Open Access
Add to BibSonomy
Abstract
A methodology for the assessment of the cerebral hemodynamic reaction to normotensive hypovolemia, reduction in cerebral perfusion and orthostatic stress leading to ischemic hypoxia and reduced muscular tension is presented. Most frequently, the pilots of highly maneuverable aircraft are exposed to these phenomena. Studies were carried out using the system consisting of a chamber that generates low pressure around the lower part of the body - LBNP (lower body negative pressure) placed on the tilt table. An in-house developed 6-channel NIRS system operating at 735 and 850 nm was used in order to assess the oxygenation of the cerebral cortex, based on measurements of diffusely reflected light in reflectance geometry. The measurements were carried out on a group of 12 active pilots and cadets of the Polish Air Force Academy and 12 healthy volunteers. The dynamics of changes in cerebral oxygenation was evaluated as a response to LBNP stimuli with a simultaneous rapid change of the tilt table angle. Parameters based on calculated changes of total hemoglobin concentration were proposed allowing to evaluate differences in reactions observed in control subjects and pilots/cadets. The results of orthogonal partial least squares-discriminant analysis based on these parameters show that the subjects can be classified into their groups with 100% accuracy.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Training of military pilots and astronauts is aimed to prevent a G-force induced loss of consciousness [1,2]. High acceleration of the body in longitude direction (Gz) leads to draining the blood away from the brain and cause sudden cerebral hypoxia [3,4]. Such a situation may lead to loss of vision or even consciousness of the pilot and fatal accidents in consequence. The training and exams of the pilots were introduced and designed to increase immunity for the acceleration. It was shown that the trained individuals could keep consciousness up to 9Gz, while non-trainees can pass out only between 3 Gz and 4 Gz [5,6].
To introduce the pilot into the environment with high G-force in laboratory conditions, the centrifuges are developed and used in military air force units. These devices enable to get large rotational speeds and G-force up to 20Gz. Nowadays, centrifuges are combined with modern flight simulators allowing to mimic real flight conditions. Centrifuges are bulky and expensive instruments and their use is related to high maintenance costs. Moreover, the conditions of the tests are not safe for the investigated subject. In case of bad tolerance of the subject to the acceleration provided by the centrifuge and necessity to stop the test rapidly, the access to the subject is very limited and the test may lead to serious cerebral perfusion disorders. Therefore, there is a need to develop cheaper and simpler methods for screening tests of the pilots and pilot candidates. The tilt table allowing for rapid changes of the inclination of the whole human body [4,7,8] can be used which can mimic the acceleration from + 1Gz to -1Gz. With additional lower body negative pressure (LBNP) chamber installed on the table, these values can be even higher [3,9]. The LBNP introduced an additional load to the cardiovascular system and is an established technique to simulate gravitational stress [10] and manipulate baroreceptors [11].
In this paper, we analyze cerebral hemodynamic responses to the orthostatic stress of the group of pilots and healthy volunteers (candidates for the pilots) examined on the tilt table with the LBNP chamber. The examination station called ORTHO-LBNP is shown in [12,13]. It is equipped with a tilting table allowing for rapid adjustment of the whole-body inclination angle from head-upright-tilt (HUT) to head-down-tilt (HDT) position. The LBNP chamber mounted on the table allows for the accumulation of blood in the vessels located in lower body parts to simulate the desired state of cerebral hypoxia with additional verticalization.
Assessment of cerebral oxygenation, one of the most important parameters in the monitoring of consciousness, is realized by the 6 channel near-infrared spectroscopy (NIRS) system, designed and constructed by the Nalecz Institute of Biocybernetics and Biomedical Engineering, Warsaw, Poland. NIRS is an optoelectronic technique which makes use of relatively low absorption of the tissue in a wavelength range between 600 and 900 nm. It allows for relatively deep penetration of the light into the tissue and back to the surface, still enabling to be detected [14,15]. Due to specific absorption spectra of oxy- and deoxyhemoglobin in this range, it is possible to assess changes of concentration of these chromophores in time or determine the oxygen saturation of the tissue under investigation [16,17]. Moreover, due to many advantages (non-invasiveness, real-time monitoring, simple operation, relatively low price), NIRS-based optical methods are extensively tested as potential research tools for neurophysiological studies as well as considered for application in clinical diagnostics and monitoring of medical therapy. NIRS techniques are used in clinical trials, including validation of these techniques in assessment and monitoring of oxygenation and blood supply to the brain cortex. Clinical trials on statistically representative groups of patients concerned in particular the monitoring of ischemic hypoxia during carotid endarterectomy [18,19], cardiac and general surgery [20,21] and monitoring of patients with brain damage [22–24]. The majority of these studies show a significant correlation between the changes in oxygenated (HbO2) and reduced (Hb) hemoglobin concentration and brain tissue saturation (StO2) measured with NIRS instruments with parameters of cerebral oxygenation obtained by invasive techniques such as oxygen partial pressure in the cerebral tissue and blood oxygen saturation in the jugular veins.
In several studies, monitoring of the brain oxygenation of military pilots during the flight or in a flight simulator was reported [25–29]. In these studies changes in oxygen concentration in the cockpit were also presented. Of particular interest are studies on pilots in real flight conditions in which high acceleration/deceleration events occur and their influence on brain oxygenation was monitored [30–32]. Studies have shown that the changes in the pilot’s cerebral oxygenation can be monitored by the assessment of changes in concentration of oxy- and deoxyhemoglobin and tissue saturation index using NIRS devices. Changes in signals measured by the NIRS technique are strongly related to the course of changes in aircraft accelerations during the flight. It was concluded that the application of near-infrared spectroscopy, in particular, the measurement of HbO2 concentration, enables effective monitoring the cerebral oxygenation status of military aircraft pilots during the flight [30].
The present study aims to verify the usefulness of the NIRS technique applied during ORTHO-LBNP tests, i.e. orthostatic stress and pooling of blood in the lower extremities, in the assessment of the tolerance of the subject to the hypovolemic hypoxia. Moreover, we aimed to propose the cerebral hemodynamic-related parameters that could allow us to describe tolerance to overloads and enable assessment of the effectiveness of the regulatory mechanism, responsible for maintaining the correct level of blood flow in the cerebral tissue. These parameters will be applied in the analysis of data obtained in military pilots/cadets (study group) and healthy volunteers (candidates for the military pilots, control group). Finally, we will apply a statistical method for analysis of the proposed parameters and show that the subjects can be classified depending on their tolerance to hypovolemic hypoxia.
2. Methods
2.1 Instrumentation
The 6-channel near-infrared spectroscopy system for high maneuverable aircrafts pilot studies during ORTHO-LBNP experiments (NIRSI) was constructed. The methodology is based on measurement and analysis of diffusely reflected light in the head tissues at 2 wavelengths from the near-infrared spectral range. NIRSI was developed and implemented into the ORTHO-LBNP system [13] and is presented in Fig. 1. The NIRS setup consist of the emitting module, detecting module, signal processing and communication module and power supply, Fig. 1(b). The whole system is closed in an electrically shielded housing. The NIRSI instrument also includes a system of fixing the measuring optodes on the head of the subject.
