PortableNear-InfraredTechnologiesandDevicesfor...

Review ArticlePortable Near-Infrared Technologies and Devices forNoninvasive Assessment of Tissue Hemodynamics

Lin Hou1 Yinqiu Liu1 Lixia Qian2 Yucong Zheng2 Jinnan Gao2 Wenxing Cao3

and Yu Shang 1

1North University of China No 3 Xueyuan Road Taiyuan 030051 China2Shanxi Dayi Hospital No 99 Longcheng Street Taiyuan 030032 China3University of Electronic Science and Technology of China No 2006 Xiyuan Street Chengdu 611731 China

Correspondence should be addressed to Yu Shang yushangnuceducn

Received 18 July 2018 Revised 24 December 2018 Accepted 14 January 2019 Published 12 February 2019

Academic Editor Valentina Hartwig

Copyright copy 2019 LinHou et al+is is an open access article distributed under the Creative Commons Attribution License whichpermits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Tissue hemodynamics including the blood flow oxygenation and oxygen metabolism are closely associated with many diseasesAs one of the portable optical technologies to explore human physiology and assist in healthcare near-infrared diffuse opticalspectroscopy (NIRS) for tissue oxygenation measurement has been developed for four decades In recent years a dynamic NIRStechnology namely diffuse correlation spectroscopy (DCS) has been emerging as a portable tool for tissue blood flow mea-surement In this article we briefly describe the basic principle and algorithms for static NIRS and dynamic NIRS (ie DCS)+enwe elaborate on the NIRS instrumentation either commercially available or custom-made as well as their applications tophysiological studies and clinic +e extension of NIRSDCS from spectroscopy to imaging was depicted followed by in-troductions of advanced algorithms that were recently proposed+e future prospective of the NIRSDCS and their feasibilities forroutine utilization in hospital is finally discussed

1 Introduction

Many diseases would lead to abnormal status in micro-vasculature blood flow oxygenation and oxygen meta-bolism [1 2] For example ischemic stroke causes reductionin both blood flow and oxygenation [3] Malignant tumorleads to higher blood flow higher oxygen consumption rateand lower oxygen status [4 5] For the people with ob-structive sleep apnea large fluctuations in microvasculatureblood flow and total hemoglobin concentration were dis-covered [6]

+ere are few technologies currently available to non-invasively monitor the oxygen status in tissue PO2 is atechnology capable of quantifying the oxygen pressure via atiny needle [7] +is invasive pattern of measurementhowever is unfavorable to healthy population As a con-ventional technology for tissue blood oxygenation mea-surement including the oxy- deoxy- total hemoglobinconcentration ([HbO2] [Hb] and THC) and oxygen saturation

(StO2) near-infrared diffuse optical spectroscopy (NIRS)has been developed for more than four decades since 1970s[8] NIRS oximeter was often utilized in clinical andphysiological studies [8ndash10] for detecting the relative deeptissue (up to several centimeters) owing to the low tissueabsorption in near-infrared (NIR) range (650 to 900 nm)However the oxygenation alone only reflects the staticbalance between the oxygen supply and consumption [11]Blood flow is such a parameter that elucidates how fast andefficient the oxygen is carried in the blood and supplied tothe tissue Studies show that blood flow is highly sensitive topathophysiological alteration [12] thus could be a potentialindicator for early detection of many diseases Additionallythe blood flow reflects the reactivity patterns of vascularvasodilatation [3] thus assisting in evaluation of vesselfunctions +erefore investigation of blood flow as well asits correlations with pathophysiological consequences hassignificant implications for diagnosis and medicationsFurthermore the combination of tissue blood flow and

HindawiJournal of Healthcare EngineeringVolume 2019 Article ID 3750495 11 pageshttpsdoiorg10115520193750495

oxygenation allows for estimate of the oxygen metabolic rate[11 13] a valuable parameter closely associated with tissuepathology

At present the low-cost modalities to noninvasively andfast measure the blood flow in deep tissue are not estab-lished Methods currently available for assessment of bloodflow include Doppler ultrasound [14] laser Doppler [15]and perfusion-weighted magnetic resonance imaging (per-fusion MRI) [16] Doppler ultrasound can only assess theblood velocity in single and large vessel [14 17] Manydiseases however are relevant to the abnormal blood flow inmicrovasculature level rather than macrovasculature level[1 2] While perfusion MRI can provide the microvascu-lature blood flow at high spatial resolution the high cost andlow mobility of MRI instrument preclude its wide use forroutine disease screening Laser Doppler can only detectblood flow at superficial tissue (sim1mm depth) [18] thusunable to probe the diseases mostly originated from deeptissues In summary the routine use of the above techniquesis limited due to availability expense and the difficulty ofmaking continuous measurements

In recent years a dynamic NIRS technology namelydiffuse correlation spectroscopy (DCS) has gained rapiddevelopment [19ndash23] DCS quantifies the speckle fluctua-tions of diffuse light which are sensitive to the motion of redblood cells in the microvasculature thus permitting fastassessment of blood flow in deep tissues (up to severalcentimeters)

+e article will briefly describe the technology and de-vices for NIRS and DCS as well as their applications tophysiological studies and clinic +e extension of NIRS fromspectroscopy to imaging was depicted followed by in-troductions of advanced algorithms that were recentlyproposed +e future prospective of the NIRS and DCS andtheir feasibilities for routine use in hospital was finallydiscussed

2 NIRS Principle and Devices

When being injected into the biological tissues that areconsidered as turbid medium the near-infrared lighttransports through the medium in ldquodiffusiverdquo manner [2]due to tissue absorption (mainly red blood cells) andscattering (mainly organelles and mitochondria) [1 24] +eaim of NIRS technology is to separate the tissue absorptionfrom scattering through which the hemoglobin concen-tration is derived Over the years two approaches namelythe time-resolved spectroscopy [25 26] and frequency-modulated spectroscopy [27 28] have been developed toaccount for this separation For the time-resolved approacha pulsed laser (a few of picoseconds) is adopted and the time-of-flight of photons is recorded by a digital time-correlatedsingle photon counting (TCSPC) board For frequency-modulated approach the light is modulated at high fre-quency (around 100MHz) [29] so that the phase ofmodulated light escaping out from the tissue could be traced

Both these two approaches were commercialized withnumerous applications [30] particularly for neurologicalstudies Either the time-resolved or frequency-modulated

approach however requires complicated hardware to ma-nipulate the light substantially increasing the cost ofinstrument

In practice the NIRS was often simplified by usingcontinuous-wave (CW) light instead of pulsed or frequency-modulated light In CW oximeter the tissue scattering (μs)was assumed to be a constant number and only the relativechange of tissue absorption coefficient needs to be de-termined by the modified BeerndashLambert law

lnI t0( 1113857

I(t)1113888 1113889 Δμa(λ) middot L

εHbO(λ)Δ HbO21113858 1113859 + εHb(λ)Δ[Hb]( 1113857 middot L

(1)

where I(t0) and I(t) are the light intensity detected at time t0and t respectively Δμa(λ) is the relative change of theabsorption coefficient at wavelength λ and L is the meanphoton length between a specific source-detector (S-D) pairBy collecting the light at two wavelengths the oxygenationchanges can be calculated as [31ndash33]

Δ HbO21113858 1113859 εHb λ1( 1113857Δμa λ2( 1113857minus εHb λ2( 1113857Δμa λ1( 1113857

εHb λ1( 1113857εHbO λ2( 1113857minus εHbO λ1( 1113857εHb λ2( 1113857

Δ[Hb] εHbO λ2( 1113857Δμa λ1( 1113857minus εHbO λ1( 1113857Δμa λ2( 1113857


(2)

where Δ[HbO2] and Δ[Hb] are the relative change of oxy-and deoxy-hemoglobin concentration and εHbO(λ) andεHb(λ) are their extinction coefficients respectively

Figure 1 illustrates a typical scheme of CW oximeterBriefly a dual-wavelength LED at NIR range was controlledby transistor-transistor logic (TTL) signals injecting thephotons at each wavelength into the tissue alternately +eescaped photons were collected in parallel by all opticaldetectors and transferred to the computer wherein themodified BeerndashLambert law was applied yielding time-course oxygenation data

