Which neuroimaging method helps provide information about brain function by monitoring oxygen levels in the blood flow?

Near-Infrared Spectroscopy (NIRS)

Near-infrared spectroscopy (NIRS), which is a noninvasive tool, gives a venous weighted estimate of the regional cerebral oxygen saturation (rSO2). NIRS has been used to test for adequate brain protection during aortic arch surgery under deep hypothermic circulatory arrest (DHCA).71 NIRS changes preceded changes in ICP in patients having delayed traumatic hematomas.72 In patients with carotid artery occlusion, oxyhemoglobin saturation at rest measured by NIRS could discriminate symptomatic from asymptomatic patients.73 NIRS has been used to test CBF autoregulation too.74 An NIRS-based index, called total hemoglobin reactivity, was correlated with similar indices derived from ICP waveform75 and TCD.76

Abstract

Near-infrared spectroscopy (NIRS) is an emerging noninvasive monitoring modality based on chromophore absorption of infrared light. There are four different types of NIRS system—continuous, time domain, frequency domain, and functional—and numerous devices that are now commercially available. These devices differ by the use of different technical components to obtain continuous NIRS values. Their resulting values are not necessarily interchangeable, and few studies have compared them.

Cerebral NIRS is an indirect indicator of perfusion adequacy. Therefore, it allows continuous information on oxygen supply-versus-demand balance. NIRS is indicated when continuous monitoring of cerebral tissue perfusion is considered. Therefore, the application of NIRS covers many procedures including cardiac and noncardiac surgeries but also applications in intensive care and emergency medicine in other clinical and research areas. Furthermore, there is a growing interest in the use of somatic NIRS because both cerebral and somatic desaturation have been associated with worse outcome. The somatic component could serve as an earlier warning of impaired perfusion. There are several limitations in the use of NIRS monitoring and pitfalls that are discussed in detail. As any technology, NIRS is being used as part of a multimodal strategy because abnormal NIRS indicates a perfusion problem but does not provide the precise mechanism. Future studies should explore the impact of this approach on clinical outcome.

Near-Infrared Spectroscopy

NIRS provides a real-time continuous measurement of regional cerebral blood oxygenation and indirect blood flow. NIRS uses the principle of reflectance spectrophotometry.52

NIRS has been used to measure rSO2 in a variety of clinical conditions. In patients with SAH it is used during clipping of aneurysm, early detection of vasospasm. It is also used for managing head-injured patients, carotid endarterectomy, cardiac surgery during cardiopulmonary bypass. There are limitations of NIRS use. The absorption by extra skull tissues, presence of hematomas, other pigments such as melanin, bilirubin can give false values. The skull thickness as well as hemoglobin concentration also interferes with the accuracy of monitoring.

D Near-Infrared Spectroscopy

Near-infrared spectroscopy (NIRS) is a technique based on two principles. First, light in the near-infrared zone can pass through the thin skin, bone, and other tissues of the infant. Second, the appropriate choice of near-infrared wave-lengths allows interpretation of changes in light absorption that reflect oxygenation. NIRS can provide crucial information on cerebral hemoglobin oxygen saturation, cerebral blood flow, cerebral blood volume, cerebral oxygen delivery, cerebral venous oxygen saturation, and cerebral oxygen utilization.29 No ionizing radiation is employed, the apparatus is portable, and continuous monitoring is possible, which makes it a practical tool for assessing newborns. Although NIRS is still primarily used as a research tool, several studies have focused on obtaining reference ranges for certain NIRS measures for preterm and term infants.99–101 NIRS has also been used to study cortical pain responses in neonates.102 In addition, functional NIRS (fNIRS) has been developed to examine brain activation. NIRS and functional NIRS have been applied to healthy and at risk neonates.103,104