Fig. 1. a) 6-channel NIRSI system integrated into the ORTHO-LBNP station. b) The NIRSI setup. 1 – emitting module; 2 – detecting module; 3 – signal processing and communication module; 4 – power supply.
Emission module:
The device is equipped with emission modules consisting of 6 modified LEDs operating at wavelengths of λ1=735 nm (M735F1, Thorlabs, USA) and another 6 modified LEDs operating at λ2 = 850 nm (M850F2, Thorlabs, USA). The diodes are powered by the LEDD1B LED system (Thorlabs, USA).
Each LED is controlled with a separate driver unit which generated a current of up to 1.2 A. The light emitted at each wavelength (735 and 850 nm) is coded by generating the rectangular waves of filling factors 40% and 30%, respectively. The rectangular waves of the signal which drives LEDs operating at λ2 have 55 ms delay in relation to the rectangular wave driving the LEDs of λ1. The total period of both waveforms is 0.1 s resulting in the measurement frequency of 10 Hz. This procedure provides a sequential, alternating generation of the light at two wavelengths at a single point of emission. As a result, the filling of the controlling signal allows the coding of the used wavelengths. The delay of the control signals allows monitoring a background signal that is used to eliminate the influence of external light on the measured optical signals.
Light transmission:
The light is transmitted into a human head with the use of 2 bifurcated fiber bundles for delivery of the light from LEDs to positions marked as 1-X-L and 2-X-P located on the forehead of the subject, (see Fig. 2). Another set of 4 quadfurcated fiber bundles is used to deliver the light from two pairs of LEDs to positions marked as 3/4-X-L and 5/6-X-P located above the motor cortex of the subject. This configuration is used in order to increase the total optical power delivered to the emission points located on the hairy part of the head. The emission fiber bundles (CeramOptec, Germany) allowed delivering the light at two wavelengths to all emission spots on the surface of the head.
Fig. 2. a) Source-detector pairs (optodes) located on a surface of the subjects head; b) schematic locations of the optodes: o – detecting point, x – emitting point.
Diffusely reflected photons were collected from a human head and delivered to a detecting unit with the use of fiber bundles (CeramOptec, Germany).
The pairs of the source and detecting fiber bundles were positioned on forehead separated by a distance of 3 cm and 2 cm, and above motor cortex separated by a distance of 3 cm (Fig. 2(b)) in such a way that C3 and C4 points were located in the middle between the emitting and detecting fiber bundles tips positions.
Both emission and detection fiber bundles (the optodes), with a length of 2 m consist of pure fused silica fibers with core/cladding 185/200µm, numerical aperture NA = 0.22 and the effective area of approximately 7.06 mm2. The bundles have a bent into a 90-deg curve tip on the patient side to sit in the cap. The optodes were mounted in an optode holder which, in turn, was integrated into a standard EEG cap (EasyCap, Germany, Fig. 2(a)). The optode holder that passed through the cap is a designed and 3D printed thin plastic pad with the holes to insert the optodes.
The optode tips protruding through the holder in the cap can slide in and out of the cap to match the curvature of the head. To provide a safe and comfortable surface that sits against the scalp, a 1 cm layer of the biocompatible, medical, soft foam is glued to the patient-sided surface of the pad. The EEG cap maintains the coupling between the fiber bundles and the scalp. Fibers are inserted into the holes in the pad and kept perpendicular to the scalp by a plastic rigid stripping and spacers. To improve the coupling of the fibers with the scalp, the fiber tips extend for several mm through the interior of the cap to facilitate the combing of the fibers through the hair. Lastly, the cap is held in place against the head with straps under the chin of the subject. The optodes fixed in such a way ensured that the head remained motionless in relation to the rest of the body.
These solutions allowed us to eliminate most of the motion artefacts in the measured optical signal. However, we found in the signal a high amplitude spikes related to the tilt table motion. These artefacts were easily recognizable, occurred only in the time moment of the tilt table taking its boundary position (70° or -30°) and did not affect the parameters derived from the signals related to the hemodynamic response.
Detection module:
The detecting unit is a photodiode S2386-5K (Hamamatsu, Japan) combined with a custom-made preamplifier (Galwes, Poland).
Signal processing and communication module:
A module has been developed for the digital processing of a signal registered by the detection module. The module provides power supply for the detection module and is responsible for the generation of signals controlling the operation of LEDs (LEDs control waveforms), conversion of photodiodes analog signals into digital form, decoding of the detected 2-wavelengths signals taking into account the LED control rectangle-shape waveforms.
The microcontroller included in the module is responsible for controlling the operation of the device and communication with the ORTHO-LBNP station.
The NIRSI setup is powered with power supply PM60-32A (PowerGate LLC) approved for medical applications. The system is enclosed in a propacPRO 3U 84HP D266 housing (nVent Schroff GmbH, Germany) which provides electromagnetic protection.
2.2 Measurement procedure and protocol
Measurements were carried out after obtaining the informed consent of each studied subject according to the approval of the Ethical Committee at the Military Institute of Aviation Medicine in Warsaw, Poland (Decision No. 18/2015) and after registration of the study and measurement protocols at https://clinicaltrials.gov/ (number NCT03354234).
During the ORTHO-LBNP study, changes in brain oxygenation were evaluated in response to two types of stimuli - reduced pressure around the lower part of the body and body tilting, for a group of 12 military pilots and cadets of the Military Academy of Air Force (WSOSP) in Dęblin (study group) and 12 healthy volunteers (control group). Both groups include 2 women and 10 men; age: 23.9 ± 2.1 and 18 ± 0 years (p < 0.001), weight: 79.46 ± 14.0 and 72.3 ± 10.2 kg (p = 0.088), height: 178.3 ± 7.3 and 178.4 ± 6.1 cm (p = 0.663) for studied and control groups, respectively. P values are obtained by the Mann-Whitney test.
The measurement protocol consisted of three main phases (Fig. 3):
- 1. Tilting with a speed of 40°/s: a rapid change of the tilting angle from the initial position (0°) to -30° (HDT) and immediately afterwards HUT to 70°; holding in this position for 60 seconds (ORTHO phase) and return to the rest position (180 seconds).
- 2. The rapid drop of pressure around the lower part of the body down to -100 mmHg, maintaining this pressure for 60 seconds (LBNP phase) while the subject is in the 0° position and release of the pressure - return to the initial conditions (180 seconds).
- 3. Tilting with the speed of 40°/s with simultaneous pressure reduction down to -70mmHg; three repetitions of changes in position (+70 ° and -30 °) and pressure (-100mmHg and 0) with 30 seconds in each position (phase ORTHO + LBNP), and return to the rest position (180 seconds).
Fig. 4. Selected example of changes in the concentration of total hemoglobin (ΔCHbTot) calculated for the entire course of the experiment consisting of phase ORTHO, phase LBNP and triple-phase ORTHO + LBNP: the tilting combined with a sudden drop of pressure around the lower part of the body. The ΔC1-9 parameters are calculated as the maximum amplitudes of CHbTot changes as a response to individual stimuli.