+e reason for using CW light as the source light of NIRStechnology is that the tissue scattering (majorly contributedfrom organelles and mitochondria) changes much less thanthe tissue absorption in most of physiological status +usthe tissue blood oxygenation would still been properlyextracted by the CW tissue oximeter Moreover this low-cost device could minimize the cross-talk between the ab-sorption and scattering

+e CW tissue oximeter was also commercialized bydifferent manufacturers (eg NIRO Hamamatsu IncINVOS Somanetics Inc) and has various clinical applica-tions At CW working mode the absolute value of tissueabsorption coefficient (μa) could be extracted by the near-infrared light collected from multiple source-detector (S-D)separations according to the theory of spatial-resolvedspectroscopy [34] +is theory permits generation of aslope through multiple S-D separations of optical data fromwhich the tissue absorption will be determined In generalmore S-D separations perform better in calculating ab-sorption coefficient (μa) with stable and accurate values Forexample four S-D separations (20 to 35 cm) were used in a

2 Journal of Healthcare Engineering

commercial tissue oximeter (OxiplexTS ISS Inc) On theother hand however more S-D separation leads to highercost of instrumentation and increases the size of opticalprobe Hence two S-D separations ie the minimal numberto determine a slope were used more frequently in com-mercial instruments (eg NIRO Hamamatsu Inc INVOSSomanetics Inc)

Despite the wide use of CW light it is worthy to note thatthe tissue absorption coefficient (μa) and reduced scatteringcoefficient (μsprime) cannot be uniquely determined from theCW tissue oximeter alone regardless of the number of S-Dseparations being used [35] To separate these two co-efficients time-resolved spectroscopy or frequency-modulated spectroscopy must be adopted

3 DCS Principle and Devices

DCS is originated from the technology of dynamic lightscattering [36 37] in which single scattering event is allowedwithin thin biological sample To account for in vivo mea-surement in bulk tissues or organs advanced mathematicalprocedures were proposed to model the multiple scatteringevent namely diffuse correlation spectroscopy (DCS)[20 38] or diffusing wave spectroscopy (DWS) [39 40] Asdepicted in Figure 2 a long coherence length (gt5m) laser innear-infrared range injects photons into the tissue via opticalfiber +e photons travel through the tissue and are scatteredback to the detector via a single mode fiber +ere arefluctuations in the intensity of escaped photons which is dueto the moving of scatters (mainly red blood cells) in themicrovasculature system A digital correlator takes the de-tector output and uses photon arrival times to compute thelight intensity autocorrelation function From the normal-ized intensity autocorrelation function the electric fieldtemporal autocorrelation function G1(τ) which satisfies thecorrelation diffusion equation in turbid medium [20 38] iscalculated Approximately the normalized electric fieldtemporal function g1(τ) at early delay time decays expo-nentially with τ (delay time) +e blood flow index (BFI)information is extracted by fitting the curve of temporalfunction g1(τ) whose decay rate depends mainly on themotion of red blood cells [20 21 41]

+e current optical probes for DCSDCT are designedwith source and detector fibers confined by a flat foam padrestricting the use of this technology for tissues with differentgeometries Moreover the requirement in temporalspatialresolution varies across different tissues and organs

As a relative new technology there are commercialdevices for DCS (eg Neuro-Monitor HemoPhotonics IncSpain) Similar to tissue oximeter this commercial DCSDCT device is primarily utilized for neurological functionevaluation limiting its application for other diseases

4 Integrated System and Hybrid Optical Probe

Either the NIRS or DCS when being used alone couldprovide only single hemodynamic parameter (oxygenationor blood flow) At early time the NIRS and DCS deviceswere applied respectively to the biological tissues in orderto obtain comprehensive hemodynamic information Underthis framework however it is difficult to measure the ox-ygenation and blood flow from the same tissue because theNIRS and DCS probes cannot be physically overlappedMore importantly since the two devices are controlledindependently the source light from one device in-terferences with the signal collection by another device if thetwo probe are placed closely on the tissue In recent yearsattempts have been made to combine the DCS with NIRSdevices to form an integrated system for simultaneousmeasurement of blood flow and oxygenation at microvas-culature level also permitting estimate of oxygen metabolicparameters (eg oxygen consumption rate) [11 42] Suchintegrated NIRSDCS system for tissue oxygen consumptionmonitoring is already commercially available (MetaOx ISSInc USA) Prior to the commercial products the integratedsystem has been constructed by various research groups

Figure 3 shows a typical integrated optical systemcombing the tissue oximeter and flowmeter Briefly thedual-wavelength NIR LED (oximeter) and the long-coherence NIR laser are turned onoff by the TTL switch(Figure 3(a)) launching the photon into biological tissuesalternately +e photons scattered back from the tissue arecollected by either optical detectors (oximeter) or singlephoton detectors (flowmeter) In this pattern the two de-vices work onoff in sequential pattern

Control panel

AD board

Dual-wavelengthNIR LED Optical detectors

TTL Light intensity

Injected photons

Escaped photons

Tissue

Figure 1 Graphic illustration of a NIRS tissue oximeter

Control panel

AD board

Long-coherenceNIR laser

Single photondetectors

TTL

Injected photons

Escaped photons

Digital correlator

Autocorrelationfunction

Figure 2 Graphic illustration of a tissue flowmeter

Journal of Healthcare Engineering 3

To comply with the integrated system one needs ahybrid optical probe containing all source and detectorfibers in place +is configuration of fiber setup ensures theblood flow and oxygenation information measured by DCSand NIRS to cover mostly the same volume in the tissue Asmentioned earlier the sources of DCS and NIRS are turnedonoff sequentially in order to avoid the light interferencebetween the two technologies

For construction of hybrid optical probe a foam pad isused to house both NIRS and DCS fibers (Figure 3(b)) thatare originally distributed in two separated pads Within thepad the fibers are arranged to form multiple-distancesource-detector pairs for both NIRS and DCS measure-ments +is all-in-one probe has many advantages such ascompact size easy to tape and enable detection of bloodflow and oxygenation in an overlapped tissue volume asstated earlier making it possible to estimate the metabolicparameters (eg the oxygen consumption rate) Moreoverthe multiple separations provide hemodynamic informationfrom different layers of tissues (eg scalp skull and braincortex) which can be separated by using the advancedalgorithms

+e design of optical probe would be adaptable to dif-ferent tissues or organs For the brain studies there is littledeformation on the head when subject to contact pressure

Hence the materials with good elasticity and stiffness(eg silicon glue and polymer foam) are often used to housethe S-D fibers which would adapt the small curvature of thehuman head in contact pattern For measurement of thebreast tumor in which large tissue deformation occursnoncontact pattern of measurement is preferred in order toadapt to the large curvature of breast For such clinicalapplications a lens system is designed to focus the sourceand detector on the tissue surface [4] permitting noncontactcollection of optical signals To minimize the in-strumentation cost the lens system will be controlled to scanover the tissue surface greatly reducing the number ofoptical components (laser and detectors) [4]

+e integrated NIRS system and the hybrid opticalprobe depicted above have advantages of high-throughputportability and easy-to-use which have high flexibility to beused in different tissue type disease category and clinicalsetting +e successful carrying out of the hemodynamicmeasurements promote the translation of DCSDCT intoclinic making it possible to routinely use this technology forearly detection and therapeutic evaluation of a variety ofdiseases

Table 1 lists the physiological or clinical studies whereinthe integrated NIRSDCS instrument and hybrid probe wereused Among these a small portion of studies reports the

TTLswitch

40

(b)

Probe

Single photon detectors

Long-coherence NIR laser

Optical detectors

Control panel

(a)

(c)

Dual-wavelength NIRLED

400

300

200

100

0

Pre-revascularization

Post-revascularizan

Arterial occlusion

0830 0900 0930 1000 1030 1100

0830 0900 0930 1000 1030 1100

20

0

ndash20

ndash40

rBF

()

microM

∆[HbO2]∆[Hb]

Time

Figure 3 Graphic illustration of the integrated optical system (a) and the hybrid optical probe on human lower limb (b) +e application ofthis system to simultaneously monitor the blood flow and oxygenation changes during arterial revascularization is exhibited (c)


improvements on the optical instrumentation or probe de-sign For example Shang et al [32] combined a commercialfrequency-domain NIRS oximeter and a dual-length DCSflowmeter to form an integrated system Accordingly a hy-brid probe that houses the NIRS fibers and DCS fibers in thesame foam was designed In another study Li et al [62]designed a noncontact probe which focuses the source and

detector light on the surface of the target tissue via lenssystem +is probe permits simultaneous monitoring of theblood flow and oxygenation without contact with the tissues+is noncontact pattern of measurement is suitable to be usedin deformable tissues such as human breast Both contact andnoncontact systems were generally validated through themanipulation protocol of temporary cuff occlusion on the

