Cognitive fatigue due to exercise under normobaric hypoxia

Ethics statements

This study was approved by the Institutional Ethics Committee of the University of Tsukuba, Faculty of Health and Sport Sciences (approval number: Tai 025–120) and was conducted in accordance with the latest version of the Declaration of Helsinki. The study participants provided written informed consent for participation and publication of their details.


Fourteen right-handed young adults (12 men and 2 women) participated in this study (Table 2). The sample size was determined by assuming that the effects of exercise in hypoxic environments would be similar to those in our previous study8. All participants were Japanese native speakers, healthy, and naive to the experimental procedures for which they volunteered. None of the participants had a history of neurological, psychiatric, or respiratory disorders or a disease requiring medical care. All the participants had normal or corrected-to-normal vision and normal color vision. All participants were asked to refrain from exercise and the consumption of alcohol and caffeine for at least 24 h prior to each experiment to control for external factors that could affect cardiovascular and executive functions. Post-hoc sensitivity analysis performed based on this sample with 80% power and 0.05 alpha demonstrated sufficient sensitivity to detect repeated-measures effects exceeding f = 0.40 and t-test differences exceeding d = 0.81 (with a two-tailed alpha), as computed using G*Power (

Table 2 Participants’ characteristics.

Experimental procedures

On the first day, participants underwent a graded exercise test to measure their \(\dot\textV\textO_2\textpeak\) and determine the appropriate individual intensity for moderate exercise, which was defined as 50% of a participant’s \(\dot\textV\textO_{2\textpeak}\) based on the American College of Sports Medicine’s classification of physical activity intensity36. The detailed procedures for the graded exercise test were the same as those in our previous study8. The participants practiced the CWST twice before being subjected to the main experimental conditions. A few days after the first visit, two main experimental conditions were conducted in a single-blind (participant being blinded) crossover study design: exercise under moderate normobaric hypoxia (HE) or ME, which improved the inhaled oxygen concentration to suppress SpO2 during exercise. All participants participated in both the HE and ME conditions, each on separate days, with the order counterbalanced across participants (Fig. 5). In both conditions, participants underwent the CWST before and 15 min after 10 min of moderate-intensity exercise on a recumbent cycle ergometer (Strength Ergo 240 W, Mitsubishi Electric Corp., Tokyo, Japan) at 60 revolutions per minute, based on our previous study methods8. Cortical hemodynamic changes in the l-DLPFC were monitored using fNIRS while the participants performed the CWST.

Figure 5
figure 5

(A) The two conditions, breathing a moderately hypoxic gas (hypoxic exercise [HE]) and breathing a milder hypoxic gas, during which oxygen was added to the moderately hypoxic gas to maintain the oxygen saturation (SpO2) level during milder hypoxic exercise (ME). Cortical hemodynamic changes were monitored with functional near-infrared spectroscopy (fNIRS) while participants performed the Stroop task. HR, heart rate. (B) In both conditions, the exercise and color-word Stroop task (CWST) were performed on a recumbent cycle ergometer. Hypoxic gas (\(\textF_\textIO__2\) = 0.135) stored in a Douglas bag was inhaled through a mask. (C) In the ME condition, SpO2 was adjusted during exercise by adding oxygen gas with humidity to prevent the participant’s throat from drying out to the hose connecting the Douglas bag to the mask.

In the HE condition, participants breathed hypoxic gas, as in our previous study8, which was a mixture of 13.5% oxygen and 0.03% carbon dioxide in nitrogen (\(\textF_\textIO__2\) = 0.135; equivalent to an altitude of approximately 3,500 m) through a mask connected to a Douglas bag. In the ME condition, as in the HE condition, participants breathed a moderately hypoxic gas during the CWST, but only during exercise; they breathed a milder hypoxic gas (\(\textF_\textIO__2\) = 0.161 ± 0.018) to regulate the hypoxic gas so their SpO2 did not decrease from the resting state. To adjust the oxygen concentration, oxygen was injected directly into the hose while monitoring the SpO2 during exercise. Expired air was exhausted directly outside the mask so participants did not re-breathe it. The participants were exposed to a moderately hypoxic gas 10 min before the pre-Stroop session while sitting on a cycle ergometer. HR was monitored by a heart rate monitor (V800, Polar Electro, Kempele, Finland), SpO2 was monitored by a pulse oximeter (OLV-3100, Nihon Kohden, Tokyo, Japan) placed on the left earlobe, and exhaled gas was monitored every minute by a gas analyzer (Aeromonitor AE-310S; Minato Medical Science, Osaka, Japan). The participants’ RPE37 were recorded before exposure to hypoxia, every minute during exercise, and before the CWST.

