Effects of meditation or walking as a microbreak on fatigue recovery during simulated air traffic control tasks

Abstract

Air traffic control (ATC) is a demanding domain characterized by intense operational demands, including high workload and task complexity that act as significant contributors to operator fatigue alongside other physiological and psychological factors. This study investigates the effects of different types of microbreak interventions on post-task fatigue recovery among air traffic controllers (ATCOs). Using a simplified simulated ATC task, we recruited ATC-naïve participants to evaluate two types of microbreaks—meditation and walking—as well as a control condition without microbreaks. We recorded participants’ self-perceived fatigue, eye movements, heart rate variability (HRV), critical flicker fusion (CFF), heart rate, blood pressure, and performance metrics during the simulated task. The results indicate that taking a microbreak is effective in promoting subjective fatigue recovery. Furthermore, microbreaks had a significant effect on the preservation of task performance, with their effectiveness differing across microbreak types. The walking microbreak was particularly effective in reducing operational errors, whereas the meditation microbreak showed a non-significant trend toward mitigating declines in reaction time. The findings suggest that both microbreak types are beneficial for fatigue recovery, but walking microbreaks may confer additional benefits for sustaining operational performance.

Introduction

Air traffic control (ATC) is one of the most demanding and challenging domains, requiring operators to maintain sustained attention, highly precise decision-making, and rapid reactions in a fast-paced, high-stakes environment. Fatigue, characterized by physical and/or mental weakness and exhaustion, can lead to diminished mental and/or physical performance¹⁾. Fatigue has been defined as feeling drowsy, worn out, or exhausted^{1, 2)}. The International Civil Aviation Organization (ICAO) defines fatigue as a physiological state of diminished mental or physical performance capacity that can affect a person’s alertness and capacity to carry out safety-related operational duties³⁾.

Air traffic controllers (ATCOs) fatigue

ATCOs must simultaneously monitor several planes at once and issue accurate instructions at the appropriate time. ATCO tasks are cognitively demanding, requiring sustained attention and high-precision information processing⁴⁾. Numerous studies have investigated the underlying factors contributing to ATCO fatigue, revealing that it is primarily associated with irregular shift patterns, night shifts, and inadequate rest periods. For example, Peukert et al. found that an excessive number of consecutive working days and a rapid succession of night and early shifts greatly increased subjective fatigue⁵⁾. Roach et al. underscored that irregular work hours frequently result in sleep deprivation and altered circadian rhythms for ATCOs, culminating in chronic sleep debt, which is a significant contributor to fatigue⁶⁾. ATCO fatigue can lead to delayed responses and slower information processing¹⁾.

This fatigue poses a serious threat to ATCO performance and aviation safety. A UK review (1980–1988) found that civil controller errors contributed to 21.5% of reported aviation incidents⁷⁾, whereas another review showed that 13% of ATCO incident reports cited fatigue-related factors⁸⁾. Chen et al. found that nearly 50% of surveyed controllers reported feeling fatigued after work due to shift work, heavy workload, and stress⁹⁾. With global air traffic expected to grow, and passenger numbers projected to exceed 7 billion by 2035^{9, 10)}, the workload and stress on ATCOs are expected to rise, further exacerbating fatigue and its consequences.

Fatigue recovery

Current regulations and research

Research in non-aviation fields has established a broad consensus on effective fatigue countermeasures. A large body of empirical studies shows that methods such as sleep and napping, regular rest breaks, microbreaks, and exercise or relaxation interventions are widely considered effective for alleviating work-related fatigue and improving cognitive performance. For instance, planned napping has been proven to reduce sleepiness and improve performance among night-shift workers across various professions, including healthcare, manufacturing, and utilities^{11,12,13,14,15)}. Regular rest breaks and microbreaks have been found to restore attention and self-control, and improve productivity among a wide range of workers^{16, 17)}. Similarly, studies consistently show that exercise and relaxation interventions have significant effects on reducing work-related fatigue. Research on call center operators found that stretching and joint mobilization exercises can effectively reduce muscle discomfort and fatigue levels¹⁸⁾. For the general working population, exercise and relaxation therapies have been shown to significantly reduce occupational stress and fatigue¹⁹⁾. Moreover, in the medical field, studies have confirmed that surgeons who implement microbreaks with light exercise during long surgeries effectively reduce subjective fatigue and maintain operational accuracy^{20, 21)}.

While these methods are well established as effective in the general population, studies on “in-shift” fatigue recovery in high-attention-demanding occupations, such as ATC, remain limited. The ICAO Manual for the Oversight of Fatigue Management Approaches focuses primarily on sleep and scheduling but also emphasizes the importance of in-shift breaks for preventing fatigue accumulation and improving performance¹⁰⁾. However, ICAO Annex 11²²⁾ states that governments are required to implement fatigue management measures, whereas the introduction of a fatigue risk management system (FRMS) remains optional, resulting in a lack of standardized fatigue management methods.