Table 1 Studies wherein the integrated NIRSDCS instrument and hybrid probe were utilized

Year Ref no Authors Target tissue Measured variables Physiological manipulationdiseases2001 [20] Cheung et al Rat brain BFI hemoglobin StO2 Hypercapnia

2003 [42] Culver et al Rat brain BFI hemoglobin StO2 oxygenconsumption rate Ischemia

2004 [43] Durduran et al Human brain BFI hemoglobin change oxygenconsumption rate Cortex activation

2005 [44] Yu et al Human skeletal muscle BFI hemoglobin StO2 oxygenconsumption rate Plantar flexion exercise

2006 [45] Sunar et al Human tumor BFI hemoglobin StO2 Headneck tumor2006 [46] Yu et al Human tumor BFI StO2 Prostrate tumor2007 [47] Sunar et al Mouse tumor BFI StO2 Melanoma tumor2007 [48] Zhou et al Human tumor BFI hemoglobin Breast tumor

2009 [32] Shang et al Human skeletal muscle BFI hemoglobin change Cuff occlusion onupper limb

2009 [49] Durduran et al Human brain BFI hemoglobin StO2 Stroke2009 [50] Zhou et al Piglet brain BFI hemoglobin StO2 Closed head injury

2010 [51] Durduran et al Human infant brain BFI hemoglobin oxygenconsumption rate Congenital heart defects

2010 [52] Mesquita et al Mouse skeletal muscle BFI StO2 oxygenconsumption rate Hindlimb ischemia

2010 [13] Roche-Labarbe et al Human infant brain BFI hemoglobin StO2 oxygenconsumption rate Premature

2010 [53] Edlow et al Human infant brain BFI hemoglobin Posture change2010 [54] Kim et al Human brain BFI hemoglobin change Critically brain-injured2011 [3] Shang et al Mouse brain BFI hemoglobin change Ischemia2011 [55] Yu et al Human skeletal muscle BFI hemoglobin change Ischemia2011 [33] Shang et al Human brain BFI hemoglobin change Ischemia2012 [56] Cheng et al Human brain BFI hemoglobin change Posture change2012 [57] Dong et al Human tumor BFI hemoglobin change Headneck tumor

2012 [58] Gurley et al Human skeletal muscle BFI hemoglobin StO2 oxygenconsumption rate Handgrip exercise

2012 [11] Shang et al Human skeletal muscle BFI hemoglobin StO2 oxygenconsumption rate Fibromyalgia

2013 [59] Buckley et al Human infant brain BFI StO2 oxygenconsumption rate Cardiac surgery

2013 [60] Mesquita et al Human skeletal muscle BFI StO2 Peripheral artery disease

2013 [61] Mesquita et al Human brain BFI hemoglobin oxygenconsumption rate

Transcranial magneticstimulation

2013 [62] Li et al Human skeletal muscle BFI hemoglobin change Cuff occlusion onupper limb

2013 [63] Shang et al Human skeletal muscle BFI hemoglobin change Electrical stimulation

2014 [64] Jain et al Human infant brain BFI StO2 oxygenconsumption rate Congenital heart disease

2014 [65] Cheng et al Human brain BFI hemoglobin change Posture change2014 [6] Hou et al Human brain BFI hemoglobin change Sleep apnea

2015 [66] Henry et al Human skeletal muscle BFI hemoglobin StO2 oxygenconsumption rate Plantar flexion exercise

2016 [12] Dong et al Human tumor BFI hemoglobin change Headneck tumor

2017 [67] Baker et al Human skeletal muscle BFI hemoglobin StO2 oxygenconsumption rate Peripheral artery disease

2018 [68] Ko et al Swine brain BFI hemoglobin StO2 oxygenconsumption rate Deep hypothermic


limb which reduces the blood supply to the skeletal musclewhere the blood flow and oxygenation are measured

In addition to the instrumentation improvements asmentioned above the majority of studies using the in-tegrated system focused on the physiological or clinicalapplications For the physiological studies a variety of ex-perimental manipulation were designed to challenge thetarget tissues such as the hypercapnia to rat brain [20]cortex activation to human brain [43] plantar flexion orhandgrip exercise to human skeletal muscles [58 66]posture change to human brain [53 56 65] and the hy-pothermic to swine brain [68]

+ese physiological manipulations alter the amount oftissue hemodynamics (blood flow oxygenation and oxygenconsumption rate) making it convenient to detect the po-tential defects or diseases (eg aging autoregulation im-pairment and vasovagal syncope)

For the clinical or preclinical studies the integratedsystem was primarily used to probe ischemic diseases ortumor For these purposes animal models were often used tocreate the relevant diseases including the cerebral ischemiaon rat and mouse [3 42] hindlimb ischemia on the mouse[52] and melanoma tumor on the mouse [47] +e outcomesderived from the animal studies would assist the un-derstanding of pathophysiological mechanism and im-provements of therapeutic approaches In addition to theanimal models the integrated systemwas also directly appliedto clinic due to the advantages of noninvasiveness and easy-to-use For example Sunar et al [47] and Dong et al [12 57]longitudinally monitored the blood flow and oxygenationchanges in the headneck tumor over weeks of chemo-radiation therapy and found these hemodynamic changeswere associated with pathophysiological outcomes Yu et al[46] and Zhou et al [50] assessed the hemodynamic changesin prostate and breast tumors during photodynamic therapyand chemotherapy respectively Durduran et al [49] in-vestigated the hemodynamic responses to posture change inthe stroke patients and found that the response pattern wassignificantly affected by the diseases Shang et al [33] and Yuet al [55] on the other hand evaluated the integrated systemfor intraoperative monitoring of the carotid and femoralarterial surgeries respectively +e experiment results verifiedthe high sensitivity of the integrated NIRSDCS technologyfor ischemic detection during the vascular surgeries

As a graphic example Figure 3(c) shows the musclehemodynamics (blood flow and oxygenation) during arterialvascularization [55] It is clearly seen that the arterialclamping causes acute ischemia and hypoxia on the muscletissue ie decrease in blood flowΔ[HbO2] and increase inΔ[Hb] Following the release of cuff occlusion there ishyperemic response elevating the blood flow to a peak value+ereafter both the blood flow and oxygenation recovergradually towards their baseline Additionally the post-revascularization improvements in tissue hemodynamicswere detected +is study demonstrates the high sensitivityof NIRSDCS in monitoring the intraoperative interventionsas well as the acute therapeutic assessments

In addition there are reports of hemodynamic assess-ment on other diseases including the congenital heart

defects [51 64] premature [13] critically brain-injured [54]fibromyalgia [11] and peripheral artery disease [60 67]+ose clinical studies validated the wide usage of integratedNIRSDCS system for various diseases

5 Imaging Modalities for NIRS and DCS

+eNIRS andDCS provide time variation of oxygenation andblood flow respectively In order to obtain the spatial contrastof the tissue hemodynamics there is a need to extend theNIRS and DCS from spectroscopy to tomography termed asdiffuse optical tomography (DOT) and diffuse correlationtomography (DCT) respectively +e DOTDCT share thesame physical principle of NIRSDCS except that far moreS-D pairs are needed to perform image reconstruction Asreported in literature a high-density DOT system consists of96 sources and 92 detectors providing the optical mea-surements from more 1200 S-D pairs [69] Similarly a rep-resentative DCTdevice consists of 2 laser diodes and 15 APDand with 15 steps of line scanning generating a total of 450autocorrelation g1(τ) curves [70] In addition to the hardwareconfiguration the algorithms adopted for image re-construction is critical in determining imaging quality ofblood flow or oxygenation which will be discussed in Section6 Apparently a large number of optical components (laserdetector etc) are required for DOT or DCT substantiallyelevating the cost of instrumentation A few efforts have beenmade to lower down the physical S-D measurements Forexample the line or rotated scanning could efficiently reducethe number of optical components but needs the optical probeto operate in noncontact pattern +e noncontact measure-ment however is susceptible to the motion artifacts andambient light +e optical switch which enables sharing thelaser or detector among multilocation measurements couldbe a good alternative option to save the instrument costNevertheless a number of hardware source and detector pairsare still needed for the imaging of tissue hemodynamicsmaking the instrument no longer portable +us the he-modynamic imaging is not the focus of this article

