1 Introduction

Improving human intellectual attributes has become an important concern in fast-paced work environments (Kianto et al. 2019; Simonenko 2018). Complex work environment factors (e.g., stressors, workload, conflicts) should be balanced by recovery-enhancing processes to create employee well-being within a company (Sonnentag et al. 2017; Sonnentag 2018). Some organizations and companies have applied lunch break times during more than 6–8 h of daily work (JETRO 2023; OSHA Education Center 2023; UK Government 2023). However, it is difficult to maintain optimal working performance for the 3–4-h period before and after the lunch break time. Consequently, the micro-break concept is offered to maintain stable work performance without taking excessive working time.

Micro-breaks (MB) are defined as short breaks from work for momentary work recovery (Trougakos and Hideg 2009). MB is an individual strategy to maintain or improve occupational well-being during work (Zacher et al. 2014). Although no exact standard exists for MB duration (Kim et al. 2018), several reports of the literature suggest that a short amount of time of even a few seconds or minutes is sufficient for the MB duration, without adversely affecting productivity (Ibes et al. 2018; Kitayama et al. 2022; Lee et al. 2015). Several activities can be designated as MB, including relaxation activities (e.g., listening to music, meditating, taking a short nap), nutrition-intake activities (e.g., drinking water, snacking), socialization (e.g., chatting, contacting friends or family), and other activities (e.g., look out a window, walking, stretching) (Fritz et al. 2011; Kim et al. 2018).

Earlier studies have investigated MB effects on work performance. A survey study involving 71 call operators indicated that cognitive MB improved their positive affect, which led to higher sales achievement (Kim et al. 2018). Another survey study involving 120 respondents indicated that the desire for self-reward triggers individuals to take MB (Bosch and Sonnentag 2019). The attention of assembly workers was improved by taking MB during work, as shown by EEG analysis (Mijović et al. 2015). A preliminary study conducted by Kitayama et al. (2022) showed that, on average, the condition with systematic MB during work positively affected the performance of three participants compared to conditions without breaks. Biwer et al. (2023) described that a systematic break during work raised efficiency in terms of work completion and that it presents benefits mood restoration compared to self-regulated breaks. Other studies conducted by Conlin et al. (2021) corroborated the beneficial results of taking MB during work to restore performance for both relaxation and expert activities compared to a no MB condition. Additionally, Lee et al. (2020) described that using the pause in a computerized simulation game for the medical students caused a higher cognitive load and performance than without the pause. It was also described that not taking frequent breaks can lead to accumulation of fatigue symptoms in work-from-home office cases (Cropley et al. 2023). Hunter and Wu (2016) further suggest that taking more frequent short breaks rather than postponing the break to a later time can better promote recovery.

However, investigations of systematic MB intervention during cognitive work under controlled experimentation with quantitative and qualitative measurement conditions have rarely been reported. Consequently, this study was conducted using controlled laboratory experiments to contribute objective evidence supporting the investigation of systematic MB on intellectual concentration during intellectual work.

2 Methods

To examine effects of micro-breaks (MB) on intellectual concentration work, a controlled laboratory experiment was conducted with experiment participants. Quantitative and qualitative measurement evaluations were employed, as described by Ayres et al. (2021) indicating that the combination of quantitative and qualitative measurements is effective for detecting the shift correlates of cognitive work tasks. The experiment was designed with certain procedures, specific tasks, and equipment. Details of the experiment procedure, cognitive tasks, measurement, experiment environment, participants, and data analysis are described in the following paragraph.

2.1 Experiment procedure

For this study, the independent variable was systematic MB intervention during the cognitive task. The answering time for each question, the intellectual concentration performance such as the concentration time ratio (CTR) score, and the subjective rating questionnaire were set as dependent variables. As the MB intervention, a blank gray screen appeared for 20 s repeatedly, immediately after a problem was answered after a 7.5-min task. The MB paused the cognitive task. All participants performed cognitive tasks under two conditions: with MB intervention and without MB intervention.

In this study, the experiment was conducted in a controlled laboratory environment and procedures to minimize factors that might interfere with the results except the MB intervention. Based on the rigorous procedures for carrying out the experiment, the results obtained were expected to portray a causal relationship created due to the presence of MB intervention. Repeated measures or within-group experiment designs were conducted to ensure the same subject complied with the tasks concerning the systematic MB intervention. We expect by adopting repeated measures or within-group experiments, we could discover whether the results obtained are due to the provision of systematic MB or not, and not because of the differences in individual group competencies.