Few studies have examined the effectiveness of in-shift break strategies for ATCOs. Lesiuk²³⁾ reported that 15–20 min of listening to music or resting significantly reduced anxiety but did not produce measurable changes in physiological indicators. Signal et al.²⁴⁾ demonstrated that a 40-min nap taken during night shifts significantly improved alertness and task performance.

Microbreak

In this study, microbreaks are defined as breaks lasting less than 10 min, and prior research across multiple occupational domains has demonstrated their effectiveness in mitigating fatigue and supporting performance. Experimental studies have shown that performance on mental addition tasks improves with two 3-min breaks every 50 min¹⁶⁾, while meat-processing workers exhibit enhanced performance when provided with 9-min breaks every 51 min²⁵⁾. Similarly, studies in surgical settings indicate that surgeons who incorporate brief in-task breaks experience reduced fatigue and maintain higher levels of performance²⁰⁾. Albulescu et al.²⁶⁾ further demonstrated that microbreaks help sustain vigor and attenuate fatigue over time.

With respect to specific microbreak modalities, sitting meditation has been shown to alleviate mental fatigue and reduce stress and anxiety^{19, 23)}. Other microbreak activities—including listening to music²³⁾, walking^{16, 20)}, exercising^{18, 20)}, reading, and watching television¹⁶⁾ have been shown to alleviate physical fatigue and reduce stress and anxiety. Previous studies indicate that eyestrain is the most frequently reported symptom of fatigue among controllers⁹⁾. Although eyestrain is a physiological sensation, in the context of ATC, it is closely associated with mental fatigue. The sustained cognitive workload and continuous visual attention required in ATC operations contribute to reduced self-reported ocular comfort, making eyestrain an important subjective manifestation of the overall mental fatigue state. However, music-based interventions require individualized music selection and do not produce effects on anxiety or physiological measures that differ significantly from those of sitting meditation²³⁾. Therefore, considering the operational characteristics of ATC and the feasibility of workplace implementation, meditation (mental relaxation) and walking (physical relaxation) were selected as the microbreak interventions to be systematically compared in this study.

Regarding the duration of microbreaks, an optimal length may exist, but it has not yet been established^{17, 26)}. To maintain concentration and prevent fatigue accumulation, Japanese ATCOs are required to take multiple short breaks during their shifts²⁷⁾.

Standardized measures of fatigue

Fatigue is a multidimensional construct with complex characteristics. Three primary approaches are used to assess fatigue: subjective reports, physiological indicators, and performance measures. Employing all three assessment methods helps minimize bias in evaluating fatigue^{28, 29)}.

Specifically, subjective rating scales, such as the Visual Analog Scale (VAS), are widely used as straightforward tools to quantify fatigue intensity in both clinical and experimental settings³⁰⁾. The VAS has been validated as a reliable instrument for assessing both fatigue and energy levels³¹⁾. These subjective assessments are complemented by physiological indicators, including critical flicker fusion (CFF) and heart rate variability (HRV). CFF is a validated measure for detecting changes in central nervous system load and visual fatigue³²⁾, whereas HRV reflects autonomic regulation and is widely applied in fatigue and stress research^{33, 34)}. In human factors research, objective performance metrics, including reaction time and task accuracy, have been validated as reliable indicators of cognitive impairment^{33, 34)}.

In this study, we employed these validated tools to assess subjective, physiological, and performance-based fatigue, ensuring a comprehensive and reliable evaluation.

Purpose and hypotheses

Based on the preceding evidence, workplace microbreaks may facilitate fatigue recovery among ATCOs. This study aimed to investigate the effects of two microbreak types—meditation and walking—on fatigue recovery, with the goal of identifying a practical and effective short-term break for implementation in the ATC workplace. In addition to self-reported fatigue, the study evaluated physiological and performance indicators, including eye movements, HRV, CFF, heart rate, and blood pressure. Based on the above rationale, the following hypotheses were formulated:

H1: Microbreaks mitigate a portion of the fatigue induced by ATC tasks.

H2: Individuals who take microbreaks exhibit superior task performance compared to those who do not.

Subjects and Methods

Participants

The study recruited 44 participants (29 males and 15 females) aged 20–35 yr (mean age=24.9, SD=2.7). Participants were nonexpert individuals with no prior ATC experience. This study was designed as a fundamental investigation of fatigue, aiming to examine core cognitive processes relevant across populations. In human factors and fatigue research utilizing simulated ATC tasks, the use of nonexpert participants is a widely accepted and validated methodology^35,36,37,38). The primary advantage of this approach is that it allows investigators to focus on core cognitive mechanisms associated with fatigue, rather than on the specific operational skills of ATCOs. This approach enabled clearer observation of fatigue effects on fundamental cognitive functions (e.g., sustained attention and information processing), thereby enhancing the generalizability of the findings.

The study was conducted with approval from the Institute of Systems and Information Engineering Research Ethics Committee of the University of Tsukuba (approval number: 2022R606-1). Informed consent was obtained from all participants, who were provided with information regarding the study’s purpose, potential risks, and anticipated benefits. Participants were also informed that participation was voluntary and that they could withdraw at any time without any consequences.