6 Advanced Algorithms for NIRS and DCS

Apart from instrumentation the reconstruction algorithmof DCSNIRS also contributes greatly to the accuracy androbustness of oxygenation and blood flow values which ishighly crucial to the detective sensitivity of diagnosticoutcomes [31 71] For DCS as stated earlier the conven-tional algorithm is based on the analytical solution of apartial differential equation (PDE) [20 21 41] However theanalytical solution of PDE is dependent on tissue geometryand usually has complicated mathematical forms For thepurpose of easy implementation the tissue geometry is oftenassumed as semi-infinite +ere are also analytical solutionsto PDE for other regular tissue geometries such as cylinderand sphere [41 72] However those solutions are rarelyadopted for practical applications due to the complicatedand tedious computations In recent years a finite elementmethod (FEM) for DCT was developed based on similarapproach for NIRS [70] Although the FEMwas proved to be


efficient in reconstruction of flow value in tissue withcomplicated geometry it is difficult to fully utilize auto-correlation data thus susceptible to data noise

For sufficient use of the DCSDCT data a mathe-matical approach namely the Nth-order linear algorithm(ie NL algorithm) was developed for extracting BFIvalue Unlike the analytical solution or FE method the NLalgorithm does not seek for PDE solution Instead it wasoriginated from a combination of a Nth-order Taylorpolynomial and the integral form of g1(τ) After a series ofmathematical procedures the BFI is extracted by iterationof linear regressions ultimately reaching the followingexpressions [73]

g1(τ j)minus 1 τ 1113944n

i1A(i j)αDB(i) (3)

g1(τ j)minus 1

minus 1113944N

k2

1113936Qp1w(p j) minus21113936

ni1k

20(i)αDB(i)s(i p j)μs

prime(i)1113872 1113873k

kτk

τ 1113944n

i1A(i j)αDB(i)

(4)

Equations (3) and (4) are called first-order and Nth-order linear (NL) algorithm respectively Here α DB(i) isthe BFI of the ith tissue component w(p j) is the nor-malized weight of photon package collected by jth detectorand s(i p j) is the path length of pth photon packagetraveled within ith tissue component and collected by jthdetector +e coefficient of matrix A(i j) has the followingexpression

A(i j) 1113944

Q

p1minus2w(p j)k

20(i)s(i p j)μs

prime(i) (5)

+e key parameter contained in equations (3)ndash(5) iss(i p j) which is determined by the tissue geometry andheterogeneity indicating that both of these two features arefully taken into consideration in the NL algorithm +eaccuracy of the NL algorithm on various geometries andtissue heterogeneity was validated in DCS through com-puter simulations as well as animal experiment [73 74]

+e similar procedure for heterogeneous tissue is alsoapplicable to NIRS leading to the following equations tocalculate the changes of chromophore concentration[31 75]

OD1

ODN

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠

s11 s1M

sN1 sNM

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠

Δμa1

ΔμaM

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠ (MleN)

(6)

where ODi is the optical density from ith S-D measurementsij is the photon path length within jth tissue measured fromith S-D pairM andN are the numbers of tissue and S-D pair

respectively andΔμaj is the relative change of the absorptioncoefficient at jth tissue For convenience in matrix operationequation (6) can be abbreviated as

OD sΔμa (7)

Here s is the matrix of photon path lengthWhenMNthe changes of absorption coefficient can be calculatedaccording to the following equation

Δμa sminus1OD (8)

It is possible to extract the hemodynamic information athigher accuracy by using the optical signals from moremeasurements (ie NgtM) In this situation the least-square fitting method can be used to calculate thechanges of absorption coefficient at different tissues

Δμa sTs1113872 1113873minus1

sTOD (9)

Due to the correlation of optical signals among differentS-D pairs the matrix s is often ill-posed ie susceptible tothe data noise +e noise influence on oxygenation mea-surement could be partially alleviated by introducing reg-ularization approaches One of the widely-used approachesis Tikhonov regularization with which the final form ofhemoglobin calculation is the following

Δμa sTs + α2I1113872 1113873

minus1s

TOD (10)

Here α is the constant number controlling the balancebetween the equation fidelity and solution robustness

7 Clinical Applications

Over the years the NIRS and DCS for tissue oxygenationand blood flow measurements have been extensivelyvalidated to other modalities and been translational tovarious physiological and clinical studies such as as-sessment of autoregulation capability [56] therapeuticmonitoring of ischemic stroke [33] and early detection oftumors [4] +ese applications are summarized in Table 1and briefly reviewed in Section 4 Some representativeapplications are introduced with details in the followingsubsections

71 Assessment of Autoregulation Capability Studies haveshown that many diseases such as ischemic stroke ob-structive sleep apnea vasovagal syncope or Alzheimerrsquosdisease may cause deficit in autoregulation capability whichcould be assessed by alteration of both blood pressure andcerebral hemodynamics For example change of posturecould alter both blood pressure and cerebral hemodynamicsHowever the changing pattern of cerebral hemodynamics inpatients dependent on the autoregulation capability maydiffer significantly from those in healthy subjects fromwhich we can detect the neurological deficit +e technol-ogies of NIRS and DCS were well established to measure thecerebral hemodynamics and the blood pressure could belongitudinally assessed by a finger plethysmograph at the


pulse frequency (sim1Hz) +e physiological protocols toinduce the cerebral hemodynamics include enforcedbreathing at 01Hz cuff occlusion on both side of lowerlimbs posture transition from squatting to standing and thehead-up-tilting (HUT) Among these the 70-degree HUTwas adopted widely in physiological studies For most of thepeople it was found that there are spontaneous low fre-quency oscillations (LFOs) around 01Hz in some physio-logical parameters including the blood pressure andhemodynamics and the LFOs could be affected by theimpairment in autoregulation capability For rapid andnoninvasive assessment of the autoregulation capability ahybrid near-infrared diffuse optical instrument was con-structed to simultaneously detect LFOs of cerebral bloodflow (CBF) and oxygenation ([HbO2] [Hb] and THC) [56]and a finger plethysmograph was used to detect mean ar-terial pressure (MAP) +e coherence outcomes of LFOsbetween MAP and each of the hemodynamic variables showthat the CBF [HbO2] and THC were reliable hemodynamicparameters in detecting LFOs [56] which has clinical im-plication in early diagnosis of cerebral impairment caused byischemic stroke obstructive sleep apnea vasovagal syncopeor Alzheimerrsquos disease

724erapeuticMonitoring of Ischemic Stroke During manyhead and neck surgeries such as carotid endarterectomy themajor arteries such as common carotid artery internalcarotid artery and middle cerebral artery are temporarilyoccluded for blooding prevention and operating conve-nience On the other hand this procedure increases the riskof ischemic stroke +erefore real-time monitoring of ce-rebral hemodynamics is critical for minimizing the intra-operative stroke and acute assessment of surgical outcomesFor this purpose the DCS and NIRS were combined in aportable device namely DCS flow-oximeter [32] As a pilotstudy of intraoperative assessment two fiber-optic probeswere taped on both sides of forehead for cerebral hemo-dynamicmeasurements [33] Meanwhile the brain wave wasmonitored by electroencephalogram (EEG) +e cerebralblood flow and oxygenation were simultaneously monitoredon twelve patients throughout entire period of carotidendarterectomy +e results verified the feasibility andsensitivity of NIRS and DCS in probing transient cerebralischemia due to arterial clamping and postsurgical hyperperfusion [33] More importantly the comparison outcomesshow that the CBF responses to arterial clamping measuredby DCS were faster larger and more sensitive than EEGresponses +is study suggests the potential of NIRSDCStechnologies for routine use in surgical rooms

73 Early Detection of Breast Tumors +e blood flow whichreflects the dynamic demands of oxygen in tissue is highlysensitive to the growth of the tumor with extraordinary highmetabolic rate Hence monitoring of microvasculatureblood flow along with oxygenation measurement will probethe malignant tumor earlier than the morphological imagingmodality such as ultrasound CT orMRI+e portable NIRS

technology has been used to early diagnosis a few cancersmajorly the breast tumors

