc /
[35] J. W. K ug, G. Rose, D. S uch , G. Cli o d, and J. Os e , “Limi a ions
o VCG based ga ing me hods in ul a high ield ca diac MRI,” J.
Ca dio asc. Magn. Reson., ol. 15, no. S1, pp. 1–2, 2013.
[36] C. J. Snyde e al., “Ini ial esul s o ca diac imaging a 7 esla,” Magn.
Reson. Med., ol. 61, no. 3, pp. 517–524, 2009.
[37] D. Abi-Abdallah, V. Robin, A. D ochon, and O. Fokapu, “Al e a ions
in human ECG due o he Magne oHyd oDynamic e ec : A me hod o
accu a eR peak de ec ion in hep esenceo high MHD a i ac s,”inP oc.
29 h Annu. In . Con . IEEE Eng. Med. Biol. Soc., 2007, pp. 1842–1845.
LADROVA e al.: MONITORING AND SYNCHRONIZATION OF CARDIAC AND RESPIRATORY TRACES 219
[38] B. Sie e s, M. Wiesne , N. Ki ia, U. Speise , S. Schoen, and R. H.
S asse , “In luence o he igge echnique on en icula unc ion
measu emen s using 3-Tesla magne ic esonance imaging: Compa ison
o ECG e sus pulse wa e igge ing,” Ac a Radiologica, ol. 52, no. 4,
pp. 385–392, 2011.
[39] T. F auen a h, S. Koze ke, P. Boesige , and T. Niendo , “Ca diac ga ing
ee o in e e ence wi h elec o-magne ic ields a 1.5T., 3.0T. and 7.0T,”
in P oc. Annu. Mee ing In . Soc. Magn. Reson. Med., 2008, p. 207.
[40] F. Odille, C. Pasquie , R. Abache li, P.-A. Vuissoz, G. P. Zien a a, and
J.Felblinge , “Noise cancella ion signal p ocessing me hod and compu e
sys em o imp o ed eal- ime elec oca diog am a i ac co ec ion du -
ing MRI da a acquisi ion,” IEEE T ans. Biomed. Eng., ol. 54, no. 4,
pp. 630–640, Ap . 2007.
[41] J. Os e , O. Pie quin, R. Abache li, M. K aeme , and J. Felblinge , “Inde-
penden componen analysis-based a e ac educ ion: Applica ion o he
elec oca diog am o imp o edmagne ic esonanceimaging igge ing,”
Physiol. Meas., ol. 30, no. 12, pp. 1381–1397, 2009.
[42] J. W. K ug, G. Rose, G. D. Cli o d, and J. Os e , “ECG-based ga ing in
ul a high ield ca dio ascula magne ic esonance using an independen
componen analysis app oach,” J. Ca dio asc. Magn. Reson., ol. 15,
no. 1, pp. 1–13, 2013.
[43] J. Os e , O. Pie quin, M. K aeme , and J. Felblinge , “Nonlinea Bayesian
il e ing o denoising o elec oca diog ams acqui ed in a magne ic
esonance en i onmen ,” IEEE T ans. Biomed. Eng., ol. 57, no. 7,
pp. 1628–1638, Jul. 2010.
[44] D.Abi-Abdallah,E.Chau e ,L.Bouche -Fak i,A.Ba ailla d,A.B igue ,
and O. Fokapu, “Re e ence signal ex ac ion om co up ed ECG using
wa ele decomposi ion o MRIsequence igge ing:Applica ion osmall
animals,” Biomed. Eng. Online, ol. 5, no. 1, pp. 1–12, 2006.
[45] D. Abi-Abdallah, A. D ochon, V. Robin, and O. Fokapu, “Ca diac and
espi a o y MRI ga ing using combined wa ele sub-band decomposi ion
and adap i e il e ing,” Ann. Biomed. Eng., ol. 35, no. 5, pp. 733–743,
2007.
[46] M. Schmid , J. W. K ug, A. Gie s o e , and G. Rose, “A eal- ime QRS
de ec o based on highe -o de s a is ics o ECG ga ed ca diac,” Compu .
Ca diol. 2014, ol. 2014, no. 41, pp. 733–736, 2014.
[47] D. S äb, J. Roessle , K. O’B ien, C. Hamil on-C aig, and M. Ba h, “ECG
igge ing ul a-high ield ca dio ascula MRI,” Tomog aphy, ol.2,
no. 3, pp. 167–174, 2016.
[48] L. Edenb and , A. Hous on, and P. W. Mac a lane, “Vec o ca diog ams
syn hesized om 12-lead ECGs: A. new me hod applied in 1792 heal hy
child en,” Pedia . Ca diol., ol. 15, no. 1, pp. 21–26, 1994.