Simulator

The simulator employed in this study was APP Control Lite (version 3.7; APPcalyptus UG, Frankfurt, Germany), a high-fidelity ATC simulator designed to closely replicate the approach control radar display and integrate the primary functions of the approach control communications position. This study, designed as a fundamental investigation of fatigue, utilized the simulator to generate a high cognitive load environment by simulating ATC operations. The simulator has been employed in previous studies as a training tool for novice ATC operators³⁹⁾. The simulated environment featured two runways for takeoff and two for landing centrally located within the sector, with seven waypoints positioned along the sector boundaries to allow aircraft entry and exit. Participants provided verbal instructions to pilots to adjust aircraft altitude, airspeed, and heading, ensuring that each aircraft reached its designated waypoint at the target altitude. Participants used mouse clicks to access specific information, including aircraft flight number, type, destination, target altitude, current altitude, and current airspeed. The system generated a warning when horizontal separation between aircraft fell below 5 NM and vertical separation was less than 2,000 ft. A loss of separation was determined when horizontal separation fell below 3 NM and vertical separation was less than 1,000 ft (Fig. 1).

Fig. 1.

Interface of the APP Control Lite (version 3.7) used for simulated air traffic control tasks.

Task performance measures

To evaluate participants’ operational efficiency and safety awareness, task performance was measured using two indicators: loss-of-separation events and reaction duration. A loss-of-separation event was recorded whenever horizontal separation between aircraft fell below 3 nautical miles and vertical separation dropped below 1,000 feet, according to simulator settings. Reaction duration was defined as the time interval between the entry of a new aircraft into the sector and the participant’s first fixation on it. This metric was derived from data collected using Tobii Pro Glasses 3 (Tobii AB, Stockholm, Sweden), a device capable of accurately correlating eye-tracking timestamps with aircraft entry times, thereby providing a precise estimate of each participant’s detection time for incoming aircraft.

Experimental design

A between-subjects experimental design was adopted to prevent potential learning and residual fatigue effects associated with within-subjects designs, thereby ensuring the internal validity of the findings. A substantial body of research indicates that, in fatigue and recovery experiments, residual fatigue from a prior condition can compromise subsequent task performance, thus affecting the validity of the results^40,41,42). For example, in sleep deprivation and recovery studies, residual fatigue has been shown to affect alertness and cognitive performance^{37, 43, 44)}. Similarly, research on shift workers indicates that residual fatigue can systematically impair multitasking performance⁴⁵⁾. Therefore, a between-subjects design was considered the most appropriate approach for this type of research. Participants were randomly assigned to Groups A, B, and C (15, 15, and 14 participants, respectively) to receive a meditation microbreak, a walking microbreak, or non microbreak during the experiment. Random assignment ensured that the groups were balanced on key characteristics, including gender (Group A: 10 males/5 females; Group B: 10 males/5 females; Group C: 9 males/5 females) and time of day (morning/afternoon), thereby minimizing potential confounding variables. Participants in Group A were instructed to sit in a chair, close their eyes, and focus on their breathing. If focusing on the breath proved difficult, they were encouraged to count silently coordinated with their breath. Meditation took place in an office environment, where ambient sounds of colleagues working on computers were present. Participants in Group B were instructed to walk in an indoor circular corridor adjacent to the lab. The corridor featured large floor-to-ceiling windows, providing an unobstructed view of the outdoors.

Scenario

During each ATC task, approximately 23 aircraft entered the sector at random from either waypoints or runways. Participants provided verbal commands regarding aircraft altitude, speed, and heading to guide each aircraft to its destination at the target altitude. Participants were instructed to maintain proper aircraft separation and avoid triggering separation warnings. If a warning occurred, participants were required to direct aircraft to restore safe separation promptly.

Procedure

As shown in Fig. 2, following a 30-min learning period on the ATC simulator, participants took a 3- to 5-min break before beginning the 30-min formal ATC Task 1. Participants’ skills were assessed by observing their performance during the final 10 min of the training task following the learning period. The pass criterion required that no loss-of-separation events occurred during this period. Participants who did not meet this criterion received an additional 15 min of training. If participants still failed to meet the standard, their data were excluded from analysis. Participants in the meditation and walking groups took a 10-min microbreak after Task 1, whereas participants in Group C did not take a break. This was followed by the 30-min formal ATC Task 2. To prevent learning effects, aircraft names and routes were randomized for all formal tasks, ensuring that they did not repeat those used during the training session.

Fig. 2.

Experimental procedure.

Measurements

The independent variable in this experiment was the type of microbreak implemented between two simulated ATC tasks: meditation, walking, or non microbreak. Dependent measures included subjective assessments, eye movement metrics, CFF threshold, heart rate (beats per minute), HRV, and blood pressure. Fatigue can be evaluated through self-assessment as well as physiological responses, including fixation, saccades, blinks, pupil diameter, CFF, heart rate, HRV, and blood pressure^{1, 46,47,48,49,50,51,52,53)}.