In an earlier study a hand-held optical probe containingfour S-D pair fibers was used to scan over the breasthorizontally and vertically [5] +e optical data weretransferred into a DCS device for extracting the blood flowindex (BFI) at different locations In the women with breastcancer a high contrast in BFI was found between malignanttumor and the normal tissue demonstrating the high sen-sitivity of DCS in detecting the abnormal blood flow causedby the cancer However the breast tissue is soft and sus-ceptible to contact pressure Hence it is difficult to obtainthe true BFI values in natural posturer Recently a non-contact probe is constructed through which the lightemitted into (or escaped out from) the tissue is delivered bythe lens system [76] rather than directly contacting with theskin +e noncontact pattern of measurement efficientlyavoids the breast deformation caused by the contact pres-sure Moreover the motor-controlled line scanning systempermits data acquisition from a large number of S-D pairsmaking it possible to reconstruct the breast tumor imaging[70] Similar to the contact-measurement study mentionedabove the imaging from the women with malignant tumorexhibits high spatial contrast (59- and 100-fold) betweenthe malignant tumor and normal tissue [4]

+ese studies indicate the potential of NIRSDCS sys-tems for noninvasive and early diagnosis of the breast tu-mors which is greatly beneficial to the healthcare of middleand aging women

In addition to the clinical applications mentioned abovethe NIRSDCS has been widely translated to assess the tissuehemodynamic status for other diseases or medical in-terferences including the peripheral artery disease (PAD)[60] fibromyalgia [11] radiation therapy on headnecktumor [12] and tissue transfer flaps [77] Due to thelength limit the readers are encouraged to see references formore details

8 Challenges and Future Prospective

Although some NIRS commercial products have been uti-lized for physiological and disease studies there are somechallenges for its wide translation to clinic For example theabsolute value of oxygenation parameters ([HbO2] [Hb]THC and StO2) varies over a large range in healthy pop-ulation making it difficult to set a threshold to separate thediseased tissues and normal tissues In practice thereforethe oxygenation contrast either temporally or spatially wasoften used for abnormal detection Likewise the standardprocedure to detect abnormal tissue with DCS flow mea-surement has not been established Besides a few factorssuch as the ambient noise the curvature of tissue geometryand the contact pressure between optical probe and tissuesurface will affect the NIRSDCS technology for hemody-namic measurement

In summary this article provides a review of the staticand dynamic NIRS technologies in comprehensive aspectsincluding the physical principle the instrumentation thealgorithms for hemodynamic calculation and the applications


in physiological and clinical measurement Over four decadedevelopment the NIRS technology has demonstrated ex-cellent feasibility in exploring the microvasculature oxygenkinetics and pathophysiological mechanism +e success ofhemodynamic measurement greatly promoted the trans-lation of NIRS technology to clinic However the accuracyand efficiency of this technology for early disease detectionand therapeutic evaluation is heavily dependent on the tissuetype disease category and clinical environment +us thereare still urgent needs to develop utility-driven strategy forroutine application of NIRSDCS in clinic Ultimately weanticipate that the portable inexpensive NIRSDCS systemswill be widely commercialized for routine utilization inclinical rooms and patientrsquos bedside

Conflicts of Interest

+e authors declare that there are no conflicts of interest

Acknowledgments

+is work was partially supported by the National NaturalScience Foundation of China (61771433 and 61671413) Na-tional Key Research and Development Program of China(2016YFC0101605) National Key Scientific Instrument andEquipment Development Project of China (2014YQ24044508)and OIT Program of Shanxi Province

References

[1] Y Shang T Li and G Yu ldquoClinical applications of near-infrared diffuse correlation spectroscopy and tomography fortissue blood flow monitoring and imagingrdquo PhysiologicalMeasurement vol 38 no 4 pp R1ndashR26 2017

[2] T Durduran R Choe W B Baker and A G Yodh ldquoDiffuseoptics for tissue monitoring and tomographyrdquo Reports onProgress in Physics vol 73 no 7 article 076701 2010

[3] Y Shang L Chen M Toborek and G Yu ldquoDiffuse opticalmonitoring of repeated cerebral ischemia in micerdquo OpticsExpress vol 19 no 21 pp 20301ndash20315 2011

[4] L He Y Lin C Huang D Irwin M M Szabunio and G YuldquoNoncontact diffuse correlation tomography of human breasttumorrdquo Journal of Biomedical Optics vol 20 no 8 article086003 2015

[5] T Durduran R Choe G Yu et al ldquoDiffuse optical mea-surement of blood flow in breast tumorsrdquo Optics Lettersvol 30 no 21 pp 2915ndash2917 2005

[6] Y Hou Y Shang R Cheng et al ldquoObstructive sleep apnea-hypopnea results in significant variations in cerebral hemo-dynamics detected by diffuse optical spectroscopiesrdquo Physi-ological Measurement vol 35 no 10 pp 2135ndash2148 2014

[7] N Khan B B Williams H Hou H Li and H M SwartzldquoRepetitive tissue pO2 measurements by electron para-magnetic resonance oximetry current status and future po-tential for experimental and clinical studiesrdquo Antioxidants ampRedox Signaling vol 9 no 8 pp 1169ndash1182 2007

[8] F Jobsis ldquoNoninvasive infrared monitoring of cerebral andmyocardial oxygen sufficiency and circulatory parametersrdquoScience vol 198 no 4323 pp 1264ndash1267 1977

[9] M Wolf M Ferrari and V Quaresima ldquoProgress of near-infrared spectroscopy and topography for brain and muscle

clinical applicationsrdquo Journal of Biomedical Optics vol 12no 6 article 062104 2007

[10] M Ferrari M Muthalib and V Quaresima ldquo+e use of near-infrared spectroscopy in understanding skeletal musclephysiology recent developmentsrdquo Philosophical Transactionsof the Royal Society A Mathematical Physical and EngineeringSciences vol 369 no 1955 pp 4577ndash4590 2011

[11] Y Shang K Gurley B Symons et al ldquoNoninvasive opticalcharacterization of muscle blood flow oxygenation andmetabolism in women with fibromyalgiardquo Arthritis Researchamp 4erapy vol 14 no 6 p R236 2012

[12] L X Dong M Kudrimoti D Irwin et al ldquoDiffuse opticalmeasurements of head and neck tumor hemodynamics forearly prediction of chemoradiation therapy outcomesrdquoJournal of Biomedical Optics vol 21 no 8 article 0850042016

[13] N Roche-Labarbe S A Carp A Surova et al ldquoNoninvasiveoptical measures of CBV StO2 CBF index and rCMRO2 inhuman premature neonatesrsquo brains in the first six weeks ofliferdquoHuman Brain Mapping vol 31 no 3 pp 341ndash352 2009

[14] M E Tschakovsky J K Shoemaker and R L HughsonldquoBeat-by-beat forearm blood flow with Doppler ultrasoundand strain-gauge plethysmographyrdquo Journal of AppliedPhysiology vol 79 no 3 pp 713ndash719 1995

[15] J L Saumet D L Kellogg W F Taylor and J M JohnsonldquoCutaneous laser-Doppler flowmetry influence of underlyingmuscle blood flowrdquo Journal of Applied Physiology vol 65no 1 pp 478ndash481 1988

[16] G Yu T F Floyd T Durduran et al ldquoValidation of diffusecorrelation spectroscopy for muscle blood flow with con-current arterial spin labeled perfusion MRIrdquo Optics Expressvol 15 no 3 pp 1064ndash1075 2007

[17] C Jansen E M Vriens B C Eikelboom F E VermeulenJ van Gijn and R G Ackerstaff ldquoCarotid endarterectomywith transcranial Doppler and electroencephalographicmonitoring A prospective study in 130 operationsrdquo Strokevol 24 no 5 pp 665ndash669 1993

[18] D T Ubbink M J H M Jacobs G J Tangelder D W Slaafand R S Reneman ldquo+e usefulness of capillary microscopytranscutaneous oximetry and laser Doppler fluxmetry in theassessment of the severity of lower limb ischaemiardquo In-ternational Journal of Microcirculation vol 14 no 1-2pp 34ndash44 2008