Behavioral measurements

The CWST1,2,3,8,15,38,39,40,41,42 was adopted in an event-related design. The CWST, which included two rows containing letters or words, was presented on a screen, and the participants were instructed to decide whether the color of the letters or words in the top row corresponded to the color name presented in the bottom row. Participants pressed a “yes” or “no” button with their right forefinger or middle finger to respond. The RT and error rates were measured.

The CWST consisted of three trials, including 10 neutral, 10 congruent, and 10 incongruent trials. For neutral trials, the top row contained sets of X’s (XXXX) written in red, blue, green, or yellow, and the bottom row contained the words “RED,” “BLUE,” “GREEN,” or “YELLOW” written in black. For congruent trials, the top row contained the words “RED,” “BLUE,” “GREEN,” or “YELLOW” written in a color congruent with that of the bottom row. For incongruent trials, the color word in the top row was written in an incongruent color to produce interference between the color of the word and the color name. All words were written in Japanese. The correct answer rate assigned to “yes” and “no” was 50%. Each stimulus was separated by an inter-stimulus interval showing a fixation cross for 9–13 s to avoid predicting the timing of the subsequent trial1,2,8,15,38,39,40,42. The stimulus remained on the screen until a response was given or for 2 s. In the present study, Stroop interference, a specifically defined cognitive process, was used to elucidate the effect of an acute bout of moderate exercise under hypoxic conditions on executive function. Therefore, the (incongruent-neutral) contrast, which is assumed to represent Stroop interference, was calculated.

fNIRS measurements

We used a multichannel fNIRS optical topography system (ETG-7000, Hitachi Medical Corporation, Chiba, Japan) set with two wavelengths of near-infrared light (785 and 830 nm). We analyzed the optical data from fNIRS based on the modified Beer–Lambert law43, as previously described44. This method allowed us to calculate signals reflecting the oxy-Hb, deoxy-Hb, and total hemoglobin concentration changes, calculated in millimolar-millimeters (mMmm)44. The composition of the fNIRS probe holder and its placement followed the same procedure used in our previous studies1,2,3,8,38,39,40,42,45. To register fNIRS data in the Montreal Neurological Institute (MNI) space, we used virtual registration46,47. Briefly, this method enables the placement of a virtual probe holder on the scalp by stimulating the holder’s deformation and registering probes and channels onto a reference brain in a preconstructed magnetic resonance imaging database48,49. We probabilistically estimated the MNI coordinate values for the fNIRS channels to obtain the most likely estimate of the location of the given channels for the group of participants and the spatial variability associated with the estimation50,51.

Analysis of the fNIRS data

In this study, neural activity related to the CWST was evaluated by examining changes in oxy-Hb, as shown in our previous studies1,2,3,8,38,39,40,42,45. Individual timeline data for the oxy-Hb signal of each channel were preprocessed with a band-pass filter using a cut-off frequency of 0.04 Hz to remove baseline drift and 0.3 Hz to filter out heartbeat pulsations. Channel-wise and subject-wise contrasts were obtained by calculating the inter-trial mean of differences between the oxy-Hb signals of the peak (6–10 s after trial onset) and baseline (0–2 s before trial onset) periods based on our study8. The contrasts were calculated as prefrontal activation elicited by a cognitive task, and the contrasts obtained were subjected to second-level random-effects group analysis.

Based on a method widely used in anatomical labeling systems, such as the LBPA4049, channels 13, 14, 16, and 17 were combined to analyze l-DLPFC activity, which was found to decrease after moderate-intensity exercise under hypoxic conditions in a previous study8.

Statistical analysis

The HR, RPE, and SpO2 were subjected to repeated measures two-way ANOVA with condition (HE/ME) and time (before exposure to hypoxia/pre-Stroop/during exercise/post-Stroop) as within-subject factors. The RT and error rate were subjected to repeated measures three-way ANOVA with trial (incongruent/neutral), condition (HE/ME), and session (pre/post) as within-subject factors to examine whether the general tendencies for the Stroop task could be reproduced in all conditions. The Stroop effect associated with acute moderate exercise on all outcome measures was analyzed using repeated measures two-way ANOVA with condition (HE/ME) and session (pre/post) as within-subject factors. When a significant F-value was obtained, a post hoc test using the Bonferroni method for multiple corrections was applied to identify significant differences among the mean values.

Moreover, to clarify the relationships of physiological parameters (SpO2, ETCO2, and \(\dotV_\textE\)) during exercise with executive performance and task-related brain activation, we conducted parametric Pearson correlation analyses.

All data are presented as mean ± standard error. Statistical significance was set a priori at P < 0.05 for all comparisons. Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 26 (SPSS Inc., Chicago, IL, USA).

Health and Fitness