2.3 Data analysis
An algorithm for calculating changes in hemoglobin concentrations from optical signals measured with the NIRSI device was developed based on the intensity of light re-emitted from the tissue. Changes in oxy- and deoxyhemoglobin concentration ΔCHbO2 and ΔCHb were calculated using changes in light attenuation ΔA and molar extinction coefficient for HbO2 and Hb [33] according to the modified Lambert-Beer law [34]. The analysis of the measurement data was carried out using the scripts developed in the Matlab environment (The MathWorks, Inc., USA).
In order to perform statistical analysis, nine parameters which describe responses of the cerebral circulation to individual stimuli were defined (Fig. 4), based on the amplitudes of changes in concentrations of total hemoglobin (ΔCHbTot) and are the maximum amplitudes of ΔCHbTot as a response to individual stimuli (ORTHO: ΔCHUT and ΔCHDT corresponding to HUT and HDT positions, LBNP: ΔCLBNP or double stimuli ORTHO + LBNP: ${\Delta C}_{\textrm{HUT}}^{\textrm{LBNP}}$).
2.4 Statistical analysis
First, univariate data analysis was performed for all amplitude-based parameters (variables) using a one-way analysis of variance (ANOVA). Next, multivariate analysis (MVA) was performed. Orthogonal partial least squares-discriminant analysis (OPLS-DA) was applied [35]. Centring and univariate variance (UV) scaling [36] were performed on data based on NIRS measurement results in order to maximize the covariance between the independent variables.
The ellipse drawn on the scores plot represents the Hotelling T-square test with 95% confidence in the model. Scores inside the Hotelling T-square ellipse represent observations that are close to each other on the plot are more similar than observations distant from each other. The separation of groups of observations is seen on the plot if one group of scores can be clustered and distinguished from another clustered group of scores in the predictive component (horizontal axis of the plot). Distance between scores in the vertical direction is indicative of within-group variation. This method allowed us to construct a classification model to distinguish between studied subjects (pilots) and subjects from the control group. The model performance was described by R2, which describes the total variation in the data and estimates goodness of fit, and Q2, which is an internal cross-validation parameter and represents an estimating of the goodness of the prediction. We adopted the limit of good predictive capability for the OPLS-DA model, as described elsewhere [37]. The model was positively validated using CV ANOVA test (Fisher test).
Moreover, to identify the most important of the nine variables available for the analysis, we have build Parameter Sets consisting of a different number of parameters related to the brain oxygenation changes. Variable Importance in the Projection (VIP) values were calculated for these different Parameters Sets. Variables with VIP values higher than one were considered significant for discrimination between groups, and higher VIP values indicated the greater contribution of the variable to the discrimination between groups. The Fishers test was used to assess the significance of group separation and the validity of the method. P values < 0.05 were considered significant.
3. Results
A selected example of the estimated changes in hemoglobin concentrations during the whole experiment is presented in Fig. 5. In the first phase of the experiment (ORTHO phase), the cardiovascular response was tested in orthostatic stress conditions, associated with a sudden change in the subject body position from horizontal to vertical with simultaneous change of the overload axis (change of the body position from 0°to HDT (-30°) and HUT (+70°)). It was observed that in the studied subjects, the total hemoglobin concentration CHbTot increases when the subject is in the HDT (-30°) position and decrease after tilting angle change up to HUT (+70°) position. In all subjects, in both groups (study and control), changes in hemoglobin showed a similar pattern of behaviour during the orthostatic test.
Fig. 5. Upper panel: selected example of changes in the concentration of oxy- (ΔCHbO2, red line), deoxy- (ΔCHb, blue line) and total hemoglobin (ΔCHbTot, black line) concentrations observed during the entire course of the experiment consisting of ORTO phase, LBNP phase and three repetitions of ORTO + LBNP phase: tilting combined with sudden drop of a pressure around the lower part of the body. Measurement was carried out using a NIRSI device for the optode positioned above the left hemisphere of the subject’s forehead and 3 cm of source-detector separation. Bottom panel: changes in tilting angle and changes in LBNP chamber pressure.
In the second phase of the experiment (LBNP phase), the cerebral oxygenation response was analyzed in normotensive hypovolemia which leads to ischemic hypoxia. In a healthy subject, as a result of a sudden decrease of the pressure around the lower part of the body, the CHbTot decreases. This phenomenon is caused by a decrease in the volume of circulating blood in the upper part of the body (including the brain). A decrease in total hemoglobin and correlated oxygenated and deoxygenated hemoglobin concentration decrease is observed as a result of a change in cerebral blood volume (see Fig. 5). However, not in all of the subjects of both groups, the changes in total hemoglobin concentration were in line with this typical pattern during LBNP phase of the experiment. In some subjects, no changes in total hemoglobin concentration were observed. In other subjects, an increase in total cerebral hemoglobin concentration was noticed. This effect is associated with an increase in blood volume in the upper part of the body and is paradoxical. For three subjects in which CHbTot increase was observed the ${\Delta C}_{\textrm{3}}^{\textrm{LBNP}}$ parameters were excluded from further statistical analysis.
In the third phase of the experiment, along with decreasing of the pressure around the lower half of the body, the orthostatic stress was an additional stimulus associated with a sudden change in the subject body position from horizontal to vertical and vice versa (ORTHO + LBNP phase). For all subjects of both groups, a decrease in total and oxygenated hemoglobin concentrations was observed at a constant level of deoxyhemoglobin or small changes of its concentration as a response to the change of the tilting angle up to + 70° and simultaneous sudden drop of chamber pressure. After reaching the -30° position and the increase in chamber pressure, the hemoglobin concentrations returned to their baseline values (Fig. 5).
For all three phases of the experiment, the observed changes in CHbTot were within a wide range of amplitudes (0-100 µM) for 5 from 6 measuring channels. The analysis of the data for channel 5 (see Fig. 2(b) for optode location) was impossible because the optical signal for this channel was not registered correctly in the study group (pilots) due to photodiode power failure.
Results of the calculation of parameters ΔC1-9 for all subjects in the study and control group are presented in the form of box-charts in Fig. 6. The box and whisker plot for all calculated parameters ΔC1-9 are presented separately for source-detector separations of 3 cm and 2 cm, for the study group (pilots - red and pink plot, respectively) and the control group (healthy volunteers - blue and cyan plot). The changes related to a hemodynamic stimulus are larger for shorter source-detector (r = 2cm) separation than for the longer one (r = 3cm). This effect is related to the proper function of the brain autoregulation mechanism which does not allow for large changes in cerebral blood flow when the driving pressure changes are large. The differences in amplitude of CHbTot changes can be observed between the study and the control groups. In the case of measurements carried out for a source-detector separation of 3 cm, the values of the parameters ΔC associated with reduced blood flow to the brain (parameters ΔC2,HUT, ${\Delta C}_\textrm{3}^{\textrm{LBNP}}$, ${\Delta C}_{\textrm{5,HUT}}^{\textrm{LBNP}}$, ${\Delta C}_{\textrm{7,HUT}}^{\textrm{LBNP}}$ and ${\Delta C}_{\textrm{9,HUT}}^{\textrm{LBNP}}$ related to LBNP and ORTHO + LBNP during HUT stimuli) are noticeably higher in the control group than in the study group (Fig. 6, left panel). However, in the case of r = 2cm, these differences are even more pronounced (Fig. 6, right panel). Moreover, it can be seen that for parameters related the ORTHO + LBNP phase during the simultaneous influence of the orthostatic stress and ischemic hypoxia conditions, the dispersion of parameters (ΔC4-9) is greater than parameters dispersion related to a single stimulus (ΔC1,HDT, ΔC2,HUT - ORTHO or ${\Delta C}_\textrm{3}^{\textrm{LBNP}}$ - LBNP) for both groups.