[19] G Maret and P E Wolf ldquoMultiple light scattering fromdisordered media +e effect of brownian motion of scat-terersrdquo Zeitschrift fur Physik B Condensed Matter vol 65no 4 pp 409ndash413 1987

[20] C Cheung J P Culver K Takahashi J H Greenberg andA G Yodh ldquoIn vivocerebrovascular measurement combiningdiffuse near-infrared absorption and correlation spectros-copiesrdquo Physics in Medicine and Biology vol 46 no 8pp 2053ndash2065 2001

[21] D Irwin L Dong Y Shang et al ldquoInfluences of tissue ab-sorption and scattering on diffuse correlation spectroscopyblood flow measurementsrdquo Biomedical Optics Express vol 2no 7 pp 1969ndash1985 2011

[22] D R Busch J M Lynch M E Winters et al ldquoCerebral bloodflow response to hypercapnia in children with obstructivesleep apnea syndromerdquo Sleep vol 39 no 1 pp 209ndash216 2016

[23] L He Y Lin Y Shang B J Shelton and G Yu ldquoUsing opticalfibers with different modes to improve the signal-to-noiseratio of diffuse correlation spectroscopy flow-oximetermeasurementsrdquo Journal of Biomedical Optics vol 18 no 3article 037001 2013


[24] G Yu ldquoNear-infrared diffuse correlation spectroscopy incancer diagnosis and therapy monitoringrdquo Journal of Bio-medical Optics vol 17 no 1 article 010901 2012

[25] X Intes J Chen V Venugopal and F Lesage ldquoTime-resolveddiffuse optical tomography with patterned-light illuminationand detectionrdquo Optics Letters vol 35 no 13 pp 2121ndash21232010

[26] B Chance J S Leigh H Miyake et al ldquoComparison of time-resolved and -unresolved measurements of deoxyhemoglobinin brainrdquo Proceedings of the National Academy of Sciencesvol 85 no 14 pp 4971ndash4975 1988

[27] M Dehaes A Aggarwal P-Y Lin et al ldquoCerebral oxygenmetabolism in neonatal hypoxic ischemic encephalopathyduring and after therapeutic hypothermiardquo Journal of Cere-bral Blood Flow amp Metabolism vol 34 no 1 pp 87ndash94 2013

[28] J H Choi S Lee M Lee D Koh and B M Kim ldquoCerebralhemodynamic responses to seizure in the mouse brain si-multaneous near-infrared spectroscopy-electroencephalographystudyrdquo Journal of Biomedical Optics vol 15 no 3 article 0370102010

[29] G Yu T Durduran D Furuya J H Greenberg andA G Yodh ldquoFrequency-domain multiplexing system for invivo diffuse light measurements of rapid cerebral hemody-namicsrdquo Applied Optics vol 42 no 16 pp 2931ndash2939 2003

[30] K Yoshitani M Kawaguchi K Tatsumi K Kitaguchi andH Furuya ldquoA comparison of the INVOS 4100 and the NIRO300 near-infrared spectrophotometersrdquo Anesthesia amp Anal-gesia vol 94 no 3 pp 586ndash590 2002

[31] G Strangman M A Franceschini and D A Boas ldquoFactorsaffecting the accuracy of near-infrared spectroscopy con-centration calculations for focal changes in oxygenation pa-rametersrdquo Neuroimage vol 18 no 4 pp 865ndash879 2003

[32] Y Shang Y Zhao R Cheng L Dong D Irwin and G YuldquoPortable optical tissue flow oximeter based on diffuse cor-relation spectroscopyrdquo Optics Letters vol 34 no 22pp 3556ndash3558 2009

[33] Y Shang R Cheng L Dong S J Ryan S P Saha and G YuldquoCerebral monitoring during carotid endarterectomy usingnear-infrared diffuse optical spectroscopies and electroen-cephalogramrdquo Physics in Medicine and Biology vol 56 no 10pp 3015ndash3032 2011

[34] H Liu D A Boas Y Zhang A G Yodh and B ChanceldquoDetermination of optical properties and blood oxygenationin tissue using continuous NIR lightrdquo Physics in Medicine andBiology vol 40 no 11 pp 1983ndash1993 1999

[35] S R Arridge and W R B Lionheart ldquoNonuniqueness indiffusion-based optical tomographyrdquo Optics Letters vol 23no 11 pp 882ndash884 1998

[36] W Brown Dynamic Light Scattering 4e Method and SomeApplications Clarendon New York NY USA 1993

[37] A B Leung K I Suh and R R Ansari ldquoParticle-size andvelocity measurements in flowing conditions using dynamiclight scatteringrdquoApplied Optics vol 45 no 10 pp 2186ndash21902006

[38] D A Boas L E Campbell and A G Yodh ldquoScattering andimaging with diffusing temporal field correlationsrdquo PhysicalReview Letters vol 75 no 9 pp 1855ndash1858 1995

[39] D J Pine D A Weitz P M Chaikin and E HerbolzheimerldquoDiffusing wave spectroscopyrdquo Physical Review Lettersvol 60 no 12 pp 1134ndash1137 1988

[40] J Li M Ninck L Koban T Elbert J Kissler and T GislerldquoTransient functional blood flow change in the human brainmeasured noninvasively by diffusing-wave spectroscopyrdquoOptics Letters vol 33 no 19 pp 2233ndash2235 2008

[41] D A Boas and A G Yodh ldquoSpatially varying dynamicalproperties of turbid media probed with diffusing temporallight correlationrdquo Journal of the Optical Society of America Avol 14 no 1 pp 192ndash215 1997

[42] J P Culver T Durduran D Furuya C CheungJ H Greenberg and A G Yodh ldquoDiffuse optical tomographyof cerebral blood flow oxygenation and metabolism in ratduring focal ischemiardquo Journal of Cerebral Blood Flow ampMetabolism vol 23 no 8 pp 911ndash924 2016

[43] T Durduran G Yu M G Burnett et al ldquoDiffuse opticalmeasurement of blood flow blood oxygenation and meta-bolism in a human brain during sensorimotor cortex acti-vationrdquo Optics Letters vol 29 no 15 pp 1766ndash1768 2004

[44] G Yu T Durduran G Lech et al ldquoTime-dependent bloodflow and oxygenation in human skeletal muscles measuredwith noninvasive near-infrared diffuse optical spectros-copiesrdquo Journal of Biomedical Optics vol 10 no 2 article024027 2005

[45] U Sunar H Quon T Durduran et al ldquoNoninvasive diffuseoptical measurement of blood flow and blood oxygenation formonitoring radiation therapy in patients with head and necktumors a pilot studyrdquo Journal of Biomedical Optics vol 11no 6 article 064021 2006

[46] G Yu T Durduran C Zhou et al ldquoReal-time in situmonitoring of human prostate photodynamic therapy withdiffuse lightrdquo Photochemistry and Photobiology vol 82 no 5pp 1279ndash1284 2006

[47] U Sunar S Makonnen C Zhou et al ldquoHemodynamic re-sponses to antivascular therapy and ionizing radiationassessed by diffuse optical spectroscopiesrdquo Optics Expressvol 15 no 23 pp 15507ndash15516 2007

[48] C Zhou R Choe N Shah et al ldquoDiffuse optical monitoringof blood flow and oxygenation in human breast cancer duringearly stages of neoadjuvant chemotherapyrdquo Journal of Bio-medical Optics vol 12 no 5 article 051903 2007

[49] T Durduran C Zhou B L Edlow et al ldquoTranscranial opticalmonitoring of cerebrovascular hemodynamics in acute strokepatientsrdquo Optics Express vol 17 no 5 pp 3884ndash3902 2009

[50] C Zhou S A Eucker T Durduran et al ldquoDiffuse opticalmonitoring of hemodynamic changes in piglet brain withclosed head injuryrdquo Journal of Biomedical Optics vol 14no 3 article 034015 2009

[51] T Durduran C Zhou E M Buckley et al ldquoOptical mea-surement of cerebral hemodynamics and oxygen metabolismin neonates with congenital heart defectsrdquo Journal of Bio-medical Optics vol 15 no 3 article 037004 2010

[52] R CMesquita N Skuli M N Kim et al ldquoHemodynamic andmetabolic diffuse optical monitoring in a mouse model ofhindlimb ischemiardquo Biomedical Optics Express vol 1 no 4pp 1173ndash1187 2010