Near-Infrared Spectroscopy

NIRS techniques allow the temporal measurement of chromophore (absorbing molecules) concentration changes within tissue. The chromophores of primary interest for NIRS techniques are oxyhemoglobin (HbO2), deoxyhaemoglobin (HHb) and cytochrome c oxidase, as the concentration of these varies with oxygenation and metabolic status of tissue (Elwell, 1995). Each chromophore has a unique absorption spectrum, where the specific extinction coefficient (ε) is expressed as a function of the wavelength (Horecker, 1943; Pellicer and Bravo, 2011). Concentration changes are calculated using an attenuation change of the measurement of light (typically using 2 or more wavelengths, between 650 and 950 nm) and by solving a linear equation based on the modified Beer-Lambert law. These temporal measurements are then used to calculate concentration changes in blood volume (total hemoglobin, tHb = HbO2+HHb), allowing the assessment of tissue oxygenation (saturation/perfusion as the ratio of HHb/tHb), a measure of oxidative metabolism (cytochrome C), and hence inferring blood flow.

Three main types of NIRS systems have been developed, continuous wave (CW), frequency domain (FD) and time domain (TD) NIRS, each with various levels of complexity and distinct types of information they can retrieve (Bakker et al., 2012; Ferrari and Quaresima, 2012; Lange and Tachtsidis, 2019). The key differences between these technologies is linked to how the light is emitted and detected, either with or without a measurement of the light scattering properties of the photon as it travels through the tissue (see Fig. 3B). The CW-NIRS instrument is the most simplistic of these, relying solely on measuring the attenuation of light from the source emitter to the photodiode detector. The light scatter is assumed to be homogenous between the emitter and detector points, using a standardized parameter to estimate the distance light has traveled (i.e., the differential pathlength factor). Consequently, CW-NIRS is a fast and relatively inexpensive method, yet cannot provide absolute chromophore concentration values as there is no way to distinguish between intensity changes due to scatter and changes due to absorption. This means that hemodynamic data derived from CW-NIRS systems are taken as relative change values from a preceding “normal” reference baseline, which makes baseline hemoglobin volume comparisons between participants (and studies) problematic. Furthermore, the lack of information about where the light has traveled between the emitter and detector means that the attenuated light detected will be affected by all tissue/fluid traveled through [i.e., extracranial (skin, skull) and intracranial (CSF, cortex)]. To address this issue, systems and algorithms such as spatial resolved spectroscopy have been developed that analyze the light signal from a series of multi-distance (3–4 cm) detectors (Suzuki et al., 1999). This allows for some separation of the superficial signal from that of deeper tissue signal changes via detector-weighted reconstructed parameters, providing a relative measure of tissue oxygenation (TOI) and hemoglobin (THI) indices in deeper tissues (i.e., 2–3 cm from the detector placement on the surface).

FD-NIRS systems use a similar light source as CW, however the intensity is modulated at very high frequencies (MHz), typically at 100 MHz. This high frequency modulation allows measurement of both light attenuation and phase shift (Fig. 3B) to provide a more quantified assessment of the tissue optical properties. The phase change of light is analogous to the mean-time of flight of photons and therefore indicates the level of scattering the light has undergone and therefore helps distinguish between intensity changes from scatter and absorption events (Pogue et al., 1997); although deriving absolute values is still problematic and requires system calibration against known phantoms and models. To support this methodology, the required detectors are more complex and thus expensive relative to the CW systems.

TD-NIRS is the most complicated type but gives the greatest level of depth level sampling for the emitted light. The highly specialized detectors count photons from repeated measures due to a single laser pulse (10s of picoseconds) over time, allowing more accurate estimations of light propagation through the imaged tissue. Specifically, the spread of the measured light attenuation (temporal spread point function) indicates the amount of both absorption and scatter the light as undergone, with the photons arriving later assumed to have traveled further and thus deeper – which is particularly important for brain targeted measures when trying to distinguish intracranial from extracranial sources for chromophore measures. However, the highly specialized light source and detector instrumentation makes the TD systems significantly more expensive and are therefore not widely used, which is reflected by their limited commercial availability.