Subjective fatigue was assessed using a method similar to a VAS, implemented via the Windows version of FHM Safety (National Institute of Occupational Safety and Health, Japan, Kiyose, Tokyo, Japan) software. Participants indicated their perceived level of fatigue by dragging a slider on a computer screen using a mouse. The slider was anchored with “completely fine” at the left end and “completely fatigued” at the right end (Fig. 3a). The slider position was quantified as a numerical value from 0 to 100, with higher values indicating greater fatigue. Subjective assessments were conducted before Task 1, after Task 1, before Task 2, and after Task 2.

Fig. 3.

Screenshot of FHM Safety software (Windows version).

Subjective assessments are known to be influenced by motivation, personal factors, experience, and training⁵⁴⁾. Therefore, physiological measurements were conducted concurrently. During Tasks 1 and 2, eye movement metrics—including pupil diameter, fixation duration and frequency, saccade duration and frequency, and blink duration and frequency—were recorded using Tobii Pro Glasses 3. These metrics served as objective indicators of participants’ cognitive state and fatigue levels^{9, 47,48,49, 51, 52, 55)}.

The critical flicker fusion (CFF) was measured using FHM Safety (Windows version) software (Fig. 3b). The test was conducted in a well-lit office environment, with participants seated approximately 45 cm from the computer screen. During each trial, one of the four directional arrows on the screen flickered at a progressively decreasing frequency. Participants pressed the corresponding arrow key as soon as they identified the flickering arrow. The software recorded the flickering frequency (Hz) at which participants made a correct response. Each participant completed five trials. If a participant made three or more errors within these trials, additional trials were administered until three correct responses were achieved in the most recent five trials. The final CFF value was the average frequency of the most recent five correct responses. This metric is widely used in ergonomics and occupational health to assess fatigue⁵⁶⁾. A higher CFF, indicating the ability to perceive faster flicker frequencies, corresponds to lower fatigue levels⁵⁷⁾.

Continuous heart rate was recorded using the Polar Verity Sense to calculate HRV. An increase in HRV corresponds to higher fatigue levels^{47, 58)}. Heart rate (beats per minute) and blood pressure were measured using the automatic blood pressure monitor HEM-1000 Spot Arm before and after Tasks 1 and 2. A decrease in heart rate is associated with fatigue induced by ATC tasks^{9, 59)}.

Data analysis

A two-way ANOVA with one repeated measure was conducted to examine the effects on self-perceived fatigue, pupil diameter, fixation duration and count, saccade duration and count, eyes-closed duration and count, CFF, heart rate, systolic blood pressure (SBP), and diastolic blood pressure (DBP), HRV, and reaction duration to detect aircraft entering the sector, from before to after the microbreak. The significance threshold was set at p<0.05. A χ² test with pairwise comparisons was conducted to examine changes in loss-of-separation counts among the three groups.

Results

The statistical results for all measured indicators are summarized in Table 1.

Table 1. Results of repeated measures ANOVA for all indicators.

Measured indicator		Effect type	F value	Sig. (p)	Partial Eta squared (η2)	Sphericity assumption
Self-perceived fatigue		Interaction	1.153	0.337	0.053	Huynh-Feldt corrected
		Main Effect of Moment	11.169	<0.001**	0.214	Huynh-Feldt corrected
		Main Effect of Type	0.074	0.929	0.004	Not applicable
Eye movement indicators
	Left Pupil diameter	Interaction	0.007	0.993	0.000	Not applicable (only 2 levels)
		Main Effect of Moment	6.371	0.016*	0.144	Not applicable (only 2 levels)
		Main Effect of Type	2.263	0.118	0.106	Not applicable
	Right Pupil diameter	Interaction	1.106	0.341	0.055	Not applicable (only 2 levels)
		Main Effect of Moment	5.912	0.020*	0.135	Not applicable (only 2 levels)
		Main Effect of Type	0.870	0.427	0.044	Not applicable
	Fixation duration	Interaction	0.313	0.733	0.016	Not applicable (only 2 levels)
		Main Effect of Moment	4.442	0.042*	0.105	Not applicable (only 2 levels)
		Main Effect of Type	0.838	0.440	0.042	Not applicable
	Fixation count	Interaction	0.467	0.631	0.024	Not applicable (only 2 levels)
		Main Effect of Moment	0.257	0.615	0.007	Not applicable (only 2 levels)
		Main Effect of Type	0.592	0.558	0.030	Not applicable
	Saccade duration	Interaction	0.384	0.684	0.020	Not applicable (only 2 levels)
		Main Effect of Moment	0.891	0.351	0.023	Not applicable (only 2 levels)
		Main Effect of Type	0.641	0.532	0.033	Not applicable
	Saccade count	Interaction	0.464	0.632	0.024	Not applicable (only 2 levels)
		Main Effect of Moment	1.118	0.297	0.029	Not applicable (only 2 levels)
		Main Effect of Type	0.554	0.579	0.028	Not applicable
	Eyes closed duration	Interaction	1.058	0.357	0.053	Not applicable (only 2 levels)
		Main Effect of Moment	3.692	0.062	0.089	Not applicable (only 2 levels)
		Main Effect of Type	0.284	0.754	0.015	Not applicable
	Eyes closed count	Interaction	0.838	0.440	0.042	Not applicable (only 2 levels)
		Main Effect of Moment	0.000	0.995	0.000	Not applicable (only 2 levels)
		Main Effect of Type	0.659	0.523	0.034	Not applicable
Physiological indicators
	CFF	Interaction	0.964	0.448	0.045	Huynh-Feldt corrected
		Main Effect of Moment	0.243	0.848	0.006	Huynh-Feldt corrected
		Main Effect of Type	1.103	0.342	0.051	Not applicable
	Heart rate	Interaction	1.013	0.417	0.048	Huynh-Feldt corrected
		Main Effect of Moment	0.755	0.507	0.019	Huynh-Feldt corrected
		Main Effect of Type	0.680	0.512	0.033	Not applicable
	HRV	Interaction	2.992	0.061	0.127	Not applicable (only 2 levels)
		Main Effect of Moment	7.877	0.008**	0.161	Not applicable (only 2 levels)
		Main Effect of Type	0.069	0.934	0.003	Not applicable
	SBP	Interaction	1.395	0.222	0.065	Sphericity assumed
		Main Effect of Moment	1.579	0.198	0.038	Sphericity assumed
		Main Effect of Type	0.414	0.664	0.020	Not applicable
	DBP	Interaction	1.124	0.353	0.053	Sphericity assumed
		Main Effect of Moment	4.468	0.005**	0.100	Sphericity assumed
		Main Effect of Type	0.024	0.977	0.001	Not applicable
Performance Indicators
	Reaction duration	Interaction	3.752	0.033*	0.165	Not applicable (only 2 levels)
		Main Effect of Moment	0.647	0.426	0.017	Not applicable (only 2 levels)
		Main Effect of Type	0.794	0.459	0.040	Not applicable