Fig. 6. Box and whisker plot for all calculated parameters ΔC1-9, presented separately for the source-detector separations r = 3 and 2 cm for pilots N = 12 (red and pink plot) and healthy volunteers N = 12 (blue and cyan plot). The ΔC subscripts indicate the applied stimuli: 1,2 – ORTHO, 3 – LBNP, 4-9 – ORTHO + LBNP. The range of values of large rectangles: 25% -75%, the horizontal line inside the rectangle determines the median value. Squares represent the mean values of change in total hemoglobin concentration.
Since an amplitude of ΔCHbTot in the case of r = 2 cm is relatively larger than for r = 3cm, it was decided to provide the new parameter, which is based on the difference in the amplitudes of CHbTot changes occurring in the brain cortex and CHbTot changes occurring mostly in the extracerebral layers (skin and skull), and, thus, representing the changes of the hemodynamic parameters in the cerebral cortex. The values of new parameters Δ(ΔC1-9) were calculated by subtracting from ΔC value calculated for r = 3cm the corresponding ΔC value calculated for r = 2cm (i.e. ΔCHbTot, 3cm – ΔCHbTot, 2cm). The calculated parameters for all subjects in the study group and the control group are presented as box-charts in Fig. 7. Similarly to the previous case (presented in Fig. 6), the largest differences between the subjects from the study and control group are observed for the ORTHO + LBNP phase during the simultaneous influence of the orthostatic stress and ischemic hypoxia conditions.
Fig. 7. Box and whisker plot for parameters Δ(ΔC1-9), calculated on the basis of differences in CHbTot changes for the source-detector separations of 3 cm and 2 cm for study group N = 12 (pilots - red rectangles) and for control group N = 12 (healthy volunteers - blue rectangles). The ΔC subscripts indicate the applied stimuli: 1,2 – ORTHO, 3 – LBNP, 4-9 – ORTHO + LBNP. The range of values of large rectangles: 25% -75%, the horizontal line inside the rectangle determines the median value. Squares represent the mean values of change in total hemoglobin concentration.
3.1 Statistical analysis
Considering the estimated parameters, ΔC1-9 associated with ΔCHbTot measured at two source-detector separations and parameters Δ(ΔC1-9) describing differences between these parameters we built 4 parameters sets in order to find the best parameter set that will allow differentiating control and study group.
Parameter set 1 consisted of 9 calculated parameters ΔC1-9 averaged for the source-detector distance r = 3cm.
Parameter set 2 consisted of 18 calculated parameters ΔC1-9 averaged separately for source-detector distances of 2 cm (9 parameters) and 3 cm (9 parameters).
Parameter set 3 consisted of 9 parameters Δ(ΔC1-9) which are the differences in averaged ΔCHbtot calculated for r = 3cm and 2cm (ΔCHbTot, 3cm – ΔCHbTot, 2cm).
Parameter set 4 consisted of only 6 parameters Δ(ΔC4-9) related to ΔCHbTot, 3cm – ΔCHbTot, 2cm observed as a result of the double stimulus ORTHO + LBNP.
Univariate analysis (ANOVA) of Parameter set 1 didn’t show any statistically significant differences between all parameters. Similarly, the OPLS-DA analysis did not allow to construct a validated model for this set (p = 0.4).
ANOVA analysis of Parameter set 2 indicated significant differences between study and control groups for parameters ΔC2,HUT (p = 0.49), ${\Delta C}_\textrm{3}^{\textrm{LBNP}}$ (p = 0.025), ${\Delta C}_{\textrm{5,HUT}}^{\textrm{LBNP}}$ (p = 0.034), ${\Delta C}_{\textrm{7,HUT}}^{\textrm{LBNP}}$ (p = 0.010) and ${\Delta C}_{\textrm{9,HUT}}^{\textrm{LBNP}}$ (p = 0.012) estimated for r = 2cm. No statistically significant parameters were found for r = 3cm. OPLS-DA for Parameter set 2 allows us to construct a model that comprises only one predictive component. For this parameter set, R2 was 0.39 and Q2 was 0.27: this parameter set has a moderate data fit and weak data prediction. Model classified correctly 75% of all subjects (18 from 24) from both analyzed groups: 58% (7 from 12) of the control group and 92% (11 from 12) of a study group. The validation test (CV-ANOVA) confirmed that the model was valid (p = 0.04). The parameters differentiating the groups (VIP > 1) are ΔC2,HUT, ${\Delta C}_\textrm{3}^{\textrm{LBNP}}$, ${\Delta C}_{\textrm{5,HUT}}^{\textrm{LBNP}}$, ${\Delta C}_{\textrm{7,HUT}}^{\textrm{LBNP}}$ and ${\Delta C}_{\textrm{9,HUT}}^{\textrm{LBNP}}$.
Univariate analysis (ANOVA) for Parameter set 3 showed significant differences between groups in parameters: Δ(ΔC1,HDT), Δ(ΔC2,HUT), Δ(${\Delta C}_\textrm{3}^{\textrm{LBNP}}$), Δ(${\Delta C}_{\textrm{5,HUT}}^{\textrm{LBNP}}$), Δ(${\Delta C}_{\textrm{7,HUT}}^{\textrm{LBNP}}$) and Δ(${\Delta C}_{\textrm{9,HUT}}^{\textrm{LBNP}}$) (p < 0.001) and Δ(ΔC6,HDT) (p = 0.010). OPLS-DA for Parameter set 3 allowed to construct model consisted of one predictive and one orthogonal component. For this model, R2 was 0.73 (excellent data fit) and Q2 – 0.93 (excellent data prediction). The validation test (CV-ANOVA) confirmed that the model was valid (p < 0.0001). All subjects were correctly classified into their groups (100%) which is presented in Fig. 8(a). The parameters involved in group separation which allow distinguishing oxygenation changes significantly (VIP > 1) are Δ(ΔC2,HUT), Δ(${\Delta C}_\textrm{3}^{\textrm{LBNP}}$), Δ(${\Delta C}_{\textrm{7,HUT}}^{\textrm{LBNP}}$) and Δ(${\Delta C}_{\textrm{9,HUT}}^{\textrm{LBNP}}$).
Fig. 8. The scores plots of the two-component OPLS-DA for a) Parameter set 3 and b) Parameter Set 4 t0[1] represent within-class variation in the first orthogonal component, whereas t[1] represents between-class variation in the first predictive component. Ellipse represents Hotelling T2 with 95% confidence in score plots.
