In our preliminary study, we examined the temperature changes and the temperature distribution (entropy) of the torso during graded cycling exercise stress test using thermal imaging. We found a significant increase in entropy that was correlated with exercise intensity. Furthermore, we found a high correlation between the changes in chest entropy and a proportional increase in pulmonary ventilation (VE).
Current reports using infrared thermography during exercise have primarily focused on measuring absolute body temperature. However, these attempts have demonstrated limited effectiveness, and almost no significant correlations have been found with physiological parameters1,14. In the current study, we found no statistical changes in surface temperatures during the incremental exercise test. The considerable differences in surface temperatures during incremental exercise test can be related to changes in the ROI determination, measurement methods, and the differences in exercise test protocols1,7,15,16. Additionally, the temperature is sensitive to environmental changes, such as ambient temperature and humidity, sweating degree, and body hair amount in the ROI.
Our innovative approach of extracting advanced texture and shape features from thermal images enhances the impact of our study findings. In a recent study by Brzezinski et al., SARS-CoV-2 infection was diagnosed using thermal imaging. In this case, no significant thermal changes were found in absolute temperature values. However, a significant difference was observed in the calculated textural features (entropy), contrast, and homogeneity17. Another unrelated study also demonstrated the same concept in an animal model of fatty liver18.
The fact that mean surface temperature at the ROI did not correlate with the increase in the exercise and muscle work in the test group supports our rationale to focus on temperature distribution across organs of interest instead of a standard absolute temperature assessment. Entropy is a statistical measure that expresses the texture of an image and is less sensitive to environmental changes than to absolute temperature.
These entropy changes, manifested as a difference in the image’s texture, were visually described in previous studies. Studies on this subject described the warmer areas formed, also known as “tree shapes, hyperthermal zones, or even vessel shapes”, during the exercise test10,11,12 (Fig. 3). In thermal photography, these areas appeared brighter (warmer) and created a characteristic image reminiscent of a tree and its branches or a superficial vascular system. It has been claimed that the tree-shaped pattern represents perforator vessels1,12, and its formation is due to the opening of superficial blood vessels during exercise. There is an increase in the internal and exercising tissue temperature during exercise, which causes reflex neurogenic vasodilation. This leads to the active opening and vasodilation of the vessels in the superficial vascular system of the skin19,20,21. This study quantitatively describes the changes in the texture obtained through thermal imaging, and significant changes in the entropy measurement were found throughout the exercise.
Furthermore, we found that the increase in chest entropy was highly correlated with the increase in pulmonary ventilation during exercise. During an incremental aerobic exercise stress test, respiratory muscles in the chest are active and increase their oxygen consumption. The increase in oxygen consumption of the chest muscles in exertion can reach a rate of 10–15% of the total oxygen consumption in the body22. This increase is accompanied by changes in the heat produced by and emitted from the active muscles23. The increase in heat production in the active muscles correlates with the biogenesis of ATP production, which can double during high-intensity exercise23. Increased metabolic heat production leads to vasodilatation and the opening of superficial blood vessels in these areas. This is supported by a previous study that demonstrated that active muscle areas presented a greater blood flow and had higher temperatures than other areas, such as the chest area being warmer than the back10. Therefore, it can be assumed that the increased work of breathing by the chest muscles during exertion, which is expressed as an increase in pulmonary ventilation, led to the opening of superficial blood vessels, changing the entropy in the chest.
In conclusion, surface body temperature measured during incremental exercise stress test using IR camera showed no significant changes in skin temperature of the limbs, chest, and forehead. However, we found a significant increase in the entropy measurement throughout the exercise, which was correlated with pulmonary ventilation. The entropy findings are consistent with other studies that have described the “hot spotted areas” phenomenon. The present study is the first to describe this phenomenon quantitatively. Further studies are required to investigate whether this approach can detect ventilatory anaerobic thresholds remotely using an IR camera.
However, this study has several limitations, namely, (1) only male participants were involved in testing (because bare chest images needed to be acquired); (2) all the study participants belonged to the same age range (20 to 30 years); (3) FLIR ONE resolution limitation can interfere with results, and future research with a higher quality camera may lead to better results. (4) all the study participants were healthy and physically active. Further research with a larger sample, including more diverse populations, will be needed to apply this measurement method clinically.