Figure 1 presents the experiment procedures. The experiment was conducted where four sets of cognitive comparison tasks with 25-min duration were given to all the participants. Set I and set II were executed in the same order for each participant, with MB intervention in set I, then without MB intervention in the set II. Additionally, set III and set IV were arranged with a counterbalanced design of MB intervention among all participants to minimize the ordering effect. Additionally, the learning effect was considered to have been reduced, assuming that all participants had been familiar with the task after doing set I and set II. Later, for data analyses, only data from set III and set IV were used.

Before and after each task set, participants were given a subjective symptom questionnaire for the assignment of subjective ratings for their level of fatigue, concentration, sleepiness, blur, and dullness. Additionally, the participants answered the NASA-TLX questionnaire after each set.

Fig. 1
figure 1

Experiment procedure

2.2 Cognitive comparison tasks

Earlier studies employed comparison tasks for human reasoning research, which emphasized similarities and differences between pairing words (Farell 2022; Ichien et al. 2023). The cognitive comparison tasks adopted by Ueda et al. (2016) were used for these study experiments. The characteristics of cognitive comparison tasks are suitable to measure the intellectual concentration properties because of their unified level of difficulty for each question and because of the inability to change the answering strategies. The results approximately reflect cognitive task performance during work or study.

Figure 2 presents an example screen of the cognitive comparison task develop by Ueda et al. (2016). The half-left sides of the screen show the task questions; the half-right sides of the screen show the answering buttons. In the cognitive comparison task, two comparison categories should be solved for each question: semantic word comparison and numerical comparison. The semantic word comparison is given as two words being compared, with the respondent stating whether these two words are classified into the same category or not. Four categories are determined for each word that appears, such as a place name, animal, plant, and artificial objects. For example, when given two words “dog” and “cat”, the response should be given as the “same” category because the “cat” and “dog” both belong to the animal category. For numerical comparison, two numbers with an inequality mathematical sign between them are logically compared. Then, the respondent reports whether the inequality formula is correct or false. For example, given the two numbers “2543” and “2354” with a “<” (less than) sign of the math symbol between them, then it should be reported as a “false” statement. Combining the answers of semantic word comparison, which is “same”, and answers of numerical comparison, which is “false”, the answer box should be tapped in the “same” – “false” intersection area. After tapping the answer box, the next question appears. All questions were designed to have the same level of difficulty. All were presented in a pseudorandom order.

The cognitive comparison task was developed in the JavaScript software program (JavaScript) and was run on an iPad 2 tablet (A1395; Apple Computer Inc.). The time used for answering each question was recorded automatically in the program.

Fig. 2
figure 2

Example screen of cognitive comparison task

2.3 Measurement

2.3.1 Answering time

The performance output is measured in the form of the answering time for each question task. The answering time was recorded automatically in an experimental program. The data were downloaded as a. csv file at the end of the task. The answering time for each question counted starts when the question appears until a participant touches the square box of an answer area on the screen. The unit digit of the answering time in the experiment program is set in millisecond units.

2.4 Intellectual concentration

Intellectual abilities entailed several properties given due consideration at their level, stimulation, and potential capabilities of development in the psychology domain. In connection with intellectual terms, intellectual concentration is attributed to the ability of an individual to think, understand things, and to focus attention on some task or information. Although it is difficult to measure intellectual concentration quantitatively, one quantitative approach to measuring intellectual concentration is called the concentration time ratio (CTR) which represents the proportion of actual concentration time for a total task completion time (Ishii et al. 2018; Miyagi et al. 2013; Shimoda et al. 2013).

The CTR obtained from the average answering time of concentration state functions multiplied by the number of task questions answered, divided by the total answering time from the cognitive task work. In the CTR, a three-state model is considered, including the working state, short-term rest state, and long-term rest state (Ishii et al. 2018; Miyagi et al. 2013; Shimoda et al. 2013).