Finally, broadband NIRS is an emerging technique that allows the recovery of multiple parameters via the measurement of the NIR intensity over a wide range of wavelengths. Using the full spectrum, instead of limited wavelengths (typically 3 in CW-NIRS), it is possible to recover multiple parameters through the use of spectral unmixing, and potentially providing more sensitivity and specificity to different biomarkers; including cytochrome C oxidase, a measure of mitochondria metabolism (Bale et al., 2016). While cytochrome C oxidase is not a direct measure of flow per se, its utility for brain blood flow assessment lies in a better understanding of the underlying neuronal activity and therefore the demand for blood flow. Thus, similar to metabolic measures collected alongside BOLD MRI, a more complete picture of the neurovascular coupling process may be possible when combining NIRS measures of cytochrome C oxidase with those of oxygenated and deoxygenated hemoglobin. Indeed, the benefit of assessing derived parameters collectively is an important consideration for any physiological interpretation of the cerebral hemodynamics being assessed, as shown in Fig. 3C from the 5 parameters collected from the NIRO-200NX CW-NIRS system.

The term functional NIRS (i.e., fNIRS) is commonly used to describe how this methodology is applied to brain blood flow measurement. This functional term also acknowledges the technique's limitation to measuring relative changes in oxygenated and deoxygenated hemoglobin from cortical tissue in response to a stimulus (e.g., cognitive tasks, exercise or carbon dioxide). Despite this limitation, the functional hemodynamic responses derived from NIRS has been recognized as a useful tool in clinical applications including stroke (Yang et al., 2019), traumatic brain injury (Davies et al., 2015), and mild cognitive impairment and dementia (Yeung and Chan, 2020). Similar to MRI and TCD approaches, fNIRS can also determine cerebrovascular responsiveness to increased CO2 blood content (Miller and Mitra, 2017; Gratton et al., 2020); with the high temporal resolution of NIRS potentially being useful in targeting the kinetics of this hemodynamics response to the CO2 stimulus (Gratton et al., 2020). Common to many of these applied studies is the lack of a precise signal localization from the prefrontal cortex over which the detectors are positioned (see Fig. 3A and B). However, methods have been developed to address this issue by co-registering locations of signal with structural MRI (Gratton and Fabiani, 2009), or through the use of high-density measurement to allow full 3D tomographic recovery of these functional maps which are shown to be comparable to MRI (Eggebrecht et al., 2014).

In summary, NIRS is an attractive tool for assessing brain function due to its hemodynamic measurement basis. Other advantages of NIRS technology compared with other neuroimaging modalities include its portability, high temporal resolution (up to 40  Hz), ability to record in natural settings and relatively low cost. Long periods of recording time are achievable, it is less sensitive to head motion, safe to use and can be used easily on infants, children and clinical cohorts. However, NIRS has lower spatial resolution, lacks standard analyzing packages (though this drawback also applies to other imaging modalities) and activation and registration procedures may be inaccurate and lead to poor localization. Despite these drawbacks, the advantages of NIRS are making this technology an increasingly useful tool to assess brain health in disease states and the consideration of hemodynamic associations with cognitive function.

Near-Infrared Spectroscopy25,26

Cerebral near-infrared spectroscopy (NIRS) is an optical, noninvasive method for indirect estimation of brain oxygenation and perfusion.

NIRS for cerebral monitoring was first described by Ferrari et al., in 1985, and the first commercial available devices were introduced about 10 years later.

This technique makes use of a laser, emitting light wave frequencies within 600 and 1000 nm to enlighten the brain tissue through the skin and the skull (Fig. 2.5). The choice of this precise spectrum is done to get the maximum penetration inside the tissue. Hence the near-infrared light is transmitted several centimeters inside the brain, where it is partly absorbed by the chromophores (mainly hemoglobin) and partly scattered. Even if the scattering introduces an additional complexity, at the same time, it is also useful. In fact the brain is too large to permit transillumination, but since part of the light is scattered back to the surface, a detector placed 4–7 cm away from the emitter is able to receive this light. From this signal, by the modification of the Lambert–Beer law (to consider the scattering phenomenon), it is possible to calculate in real time the relative concentration of chromophores inside the analyzed tissue.