However, taking into account only the parameters which reflect reactions to the double stimulus ORTHO + LBNP (Δ(ΔC4-9) Parameter set 4), OPLS-DA allowed to construct a model that consisted of one predictive and three orthogonal components, for which R2 was 0.93 (excellent data fit) and Q2 was 0.91(excellent data prediction). The validation test (CV-ANOVA) confirmed that the model was valid (p < 0.0001). All subjects were correctly classified into their groups, Fig. 8(b). The parameters involved in the group separation which allow distinguishing significantly oxygenation changes variables (VIP > 1) involved in group separation are Δ(${\Delta C}_{\textrm{5,HUT}}^{\textrm{LBNP}}$), Δ(${\Delta C}_{\textrm{7,HUT}}^{\textrm{LBNP}}$) and Δ(${\Delta C}_{\textrm{9,HUT}}^{\textrm{LBNP}}$).
4. Discussion and conclusions
Lower body negative pressure (LBNP) has been used for decades to simulate orthostatic stress and the effects of decreasing of the blood volume circulating in the upper half of the human body, in particular, the effect of cerebral ischemic hypoxia [38]. The LBNP protocols and technical considerations for experimental design were reported [39]. LBNP is considered as an experimental substitute for the + Gz stress of the centrifuge, which is costly in purchase and operating.
The LBNP has been used as a model to study acute hemorrhagic shock in humans [40]. Both high and low tolerant subjects presented heart rate variability while simulating hemorrhage with LBNP [41]. A reproducibility of a continuous ramp LBNP protocol for simulating hemorrhage was reported in [42]. Physiological compensatory responses to blood loss based on arterial pressure monitoring and observation of cerebral blood flow oscillations (within high and low tolerant individuals) were shown in [43]. It was reported that blood flow velocity in the middle cerebral artery drops significantly during LBNP test and during combined LBNP and mental stress [44]. The LBNP stimulus leads to a decrease in intracranial pressure [45] as well.
Series of LBNP studies showed oxygenation measurement on muscles using the NIRS technique [43,46–48]. It was observed that muscle oxygenation correlates with the stroke volume (SV) and is the earliest indicator of progressive central hypovolemia [43]. Reductions in muscle oxygenation was visible with NIRS considerably earlier than with the standard clinical measures of HR and BP.
Table 1 summarizes the recent LBNP, ORTHO and LBNP + ORTHO studies related to the estimation of hemoglobin concentrations (CHbO2, CHb, CHbTot), changes in hemoglobin concentrations (ΔCHbO2, ΔCHb, ΔCHbTot) or tissue oxygen saturation (StO2), where measurements were carried out on the head or muscle with NIRS instrumentation.
Table 1. List of recent studies where the tissue oxygenation parameters were monitored while LBNP and/or HUT/HDT using the NIRS technique on the head and/or on the muscles. CW NIRS, FD NIRS, SRS NIRS are continuous wave, frequency-domain and spatially-resolved NIRS, respectively. M- male, F- female.
In all these studies, the changes of the CHbO2 and StO2 as a response to the three types of stimuli (LBNP, ORTHO and LBNP + ORTHO) revealed a similar polarity for head and forearm measurements. However, the amplitude of changes varies significantly with the optodes location. StO2 drops by 2.4-4.6% for the head measurements and by 3-16% for the forearm experiments. CHbO2 decreases by 0.7-2.3 µM for the head and by 12.7µM for the forearm. However, these values might not be comparable as the different experimental protocols were applied (stepwise or gradual drop of LBNP and different LBNP values). Furthermore, experiments were carried out with different NIRS devices which estimated parameters with different optodes locations and source-detector separations.
In few reports, superimposed LBNP and tilt stimuli were applied. The cerebral oxygen saturation index StO2 was obtained for 15° HDT with -25 mmHg of LBNP [49]. Cerebral blood velocity measured with the transcranial Doppler ultrasound techinque while the LBNP (-50 mmHg) in three body positions (45°HUT, supine, 45°HDT) was shown in [50]. The cerebral blood velocity decreased during LBNP in all three tilt positions. However, these changes were similar in amplitude for all body positions.
In this study, the mean change of total hemoglobin concentration (ΔCHbTot) within both groups is 15 µM and 50 µM for the source-detector separation of 3 and 2 cm, respectively. These values are significantly higher than the ΔCHbO2 reported by others. The difference between amplitudes of changes obtained at these two source-detector separations can be explained by the fact that the longer the source-detector separation r, more information from deeper cerebral compartments is present in the measured NIRS signal [51]. In the brain, the autoregulatory capacity could maintain the cerebral perfusion when changes in perfusion pressure are stimulated by LBNP. Further, the subject’s tolerance to G-force decreases when the HDT (–Gz) test is applied before the + Gz stimulus [52] as applied in this study protocol. Therefore, the higher changes in hemoglobin concentration might be expected.
Up to date studies demonstrate common pattern of hemodynamic parameters behavior as measures by the NIRS: ΔCHbO2 and/or StO2 simultaneously decrease with LBNP. This complies with results of in-flight measurements during Gz as carried out on pilots of high maneuverable aircraft [30]. The maximum reported decrease of concentration of HbO2 is about 26 µM at 7.5Gz. Similar results are presented in [25] where the maximum of ΔCHbO2 is about 23 µM and the StO2 decrease is 15%. The aplitude of change in StO2 was significantly higher in subjects who completed the whole profile of Gz changes than those who terminated the test due to G-LOC (gravity-induced loss of consciousness), whereas no significant differences were observed in CHbO2 between these two groups of subjects. Authors suggest that the observed differences between CHbO2 and StO2 may indicate high contribution of extracranial tissues and as such measured CHbO2 and StO2 may not allow prediction of the G-LOC.
The effects of the stimuli generated using the LBNP chamber within the present study are significantly smaller than those reported previously by other authors. LBNP in-vivo studies on hypovolemia as reported by Kay et al. [53] showed syncopes manifested in the form of: 1) a sudden decrease in systolic arterial pressure below 80 mmHg; 2) a sudden relative bradycardia and/or 3) the appearance of subjective presyncope symptoms, such as loss of color vision, nausea, sweating, dizziness, blurred vision or general discomfort. Subjects who went through an LBNP test at -70 mmHg without the appearance of these symptoms were classified as highly tolerant. As shown by Soller et al., ten subjects completed the LBNP test: six subjects reached cardiovascular collapse at -80 mmHg LBNP and four at -90 mmHg [43]. Also, it has been reported [50] that some participants (from total n = 13) reached presyncope with the application of LBNP (-50 mmHg) in the supine position (n = 3) and the 45°HUT position (n = 8). In the present study, the LBNP pressure applied was -100 mmHg and no symptoms described above in any of the examined subjects were observed.
This lack of strong response to the LBNP at -100 mmHg could relate to the position of the seal between the subject and the edge of the LBNP chamber. It has been shown that the reduction in central blood volume in response to the LBNP depends on the seal position: LBNP with the seal at the upper abdomen induced a large reduction in central blood volume and high increase in the heart rate as compared to the seal located at the iliac crest [54].
Furthermore, in the previous in-vivo studies, the LBNP test time have varied significantly: e.g. stimulus of 2 to 13 minutes. It is expected that a brief exposure avoid the complication of humoral responses [55]. Varying levels of the LBNP revealed quantitatively different response patterns among healthy young men [56]. High level of LBNP can has an effect on both the arterial (high pressure) and cardiopulmonary receptors and induce tachycardia, increased diastolic and reduced systolic blood pressure [57].