2.5 Subjective ratings

The workload assessment instruments used for this study are the subjective symptoms questionnaire (Sasaki and Matsumoto 2005), time lapse questionnaire, and the NASA task load index (NASA-TLX) questionnaire (Hart 2006). The subjective symptoms questionnaire was adopted for this study according to the research needs to support a distinction between the qualitative measurements of the subjective alertness feeling during the work. Regarding the symptom questionnaire, five subscales are listed: concentration, sleepiness, blurriness, dullness, and fatigue. The concentration scores measure the ability to examine one’s attention specifically during the work. The sleepiness scores measure the drowsiness level or the tendency to feel sleepy. The blurriness scores reflect the perception of the vision’s haziness. The level of boredom is adapted from the dull scores. Fatigue scores reflect the weariness condition.

In the progress questionnaire, the sleepiness, blurriness, and dullness subscale include five question points ranging from 1 “low” to 5 “high”. The total scores of sleepiness, blurriness, and dullness subscale were obtained from the summed-up scores from the question points. The time lapse questionnaire includes the subjective fatigue and concentration subscale as scores ranging from 0 “low” to 100 “high”. The NASA-TLX questionnaire consists of six subscales or load dimensions, i.e., effort (EF), frustration level (FR), mental demand (MD), physical demand (PD), overall performance (OP), and temporal demand (TD) (Hart 2006; Virtanen et al. 2021). Hart and Staveland (1987) as cited by Gawron (2000) described that NASA-TLX brings an adequate sensitive index that can diversify cognitive tasks and physical tasks of the overall workload.

3 Experiment environment

The experiment was conducted in an experiment room at the Graduate School of Energy Science, Kyoto University. A table was set in front of the participants with an iPad on top of the table and a stand behind the iPad. A chair is placed in front of the table, where the participants were able to perform the task while sitting. Figure 3 shows the experiment room layout. Boards of the partition were placed in front of the table, on the left and right sides of the table to minimize distraction while participants performed the task. The experiment room temperature was set as around 23 ± 2 °C, with humidity of 50 ± 10% RH. The lighting level was set as 500 ± 50 lx. Additionally, the CO2 concentration was approximately 550 ppm.

Fig. 3
figure 3

Layout of the experiment room

3.1 Participants

For this study, 52 participants were recruited. All were Kyoto University students who were 18–28 years old. To fulfill the objectivity and validity of the experiment’s outputs, participants are expected to follow all of the experiment’s procedures and instructions. However, among the 52 participants recruited at the beginning of the experiment, 21 participants’ data were not included in the further analysis for specific reasons: (1) not following the instructions correctly, (2) being absent from one or more than one session set in the experiment’s procedure, (3) the lowest limit of the CTR score based on the preliminary study was set at 40% (Kitayama et al. 2022; Nomura et al. 2024), CTR score that is below the lower limit was excluded which indicates that those participants not performing the task thoughtfully, (4) the CTR score between two sets with and without MB intervention is not exceed than 10% (Ueda et al. 2022) in which the differences above 10% is supposed to become the cautioned percentage that the external factors might involved apart from the MB’s intervention. The remaining 31 participants’ data (male = 20, average age = 21.2 ± 2.54 years old) were used for analysis procedures. Those participants followed the experimenter’s instructions without slacking off. They had completed all four sets of the experiment procedures. The data from 31 participants performing under two conditions in a set with MB intervention and no MB intervention indicated statistical power of 0.97, α = 0.05, and σ = 5, which were statistically sufficient for our purposes. For this experiment, ethical approval was granted by the Graduate School of Energy Science Ethics Committee at Kyoto University. Figure 4 presents a scene in which some participants are performing the experiment task.

Fig. 4
figure 4

Participants performing the experiment task

3.2 Data analysis

The quantitative data from the cognitive comparison task to be analyzed are the answering time collected by the task software and the CTR score. The data to be analyzed were extracted from set III and set IV, which respectively portray the conditions with MB intervention and without MB intervention sets in a counterbalanced design.

The CTR scores were calculated by the CTR analyzing software from the answering time data obtained for each set. In the analysis, the CTR score in set III and set IV are compared to ascertain the effects of MB intervention on the intellectual concentration performance.