[49] M. Lingman e al., “T ansien epola iza ion al e a ions domina e he
ini ial phase o an acu e an e io in a c ion - a ec o ca diog aphy s udy,”
J. Elec oca diol., ol. 47, no. 4, pp. 478–485, 2014.
[50] M. Sede holm, “The o igin o moni o ing o acu e myoca dial in a c ion
wi h con inuous ec o ca diog aphy,” J. Elec oca diol., ol. 47, no. 4,
pp. 418–424, 2014.
[51] D.Co ez,J. M. Bos, and M. J. Acke man,“Vec o ca diog aphy iden i ies
pa ien s wi h elec oca diog aphically concealed long QT synd ome,”
Hea Rhy hm, ol. 14, no. 6, pp. 894–899, 2017.
[52] J. M. Chia, S. E. Fische , S. A. Wickline, and C. H. Lo enz, “Pe o mance
o QRS de ec ion o ca diac magne ic esonance imaging wi h a no el
ec o ca diog aphic igge ing me hod,” J. Magn. Reson. Imag., ol. 12,
no. 5, pp. 678–688, 2000.
[53] N. Spiche , M. Kukuk, S. Made wald, and M. E. Ladd, “Ini ial e alua ion
o p ospec i e ca diac igge ing using pho ople hysmog aphy signals
eco ded wi h a ideo came a compa ed o pulse oxime y and elec o-
ca diog aphy a 7T. MRI,” Biomed. Eng. Online, ol. 15, no. 1, pp. 1–28,
2016.
[54] N. Spiche , S. Made wald, M. E. Ladd, and M. Kukuk, “Hea a e moni-
o ing in ul a-high- ield MRI using equency in o ma ion ob ained om
ideo signals o he human skin compa ed o elec oca diog aphy and
pulse oxime y,” Cu . Di ec ions Biomed. Eng., ol. 1, no. 1, pp. 69–72,
Sep. 2015.
[55] J. Hodel e al., “In ac anial ce eb ospinal luid spaces imaging using
a pulse- igge ed h ee-dimensional u bo spin echo MR sequence wi h
a iable lip-angle dis ibu ion,” Eu . Radiol., ol. 21, no. 2, pp. 402–410,
2011.
[56] J. W. Be hold, “His o ical e iew o mic obend ibe -op ic senso s,” J.
Ligh w. Technol., ol. 13, no. 7, pp. 1193–1199, Jul. 1995.
[57] Y. Wang e al., “In ensi y measu emen bend senso s based on pe iod-
ically ape ed so glass ibe s,” Op . Le ., ol. 36, no. 4, pp. 558–560,
2011.
[58] P. Sama ki and S. Pull eap, “Fibe op ic senso applica ions o i al
signs moni o ing: A e iew,” in P oc. 7 h In . Elec . Eng. Cong ., 2019,
pp. 1–4.
[59] L.-G. Du and and P. Piba o , “Digi al signal p ocessing o he phonoca -
diog am: Re iew o he mos ecen ad ancemen s,” C i . Re . Biomed.
Eng., ol. 23, no. 3/4, pp. 163–219, 1995.
[60] C. Will, K. Shi, F. Lu z, R. Weigel, and A. Koelpin, “In elligen signal
p ocessing ou ine o ins an aneous hea a e de ec ion using a six-po
mic owa e in e e ome e ,” in P oc. In . Symp. In ell. Signal P ocess.
Commun. Sys ., 2015, pp. 483–487.
[61] P. Podb eznik, D. Donlagic, D. Lesnik, B. Cigale, and D. Zazula, “Cos -
e icien speckle in e e ome y wi h plas ic op ical ibe o unob usi e
moni o ingo human i alsigns,”J. Biomed. Op ., ol.18,no.10,pp. 1–8,
Oc . 2013.
[62] S. Sp age , D. Donlagic, and D. Zazula, “Moni o ing o basic human i al
unc ions using op ical in e e ome e ,” in P oc. IEEE 10 h In . Con .
Signal P ocess., 2010, pp. 1–4.
[63] S. Sp age , D. Donlagic, and D. Zazula, “Es ima ion o hea a e, espi-
a o y a e and mo ion by using op ical in e e ome e as body senso ,”
in P oc. Signal Image P ocess., 2011, pp. 280–287.
[64] S. Šp age , A. Holoba , and D. Zazula, “Feasibili y s udy o hea bea
de ec ion om op ical in e e ome ic signal by using con olu ion ke nel
compensa ion,” in P oc. In . Con . Bio-Inspi ed Sys . Signal P ocess.,
2013, pp. 396–400.
[65] I. Cikajlo, C. S. Š. T. D. Z. E ja eca, and D. Šp age , “Ca diac a hy hmia
ala m om op ical in e e ome ic signals du ing es ing o sleeping o
ea ly in e en ion,” Biocybe n. Biomed. Eng., ol. 36, no. 1, pp. 267–275,
2016.
