Physiological or emotional stressors can increase symptoms of mild traumatic brain injury (mTBI) in recently injured individuals, and either type of stress can induce symptoms in individuals with mTBI who are asymptomatic when not stressed (1). For example, compared to groups of normal controls, mTBI groups showed higher heart rate (2) and decreased heart rate variability (HRV) (3, 4) as well as reduced cerebrovascular reactivity and lower end-tidal carbon dioxide (PETCO₂) trends (5, 6) in response to experimental stress. The stressors used in these experiments included exercise (3, 7), repeated breath holds and hyperventilation (5, 6), psychosocial stress (8), and mental arithmetic under conditions of noise and bright lights (2).

Hypoxia is a physiological stressor that may be readily encountered in daily life and can be relatively easily and reliably generated in a controlled laboratory setting, thus permitting precise quantification of the hypoxic physiological stressor and its effects on the volunteer. Ewing et al. (9), using an altitude chamber to generate altitude-related hypoxia, exposed a group of controls and a group of asymptomatic, concussed individuals to a simulated altitude of 3,800 m (nearly 12,500 ft). These authors reported finding a significant cognitive performance deficit in the brain-injured group at altitude, a difference which disappeared when the subjects were returned to ground level. Temme et al. (10, 11) extended the Ewing et al. study using normobaric hypoxia (NH) rather than an altitude chamber to generate hypoxic stress. NH mixes nitrogen (N₂) with mean sea level (MSL) air to reduce the percentage of inspired oxygen from the normal sea level value of 21%, a technique that is widely used in current respiratory physiology. Temme et al. found (1) at MSL, there were no differences between the mTBI and control groups in cognitive performance; (2) exposure to NH with 15.5% O₂, 14% O₂, and 13% O₂ (approximately equivalent to breathing air encountered at 8,000, 12,000, and 14,000 ft above MSL, respectively) resulted in a decrease in cognitive performance for both groups; (3) when hypoxic, the visual short-term memory of the mTBI group was significantly poorer than that of the control group; and (4) there was no difference between the groups on returning to 21% O₂, normal MSL air. These studies support the idea that NH is an effective stressor that can uncover symptoms of mTBI in individuals who are asymptomatic when unstressed.

Eye movements are a sensitive indicator of mTBI and can provide a great deal of objective diagnostic information about the concussed brain (12–15). Recently, instrumentation to measure a pupil’s response to light flashes, and the eye’s ballistic movements, saccades, was integrated into a device designed to screen for physiological impairments due to fatigue, sleep deprivation, drug and alcohol intoxication, and other physiological stressors (16–21). The device, the Functional Impairment Tester (FIT), measures ocular motility along four dimensions: pupil diameter in the dark, the latency of the pupil’s response to flashes of light, the pupil’s amplitude of constriction, and the eye’s saccadic velocity (17, 18, 20). The literature reports that these dimensions of the pupil, i.e., the pupillary light reflex (22) and the eye’s saccadic behavior (12), can be affected by mTBI. Moreover, FIT measures have demonstrated sensitivity to hypoxia (17). To date, however, the FIT has not been established as an assessment or screening tool for concussion or mTBI. Since the behavior of the eye is a standard clinical indicator of brain injury and since vision deteriorates with hypoxia, the FIT was selected as a potential non-subjective oculomotor test to differentiate between controls and mTBI subjects during NH stress. The present study extends the Temme et al. (10) cognitive finding by evaluating response differences to NH between the mTBI group and the matched healthy control (HC) group in the physiological measures of peripheral oxygen saturation (SpO₂) and pulse rate, as well as oculometric.

The volunteers and the procedures used in this study were described previously in detail by Temme et al. (10). The study was reviewed, approved, and conducted in accordance with all Federal and State laws, regulations, and standards of practice, and in accordance with all Department of Defense and U.S. Army regulations and procedures. The research protocol was approved by the Chesapeake Institutional Review Board, the U.S. Army Aeromedical Research Laboratory Human Use Committee, and the U.S. Army Medical Research and Materiel Command Human Subjects Research Review Board. The study was determined to pose a greater than minimal risk to the subjects; and as part of risk mitigation, the NH stress conditions were presented in an increasing sequence of severity with each subject being observed at a lower stress condition before being exposed to a greater stress.

Figures 1 and 2 show the group average response to each percent O₂ condition for each minute of exposure. These figures illustrate the average group response to reduced O₂ over the period of exposure. Average SpO₂ (Figure 1) for the two groups was the same at the Baseline and the Post Stress conditions. At the beginning of each NH stress condition, SpO₂ for the mTBI and HC groups was indistinguishable; but with continued hypoxic exposure, SpO₂ decreased for both groups, decreasing more for the HC group than the mTBI group. The pulse rate response (Figure 2) tended to increase for both groups from the Baseline condition through the 13% O₂ condition, but a marked drop in pulse rate is evident at the conclusion of the hypoxic exposures, as reflected in the pulse rate recorded at Post Stress.

The three levels of NH stress used in this study produced SpO₂ and pulse rate responses that differed between the mTBI and matched HC groups. However, SpO₂ and pulse rate were indistinguishable between mTBI and HC groups during Baseline, at the initiation of each NH stress condition, and during the Post NH Stress condition. Thus, consistent with previous stress and mTBI research involving physiological metrics, the mTBI group was not different from the matched HC group when not stressed.

