Assessment of dynamic cognitive performance under different thermal conditions by multiple physiological signals

Abstract

A method to evaluate cognitive performance under different thermal conditions by multiple physiological signals is proposed. Heart rate (HR), heart rate variability (RMSSD), skin conductance level (SCL), skin conductance response (SCR) magnitude and SCR total frequency were measured by a multichannel electrophysiological signal monitor, which were interpreted as thermal stress indicators that can characterize cognitive performance of subjects during cognitive stimulation tasks. In the experiment, the indoor air temperature is the independent variable, ranging from 18°C to 33°C, with a step length of three degrees. Other indoor environmental quality (IEQ) parameters are controlled variables. In each experiment, subjects subjectively voted on thermal sensation and cognitive load, and performed cognitive stimulation tasks to evaluate cognitive performance. The study found that different physiological indicators have different "thermosensitivity" zones, indicating variations in thermal attributes. Moreover, the influence mechanism of thermal conditions on the effectiveness levels of three different cognitive functions, namely identification, decision and action, were found to be diverse. Additionally, this paper utilized machine learning algorithms to construct fatigue loss function and residual interference function for cognitive performance, effectively correcting cognitive performance errors caused by fatigue and task superimposition. The feasibility of predicting cognitive performance and thermal comfort by multiple physiological indicators is validated in this study.