Evaluation of the cognitive performance under hot-humid environments using heart rate variability

. The human cognitive performance under hot and humid exposure is a critical issue for people working under hot and humid environments, for the considerations of both their health and safety. In order to find an indicator of the cognitive performance, the electrocardiogram (ECG) data were recorded continuously for 2.5 hours with portable Holter, when participants were exposed to 4 hot-humid environmental conditions and conducting different cognitive tests. The ECG data then were analysed to obtain the HRV indices that were used to establish a relationship with the cognitive test results. Temporal index pNN50 of HRV time domain indices was selected as a biomarker of the cognitive performance in this study. Results showed that the pNN50 responded significantly to the air temperature, while the influences of the cognitive tasks and the exposure time were found to be less significant. Meanwhile, the mean accuracy of the cognitive tasks was found to be positively related to the changes of mean pNN50, but a negatively related relationship between the mean reaction time and the mean pNN50 was observed. Furthermore, a positive relationship between the relative cognitive performance and the pNN50 was obtained, which meant that higher pNN50 responded to the higher cognitive performance. The potential reason could be ascribed to the changes in temperature-related allocation of the mental resources. And the low cognitive performance could be attributed to the rise of the mental fatigue due to the high sympathetic activities that corresponded to the low pNN50. The results revealed that the pNN50 could act as a biomarker of the cognitive performance under hot-humid exposure. The findings gave some implications to the real-time evaluation of the working performance and health of the people working under hot-humid environments.


INTRODUCTION
The global warming is producing more and more extreme climate events, such as the heat waves that may affect large regions and last for several days [1], leading to more frequent and long term hot and humid exposures [2]. It has been reported that extreme weather events are associated with health issues, both mentally and physiologically, including increased mortality, hospitalization rates and emergency department visits, and impaired cognitive performance [3].
Cognitive performance typically refers to the information processing of the brain and is indicated by speed, accuracy, and attentional demand [4]. It has been extensively reported that the hot-humid exposures would lead to a compromised cognitive performance (e.g. slower reaction time and more errors). [5,6], although these observational studies are usually small in size and/or conducted in younger or selective populations. In addition, Raab's study has revealed that the impaired cognitive performance was resulted from the heart rate variability (HRV) changes caused by the environmental temperature [7]. Other studies [8,9,10] on cognitive performance using HRV also supported Raab's conclusion.
As for the reasons for the compromised cognitive performance, a few possibilities have been put forward, such as the impaired brain functioning caused by the hot environment [11,12], the reduction in cerebral blood flow [13] and the accompanying mental fatigue [14,15], as well as the attenuated attention, working memory, information retention and processing of the human body [16].
A growing body of studies has been conducted in recent years to reveal and evaluate the changes in cognitive performance in hot-humid exposure [17][18][19]. However, only a few cognitive tasks have been frequently used in these studies. What's more, the exposure temperature and the exposure time were not enough to simulate the real working settings that usually contain continuous multi-cognitive tasks and long-time exposure, such as the military services in tropical regions. Therefore, 5 kinds of cognitive tasks and a long exposure time under high environmental temperatures were used in this study to simulate the real work settings and investigate the changes in cognitive performance. The time domain HRV index pNN50 was used to evaluate the cognitive performance. This study provides an alternative evaluation approach based on an objective physiological parameter, to acquire the cognitive performance under hot-humid exposure. 14 right-handed male subjects without cardiovascular diseases or mental disorders, aged 21.6 ± 1.8 yr, 173.1 ± 6.4 cm tall and weighing 64.8 ± 5.5 kg were recruited. They were asked to enter the environmental chamber to begin the 30min adaptation after instruments installation completed. The Holter recorded all the electrocardiogram (ECG) data immediately when they entered the chamber to measure the HRV. After the adaptation, they would be asked to start the cognitive tasks at 4 hot-humid exposure and a moderate exposure environment. All participants might quit the experiment at any session they wanted. A training was carried out before the experiment to make them understand the contents of the experiment well, especially the operations of the cognitive tasks. All participations were voluntary and paid hourly at a fixed salary. The experiment protocol was approved by the Ethics Committee.

METHODS
The participants were exposed to 4 temperatures (32, 35, 38 and 41°C) with the relative humidity (RH) of 70%. A moderate environment (26°C and RH 60%) was selected and used as a baseline for comparison. The different environments were created by an environmental chamber that could ensure the temperature deviation of ±0.5°C and RH deviation of ±3%. The ECG data were recorded by the potable Holter and then analysed by the Kubios HRV software to acquire the HRV indices. In this study, the time domain index pNN50 was selected to evaluate the cognitive performance. It is the proportion of NN50 divided by the total number of NN (R-R) intervals, in which the NN50 is the number of times successive heartbeat intervals exceed 50ms. pNN50 is closely correlated with Autonomous Nervous System activity (including sympathetic and parasympathetic activity) [20].
The cognitive tasks includ the Cueing Posner Task (T1), the Deary-liewald Task (T2), the N-back Task (T3), the Wisconsin Card task (T4), and the visual search task (T5). They correspond to the attention, choice reaction time, working memory, executive function, and attention & psychomotor speed, respectively [21]. These cognitive tests were designed on the open-access online PsytoolKit 3.3.2 platform [22]. The cognitive tasks were designed as 5 repeating sessions to simulate the sustained tasks, but the order of the sub-tasks followed the Latin square design to eliminate the potential effects that could result from the order of task.