[66] D. Zazula and S. Sp age , “De ec ion o he i s hea sound using ib e-
op ic in e e ome ic measu emen s and neu al ne wo ks,“ in P oc. 11 h
Symp. Neu al Ne w. Appl. Elec . Eng., 2012, pp. 171–176.
[67] Y.-H. Hsieh and N.-K. Chen, “Mic o ape ed Mach–Zehnde ibe in e -
e ome e o moni o ing p essu e luc ua ion and i s applica ions in pulse
a e de ec ion,” in P oc. 6 h IEEE/In . Con . Ad . In ocomm Technol.,
2013, pp. 113–115.
[68] J. Nedoma e al., “Magne ic esonance imaging compa ible non-in asi e
ib e-op ic senso s based on he b agg g a ings and in e e ome e s
in he applica ion o moni o ing hea and espi a ion a e o he hu-
man body: A compa a i e s udy,” Senso s, ol. 18, no. 11, pp. 1–30,
2018.
[69] P. Ro iz, L. Ca alho, O. F azão, J. San os, and J. Simões, “F om con-
en ional senso s o ib e op ic senso s o s ain and o cemeasu emen s
in biomechanics applica ions: A e iew,” J. Biomech., ol. 47, no. 6,
pp. 1251–1261, 2014.
[70] Ł. Dziuda, “Fibe -op ic senso s o moni o ing pa ien physiological
pa ame e s: A e iew o applicable echnologies and ele ance o use
du ingmagne ic esonanceimagingp ocedu es,” J. Biomed. Op ., ol.20,
no. 1, pp. 1–23, 2015.
[71] K. Che hana, A. S. G. P asad, S. N. Omka , and S. Asokan, “Fibe b agg
g a ing senso based de ice o simul aneous measu emen o espi a-
o y and ca diac ac i i ies,” J. Biopho on., ol. 10, no. 2, pp. 278–285,
2017.
[72] L. Dziuda, F. W. Skibniewski, M. K ej, and J. Lewandowski, “Moni-
o ing espi a ion and ca diac ac i i y using ibe b agg g a ing-based
senso ,” IEEE T ans. Biomed. Eng., ol. 59, no. 7, pp. 1934–1942,
Jul. 2012.
[73] M. Fajkus, J. Nedoma, R. Ma inek, V. Vasinek, H. Naze an, and P. Siska,
“A non-in asi e mul ichannel hyb id ibe -op ic senso sys em o i al
sign moni o ing,” Senso s, ol. 17, no. 12, pp. 1–17, 2017.
[74] Ł. Dziuda, F. W. Skibniewski, M. K ej, and P. M. Ba an, “Fibe B agg
g a ing-based senso o moni o ing espi a ion and hea ac i i y du ing
magne ic esonance imaging examina ions,” J. Biomed. Op ., ol. 18,
no. 5, pp. 278–285, 2013.
[75] D. Lo P es i e al., “Wea able sys em based on lexible FBG o es-
pi a o y and ca diac moni o ing,” IEEE Senso s J., ol. 19, no. 17,
pp. 7391–7398, Sep. 2019.
[76] Y. Haseda, J. Bone acino, H. -Y. Tam, S. Chino, S. Koyama, and H.
Ishizawa, “Measu emen o pulse wa e signals and blood p essu e by
a plas ic op ical ibe FBG senso ,” Senso s, ol. 19, no. 23, pp. 1–11,
2019.
[77] L. Dziuda, M. K ej, and F. W. Skibniewski, “Fibe B agg g a ing s ain
senso inco po a ed o moni o pa ien i al signs du ing MRI,” IEEE
Senso s J., ol. 13, no. 12, pp. 4986–4991, Dec. 2013.
[78] J. Nedoma e al., “Valida ion o a no el ibe -op ic senso sys-
em o moni o ing ca dio espi a o y ac i i ies du ing MRI exami-
na ions,” Ad . Elec . Elec on. Eng., ol. 15, no. 3, pp. 536–543,
2017.
[79] Ł. Dziuda and F. W. Skibniewski, “A new app oach o ballis oca dio-
g aphic measu emen s using ib e b agg g a ing-based senso s,” Biocy-
be n. Biomed. Eng., ol. 34, no. 2, pp. 101–116, 2014.
220 IEEE REVIEWS IN BIOMEDICAL ENGINEERING, VOL. 15, 2022
[80] L.Dziuda,J.Lewandowski, F.Skibniewski,andG.Nowicki,“Fib e-op ic
senso o espi a ion and hea a e moni o ing in he MRI en i onmen ,”
P ocedia Eng., ol. 47, pp. 1291–1294, 2012.
[81] J. Nedoma, M. Fajkus, R. Ma inek, and H. Naze an, “Vi al sign mon-
i o ing and ca diac igge ing a 1.5 esla: A p ac ical solu ion by an
MR-Ballis oca diog aphy ibe -op ic senso ,” Senso s, ol. 19, no. 3,
pp. 1–22, 2019.