The dynamics of changes in total hemoglobin ΔCHbtot as in the present study for all subjects in the LBNP phase is similar to the dynamics of hypovolemia in [53]. However, direct comparison of the absolute amplitudes of CHbTot is challenging as different NIRS techniques were used and differences in constructions of the LBNP chamber and positioning of the seal are substantial.
In this paper, we used constructed ORTHO-LBNP system [13], i.e. a prototype device consisting of the tilt table to generate orthostatic stress and the LBNP chamber for assessment of the brain ischemia. We also present results of the study, in which the main emphasis is to find the physiological parameters estimated from collected optical signals based on the NIRS technique that best correlate with the applied stimuli and enable for classifying the brain hemodynamic responses in relation to the degree of the subject’s resistance to hypovolemic ischemia. The total hemoglobin concentrations enabled classifying subjects into their groups (pilots and volunteers). The validation resulted in a 100% success rate. The same result was met with the parameters related only to the double stimulus ORTHO + LBNP. Single ORTHO and single LBNP stimuli did not revealed significant changes in the considered parameters.
However, the study is limited to the retrospective validation [58] and prediction models performance was tested on data on which the model was constructed. Prospective validation of the model should be cosidered on new independent data acquired from a different group of studies subjects [59].
Another limitation of the study is that we did not have access to any results of physical condition tests of the subjects included in the study. Thus, we do not have a reference measure of the subject’s predisposition to the experimental conditions. This study bases on the hypothesis that the active pilots are trained (also at the centrifuge) and as such are far better suited to withstand the effects of the LBNP and tilting.
The NIRS technique could be an essential tool in screening for orthostatic tolerance and training to improve G-tolerance in pilots. The NIRSI system as implemented into the ORTHO-LBNP station can benefit from further modifications and provides more parameters related directly to the brain hemodynamics, e.g. the brain tissue oxygen saturation StO2 (which needs the multi-distance NIRS approach) and/or local cerebral blood volume changes (for which monitoring of the heartbeat is needed). We have decided to use only two source-detector separations as the system is used in highly artefact generating environment (tilting mostly) and high count of optodes/fibres packed densely at one location create a high inertia mass. However, a second attempt to a development of a multi-distance optodes fixing system could be beneficial. The heartbeat oscillations are already present in the optical signals. As the cerebral arterial network is very dense, it is possible to estimate local changes in blood flow and volume with the NIRS [60]. A conclusion for the future work is to analyse the parameters to get broader view into the physiological reactivity.
Funding
National Center for Research and Development (DOBR/0052/R/ID1/2012/03).
Acknowledgements
The study was financed by the Polish National Center for Research and Development within the project DOBR/0052/R/ID1/2012/03. The authors would like to thank Łukasz Dziuda, Mariusz Krej, Paulina Baran, from Military Institute of Aviation Medicine (Warsaw, Poland) for participation in the study, Krzysztof Kowalczuk from Military Institute of Aviation Medicine (Warsaw, Poland) for medical assistance during the study and Piotr Kwaśny from ETC-PZL Aerospace Industries Sp. z o.o. (Warsaw, Poland) for the participation in ORTHO-LBNP station construction.
Measurements were carried out in the Center for Integrated Structural and Functional Research of the Central Nervous System - CNS Lab which is a joint project of a consortium of the Nalecz Institute of Biocybernetics and Biomedical Engineering of the Polish Academy of Sciences and Military Institute of Aviation Medicine (Warsaw, Poland).
Disclosures
The authors declare that there are no conflicts of interest related to this article.
References
1. E. Slungaard, J. McLeod, N. D. C. Green, A. Kiran, D. J. Newham, and S. D. R. Harridge, “Incidence of G-Induced Loss of Consciousness and Almost Loss of Consciousness in the Royal Air Force,” Aerosp. Med. Hum. Perform. 88(6), 550–555 (2017). [CrossRef]
2. K. M. Alvim, “Greyout, blackout, and G-loss of consciousness in the Brazilian Air Force: a 1991-92 survey,” Aviat Space Environ. Med. 66, 675–677 (1995).
3. A. Polese, H. Sandler, and L. D. Montgomery, “Hemodynamic responses to seated and supine lower body negative pressure: comparison with + Gz acceleration,” Aviat Space Environ. Med. 63(6), 467–475 (1992).
4. E. Szufladowicz, R. Maniewski, E. Kozluk, A. Zbiec, A. Nosek, and F. Walczak, “Near-infrared spectroscopy in evaluation of cerebral oxygenation during vasovagal syncope,” Physiol. Meas. 25(4), 823–836 (2004). [CrossRef]
5. K. V. Kumar and W. T. Norfleet, “Issues on human acceleration tolerance after long-duration space flights,” in NASA Technical Memorandum (NASA Johnson Space Center, 1992).
6. C. F. Kelly and A. H. Smith, “Chronic Acceleration Studies - Physical responses to artificial Alterations in Weight,” NASA Contractor Report CR-441 (University of California, 1966).
7. H. H. Holm, E. Bachus, V. Hamrefors, O. Melander, G. Taseveska, A. Fedorowski, and M. Magnusson, “Changes in hemodynamic parameters and cerebral oximetry during head-up tilt test in heart failure patients: the HARVEST study,” Eur. J. Heart Fail. 19(4), 466–468 (2017). [CrossRef]
8. T. Durduran, C. Zhou, B. L. Edlow, G. Yu, R. Choe, M. N. Kim, B. L. Cucchiara, M. E. Putt, Q. Shah, S. E. Kasner, J. H. Greenberg, A. G. Yodh, and J. A. Detre, “Transcranial optical monitoring of cerebrovascular hemodynamics in acute stroke patients,” Opt. Express 17(5), 3884–3902 (2009). [CrossRef]
9. D. E. Watenpaugh, G. A. Breit, T. M. Buckley, R. E. Ballard, G. Murthy, and A. R. Hargens, “Human cutaneous vascular responses to whole-body tilting, Gz centrifugation, and LBNP,” J. Appl. Physiol. 96(6), 2153–2160 (2004). [CrossRef]
10. B. T. A. Esch, J. M. Scott, and D. E. R. Warburton, “Construction of a lower body negative pressure chamber,” Adv. Physiol. Educ. 31(1), 76–81 (2007). [CrossRef]
11. Q. Fu, S. Shibata, J. L. Hastings, A. Prasad, M. D. Palmer, and B. D. Levine, “Evidence for unloading arterial baroreceptors during low levels of lower body negative pressure in humans,” Am. J. Physiol. Heart C. 296(2), H480–H488 (2009). [CrossRef]
12. A. Sobotnicki, L. Juszynski, M. Sobiech, L. Dziuda, F. W. Skibniewski, P. Kwasny, and A. Brzozowska, “Bearing position changing system, preferably of the verticalisation table,” Polish patent 224837 (2014).