Recently, new algorithms based on multiple detectors (spatially resolved NIRS) have been developed, and thanks to this technological progress, it is now possible to obtain data that gives a value of the regional saturation of O2 (rSO2). So, in a nutshell, we could compare this technique to the almost ubiquitous pulse oxymetry techniques.

Because NIRS evaluates all the chromophores within the range of the emitted light ray, the obtained value is an average value between arterial, venous, and capillary saturation; as the venous blood takes almost 70% of the cerebral volume, the changes in the venous saturation are the prevalent component in the received signal. Normal values of rSO2 have been established in 60%–80%, and a decrease below 47% has been associated with ischemia. This value has a strong correlation with other CBF monitoring techniques (EEG, TCD, and clinical assessment).

NIRS is a promising technique: in fact, it is easy to use, noninvasive, cheap enough, and has a high sensibility. The main drawbacks are the still unknown contribution to the signal from extracranial tissues and the fact that the given values are relative. Finally, practical problems are limiting the use of NIRS in an ICU setting, as scalp and facial traumas may prevent the application of the optodes, and brain lesions, like intraparenchymal hematomas, may lead to unpredictable values.

Jugular Venous Oxygen Saturation

Near infrared spectroscopy (NIRS) combined with transcranial Doppler ultrasonography (TCDU) and jugular venous oxygen saturation (SjvO2) monitoring are means for appreciating brain hemodynamics and oxygen delivery. SjvO2 monitoring requires the retrograde placement of a catheter into a jugular vein, up to the jugular bulb, to monitor the mixed cerebral venous blood and above the C1/C2 level to limit contamination by the facial venous blood.43 Side for placement should preferably be the dominant side of the brain or the side with the most prominent pathology, if any. Oxygen saturation may be measured in regularly drawn blood samples or using a fiber-optic catheter and light absorption in the red/infrared wavelength range. Although generally quite easily placed, the catheter may be the source of rare complications, including hematoma, infection, thrombosis, raised ICP, or incidental wrong vessel puncture.44

SjvO2 reflects the balance between oxygen delivery to the brain and cerebral metabolic consumption of oxygen. Normal values range between 50% and 75%. Decreases in SjvO2 can be seen in case of systemic or local oxygen supply deficiency, including hypoxia, low blood pressure, decreased CPP, embolism, or vasospasm (Table 9.2), and in case of increased oxygen consumption including hyperthermia and seizure. Causes of SjvO2 values above normal range can be classified into those related to a decrease in cerebral metabolism, restricted oxygen diffusion or extraction, shunting, increased oxygen supply, or a combination of these. For example, a global decrease in cerebral metabolism can be observed during hypothermia. Infarction or inflammation causes restriction in oxygen diffusion and extraction, as well as microvascular shunting. Increased oxygen supply is observed during hyperoxia, and hypercarbia can be responsible for substantial shunting.

Table 9.2. Causes of Abnormal Jugular Venous Oxygen Saturation Values and Proposed Diagnosis Help, and Corrective Measures

ABP, arterial blood pressure; BGA, blood gas analysis; CO, cardiac output; CPP, cerebral perfusion pressure; EEG, electroencephalography; FiO2, inspired oxygen fraction; ICP, intracranial pressure; SjvO2, jugular venous oxygen saturation; SpO2, peripheral saturation in oxygen; TCDU, transcranial Doppler ultrasonography.