[82] A. C. S. B au, C. T. Wheele , L. W. Hedlund, and G. A. Johnson, “Fibe -
op ic s e hoscope: A ca diac moni o ing and ga ing sys em o magne ic
esonance mic oscopy,” Magn. Reson. Med., ol. 47, no. 2, pp. 314–321,
2002.
[83] A. Rengle, L. Baboi, H. Sain -Jalmes, R. Sablong, and O. Beu , “Op ical
ca diac and espi a o y de ice o synch onized MRI on small animal,” in
P oc. 29 h Annu. In . Con . IEEE Eng. Med. Biol. Soc., 2007, pp. 2046–
2049.
[84] J. Nedoma e al., “A no el FBG-Based igge ing sys em o ca diac
MR imaging a 3 esla: A pilo p e-clinical s udy,” IEEE Access, ol.8,
pp. 181205–181223, 2020.
[85] J. De Jonckhee e e al., “OFSETH: Op ical ib e embedded in o echnical
ex ile o heal hca e, an e icien way o moni o pa ien unde magne ic
esonance imaging,” in P oc. 29 h Annu. In . Con . IEEE Eng. Med. Biol.
Soc., 2007, pp. 3950–3953.
[86] Z. Chen, J. T. Teo, S. H. Ng, and X. Yang, “Po able ibe op ic ballis-
oca diog am senso o home use,” P oc. SPIE, ol. 8218, 2012, A .
no. 82180X.
[87] Z. Chen, I. Ganno , J. T. Teo, S. H. Ng, and H. Yim, “Sma pillow o
hea - a e moni o ing using a ibe op ic senso ,” in P oc. SPIE, ol.7894,
2011, A . no. 789402.
[88] K. K ebbe e al., “Sma echnical ex iles based on ib e op ic senso s,”
P oc. SPIE, ol. 8421, 2012, A . no. 84212.
[89] A. G ille e al., “Op ical ibe senso s embedded in o medical ex iles o
heal hca e moni o ing,” IEEE Senso s J., ol. 8, no. 7, pp. 1215–1222,
Jul. 2008.
[90] X. Yang e al., “Tex ile ibe op ic mic obend senso used o hea bea
and espi a ion moni o ing,” IEEE Senso s J., ol. 15, no. 2, pp. 757–761,
Feb. 2015.
[91] D. Lo P es i e al., “Ca dio- espi a o y moni o ing in a che y using a
sma ex ile based on lexible ibe B agg g a ing senso s,” Senso s,
ol. 19, no. 16, pp. 1–14, 2019.
[92] D. L. P es i e al., “Ca diac moni o ing wi h a sma ex ile based on
polyme -encapsula ed FBG: In luence o senso posi ioning,” in P oc.
IEEE In . Symp. Med. Meas. Appl., 2019, pp. 1–6.
[93] Z. Chen, D. Lau, J. T. Teo, S. H. Ng, X. Yang, and P. L. Kei, “Simul a-
neous measu emen o b ea hing a e and hea a e using a mic obend
mul imode ibe op ic senso ,” J. Biomed. Op ., ol. 19, no. 5, pp. 1–11,
2014.
[94] K. Nassens ein e al., “Ca diac MRI: E alua ion o phonoca diog am-
ga ed cine imaging o he assessmen o global und egional le en ic-
ula unc ion in clinical ou ine,” Eu . Radiol., ol.22,no.3, pp. 559–568,
2012.
[95] S. Made wald e al., “7 esla ca diac imaging wi h a phonoca diog am
igge de ice,” in P oc. In . Soc. Magn. Reso. Med., 2011, p. 1322.
[96] MRI.TOOLS GmbH, “MRI.TOOLS – Inno a ion o medical imaging,”
2021. [Online]. A ailable: h ps://www.m i ools.de/index.php/p oduc s/
m i-accesso ies/easyac /
[97] A. C. La son, R. D. Whi e, G. Laub, E. R. McVeigh, D. Li, and O. P.
Simone i, “Sel -ga ed ca diac cine MRI,” Magn. Reson. Med., ol. 51,
no. 1, pp. 93–102, 2004.
[98] B. Hiba, N. Richa d, M. Janie , and P. C oisille, “Ca diac and espi a o y
double sel -ga ed cine MRI in he mouse a 7T,” Magn. Reson. Med.,
ol. 55, no. 3, pp. 506–513, 2006.
[99] G. M. Nijm, A. V. Sahakian, S. Swi yn, J. C. Ca , J. J. Sheehan, and
A. C. La son, “Compa ison o sel -ga ed cine MRI e ospec i e ca diac
synch oniza ion algo i hms,” J. Magn. Reson. Imag., ol. 28, no. 3,
pp. 767–772, 2008.