13. L. Dziuda, M. Krej, M. Smietanowski, A. Sobotnicki, M. Sobiech, P. Kwasny, A. Brzozowska, P. Baran, K. Kowalczuk, and F. W. Skibniewski, “Development and evaluation of a novel system for inducing orthostatic challenge by tilt tests and lower body negative pressure,” Sci. Rep. 8(1), 7793 (2018). [CrossRef]
14. J. Steinbrink, H. Wabnitz, H. Obrig, A. Villringer, and H. Rinneberg, “Determining changes in NIR absorption using a layered model of the human head,” Phys. Med. Biol. 46(3), 879–896 (2001). [CrossRef]
15. A. Liebert, H. Wabnitz, J. Steinbrink, H. Obrig, M. Moller, R. Macdonald, A. Villringer, and H. Rinneberg, “Time-resolved multidistance near-infrared spectroscopy of the adult head: intracerebral and extracerebral absorption changes from moments of distribution of times of flight of photons,” Appl. Opt. 43(15), 3037–3047 (2004). [CrossRef]
16. R. Maniewski, A. Liebert, M. Kacprzak, and A. Zbiec, “Selected applications of near infrared optical methods in medical diagnosis,” Opto-Electron. Rev. 12, 255–262 (2004).
17. D. A. Boas, T. Gaudette, G. Strangman, X. Cheng, J. J. Marota, and J. B. Mandeville, “The accuracy of near infrared spectroscopy and imaging during focal changes in cerebral hemodynamics,” Neuroimage 13(1), 76–90 (2001). [CrossRef]
18. C. W. Pennekamp, M. L. Bots, L. J. Kappelle, F. L. Moll, and G. J. de Borst, “The value of near-infrared spectroscopy measured cerebral oximetry during carotid endarterectomy in perioperative stroke prevention. A review,” Eur. J. Vasc. Endovasc. Surg. 38(5), 539–545 (2009). [CrossRef]
19. A. Rigamonti, M. Scandroglio, F. Minicucci, S. Magrin, A. Carozzo, and A. Casati, “A clinical evaluation of near-infrared cerebral oximetry in the awake patient to monitor cerebral perfusion during carotid endarterectomy,” J. Clin. Anesth 17(6), 426–430 (2005). [CrossRef]
20. A. Moerman, G. Vandenplas, T. Bove, P. F. Wouters, and S. G. De Hert, “Relation between mixed venous oxygen saturation and cerebral oxygen saturation measured by absolute and relative near-infrared spectroscopy during off-pump coronary artery bypass grafting,” Br. J. Anaesth 110(2), 258–265 (2013). [CrossRef]
21. H. B. Nielsen, “Systematic review of near-infrared spectroscopy determined cerebral oxygenation during non-cardiac surgery,” Front Physiol. 5, 93 (2014). [CrossRef]
22. W. Weigl, D. Milej, D. Janusek, S. Wojtkiewicz, P. Sawosz, M. Kacprzak, A. Gerega, R. Maniewski, and A. Liebert, “Application of optical methods in the monitoring of traumatic brain injury: A review,” J. Cereb. Blood Flow Metab. 36(11), 1825–1843 (2016). [CrossRef]
23. A. D. McLeod, F. Igielman, C. Elwell, M. Cope, and M. Smith, “Measuring cerebral oxygenation during normobaric hyperoxia: a comparison of tissue microprobes, near-infrared spectroscopy, and jugular venous oximetry in head injury,” Anesth Analg. 97, 851–856 (2003). [CrossRef]
24. G. Rosenthal, A. Furmanov, E. Itshayek, Y. Shoshan, and V. Singh, “Assessment of a noninvasive cerebral oxygenation monitor in patients with severe traumatic brain injury,” J. Neurosurg. 120(4), 901–907 (2014). [CrossRef]
25. K. Kurihara, A. Kikukawa, A. Kobayashi, and T. Nakadate, “Frontal cortical oxygenation changes during gravity-induced loss of consciousness in humans: a near-infrared spatially resolved spectroscopic study,” J. Appl. Physiol. 103(4), 1326–1331 (2007). [CrossRef]
26. O. Eiken, M. E. Keramidas, N. A. S. Taylor, and M. Grönkvist, “Intraocular pressure and cerebral oxygenation during prolonged headward acceleration,” Eur. J. Appl. Physiol. 117(1), 61–72 (2017). [CrossRef]
27. P. B. Benni, J. K. J. Li, B. Chen, J. Cammarota, and D. W. Amory, “Correlation of NIRS Determined Cerebral Oxygenation with Severity of Pilot + Gz Acceleration Symptoms,” in Oxygen Transport to Tissue XXIV, J. F. Dunn and H. M. Swartz, eds. (Springer, US, 2003), pp. 381–389.
28. H. Ayaz, M. P. Çakir, K. Izzetoğlu, A. Curtin, P. A. Shewokis, S. C. Bunce, and B. Onaral, “Monitoring expertise development during simulated UAV piloting tasks using optical brain imaging,” in Proceedings of IEEE Aerospace Conference (IEEE, 2012), 1–11.
29. A. Onozawa, A. Kikukawa, and Y. Miyamoto, “A new evaluation method for + Gz tolerance with loratadine by using a near-infrared spectroscopy,” Dyn. Med. 7(1), 3 (2008). [CrossRef]
30. A. Kobayashi and Y. Miyamoto, “In-flight cerebral oxygen status: continuous monitoring by near-infrared spectroscopy,” Aviat Space Environ. Med. 71, 177–183 (2000).
31. A. Kikukawa, A. Kobayashi, and Y. Miyamoto, “Monitoring of pre-frontal oxygen status in helicopter pilots using near-infrared spectrophotometers,” Dyn. Med. 7(1), 10 (2008). [CrossRef]
32. A. Kobayashi, A. Tong, and A. Kikukawa, “Pilot cerebral oxygen status during air-to-air combat maneuvering,” Aviat Space Environ. Med. 73, 919–924 (2002).
33. S. A. Prahl, “Tabulated Molar Extinction Coefficient for Hemoglobin in Water,” http://omlc.ogi.edu/spectra/hemoglobin/summary.html (1999).