To grasp the accuracy for the cognitive performance in each set more intuitively, and to facilitate more detailed comparison of performances in a set with MB intervention and without MB intervention, the answering time was divided into seven time blocks for every four minutes in a time series horizon. Figure 5 shows a detailed time frame grouping in a set with MB intervention and without MB intervention. Later in the analysis section, these four minute time frame groupings are stated as the phase factors. In a phase factor, to grasp the intuitive comparable data aligned with the MB effect, the included phase starts from block numbers 3 to 7 for the condition with MB intervention and block numbers 3 to 7 for the no-MB condition. The actions taken to correct the earlier answer by participants, as indicated by the undo movements, were removed from answering time data analysis.

Fig. 5
figure 5

Time frame grouping for a set with and without MB intervention

Statistical analysis was used to compare the quantitative data of the answering time and the qualitative data from the questionnaire results. Repeated measures ANOVA (RM-ANOVA) was applied to analyze each factor’s significance in the quantitative data and qualitative data. Minitab statistical software was used for the analyses (Minitab, LLC.).

4 Results

4.1 Analysis of answering time

The answering time in each block was found to be significantly different in the conditions with and without MB intervention (p < 0.001). Details of the answering time in each block and significant differences are shown in Fig. 6. Tukey’s post hoc analysis was applied to ascertain which block contributes to the significant difference. The contrast significant difference was found in block 3 with the MB condition to block 4 with the no-MB condition. This result indicates that the first MB for 20 s after 7.5 min of intellectual work can help to improve performance and can result in contrasting performance after 12 min of work without the MB condition. A significant difference was also found in block 5 with the MB condition to block 5–7 with the no MB condition. In block 5 with MB, the second MB was induced for 20 s. It restored the intellectual performance and caused a significant difference in answering time after 16 min of intellectual work without MB.

Fig. 6
figure 6

Answering times for each block and indicators of significant difference

4.2 Analysis of concentration time ratio (CTR)

The CTR score in the condition with MB intervention was then compared with that obtained in the condition without MB. Figure 7 shows the average CTR score of 31 participants’ data with MB intervention and without MB intervention. The CTR scores in a set with MB intervention and in a set without MB intervention follow a normal distribution based on an Anderson–Darling normality test with a p value 0.076 in an MB condition and 0.254 in a condition without MB. Because the CTR score for both conditions follows the normal distribution, the statistical two sample t-test conducted. The t test results revealed no significant difference between the set with MB intervention and the set without MB intervention (p = 0.721). Even though no significant difference was found from comparison of the CTR scores in the conditions, the CTR score with MB was not reduced despite a 1-min break taken under the MB condition.

Fig. 7
figure 7

Average CTR scores of 31 participants’ data in a set with MB intervention and a set without MB intervention

From 31 participants’ data, 17 participants showed better CTR scores in a condition with MB intervention than with no MB intervention. Results showed that more than half of the participants obtained a higher concentration level in a condition with MB intervention.

4.3 Subjective ratings

4.3.1 Subjective symptoms questionnaire

Figures 8 and 9 present the subjective symptoms questionnaire scores. ANOVA analysis results indicated significant difference for the questionnaire time factor (p < 0.001). The intervention factor was found to have no significant difference (p = 0.773), which was true also for the interaction between factor (intervention and questionnaire time, p = 0.962).

Fig. 8
figure 8

Subjective symptoms questionnaire in conditions with and without MB.

Fig. 9
figure 9

Subjective symptoms questionnaire results for each subscale

4.4 NASA-TLX questionnaire

Figure 10 shows the average of NASA-TLX weighted scores. The paired t-test analysis indicates a significant difference between results obtained in a condition with MB intervention and without MB intervention (p < 0.05). This result has pinpointed the participants feeling that, in the cognitive comparison task, the mental workload was significantly higher in the condition without MB intervention.

Fig. 10
figure 10

Weighted score of NASA-TLX.

5 Discussion

Answering time analysis findings indicate that intellectual performance measures for conditions with and without MB intervention were significantly different (p < 0.001). On average, the performance under MB intervention shows less degradation than the performance achieved without MB intervention shown by the sequences of answering time pattern. When the MB was induced, the performance improved slightly from that found for the previous block (e.g., answering time was improved in block 5 in MB condition compared to block 3 in MB condition and block 7 in MB condition compared to block 6 in MB intervention). However, that improvement effect might decay over time. In contrast, the set without MB intervention showed an increasing trend of the answering time gradually from block 3 to block 7, which indicates cognitive performance degradation over time.