[100] M. E. C owe e al., “Au oma ed ec ilinea sel -ga ed ca diac cine imag-
ing,” Magn. Reson. Med., ol. 52, no. 4, pp. 782–788, 2004.
[101] T. A. Sp aggins, “Wi eless e ospec i e ga ing: Applica ion o cine
ca diac imaging,” Magn. Reson. Imag., ol. 8, no. 6, pp. 675–681,
1990.
[102] R. D. Whi e, C. B. Paschal, M. E. Clampi , T. A. Sp aggins, and
G. W. Lenz, “Elec oca diog aph-independen , ‘wi eless’ ca dio ascula
cine MR imaging,” J. Magn. Reson. Imag., ol. 1, no. 3, pp. 347–355,
1991.
[103] N. Gai and L. Axel, “Co ec ion o mo ion a i ac s in linog am and p o-
jec ion econs uc ion MRI using geome y and consis ency cons ain s,”
Med. Phys., ol. 23, no. 2, pp. 251–262, 1996.
[104] C.J.Ha dy,L.Zhao,X.Zong,M.Sa ana han,andE.K. Yucel,“Co ona y
MR angiog aphy: Respi a o y mo ion co ec ion wi h BACSPIN,” J.
Magn. Reson. Imag., ol. 17, no. 2, pp. 170–176, 2003.
[105] M. Kachel ieß, D.-A. Senns , W. Maxlmose , and W. A. Kalende ,
“Kymog am de ec ion and kymog am-co ela ed image econs uc ion
om subsecond spi al compu ed omog aphy scans o he hea ,” Med.
Phys., ol. 29, no. 7, pp. 1489–1503, 2002.
[106] J. M. Rubin, J. B. Fowlkes, M. R. P ince, R. T. Rhee, and T. L. Chene e ,
“Dopple USga ingo ca diacMRimaging,”Acad. Radiol., ol.7,no.12,
pp. 1116–1122, 2000.
[107] F. Ko ding e al., “Dopple ul asound compa ed wi h elec oca diog am
andpulseoxime yca diac igge ing:A pilo s udy,”Magn. Reson. Med.,
ol. 74, no. 5, pp. 1257–1265, 2015.
[108] F. Ko ding e al., “Dopple ul asound igge ing o ca dio ascula MRI
a 3T in a heal hy olun ee s udy,” Magn. Reson. Med. Sci., ol. 16, no. 2,
pp. 98–108, 2017.
[109] Ul asound Supply, “Ul asound Supply - Ul asounds, C-A ms,
P obes,” 2021. [Online]. A ailable: h ps://www.ul asoundsupply.
com/p oduc s/ul asound-p obes/ge-p obes/pa6-8-ca diac-sec o -
p obe/
[110] A.Taebi,B.Sola ,A.Boma ,R.Sandle ,andH.Mansy,“Recen ad ances
in seismoca diog aphy,” Vib a ion, ol. 2, no. 1, pp. 64–86, 2019.
[111] O. T. Inan e al., “Ballis oca diog aphy and seismoca diog aphy: A
e iew o ecen ad ances,” IEEE J. Biomed. Heal h In o ma ., ol. 19,
no. 4, pp. 1414–1427, Jul. 2015.
[112] G. Cosoli, L. Casacandi ella, E. P. Tomasini, and L. Scalise, “Hea a e
assessmen by means o a no el app oach applied o signals o di e en
na u e,” J. Phys.: Con . Se ., ol. 778, no. 1, pp. 1–8, 2017.
[113] F. Land eani e al., “Bea - o-bea hea a e de ec ion by sma phone’s
accele ome e s: Valida ion wi h ECG,” in P oc. 38 h Annu. In . Con .
IEEE Eng. Med. Biol. Soc., 2016, pp. 525–528.
[114] M. J. Tadi e al., “A eal- ime app oach o hea a e moni o ing using a
hilbe ans o m in seismoca diog ams,” Physiol. Meas., ol. 37, no. 11,
pp. 1885–1909, 2016.
[115] A. Taebi, A. J. Boma , R. H. Sandle , and H. A. Mansy, “Hea a e mon-
i o ing du ing di e en lung olume phases using seismoca diog aphy,”
in P oc. Sou heas Con, 2018, pp. 1–5.
[116] M. J. Tadi, E. Leh onen, O. Lahdenoja, M. Pankaala, and T. Koi is o,
“An adap i e app oach o hea bea de ec ion based on S- ans o m in
seismoca diog ams,” in P oc. 38 h Annu. In . Con . IEEE Eng. Med. Biol.
Soc., 2016, pp. 2370–2373.
[117] J. Wahls om e al., “A hidden ma ko model o seismoca diog aphy,”
IEEE T ans. Biomed. Eng., ol. 64, no. 10, pp. 2361–2372, Oc . 2017.