Interpretation of SjvO2 values may be uneasy in some instances. It is a global measure and it may miss eventually occurring focal brain ischemia.45 When brain tissue death is massive, SjvO2 can be high as a consequence of shunting. Classical causes of wrong SjvO2 values are displacement of the catheter, thrombus around its tip, and contamination by facial venous blood. Anesthesia may also have an influence on SjvO2 values. Transient desaturations have been reported under propofol anesthesia, more frequently in normothermic than in hypothermic patients,46 while SjvO2 increases have been observed under sevoflurane anesthesia.47 This is due to the different properties of those agents on cerebral vasculature, propofol being vasoconstrictive, and sevoflurane vasodilating. Side of recording may also be important, insofar as cerebral venous drainage usually predominate on one side through the dominant jugular bulb.48

SjvO2 has mostly been used to guide therapy in the intensive care unit for traumatic brain injury and subarachnoid hemorrhage patients. However, its intraoperative use during intracranial neurosurgery allows commonly detecting desaturation episodes,48 and its combination with TCDU49 may impact on clinical decision-making to optimize cerebral physiology. Large randomized clinical trials investigating the impact of intraoperative SjvO2 monitoring on patient outcome are lacking.

Exercise physiology

Impaired oxygen extraction by exercising muscle can be detected by near-infrared spectroscopy or venous blood gas measurements after forearm nonischemic testing. Both methods can detect the degree of deoxygenation of hemoglobin. PO2 rises paradoxically in patients with mitochondrial myopathy or PEO, and the degree of rise reflects the severity of oxidative impairment.31,32 By 31P-magnetic resonance spectroscopy, the ratio of phosphocreatine to inorganic phosphate (PCr to Pi) can be measured in muscle at rest, during exercise, and during recovery. In patients with mitochondrial dysfunction, PCr-to-Pi ratios are lower than normal at rest, decrease excessively during exercise, and return to baseline more slowly than normal.33,34 However, exercise magnetic resonance spectroscopy (MRS) should be used by specialized centers.

8.3.2 Blood flow

One of the easiest changes to measure in animals, or especially in humans, that occurs after tPBM, is changes in cerebral blood flow and changes in oxygenation. This is because near-infrared spectroscopy is a noninvasive technique that has shown rapid growth in recent years. In fact, Wang et al. (2016) carried out this measurement on the forearms of human volunteers treated with a 1064 nm laser. They found that PBM induced significant increases of CCO concentration (Delta[CCO]) and oxygenated hemoglobin concentration (Delta[HbO]) on the treated site as the laser energy dose accumulated over time. A strong linear relationship between Delta[CCO] and Delta[HbO] was observed indicating a response in both oxygen supply and blood volume. Schiffer et al. (2009) tested tPBM using an 810 nm LED applied to the forehead for major depression and anxiety, and used an INVOS commercial system from Somanetics (Troy, MI) to measure cerebral hemoglobin (cHb) in left and right frontal and rCBF, in addition to the device’s usual oxygen saturation output.

It has been suggested that the release of NO as a result of PBM is responsible for the increased cerebral blood flow (Lee et al., 2017). NO is a major neuronal signaling molecule which, among other functions, possesses the ability to trigger vasodilation. To do so, it first stimulates soluble guanylate cyclase to form cyclic-GMP (cGMP). The cGMP then activates protein kinase G, leading to the reuptake of Ca2+ and the opening of calcium-activated potassium channels. Due to the subsequent fall in concentration of Ca2+, myosin light-chain kinase is prevented from phosphorylating the myosin molecule, causing the smooth muscle cells in the lining of blood vessels and lymphatic vessels to become relaxed (Charriaut-Marlangue et al., 2013). This vasodilation then promotes improved circulation, which in turn leads to improved cerebral oxygenation in a similar manner to that observed with pulsed electromagnetic fields (Bragin et al., 2015).

Brain Monitoring

Advances in brain monitoring techniques and the development of novel technologies have not only improved our understanding of both the healthy and injured brain but also allow for individualized patient management. We briefly review a few of these techniques.