In mammals, SpO₂ normally decreases with reduced inspired O₂ concentrations. The findings reported here showed the mTBI group had SpO₂ values that were consistently higher, i.e., less reduced, than the HCs throughout each NH stress condition. Furthermore, at a constant percent O₂, the falloff in SpO₂ over time was less for the mTBI than for the HC group (Figure 1). These findings, the first to compare SpO₂ between mTBI and HC groups under NH in humans, may seem counterintuitive. However, it may be noted that a normal response to hypoxia is an increase in ventilation, the hypoxic hyperventilatory response (HVR). An increased HVR in the mTBI group could account for the increase in SpO₂ relative to the HC group and such a differential effect on the HVR may reflect changes associated with mTBI in the autonomic and cardiovascular systems (30). Furthermore, the observation that pulse rate trended lower in the hypoxic mTBI subjects than in hypoxic HC would also be consistent with an increased HVR in the mTBI group. The report by Len et al. (5) that PETCO₂ is lower in the mTBI group than HC group when breath holding and when hyperventilating seems consistent with this idea. Unfortunately, a broader set of respiration measures (e.g., ETO₂, respiration rate, minute ventilation, and total volume of gas expired per condition) was not collected in either the present study nor reported by Len et al. Exploring the relationships among blood oxygenation and respiration in those with a history of mTBI who are currently asymptomatic and appear to have fully recovered from injury is necessary to further explore the latent impact of brain injury.

On the other hand, the SpO₂ differences reported here could possibly occur because of differences in physical fitness between the two subject groups, with the mTBI group being more physically fit than the HC group. However, this seems unlikely since the HCs were selected to ensure the groups did not differ statistically along the dimensions of age, height, BMI, systolic and diastolic blood pressure, pulse rate, or respiration rate; moreover, the pulse rate response was not significantly different between groups.

It may be noted that the pulse rate recorded during Post Stress (Figure 2), while not statistically different between the mTBI and HC groups, was statistically significantly less than the pulse rate recorded during the Baseline, despite the fact that neither condition posed any hypoxic stress. This significant decrease in pulse rate immediately following the conclusion of the 13% O₂ is consistent with previous findings reported in the literature. Roche et al. (31) measured cardiovascular response in a group of healthy volunteers before, during, and after a 15-min hypoxic exposure with 11% O₂ stress and reported that heart rate increased 19% over the initial baseline during hypoxia; but more interesting, during the 15-min recovery period heart rate decreased 10% below baseline. The methods used by Roche et al., which included measures of HRV, assessed the autonomic nervous system and pointed to an exaggerated sympathetic response to NH followed by an enhanced parasympathetic response during the recovery period, which is completely consistent with the findings reported here.

The FIT showed sensitivity to hypoxia as noted in Table 8; the main effects for NH stress were significant for diameter and response amplitude of the pupil and saccadic velocity. The main effect for NH stress was consistent with the study by Cymerman et al. (17), which showed changes in pupil diameter, response amplitude, and latency but without changes in saccadic velocity following 3 h of continuous exposure to an altitude of 4,300 m (14,107 ft). However, since the FIT failed to show differences between the mTBI and HC groups it is possible that there are, in fact, no differences in the ocular motor behavior between the two subject groups; but based on the emerging scientific literature describing oculometrics in mTBI (12, 22, 32–34), it seems more likely that the FIT was merely insufficiently sensitive to differences that do exist. It should be noted that the FIT was designed as a non-invasive, oculomotor, performance-based fitness for duty impairment screener, and as such, the FIT may simply have been inadequate for the purposes of this research.

Normobaric hypoxia did not produce undesirable physiological outcomes in the mTBI subjects in this study. SpO₂ and pulse rate values did change with increased NH; however, values did not approach limits considered to be dangerous to health. Also, no trial had to be stopped due to health concerns or extreme discomfort experienced by the subjects. Unlike the breathing behavior alterations used by Len et al. (5, 6), the ROBD produces a sustained physiological stress and affords the opportunity to administer additional tests under a range of constant NH stress levels. These effects were reversed and disappeared when the subjects were returned to sea level and room air. Our results are consistent with the findings reported by Ewing et al.; individuals with a history of mTBI may be particularly susceptible to the effects of mild or moderate hypoxia. The hypobaric hypoxic research paradigm that Ewing et al. (9) used requires a significant infrastructure to support. The NH method used here is portable, and by comparison with the hypobaric paradigm, is safe, inexpensive, and easy to create in most laboratory or clinical settings.

It is important to acknowledge limitations of the present study. One important limitation was the wide range of the time between a volunteer’s trauma and test participation, a range from 1.7 to 119.7 months. The elapsed time since injury was not controlled. Second, physical activity level (e.g., hours per week) or fitness level (VO₂ max) were not used in the matching criteria between subjects. Third, the pulse oximeter data were very limited, recorded once a minute by hand. Finally, the history of brain injury was determined by self-report. It is possible that individual subjects were not accurate about their brain injury history, a limitation that could be present in both groups.

When exposed to mild or moderate normobaric hypoxic stress conditions simulating altitudes up to 14,000 ft above MSL: (1) peripheral oxygen saturation (SpO₂) differences emerged between the mTBI and matched HCs, (2) pulse rate trended lower in the mTBI group, and (3) FIT oculometrics were not sensitive to group differences. These findings demonstrate that a relatively minor hypoxic challenge can reveal measurable differences in SpO₂ and pulse rate but not FIT measures, in otherwise asymptomatic individuals with a history of mTBI.

LT had overall responsibility of the project. PS was primary lead on data analysis and an equal contributor to manuscript writing. JB was an equal contributor on data analysis, interpretation, and manuscript writing.

The opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the U.S. Army and/or the Department of Defense.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.