[118] J. Yao, S. T idandapani, W. F. Au e mann, C. A. Wick, and P. T. Bha i,
“An adap i e seismoca diog aphy (SCG)-ECG mul imodal amewo k
o ca diac ga ing using a i icial neu al ne wo ks,” IEEE J. T ans. Eng.
Heal h Med., ol. 6, no. 1, pp. 1–11, Oc . 2018.
[119] M. J. Tadi, T. Koi is o, M. Pänkäälä, and A. Paasio, “Accele ome e -
based me hod o ex ac ing espi a o y and ca diac ga ing in o ma ion
o dual ga ing du ing nuclea medicine imaging,” In . J. Biomed. Imag.,
ol. 2014, no. 1, pp. 1–11, 2014.
[120] M. Je osch-He old e al., “The seismoca diog am as magne ic- ield-
compa ible al e na i e o he elec oca diog am o ca diac s ess moni-
o ing,” In . J. Ca diac Imag., ol. 15, no. 6, pp. 523–531, 1999.
[121] R. Ma inek e al., “A low-cos sys em o seismoca diog aphy-based
ca diac igge ing: A p ac ical solu ion o ca dio ascula magne ic es-
onance imaging a 3 esla,” IEEE Access, ol. 7, pp. 118608–118629,
2019.
[122] SIEMENS AG, “SyngoMRE11 ope a o 1710 manual - body,” [Online].
A ailable: h ps://cbbi.udel.edu/wp-con en /uploads/2017/01/Body.pd
[123] M. Zai se , J. Macla en, and M. He bs , “Mo ion a i ac s in MRI: A
complex p oblem wi h many pa ial solu ions,” J. Magn. Reson. Imag.,
ol. 42, no. 4, pp. 887–901, 2015.
[124] J. P. Ea ls, V. B. Ho, T. K. Foo, E. Cas illo, and S. D. Flamm, “Ca diac
MRI: Recen p og ess and con inued challenges,” J. Magn. Reson. Imag.,
ol. 16, no. 2, pp. 111–127, 2002.
[125] A. D. Sco , J. Keegan, and D. N. Fi min, “Mo ion in ca dio ascula MR
imaging,” Radiology, ol. 250, no. 2, pp. 331–351, 2009.
[126] P. G. Danias and W. J. Manning, “MR na iga o s and hei use in ca diac
and co ona y imaging,” in Co ona y Magne ic Resonance Angiog aphy,
1s ed., New Yo k, NY, USA: Sp inge -Ve lag, 2002, pp. 219–227.

LADROVA e al.: MONITORING AND SYNCHRONIZATION OF CARDIAC AND RESPIRATORY TRACES 221
[127] S. K. Lemieux and G. H. Glo e , “An in a ed de ice o moni o ing
he espi a ion o small oden s du ing magne ic esonance imaging,” J.
Magn. Reson. Imag., ol. 6, no. 3, pp. 561–564, 1996.
[128] T. Noponen e al., “Spi ome y based espi a o y ga ing me hod o
ca diac PET and MRI imaging,” in P oc. IEEE Nucl. Sci. Symp. Con .
Rec., 2008, pp. 4832–4834.
[129] S. S. Vasanawala and E. Jackson, “A me hod o apid obus es-
pi a o y synch oniza ion o MRI,” Pedia . Radiol., ol. 40, no. 10,
pp. 1690–1692, 2010.
[130] D. A. Feinbe g e al., “Hyb id ul asound MRI o imp o ed ca diac
imaging and eal- ime espi a ion con ol,” Magn. Reson. Med., ol. 63,
no. 2, pp. 290–296, 2010.
[131] C.San ellie al.,“Respi a o ybellows e isi ed o mo ioncompensa ion:
P elimina y expe ience o ca dio ascula MR,” Magn. Reson. Med.,
ol. 65, no. 4, pp. 1097–1102, 2011.
[132] M. Fajkus e al., “MR ully compa ible and sa e FBG b ea hing senso :
A p ac ical solu ion o espi a o y igge ing,” IEEE Access, ol.7,
pp. 123013–123025, 2019.
[133] H. Kandpal, R. Sha ma, K. S. Madhusudhan, and K. S. Kapoo ,
“Respi a o y- igge ed e sus b ea h-hold di usion-weigh ed MRI o
li e lesions: Compa ison o image quali y and appa en di usion coe -
icien alues,” Ame . J. Roen genol., ol. 192, no. 4, pp. 915–922, 2009.
[134] J. N. Oshinski, L. Ho land, S. Mukundan, W. T. Dixon, W. J. Pa ks, and
R. I. Pe ig ew, “Two-dimensional co ona y MR angiog aphy wi hou
b ea h holding,” Radiology, ol. 201, no. 3, pp. 737–743, 1996.