Nonetheless, no significant difference was found in the CTR scores in the set with MB intervention and without MB intervention. The scores were slightly higher for a condition with MB than for a condition without MB. Additionally, more than half of the participants show higher CTR scores in the condition with MB intervention than without MB intervention. The difference in significant results between total 31 participants’ data in CTR analysis and answering time analysis might be attributable to the indulgence of MB time in the formulation. Answering time analysis does not include the pause break time in the MB condition calculation, but CTR is obtained from the average answering time of the concentration state multiplied by the number of questions answered divided by total time. However, the CTR score under the MB condition was not decreased irrespective of inclusion of a total break time of 1 min (i.e., three times MB was given with 20 s each).

Generally, the performance during the time was more stable in a condition with MB intervention than in a condition without MB. For the 25 min task set, the performance under the MB intervention showed slight improvement after the MB induced in block 3, block 5, and block 7. However, the performance in a condition without the MB intervention gradually became slower from block 3 to block 7.

The contrast significant difference was found between block 3 in the condition with MB condition to block 4, block 5, block 6, and block 7 in the condition without MB. The possibility exists that the first intervention of MB was induced too early after the 7.5-min task. It is advisable to start an MB after minute 13 at the start of block 4 in a condition without MB intervention, which might obstruct the performance degradation.

The questionnaire results suggested a significant difference in terms of the MB intervention factor for NASA-TLX, with a tendency by which, under a condition without the MB intervention, the mental workload was higher than that obtained under the condition with MB intervention. However, generally, results of the subjective symptom questionnaire indicated no significant difference between conditions with and without MB intervention. It might be true that the insertion of 20-s MB three times during the task did not significantly affect the subjective symptom feelings among the participants. Whereas, the pre-questionnaire and post-questionnaire were found to have significant differences, indicating that the participants were more likely to feel the symptom effects before and after taking a 25-min cognitive task.

5.1 Limitations and future research

In this study, several aspects could become the limitations and potential bias in which confounding variables may occur. The limited type of cognitive task in the experiment might not well represent a variety of job tasks in various fields and, thus might result differently. Cognitive task arrangement could lead to an individual’s competence (Hornung 2019). Other factors such as the limited group of participants and age group are also important to be considered. In this study, recruited participants came from a relatively uniform group of university students. Different levels of education and workers’ experience could be a forethought that could deliver a better representation of the real work environment (Kotur and Anbazhagan 2014; Quiñones et al. 1995). Individual differences and characteristics (e.g., learning style, cognitive personality style, instructional preference) might be another variable that should be taken into account (Price 2004; Sutin et al. 2019). The distinction of an individual’s characteristics and behavior might have the potency to influence the participants’ performance. The presence of intrinsic and extrinsic motivations allegedly could be causing some changes in a person’s work performance. In real work, intrinsic motivation (e.g., work enjoyment, desire to compete, self-efficacy) often appear as well as extrinsic motivation (e.g., money, awards, promotion) by any chance could have an impact on a person’s performance (Deci and Ryan 2000). Therefore, for further investigation, we recommend adopting a more holistic approach and examining different individual characteristics and group divergencies, different working tasks, and the presence of motivation.

5.2 Managerial implications

The managerial implications of this finding are the potency to implement the systematic MB at the work organization and in a daily work design to promote better work productivity. Recommendation for adopting systematic MB could be delivered formally as a management’s instruction, or implemented voluntarily by the individual who wants to attain improvement on productivity during their work. However, some considerations can be adjusted and further research such as the type of work where different job-specific tasks may have different effects. One of the ways to adopt the systematic MB more conveniently is by utilizing rest-to-work management software or applications.

6 Conclusion

To conclude, results obtained from this study demonstrate that a systematic MB has the potential benefit to stabilize worker performance and to mitigate performance degradation over time. Quantitative results obtained from the answering time analysis indicate significant differences in various time blocks related to induction of systematic MB in the 25-min cognitive task. Additionally, qualitative results of the NASA-TLX indicated that the mental workload in conditions with MB was significantly lower than in conditions without MB. As further studies, it is necessary to investigate stimuli of various types for MB intervention. Such stimuli might have benefits as a way to improve intellectual concentration. Additionally, incorporating the assorted possibilities of MB duration and interval might enrich the exploration of research correlates with the resource replenishment strategy.