34. D. T. Delpy, M. Cope, P. van der Zee, S. Arridge, S. Wray, and J. Wyatt, “Estimation of optical pathlength through tissue from direct time of flight measurement,” Phys. Med. Biol. 33(12), 1433–1442 (1988). [CrossRef]
35. Y.-W. Seo, S.-D. Lee, D.-K. Koh, and B.-M. Kim, “Partial Least Squares-discriminant Analysis for the Prediction of Hemodynamic Changes Using Near Infrared Spectroscopy,” J. Opt. Soc. Korea 16(1), 57–62 (2012). [CrossRef]
36. P. S. Gromski, H. Muhamadali, D. I. Ellis, Y. Xu, E. Correa, M. L. Turner, and R. Goodacre, “A tutorial review: Metabolomics and partial least squares-discriminant analysis – a marriage of convenience or a shotgun wedding,” Anal. Chim. Acta 879, 10–23 (2015). [CrossRef]
37. L. Sitole, F. Steffens, T. P. J. Krüger, and D. Meyer, “Mid-ATR-FTIR Spectroscopic Profiling of HIV/AIDS Sera for Novel Systems Diagnostics in Global Health,” OMICS 18(8), 513–523 (2014). [CrossRef]
38. N. Goswami, A. P. Blaber, H. Hinghofer-Szalkay, and V. A. Convertino, “Lower Body Negative Pressure: Physiological Effects, Applications, and Implementation,” Physiol. Rev. 99(1), 807–851 (2019). [CrossRef]
39. N. Goswami, J. A. Loeppky, and H. Hinghofer-Szalkay, “LBNP: past protocols and technical considerations for experimental design,” Aviat Space Environ. Med. 79(5), 459–471 (2008). [CrossRef]
40. W. H. Cooke, K. L. Ryan, and V. A. Convertino, “Lower body negative pressure as a model to study progression to acute hemorrhagic shock in humans,” J. Appl. Physiol. 96(4), 1249–1261 (2004). [CrossRef]
41. C. Hinojosa-Laborde, C. A. Rickards, K. L. Ryan, and V. A. Convertino, “Heart Rate Variability during Simulated Hemorrhage with Lower Body Negative Pressure in High and Low Tolerant Subjects,” Front Physiol. 2, 85 (2011). [CrossRef]
42. V. L. Kay and C. A. Rickards, “Reproducibility of a continuous ramp lower body negative pressure protocol for simulating hemorrhage,” Physiol Rep. 3(11), e12640 (2015). [CrossRef]
43. B. R. Soller, Y. Yang, O. O. Soyemi, K. L. Ryan, C. A. Rickards, J. M. Walz, S. O. Heard, and V. A. Convertino, “Noninvasively determined muscle oxygen saturation is an early indicator of central hypovolemia in humans,” J. Appl. Physiol. 104(2), 475–481 (2008). [CrossRef]
44. J. J. Durocher, J. R. Carter, W. H. Cooke, A. H. Young, and M. H. Harwood, “Cerebral Blood Flow Velocity During Combined Lower Body Negative Pressure and Cognitive Stress,” Aerosp. Med. Hum. Perform. 86(8), 688–692 (2015). [CrossRef]
45. W. Watkins, A. R. Hargens, S. Seidl, E. M. Clary, and B. R. Macias, “Lower-body negative pressure decreases noninvasively measured intracranial pressure and internal jugular vein cross-sectional area during head-down tilt,” J. Appl. Physiol. 123(1), 260–266 (2017). [CrossRef]
46. J. Lee, J. G. Kim, S. Mahon, B. J. Tromberg, K. L. Ryan, V. A. Convertino, C. A. Rickards, K. Osann, and M. Brenner, “Tissue hemoglobin monitoring of progressive central hypovolemia in humans using broadband diffuse optical spectroscopy,” J. Biomed. Opt. 13(6), 064027 (2008). [CrossRef]
47. Z. J. Schlader, E. Rivas, B. R. Soller, V. A. Convertino, and C. G. Crandall, “Tissue oxygen saturation during hyperthermic progressive central hypovolemia,” Am. J. Physiol. Regul. Integr. Comp. Physiol. 307(6), R731–R736 (2014). [CrossRef]
48. S. A. Bartels, R. Bezemer, F. J. de Vries, D. M. Milstein, A. Lima, T. G. Cherpanath, A. H. van den Meiracker, J. van Bommel, M. Heger, J. M. Karemaker, and C. Ince, “Multi-site and multi-depth near-infrared spectroscopy in a model of simulated (central) hypovolemia: lower body negative pressure,” Intensive Care Med. 37(4), 671–677 (2011). [CrossRef]
49. J. H. Siamwala, B. R. Macias, P. C. Lee, and A. R. Hargens, “Gender differences in tibial microvascular flow responses to head down tilt and lower body negative pressure,” Physiol Rep. 5(4), e13143 (2017). [CrossRef]
50. M. M. Tymko, C. A. Rickards, R. J. Skow, N. C. Ingram-Cotton, M. K. Howatt, and T. A. Day, “The effects of superimposed tilt and lower body negative pressure on anterior and posterior cerebral circulations,” Physiol Rep. 4(17), e12957 (2016). [CrossRef]
51. L. Wang, H. Ayaz, and M. Izzetoglu, “Investigation of the source-detector separation in near infrared spectroscopy for healthy and clinical applications,” J. Biophotonics 12(11), e201900175 (2019). [CrossRef]
52. D. D. Sheriff, I. H. Nadland, and K. Toska, “Role of sympathetic responses on the hemodynamic consequences of rapid changes in posture in humans,” J. Appl. Physiol. 108(3), 523–532 (2010). [CrossRef]
53. V. L. Kay and C. A. Rickards, “The role of cerebral oxygenation and regional cerebral blood flow on tolerance to central hypovolemia,” Am. J. Physiol. Regul. Integr. Comp. Physiol. 310(4), R375–R383 (2016). [CrossRef]
54. N. Goswami, E. Grasser, A. Roessler, D. Schneditz, and H. Hinghofer-Szalkay, “The cardiovascular response to lower body negative pressure in humans depends on seal location,” Physiol Res. 58, 311–318 (2009).
55. L. B. Panton, W. D. Franke, D. A. Bleil, S. M. Baier, and D. S. King, “Effects of resistance training on cardiovascular responses to lower body negative pressure in the elderly,” Clin. Physiol. 21(5), 605–611 (2001). [CrossRef]
56. Z. Laszlo, A. Rossler, and H. G. Hinghofer-Szalkay, “Cardiovascular changes during and after different LBNP levels in men,” Aviat Space Environ. Med. 69, 32–39 (1998).
57. M. B. Dikshit, “Lower-body suction and cardiovascular reflexes: physiological and applied considerations,” Indian J. Physiol. Pharmacol. 34, 3–12 (1990).
58. A. C. Justice, K. E. Covinsky, and J. A. Berlin, “Assessing the Generalizability of Prognostic Information,” Ann. Intern. Med. 130(6), 515–524 (1999). [CrossRef]
59. S. E. Bleeker, H. A. Moll, E. W. Steyerberg, A. R. T. Donders, G. Derksen-Lubsen, D. E. Grobbee, and K. G. M. Moons, “External validation is necessary in prediction research,” J. Clin. Epidemiol. 56(9), 826–832 (2003). [CrossRef]
60. G. Themelis, H. D’Arceuil, S. G. Diamond, S. Thaker, T. J. Huppert, D. A. Boas, and M. A. Franceschini, “Near-infrared spectroscopy measurement of the pulsatile component of cerebral blood flow and volume from arterial oscillations,” J. Biomed. Opt. 12(1), 014033 (2007). [CrossRef]
61. S. Houtman, J. M. Serrador, W. N. Colier, D. W. Strijbos, K. Shoemaker, and M. T. Hopman, “Changes in cerebral oxygenation and blood flow during LBNP in spinal cord-injured individuals,” J. Appl. Physiol. 91(5), 2199–2204 (2001). [CrossRef]
62. T. Hachiya, I. Hashimoto, M. Saito, and A. P. Blaber, “Peripheral vascular responses of men and women to LBNP,” Aviat Space Environ. Med. 83(2), 118–124 (2012). [CrossRef]
63. K. Hunt, I. Tachtsidis, K. Bleasdale-Barr, C. Elwell, C. Mathias, and M. Smith, “Changes in cerebral oxygenation and haemodynamics during postural blood pressure changes in patients with autonomic failure,” Physiol. Meas. 27(9), 777–785 (2006). [CrossRef]