[135] H. C. M. an den Bosch e al., “F ee-b ea hing MRI o he assessmen
o myoca dial in a c ion: Clinical alida ion,” Ame . J. Roen genol.,
ol. 192, no. 6, pp. W 277–W281, 2009.
[136] K. Ma sunaga, G. Ogasawa a, M. Tsukano, Y. Iwada e, and
Y. Inoue, “Use ulness o he na iga o -echo igge ing echnique o
ee-b ea hing h ee-dimensional magne ic esonance cholangiopanc e-
a og aphy,” Magn. Reson. Imag., ol. 31, no. 3, pp. 396–400, 2013.
[137] C. Klessen e al., “Magne ic esonance imaging o he uppe abdomen us-
inga ee-b ea hingT2-weigh ed u bospinechosequence wi hna iga o
igge ed p ospec i e acquisi ion co ec ion,” J. Magn. Reson. Imag.,
ol. 21, no. 5, pp. 576–582, 2005.
[138] C. J. Zech e al., “High- esolu ion MR-imaging o he li e wi h T2-
weigh ed sequences using in eg a ed pa allel imaging: Compa ison o
p ospec i e mo ion co ec ion and espi a o y igge ing,” J. Magn. Re-
son. Imag., ol. 20, no. 3, pp. 443–450, 2004.
[139] D. C. Pe e s, R. Neza a , H. Egge s, C. S ehning, and W. J. Manning, “2D
ee-b ea hing dual na iga o -ga ed ca diac unc ion alida ed agains
he 2D b ea h-hold acquisi ion,” J. Magn. Reson. Imag., ol. 28, no. 3,
pp. 773–777, 2008.
[140] B. Taouli e al., “Di usion-weigh ed imaging o he li e : Compa ison o
na iga o igge ed and b ea hhold acquisi ions,” J. Magn. Reson. Imag.,
ol. 30, no. 3, pp. 561–568, 2009.
[141] A. C. La son e al., “P elimina y in es iga ion o espi a o y sel -ga ing
o ee-b ea hing segmen ed cine MRI,” Magn. Reson. Med., ol. 53,
no. 1, pp. 159–168, 2005.
[142] N. Jin, R. J. Lewandowski, R. A. Oma y, and A. C. La son, “Respi a-
o y sel -ga ing o ee-b ea hing abdominal phase-con as blood low
measu emen s,” J. Magn. Reson. Imag., ol.29,no.4,pp. 860–868,2009.
[143] J. Liu, P. Spincemaille, N. C. F. Codella, T. D. Nguyen, M. R. P ince,
and Y. Wang, “Respi a o y and ca diac sel -ga ed ee-b ea hing ca diac
CINEimagingwi hmul iecho3Dhyb id adial SSFP acquisi ion,”Magn.
Reson. Med., ol. 63, no. 5, pp. 1230–1237, 2010.
[144] S. U ibe e al., “Whole-hea cine MRI using eal- ime espi a o y sel -
ga ing,” Magn. Reson. Med., ol. 57, no. 3, pp. 606–613, 2007.
[145] S. U ibe, P. Bee baum, T. S. Sø ensen, A. Rasmusson, R. Raza i, and
T. Schae e , “Fou -dimensional (4D) low o he whole hea and g ea
essels using eal- ime espi a o y sel -ga ing,” Magn. Reson. Med.,
ol. 62, no. 4, pp. 984–992, 2009.
[146] R. Wang e al., “Wea able espi a ion moni o ing using an in-line ew-
mode ibe Mach-Zehnde in e e ome ic senso ,” Biomed. Op . Exp.,
ol. 11, no. 1, pp. 316–329, 2020.
[147] F. Michle e al., “A clinically e alua ed in e e ome ic con inuous-wa e
ada sys em o hecon ac lessmeasu emen o human i alpa ame e s,”
Senso s, ol. 19, no. 11, pp. 1–19, 2019.
[148] H. Di, S. Sun, and Y. Che, “Respi a ion measu emen using ib e-op ic
de o ma ion senso ,” J. Mode n Op ., ol. 64, no. 6, pp. 639–645, 2016.
[149] S. Chen, F. Tan, Z. Huang, T. Yang, J. Tu, and C. Yu, “Non-in asi e sma
moni o ing sys em based on mul i-co e ibe op ic in e e ome e s,” in
P oc. Asia Commun. Pho on. Con ., 2018, pp. 1–3.
[150] Y. Zhang, Z. Chen, and H. I. Hee, “Nonin asi e measu emen o hea
a e and espi a o y a e o pe iope a i e in an s,” J. Ligh w. Technol.,
ol. 37, no. 11, pp. 2807–2814, Jun. 2019.
[151] Z. Chen e al., “Moni o ing espi a ion and ca diac ac i i y du ing sleep
using mic obend ibe senso : A clinical s udy and new algo i hm,” in
P oc. 36 h Annu. In . Con . IEEE Eng. Med. Biol. Soc., 2014, pp. 5377–
5380.