[53] B L Edlow M N Kim T Durduran et al ldquo+e effects ofhealthy aging on cerebral hemodynamic responses to posturechangerdquo Physiological Measurement vol 31 no 4 pp 477ndash495 2010

[54] M N Kim T Durduran S Frangos et al ldquoNoninvasivemeasurement of cerebral blood flow and blood oxygenationusing near-infrared and diffuse correlation spectroscopies incritically brain-Injured adultsrdquo Neurocritical Care vol 12no 2 pp 173ndash180 2009

[55] G Yu Y Shang Y Zhao R Cheng L Dong and S P SahaldquoIntraoperative evaluation of revascularization effect on is-chemic muscle hemodynamics using near-infrared diffuseoptical spectroscopiesrdquo Journal of Biomedical Optics vol 16no 2 article 027004 2011


[56] R Cheng Y Shang D Hayes S P Saha and G YuldquoNoninvasive optical evaluation of spontaneous low fre-quency oscillations in cerebral hemodynamicsrdquo Neuroimagevol 62 no 3 pp 1445ndash1454 2012

[57] L Dong M Kudrimoti R Cheng et al ldquoNoninvasive diffuseoptical monitoring of head and neck tumor blood flow andoxygenation during radiation deliveryrdquo Biomedical OpticsExpress vol 3 no 2 pp 259ndash272 2012

[58] K Gurley Y Shang and G Yu ldquoNoninvasive opticalquantification of absolute blood flow blood oxygenation andoxygen consumption rate in exercising skeletal musclerdquoJournal of Biomedical Optics vol 17 no 7 article 0750102012

[59] E M Buckley J M Lynch D A Goff et al ldquoEarly post-operative changes in cerebral oxygen metabolism followingneonatal cardiac surgery effects of surgical durationrdquo Journalof 4oracic and Cardiovascular Surgery vol 145 no 1pp 196ndash205 2013

[60] R C Mesquita M Putt M Chandra et al ldquoDiffuse opticalcharacterization of an exercising patient group with periph-eral artery diseaserdquo Journal of Biomedical Optics vol 18 no 5article 057007 2013

[61] R C Mesquita O K Faseyitan P E Turkeltaub et al ldquoBloodflow and oxygenation changes due to low-frequency repetitivetranscranial magnetic stimulation of the cerebral cortexrdquoJournal of Biomedical Optics vol 18 no 6 article 0670062013

[62] T Li Y Lin Y Shang et al ldquoSimultaneous measurement ofdeep tissue blood flow and oxygenation using noncontactdiffuse correlation spectroscopy flow-oximeterrdquo ScientificReports vol 3 no 1 p 1358 2013

[63] Y Shang Y Lin B A Henry et al ldquoNoninvasive evaluation ofelectrical stimulation impacts on muscle hemodynamics viaintegrating diffuse optical spectroscopies with muscle stim-ulatorrdquo Journal of Biomedical Optics vol 18 no 10 article105002 2013

[64] V Jain E M Buckley D J Licht et al ldquoCerebral oxygenmetabolism in neonates with congenital heart diseasequantified by MRI and opticsrdquo Journal of Cerebral Blood Flowamp Metabolism vol 34 no 3 pp 380ndash388 2013

[65] R Cheng Y Shang S Wang et al ldquoNear-infrared diffuseoptical monitoring of cerebral blood flow and oxygenation forthe prediction of vasovagal syncoperdquo Journal of BiomedicalOptics vol 19 no 1 article 017001 2014

[66] B Henry M Zhao Y Shang et al ldquoHybrid diffuse opticaltechniques for continuous hemodynamic measurement ingastrocnemius during plantar flexion exerciserdquo Journal ofBiomedical Optics vol 20 no 12 article 125006 2015

[67] W B Baker Z Li S S Schenkel et al ldquoEffects of exercisetraining on calf muscle oxygen extraction and blood flow inpatients with peripheral artery diseaserdquo Journal of AppliedPhysiology vol 123 no 6 pp 1599ndash1609 2017

[68] T S Ko C D Mavroudis W B Baker et al ldquoNon-invasiveoptical neuromonitoring of the temperature-dependence ofcerebral oxygen metabolism during deep hypothermic car-diopulmonary bypass in neonatal swinerdquo Journal of CerebralBlood Flow and Metabolism 2018 In press

[69] A T Eggebrecht S L Ferradal A Robichaux-Viehoeveret al ldquoMapping distributed brain function and networks withdiffuse optical tomographyrdquo Nature Photonics vol 8 no 6pp 448ndash454 2014

[70] Y Lin C Huang D Irwin L He Y Shang and G Yuldquo+ree-dimensional flow contrast imaging of deep tissue

using noncontact diffuse correlation tomographyrdquo AppliedPhysics Letters vol 104 no 12 article 121103 2014

[71] Y Shang ldquoAdvances in reconstruction algorithms for diffusecorrelation spectroscopy and tomographyrdquo in InternationalConference on Sensing and Imaging 2017 M Jiang N IdaA K Louis and E T Quinto Eds pp 15ndash28 Lecture Notesin Electrical Engineering 506 Springer Chengdu China2017

[72] D Boas Diffuse Photon Probes of Structural and DynamicalProperties of Turbid Media 4eory and Biomedical Applica-tions Department of Physics University of PennsylvaniaPhiladelphia PA USA 1996

[73] Y Shang and G Yu ldquoA Nth-order linear algorithm forextracting diffuse correlation spectroscopy blood flow indicesin heterogeneous tissuesrdquo Applied Physics Letters vol 105no 13 article 133702 2014

[74] Y Shang T Li L Chen Y Lin M Toborek and G YuldquoExtraction of diffuse correlation spectroscopy flow index byintegration of Nth-order linear model with Monte Carlosimulationrdquo Applied Physics Letters vol 104 no 19 article193703 2014

[75] A E Cerussi D Jakubowski N Shah et al ldquoSpectroscopyenhances the information content of optical mammographyrdquoJournal of Biomedical Optics vol 7 no 1 pp 60ndash71 2002

[76] Y Lin L He Y Shang and G Yu ldquoNoncontact diffusecorrelation spectroscopy for noninvasive deep tissue bloodflow measurementrdquo Journal of Biomedical Optics vol 17no 1 article 010502 2012

[77] C Huang J P Radabaugh R K Aouad et al ldquoNoncontactdiffuse optical assessment of blood flow changes in head andneck free tissue transfer flapsrdquo Journal of Biomedical Opticsvol 20 no 7 article 075008 2015


International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018


Active and Passive Electronic Components

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of



Journal ofEngineeringVolume 2018

SensorsJournal of



RotatingMachinery


Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation


Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

oxygenation allows for estimate of the oxygen metabolic rate[11 13] a valuable parameter closely associated with tissuepathology

At present the low-cost modalities to noninvasively andfast measure the blood flow in deep tissue are not estab-lished Methods currently available for assessment of bloodflow include Doppler ultrasound [14] laser Doppler [15]and perfusion-weighted magnetic resonance imaging (per-fusion MRI) [16] Doppler ultrasound can only assess theblood velocity in single and large vessel [14 17] Manydiseases however are relevant to the abnormal blood flow inmicrovasculature level rather than macrovasculature level[1 2] While perfusion MRI can provide the microvascu-lature blood flow at high spatial resolution the high cost andlow mobility of MRI instrument preclude its wide use forroutine disease screening Laser Doppler can only detectblood flow at superficial tissue (sim1mm depth) [18] thusunable to probe the diseases mostly originated from deeptissues In summary the routine use of the above techniquesis limited due to availability expense and the difficulty ofmaking continuous measurements

In recent years a dynamic NIRS technology namelydiffuse correlation spectroscopy (DCS) has gained rapiddevelopment [19ndash23] DCS quantifies the speckle fluctua-tions of diffuse light which are sensitive to the motion of redblood cells in the microvasculature thus permitting fastassessment of blood flow in deep tissues (up to severalcentimeters)

+e article will briefly describe the technology and de-vices for NIRS and DCS as well as their applications tophysiological studies and clinic +e extension of NIRS fromspectroscopy to imaging was depicted followed by in-troductions of advanced algorithms that were recentlyproposed +e future prospective of the NIRS and DCS andtheir feasibilities for routine use in hospital was finallydiscussed