This table presents the results of the repeated measures ANOVA, examining the effects of microbreak type (Meditation, Walking, No Microbreak) and task moment (Before Task 1, After Task 1, Before Task 2, After Task 2 or During Task 1, During Task 2) on various indicators of fatigue and performance. *p<0.05, ** p<0.01. Homogeneity of variances was verified for all between-subjects factors using Levene’s Test (p>0.05). CFF: critical flicker fusion; HRV: heart rate variability; SBP: systolic blood pressure; DBP: diastolic blood pressure.

Assumptions of repeated-measures ANOVA

Prior to analysis, the assumptions for repeated-measures ANOVA were examined. The normality of residuals was assessed using the Shapiro–Wilk test, and this assumption was met for all dependent variables (p>0.05). Sphericity was evaluated using Mauchly’s test. As shown in Table 1, the sphericity assumption was violated for most repeated-measures factors (p<0.05). Therefore, degrees of freedom were adjusted using the Huynh–Feldt correction, and the corrected p-values are reported.

Effect on self-perceived fatigue

The mean self-perceived fatigue under different conditions is shown in Fig. 4, with detailed descriptive statistics presented in Table 2.

Fig. 4.

Self-perceived fatigue values at different moments.

*p<0.05, **p<0.01 indicate a significant main effect of measurement moment based on post hoc analysis.

Table 2. Descriptive statistics of self-perceived fatigue and physiological indicators.

Group	Measurement point	Self-perceived fatigue	CFF	Heart rate	SBP	DBP
Meditation	Before Task 1	46.60 ± 23.02	46.46 ± 1.23	74.13 ± 11.80	114.87 ± 20.55	68.40 ± 13.68
	After Task 1	50.80 ± 24.38	46.49 ± 1.23	74.00 ± 10.89	119.13 ± 21.81	71.80 ± 13.32
	Before Task 2	42.80 ± 17.04	46.54 ± 0.89	72.80 ± 13.13	115.47 ± 19.49	69.13 ± 12.82
	After Task 2	52.73 ± 17.45	46.89 ± 1.36	73.27 ± 12.41	117.47 ± 19.76	70.27 ± 11.42
Walking	Before Task 1	44.13 ± 24.15	46.10 ± 1.04	77.64 ± 16.00	111.64 ± 13.14	68.14 ± 9.14
	After Task 1	53.73 ± 25.31	45.94 ± 1.77	79.36 ± 15.08	110.29 ± 12.16	69.79 ± 8.85
	Before Task 2	45.67 ± 21.83	46.09 ± 1.02	78.64 ± 16.65	112.93 ± 12.07	68.43 ± 8.38
	After Task 2	59.73 ± 23.77	45.81 ± 2.00	79.71 ± 18.08	114.64 ± 11.55	72.36 ± 8.18
No microbreak	Before Task 1	44.86 ± 20.54	46.18 ± 1.02	73.64 ± 11.04	117.79 ± 13.79	70.00 ± 9.27
	After Task 1	50.29 ± 16.55	46.47 ± 1.05	75.71 ± 11.37	116.64 ± 16.96	70.07 ± 11.24
	Before Task 2	50.29 ± 16.55	46.47 ± 1.05	75.71 ± 11.37	116.64 ± 16.96	70.07 ± 11.24
	After Task 2	56.00 ± 22.71	46.35 ± 1.41	73.00 ± 10.19	118.93 ± 17.34	71.86 ± 10.04

Mean and standard deviation values are shown for each group and measurement point for self-perceived fatigue and key physiological indicators (Critical Flicker Fusion, Heart Rate, Systolic Blood Pressure, and Diastolic Blood Pressure). All values are presented as Mean ± Standard Deviation. CFF: critical flicker fusion; SBP: systolic blood pressure; DBP: diastolic blood pressure.