[152] T. Liue al.,“Medical espi a o ymoni o ingsenso sbased on mic obend
ibe loss,” P oc. SPIE, ol. 10025, 2017, A . no. 1002517.
[153] X. Guo, Z. Li, and D. Yan, “A bio-signal moni o ing senso based on he
mic obending e ec o ibe ,” IOP Con . Se .: Ma e . Sci. Eng., ol. 612,
no. 2, pp. 1–8, 2019.
[154] H.-F. Hu, S.-J. Sun, R.-Q. L , and Y. Zhao, “Design and expe imen o
an op ical ibe mic o bend senso o espi a ion moni o ing,” Senso s
Ac ua o s A: Phys., ol. 251, no. 1, pp. 126–133, 2016.
[155] L. T. D’Angelo, S. Webe , Y. Honda, T. Thiel, F. Na bonneau, and T.
C. Lu h, “A sys em o espi a o y mo ion de ec ion using op ical ibe s
embedded in o ex iles,” in P oc. 30 h Annu. In . Con . IEEE Eng. Med.
Biol. Soc., 2008, pp. 3694–3697.
[156] A. G ille e al., “Op ical ib e senso s embedded in o medical ex iles o
moni o ing o espi a o y mo emen s in MRI en i onmen ,” P oc. SPIE,
ol. 6619, 2007, A . no. 66191R.
[157] J. Wi e al., “Sma medical ex iles wi h embedded op ical ib e senso s
o con inuous moni o ing o espi a o y mo emen s du ing MRI,” P oc.
SPIE, ol. 7653, 2010, A . no. 76533B.
[158] J. Wi e al., “Medical ex iles wi h embedded ibe op ic senso s o
moni o ing o espi a o y mo emen ,” IEEE Senso s J., ol. 12, no. 1,
pp. 246–254, Jan. 2012.
[159] D. Lo P es i e al., “A mul i-pa ame ic wea able sys em o moni o neck
mo emen s and espi a o y equency o compu e wo ke s,” Senso s,
ol. 20, no. 2, pp. 1–17, 2020.
[160] W. Ecke e al., “A no el pe iodic mac obending he e o-co e ibe op ic
senso embedded in ex ile o espi a o y mo emen s’ analysis,” P oc.
SPIE, ol. 9062, 2014, A . no. 90620D.
[161] A. Issa aye a, A. Beiseno a, D. Tosi, and C. Mola di, “Fibe -op ic based
sma ex iles o eal- ime moni o ingo b ea hing a e,” Senso s, ol.20,
no. 12, pp. 1–16, 2020.
[162] D. Lau e al., “In ensi y-modula ed mic obend ibe op ic senso o
espi a o y moni o ing and ga ing du ing MRI,” IEEE T ans. Biomed.
Eng., ol. 60, no. 9, pp. 2655–2662, Sep. 2013.
[163] W.-J. Yoo e al., “De elopmen o espi a ion senso s using plas ic op ical
ibe o espi a o y moni o ing inside MRI sys em,” J. Op . Soc. Ko ea,
ol. 14, no. 3, pp. 235–239, Sep. 2010.
[164] M. K ej, P. Ba an, and Ł. Dziuda, “De ec ion o espi a o y a e using a
classi ie o wa es in he signal om a FBG-based i al signs senso ,”
Compu . Me hods P og. Biomed., ol. 177, pp. 31–38, 2019.
[165] L. Mendes Pe ei a e al., “UTE-SENCEFUL: Fi s esul s o 3D high-
esolu ion lung en ila ion imaging,” Magn. Reson. Med., ol. 81, no. 4,
pp. 2464–2473, No . 2018.
[166] N. S. Higano e al., “Re ospec i e espi a o y sel -ga ing and emo al
o bulk mo ion in pulmona y UTE MRI o neona es and adul s,” Magn.
Reson. Med., ol. 77, no. 3, pp. 1284–1295, 2017.
[167] J. Delacos e, J. Chap inel, C. Beigelman-Aub y, D. Piccini, A. Sau y,
and M. S ube , “A double echo ul a sho echo ime (UTE) acquisi ion
o espi a o y mo ion-supp essed high esolu ion imaging o he lung,”
Magn. Reson. Med., ol. 79, no. 4, pp. 2297–2305, 2018.
[168] J. F. Heiden eich e al., “Func ional MRI o he lungs using single
b ea h-hold and sel -na iga ed ul asho echo ime sequences,” Radiol.:
Ca dio ho acic Imag., ol. 2, no. 3, pp. 1–7, Jun. 2020.
[169] M. Tibile i e al., “Mul is age h ee-dimensional UTE lung imag-
ing by image-based sel -ga ing,” Magn. Reson. Med., ol. 75, no. 3,
pp. 1324–1332, 2016.