2 NIRS Principle and Devices

When being injected into the biological tissues that areconsidered as turbid medium the near-infrared lighttransports through the medium in ldquodiffusiverdquo manner [2]due to tissue absorption (mainly red blood cells) andscattering (mainly organelles and mitochondria) [1 24] +eaim of NIRS technology is to separate the tissue absorptionfrom scattering through which the hemoglobin concen-tration is derived Over the years two approaches namelythe time-resolved spectroscopy [25 26] and frequency-modulated spectroscopy [27 28] have been developed toaccount for this separation For the time-resolved approacha pulsed laser (a few of picoseconds) is adopted and the time-of-flight of photons is recorded by a digital time-correlatedsingle photon counting (TCSPC) board For frequency-modulated approach the light is modulated at high fre-quency (around 100MHz) [29] so that the phase ofmodulated light escaping out from the tissue could be traced

Both these two approaches were commercialized withnumerous applications [30] particularly for neurologicalstudies Either the time-resolved or frequency-modulated

approach however requires complicated hardware to ma-nipulate the light substantially increasing the cost ofinstrument

In practice the NIRS was often simplified by usingcontinuous-wave (CW) light instead of pulsed or frequency-modulated light In CW oximeter the tissue scattering (μs)was assumed to be a constant number and only the relativechange of tissue absorption coefficient needs to be de-termined by the modified BeerndashLambert law

lnI t0( 1113857

I(t)1113888 1113889 Δμa(λ) middot L

εHbO(λ)Δ HbO21113858 1113859 + εHb(λ)Δ[Hb]( 1113857 middot L

(1)

where I(t0) and I(t) are the light intensity detected at time t0and t respectively Δμa(λ) is the relative change of theabsorption coefficient at wavelength λ and L is the meanphoton length between a specific source-detector (S-D) pairBy collecting the light at two wavelengths the oxygenationchanges can be calculated as [31ndash33]

Δ HbO21113858 1113859 εHb λ1( 1113857Δμa λ2( 1113857minus εHb λ2( 1113857Δμa λ1( 1113857


Δ[Hb] εHbO λ2( 1113857Δμa λ1( 1113857minus εHbO λ1( 1113857Δμa λ2( 1113857


(2)

where Δ[HbO2] and Δ[Hb] are the relative change of oxy-and deoxy-hemoglobin concentration and εHbO(λ) andεHb(λ) are their extinction coefficients respectively

Figure 1 illustrates a typical scheme of CW oximeterBriefly a dual-wavelength LED at NIR range was controlledby transistor-transistor logic (TTL) signals injecting thephotons at each wavelength into the tissue alternately +eescaped photons were collected in parallel by all opticaldetectors and transferred to the computer wherein themodified BeerndashLambert law was applied yielding time-course oxygenation data

+e reason for using CW light as the source light of NIRStechnology is that the tissue scattering (majorly contributedfrom organelles and mitochondria) changes much less thanthe tissue absorption in most of physiological status +usthe tissue blood oxygenation would still been properlyextracted by the CW tissue oximeter Moreover this low-cost device could minimize the cross-talk between the ab-sorption and scattering

+e CW tissue oximeter was also commercialized bydifferent manufacturers (eg NIRO Hamamatsu IncINVOS Somanetics Inc) and has various clinical applica-tions At CW working mode the absolute value of tissueabsorption coefficient (μa) could be extracted by the near-infrared light collected from multiple source-detector (S-D)separations according to the theory of spatial-resolvedspectroscopy [34] +is theory permits generation of aslope through multiple S-D separations of optical data fromwhich the tissue absorption will be determined In generalmore S-D separations perform better in calculating ab-sorption coefficient (μa) with stable and accurate values Forexample four S-D separations (20 to 35 cm) were used in a











Control panel

AD board


TTL Light intensity

Injected photons

Escaped photons

Tissue


Control panel

AD board



TTL

Injected photons

Escaped photons

Digital correlator










TTLswitch

40

(b)

Probe



Optical detectors

Control panel

(a)

(c)


400

300

200

100

0


Post-revascularizan

Arterial occlusion

0830 0900 0930 1000 1030 1100

0830 0900 0930 1000 1030 1100

20

0

ndash20

ndash40

rBF

()

microM

∆[HbO2]∆[Hb]

Time















































i1A(i j)αDB(i) (3)

g1(τ j)minus 1

minus 1113944N

k2

1113936Qp1w(p j) minus21113936

ni1k


prime(i)1113872 1113873k

kτk

τ 1113944n

i1A(i j)αDB(i)

(4)


A(i j) 1113944

Q

p1minus2w(p j)k

20(i)s(i p j)μs

prime(i) (5)



OD1

ODN

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠

s11 s1M

sN1 sNM

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠

Δμa1

ΔμaM

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠ (MleN)

(6)



OD sΔμa (7)


Δμa sminus1OD (8)


Δμa sTs1113872 1113873minus1

sTOD (9)


Δμa sTs + α2I1113872 1113873

minus1s

TOD (10)




















Acknowledgments


References





















































































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia











Control panel

AD board


TTL Light intensity

Injected photons

Escaped photons

Tissue


Control panel

AD board



TTL

Injected photons

Escaped photons

Digital correlator










TTLswitch

40

(b)

Probe



Optical detectors

Control panel

(a)

(c)


400

300

200

100

0


Post-revascularizan

Arterial occlusion

0830 0900 0930 1000 1030 1100

0830 0900 0930 1000 1030 1100

20

0

ndash20

ndash40

rBF

()

microM

∆[HbO2]∆[Hb]

Time















































i1A(i j)αDB(i) (3)

g1(τ j)minus 1

minus 1113944N

k2

1113936Qp1w(p j) minus21113936

ni1k


prime(i)1113872 1113873k

kτk

τ 1113944n

i1A(i j)αDB(i)

(4)


A(i j) 1113944

Q

p1minus2w(p j)k

20(i)s(i p j)μs

prime(i) (5)



OD1

ODN

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠

s11 s1M

sN1 sNM

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠

Δμa1

ΔμaM

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠ (MleN)

(6)



OD sΔμa (7)


Δμa sminus1OD (8)


Δμa sTs1113872 1113873minus1

sTOD (9)


Δμa sTs + α2I1113872 1113873

minus1s

TOD (10)




















Acknowledgments


References





















































































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia








TTLswitch

40

(b)

Probe



Optical detectors

Control panel

(a)

(c)


400

300

200

100

0


Post-revascularizan

Arterial occlusion

0830 0900 0930 1000 1030 1100

0830 0900 0930 1000 1030 1100

20

0

ndash20

ndash40

rBF

()

microM

∆[HbO2]∆[Hb]

Time















































i1A(i j)αDB(i) (3)

g1(τ j)minus 1

minus 1113944N

k2

1113936Qp1w(p j) minus21113936

ni1k


prime(i)1113872 1113873k

kτk

τ 1113944n

i1A(i j)αDB(i)

(4)


A(i j) 1113944

Q

p1minus2w(p j)k

20(i)s(i p j)μs

prime(i) (5)



OD1

ODN

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠

s11 s1M

sN1 sNM

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠

Δμa1

ΔμaM

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠ (MleN)

(6)



OD sΔμa (7)


Δμa sminus1OD (8)


Δμa sTs1113872 1113873minus1

sTOD (9)


Δμa sTs + α2I1113872 1113873

minus1s

TOD (10)




















Acknowledgments


References





















































































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia














































i1A(i j)αDB(i) (3)

g1(τ j)minus 1

minus 1113944N

k2

1113936Qp1w(p j) minus21113936

ni1k


prime(i)1113872 1113873k

kτk

τ 1113944n

i1A(i j)αDB(i)

(4)


A(i j) 1113944

Q

p1minus2w(p j)k

20(i)s(i p j)μs

prime(i) (5)



OD1

ODN

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠

s11 s1M

sN1 sNM

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠

Δμa1

ΔμaM

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠ (MleN)

(6)



OD sΔμa (7)


Δμa sminus1OD (8)


Δμa sTs1113872 1113873minus1

sTOD (9)


Δμa sTs + α2I1113872 1113873

minus1s

TOD (10)




















Acknowledgments


References





















































































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia
















Acknowledgments


References





















































































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia





























































RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


PortableNear-InfraredTechnologiesandDevicesfor...

Documents

Transcript of PortableNear-InfraredTechnologiesandDevicesfor...