Near-infrared spectroscopy (NIRS) utilizes the transparency of the scalp and skull to infrared light and the differences in absorption spectra between oxyhemoglobin and deoxyhemoglobin to quantify the local oxygen saturation of hemoglobin in the brain. The NIRS probe consists of a light source, but unlike pulse oximetry that consists of one detector, the NIRS consists of two detectors (Fig. 56.2). Light entering the detector closest to the source has passed predominantly through the scalp and skull, whereas the light entering the detector farther from the source has passed through the scalp, skull, and brain parenchyma. Comparing the differences in spectra obtained from the two detectors allows for estimation of brain tissue oxygen saturation minimizing the contribution from superficial tissues. However, current NIRS technology is unable to completely negate extracerebral contributions.51 Normal brain tissue regional oxygen saturation is 60–80% and will decrease with either increased brain oxygen consumption or decreased oxygen delivery. The latter can be due to low arterial oxygen content (e.g., low PaO2, anemia) or decreased cerebral perfusion secondary to factors such as brain edema, elevated intracranial pressure, vasospasm, vascular occlusion, or inadequate collateral blood flow. This technology is limited in that it only allows for the assessment of regional and superficial brain oxygen saturation. NIRS has been used to assess changes in cerebral oxygen delivery with patient position changes,52–54 as a possible monitor to guide selective shunting during carotid endarterectomy,55 and as a possible tool to guide anesthetic management to improve perioperative cognitive dysfunction in the elderly.56

Implantation of an oxygen sensor into brain parenchyma allows for the measurement of the local brain tissue partial pressure of oxygen (PbtO2). In normal brain, oxygen tension is 25–35 mmHg. Factors similar to those that precipitate a decrease in NIRS oxygen saturation can decrease local oxygen tension (i.e., increased metabolism, decreased arterial oxygen content, decreased cerebral perfusion). The probe is often placed in white matter to minimize potential risk to neuronal cell bodies. The gross location of the probe tip can potentially impact the reliability of monitoring in patients with focal injuries. For example, if the probe is placed distant from the injured or at-risk parenchyma, PbtO2 may be normal despite ongoing cellular hypoxia at the injury site. If the probe is placed within a necrotic region, physiologic changes are not likely to impact PbtO2. PbtO2 has been shown to be a potentially useful tool to guide outcome in patients with severe traumatic brain injury.57

Cerebral microdialysis allows for effective sampling of substances in the local extracellular fluid compartment in the brain. Common analytes include glycerol (a component of cell membrane lipids and liberated with cell lysis), glutamate (an excitatory neurotransmitter), and lactate (a marker of anaerobic metabolism). Lactate concentration is often corrected for the total amount of substrate and expressed as the lactate-to-pyruvate ratio where a value >40 suggests cerebral metabolic crisis. Other analytes can include glucose, other neurotransmitters, and potentially β-amyloid.58 As with PbtO2 monitoring, the reliability of the data obtained from cerebral microdialysis depends on the location of the tip. Cerebral microdialysis has been used to individualize management of patients with traumatic brain injury or subarachnoid hemorrhage59 and better understand the physiology of the healthy and injured brain. For example, cerebral microdialysis was recently used to demonstrate that euglycemia following traumatic brain injury may impair glucose availability to the injured brain.60 Simultaneous monitoring of multiple parameters (i.e., “multimodal monitoring”) can increase the opportunity to minimize secondary insults following brain injury.61

Assessment of the integrity of cerebral autoregulatory capacity of the brain following injury can allow for individualization of perfusion pressure targets to optimize oxygen and substrate delivery while minimizing risk for injury from hypoperfusion and hyperperfusion. With intact autoregulation, increases in systemic blood pressure should induce cerebral vasoconstriction and decrease intracranial pressure, whereas with frank autoregulatory failure and pressure passive flow, blood pressure and intracranial pressure should have a strong positive correlation. As a correlation coefficient can assume values between −1 and +1, lower values suggest intact autoregulatory capacity, whereas values closer to +1 implicate autoregulatory failure.62 Other metrics such as PbtO2,63 blood flow velocity in the cerebral vasculature determined by transcranial Doppler sonography,64 and regional cerebral oxygen saturation via NIRS65 can also be used in lieu of intracranial pressure to assess autoregulatory capacity. Autoregulatory failure as assessed by these techniques can be used to guide hemodynamic management optimizing cerebral perfusion and as a marker for poor outcome following brain injury.66–68