Changes in the mean pNN50 with air temperature
It was found in Fig. 1(a) that participants had lower pNN50 at high air temperatures (32, 35, 38 and 41°C) than in moderate environment (26°C). And the mean pNN50 kept reducing as the air temperatures elevated, which meant low pNN50 corresponding to high sympathetic activities. They would lead to enhanced autonomous thermoregulations, including the sweating and the skin vasodilatation, to counterbalance the influences of the hothumid exposure [23].
In addition, due to the fact that all the participants could only finish Session 1 at 41°C, the mean pNN50 at 41°C in Fig.1 was calculated from the results of Session 1. The lower pNN50 at 41°C corresponded to higher sympathetic activities, which in return led to the significant 'Fight or Flight' response of all participants [24]. And they finally chose to escape from the stress  Fig.1(b) detailed the relationship between the mean pNN50 and air temperature, in terms of sessions. It was observed from Fig.2 that mean pNN50 of all 5 sessions at each air temperature had few significant differences. It might mean that the tasks per se had litter influence on mean pNN50. Fig.2 illustrated the relation between the pNN50 and the exposure time. It was found that the mean pNN50 had no significant changes during the whole exposure at different air temperatures (p>0.05). The possible explanation might be that the mean pNN50 was relevant with the changes in the autonomous nervous system that mediated the thermoregulations of the human body [25]. Consequently, the mean pNN50 might have limited responses to the changes in cognitive loads.

The relationship between the mean accuracy and reaction time and pNN50
Fig.3 depicted the changes in mean pNN50 with the test results (the response time and the accuracy) of the cognitive tasks. As shown in Fig.3(a), the mean pNN50 was positively related to the mean accuracy. The mean accuracy increased as the mean pNN50 elevated. It meant that higher sympathetic activities corresponded to lower accuracy. Fig.3(b) demonstrated a negative relation between the mean response time and the mean pNN50. It implied that the lower sympathetic activities could lead to higher response speed. The possible explanations were detailed as below. The mean pNN50 reduced as the air temperature elevated, accompanying by the enhanced thermoregulations resulted from the high sympathetic activities. During that, the thermal discomfort caused by the hot-humid https://doi.org/10.1051/e3sconf/202235603016 E3S Web of Conferences 356, 03016 (2022) ROOMVENT 2022 exposure would affect the cognitive performance, bringing about low accuracy and long response time [26]. In addition, Fig.3 also revealed that shorter response time might correspond to low accuracy. The possible reason could be ascribed to the 'Fight or Flight' response [24]. Participants tried to escape from the hot-humid exposure. Therefore, they responded quickly at the cost of accuracy.
The fitting results of Fig. 3 were not good enough, due to the small sampling size. However, the results could still illustrate the relationship between the mean pNN50 and cognitive performance, since most of the data located in the 95% confidence interval. Fig.4 described the relationship between the mean pNN50 and the relative cognitive performance. The cognitive performance was calculated in terms of Ref [27]. And the relative cognitive performance was the ratio of the cognitive performance at high air temperatures and that at moderate temperature. As illustrated in Fig.4, the relative performance and the mean pNN50 were positively related. The reason lay in that the increase of the mean pNN50 corresponded to the sympathetic reduction. As a result, the autonomous thermoregulations attenuated, and more 'mental resource' could be distributed to the cognitive tasks [28]. In other words, participants might be less distracted by the physical environment, and concentrated more on the tasks. Consequently, they could acquire better performance at higher pNN50.

The relationship between the relative performance and the pNN50
In addition, there could be another explanation. During the hot-humid exposure, the human body had high autonomous thermoregulations (low pNN50), trying to strike a heat balance between the environment and the body. High sympathetic activities that accompanying by the thermoregulations might lead to mental fatigue, thus deteriorate the cognitive performance [14,15].
Finally, it was found in Fig.4 that higher cognitive performance could also be observed at low pNN50 (e.g. pNN50 =5). Provided there was no data abnormality, the potential explanation could be that the high air temperature promoted the psycho-motor functions and the arousal level [29]. As a result, the participants could act quickly and ensure the accuracy. The phenomenon could be observed in Fig.3, from which high accuracy and short response time were detected when the mean pNN50 was about 5.

CONCLUSIONS
Experiments were conducted in this paper to study the changes in cognitive performance when participants were conducting different cognitive tasks in hot-humid exposures. During the experiment, the HRV time domain index pNN50 was selected to evaluate the cognitive performance.
It was found in this study that the mean pNN50 was affected significantly by the air temperature, but affected less by the cognitive tasks (loads) or exposure time. Furthermore, a positive relation between the mean pNN50 and the accuracy, and a negative relation between the mean pNN50 and response time were observed. Finally, the study revealed that the mean pNN50 and the relative cognitive performance were positively related, the explanation of which might be ascribed to the mental resources allocation and mental fatigue caused by the changes in pNN50. Findings of this study provided some implications for the evaluation of the cognitive performance and productivity in real working settings.