A two-way ANOVA with one repeated measure was conducted to examine the effects of measurement moment and its interaction with microbreak type on self-perceived fatigue. The main effect of measurement moment on self-perceived fatigue was significant (F(2.734, 112.108)=11.169, p<0.001, η²=0.214, Huynh–Feldt corrected). The interaction between measurement moment and microbreak type was not significant (F(5.469, 112.108)=1.153, p=0.337, η²=0.053, Huynh–Feldt corrected). The main effect of microbreak type was also not significant (F(2, 41)=0.074, p=0.929) (Table 1).

Post hoc analysis with Bonferroni correction for the significant main effect of measurement moment showed that participants reported feeling fatigued after Task 1 (p<0.05) and recovered before Task 2 (p<0.05). However, participants reported feeling fatigued again after Task 2 (p<0.01). These findings indicate that overall fatigue levels changed significantly over the course of the experiment; however, these changes were not significantly affected by microbreak type, as both the main effect of microbreak type and its interaction with measurement moment were not significant.

Effects of microbreaks on eye movement and other physiological measurements

Effect on eye movement

Due to equipment malfunction, eye movement data from one participant in each of the meditation, walking, and non microbreak groups were incomplete, resulting in a sample size of 41 for eye movement analyses.

The mean values of pupil diameters, fixation duration, fixation count, saccade duration, saccade count, eyes-closed duration, and eyes-closed count under different tasks and microbreak types are shown in Fig. 5, with detailed descriptive statistics provided in Table 3. As shown in Fig. 5, left and right pupil diameters and fixation duration exhibited a trend of change over time, whereas fixation count, saccade duration, saccade count, eyes-closed duration, and eyes-closed count did not exhibit obvious trend changes.

Fig. 5.

Mean values of eye movement.

Table 3. Descriptive statistics of eye movement, physiological, and performance indicators.

Indicator	Group	Measurement point	Value
Left Pupil diameter	Meditation	Task 1	5.44 ± 0.57
		Task 2	5.39 ± 0.59
	Walking	Task 1	4.93 ± 0.56
		Task 2	4.88 ± 0.52
	No microbreak	Task 1	5.00 ± 0.90
		Task 2	4.95 ± 0.89
Right Pupil diameter	Meditation	Task 1	5.32 ± 0.47
		Task 2	5.24 ± 0.45
	Walking	Task 1	5.01 ± 0.61
		Task 2	4.91 ± 0.48
	No microbreak	Task 1	5.07 ± 0.90
		Task 2	5.07 ± 0.91
Fixation duration	Meditation	Task 1	1,579.60 ± 74.62
		Task 2	1,552.93 ± 113.45
	Walking	Task 1	1,516.51 ± 172.58
		Task 2	1,479.12 ± 227.84
	No microbreak	Task 1	1,530.18 ± 104.63
		Task 2	1,516.69 ± 129.40
Fixation count	Meditation	Task 1	2,693.21 ± 747.02
		Task 2	2,784.21 ± 866.71
	Walking	Task 1	2,662.07 ± 794.58
		Task 2	2,593.00 ± 821.60
	No microbreak	Task 1	2,899.08 ± 699.13
		Task 2	3,000.85 ± 975.36
Saccade duration	Meditation	Task 1	128.86 ± 34.13
		Task 2	137.36 ± 52.98
	Walking	Task 1	135.46 ± 54.54
		Task 2	133.97 ± 59.44
	No microbreak	Task 1	150.28 ± 56.17
		Task 2	158.11 ± 73.56
Saccade count	Meditation	Task 1	4,097.50 ± 1,426.79
		Task 2	4,460.57 ± 2,054.04
	Walking	Task 1	4,243.43 ± 2,013.00
		Task 2	4,185.50 ± 2,114.04
	No microbreak	Task 1	4,778.00 ± 1,969.31
		Task 2	5,105.23 ± 2,712.64
Eyes closed duration	Meditation	Task 1	62.17 ± 32.63
		Task 2	72.43 ± 44.25
	Walking	Task 1	79.93 ± 73.76
		Task 2	81.49 ± 71.19
	No microbreak	Task 1	77.79 ± 37.79
		Task 2	81.04 ± 35.42
Eyes closed count	Meditation	Task 1	1,050.79 ± 694.52
		Task 2	1,232.14 ± 876.49
	Walking	Task 1	2,619.64 ± 6,391.24
		Task 2	2,340.79 ± 4,849.22
	No microbreak	Task 1	1,273.62 ± 659.36
		Task 2	1,373.85 ± 714.09
HRV	Meditation	Task 1	76.66 ± 12.56
		Task 2	79.37 ± 12.96
	Walking	Task 1	76.35 ± 12.46
		Task 2	76.49 ± 12.46
	No microbreak	Task 1	76.74 ± 10.49
		Task 2	77.65 ± 9.91
Reaction duration	Meditation	Task 1	7.05 ± 3.85
		Task 2	5.84 ± 3.06
	Walking	Task 1	8.29 ± 7.27
		Task 2	8.31 ± 3.24
	No microbreak	Task 1	5.66 ± 2.98
		Task 2	8.24 ± 4.40

Mean and standard deviation values are presented for each group and measurement point for eye movement metrics and reaction duration. All values are presented as mean ± standard deviation. HRV: heart rate variability.

A two-way ANOVA indicated that the main effect of measurement moment was significant for left pupil diameter (F(1.000, 38.000)=6.371, p<0.05, η²=0.144), right pupil diameter (F(1.000, 38.000)=5.912, p<0.05, η²=0.135), and fixation duration (F(1.000, 38.000)=4.442, p<0.05, η²=0.105). However, the main effect of measurement moment was not significant for fixation count, saccade duration, saccade count, eyes-closed duration, and eyes-closed count (p>0.05). The interaction between measurement moment and microbreak type was not significant for any eye movement indicators (p>0.05). Furthermore, the main effect of microbreak type was not significant (p>0.05) for any of these indicators (Table 1). These findings suggest that significant changes in eye movement indicators were driven by task fatigue rather than microbreak type.

Effect on other physiological parameters

Due to physical reasons, blood pressure data from one participant in the walking group were not recorded; therefore, the sample size for SBP and DBP was 43.

The mean values of CFF, heart rate, SBP, and DBP at different measurement moments and across microbreak types are shown in Fig. 6, with detailed descriptive statistics provided in Table 2. Descriptive statistics for HRV are provided in Table 3.

Fig. 6.

Mean values of other physiological measurements.

**p<0.01 indicates a significant main effect of measurement moment.

A two-way ANOVA indicated that the main effect of measurement moment was significant for HRV (F(1, 41)=7.877, p<0.01, η²=0.161) and DBP (F(3, 120)=4.468, p<0.01, η²=0.100). However, the main effect of measurement moment was not significant for CFF, heart rate, and SBP (p>0.05). The interaction between measurement moment and microbreak type was not significant for any physiological indicators (p>0.05). Additionally, the main effect of microbreak type was not significant for any of these indicators (p>0.05) (Table 1). These findings suggest that changes in HRV and DBP were driven by task fatigue rather than microbreak type.

Effect on performance

Since both reaction duration and loss of separation count were analyzed from eye movement data, the sample size was also 41.

Effect on the reaction duration

The mean duration to detect aircraft entering the sector under different tasks and microbreak types is shown in Fig. 7, with detailed descriptive statistics provided in Table 3.

Fig. 7.

Mean values of reaction duration.

*p<0.05 indicates significant differences between Task 1 and Task 2 based on pairwise comparisons.

A two-way ANOVA with one repeated measure was conducted to analyze the effects of measurement moment and its interaction with microbreak type on the duration to detect aircraft. The interaction between measurement moment and microbreak type was significant (F(2, 38)=3.752, p<0.05, η²=0.165). However, the main effect of measurement moment was not significant (F(1, 38)=0.647, p=0.426, η²=0.017). The main effect of microbreak type was not significant (F(2, 38)=0.794, p=0.459) (as shown in Table 1).

Pairwise comparisons revealed that the duration to detect aircraft in Task 2 was significantly longer than in Task 1 for the non microbreak group (p<0.05). In contrast, neither the meditation nor walking groups showed a significantly longer reaction duration in Task 2 compared to Task 1. These findings indicate that while the non microbreak condition led to a significant decline in performance, walking and meditation microbreaks were effective in preventing this decline.

Effect on the loss-of-separation count

Table 4 shows the number of participants who exhibited an increase or no increase in Loss-of-Separation count in Task 2 compared to Task 1. The data revealed that in the walking group, no participant exhibited an increase in loss-of-separation count in Task 2 (0/13). In contrast, more than half of the participants in both the meditation (8/14) and non microbreak (10/14) groups exhibited an increase in loss-of-separation count.

Table 4. Loss of separation count for task 2 compared to task 1.

Group	No Increase	Increase
Meditation	6/14	8/14
Walking	13/13	0/13
No microbreak	4/14	10/14
Total	23/41	18/41

This table summarizes the number of participants in each group who experienced an increase in loss of separation errors from Task 1 to Task 2.

χ² tests were conducted to compare the changes in loss-of-separation count among the three groups. Pairwise comparisons, as summarized in Table 5, revealed that both the meditation group (χ²=10.556, df=1, p<0.005) and the non microbreak group (χ²=14.748, df=1, p<0.001) exhibited significantly higher increases in Loss-of-Separation count compared to the walking group. There was no significant difference between the meditation and non microbreak groups (χ²=0.622, df=1, p=0.693).

Table 5. χ² test results for loss of separation count.

Comparison	χ2 value	df	Sig. (p)
Meditation group vs. Walking group	10.556	1	<0.005**
No Microbreak group vs. Walking group	14.748	1	<0.001***
Meditation group vs. No Microbreak group	0.622	1	0.693

This table presents the results of chi-square tests comparing the change in loss of separation errors between the different microbreak groups. **p<0.01, ***p<0.001.

Discussion

Based on the results of our experiment, the effects of microbreaks on fatigue recovery and performance were multifaceted. Our analysis confirmed that self-perceived fatigue changed significantly across task stages, but these changes were not affected by microbreak type. Specifically, participants in all groups experienced a brief recovery of subjective fatigue during the rest period, indicating that a short break alone can help alleviate feelings of fatigue. This finding supports H1, indicating partial fatigue recovery. These observations align with prior research indicating that short breaks, naps, and light exercise effectively reduce occupational fatigue^{11,12,13,14,15,16,17,18,19,20,21, 25, 26)}. For example, microbreaks have been shown to sustain vigor¹⁶⁾, and brief breaks during surgical procedures have alleviated subjective fatigue²¹⁾. In general, temporary task disengagement appears sufficient to promote subjective recovery, regardless of the break type. However, the effect is limited, as subjective fatigue levels increased again after Task 2, reaching levels comparable to those observed after Task 1.

Physiological measures mirrored these trends. Although microbreak type did not significantly affect physiological indicators, this is consistent with previous ATC studies^{23, 59)}. Eye movement metrics (e.g., fixation duration, pupil diameter) and physiological indicators (e.g., HRV) exhibited patterns similar to those reported in cognitive fatigue literature^{47, 51, 52)}, suggesting that microbreaks provide only modest relief from physiological fatigue accumulation. These results are consistent with prior findings in ATCOs, where 15–20 min interventions, such as listening to music or resting, did not significantly alter physiological indicators²³⁾. This indicates that fatigue accumulated during cognitively demanding tasks may require longer or more intensive interventions to achieve measurable physiological recovery. While brief 10-min exercise sessions reduce physical fatigue in other occupations¹⁸⁾, such short-term activity may be insufficient to reverse central nervous system load in cognitively demanding tasks, such as ATC simulations.

Despite the limited effect of microbreaks on subjective and physiological fatigue recovery, they played a significant role in maintaining task performance. In the non microbreak condition, participants’ reaction duration increased significantly in Task 2, reflecting a decline in performance due to fatigue. A key contribution of this study is its differentiated analysis of the effects of two microbreak types on operational performance, providing comparative evidence to guide microbreak selection in purely cognitive, high-load tasks. Meditation microbreaks effectively prevented a significant increase in reaction duration. While prior studies have primarily highlighted meditation’s role in alleviating occupational stress¹⁹⁾ and psychological anxiety²³⁾, our findings suggest that such psychological relief may also help maintain objective cognitive speed. However, despite preserving reaction time, the error rate (loss-of-separation) in the meditation group remained significantly higher than in the walking group. In contrast, walking microbreaks demonstrated broader benefits, preventing an increase in reaction duration while significantly reducing critical error rates. These results support H2, indicating that specific microbreaks outperform the no-break condition in maintaining performance, with walking showing a clear advantage. Light physical activity during walking may be more effective than sedentary meditation in helping operators rebuild situational awareness and sustain attention during complex ATC tasks^{16, 20)}. Consequently, walking microbreaks appear to offer a unique and superior advantage for ensuring operational safety in safety-critical ATC scenarios.

These findings provide practical guidance for fatigue management in the ATC domain. Microbreak strategies can be tailored to operators’ specific needs: meditation may be prioritized to enhance cognitive speed, whereas walking may be preferred when the prevention of operational errors and the maintenance of safety are paramount.

This study has several limitations. First, participants were nonprofessional ATCOs, and the task environment was simplified, which may limit generalizability. Second, the measurement process for the non microbreak group may have unintentionally provided a brief rest period, potentially affecting the evaluation of microbreak effects. Future research should employ more realistic experimental designs involving professional personnel in controlled settings.

In summary, microbreaks represent an effective short-term strategy for reducing fatigue among ATCOs. Although a 10-min break is insufficient to fully eliminate accumulated physiological workload, it significantly alleviates subjective fatigue and helps maintain task performance, particularly in preventing critical errors. The superior performance of walking over meditation in error prevention underscores the importance of tailoring rest methods to task demands and provides new evidence for fatigue mitigation in high-pressure occupational settings.

Funding

This work was supported by JST SPRING, Grant Number JPMJSP2124.

Conflict of Interest

None.

Funding Statement

This work was supported by JST SPRING, Grant Number JPMJSP2124.