Awareness about limitations

Like any other physiological signal, EDA has limitations that make it unsuitable for certain HCI paradigms. EDA is a slowly changing signal—SCRs to a stimulus or arousal-induced tonic changes have a delay of 1 to 3 seconds [12]. Adding the computational delay for detecting an SCR or a SCL shift might introduce even further lag. Therefore, EDA may not be suitable for HCI applications that require real-time responses such as modalities for game interactions or biofeedback for drivers. Further, a core topic of HCI research is the recognition of physical and mental activity in everyday settings. Studies that investigate these are traditionally conducted in laboratory settings followed by in-the-wild setups utilizing medical-grade devices and mostly apply no restriction on participants’ activities and movements and in most cases tasks include intense movements and activities. On the other hand, the correct sensor placement and signal processing approaches are crucial for obtaining meaningful EDA signals. Consequently, a potentially incorrect placement of the sensors, or displacement during a study (which is likely in the presence of intense movements), and the resulting very low signal-noise-ratio (SNR) through motion artifacts increase the risk of false signal readings [20, 21]. In addition to motion artifacts, during intense physical activities, the human body increases sweating to lower the body temperature, which also affects the EDA signal. Furthermore, due to the realistic tasks in most of in-the-wild setups, HCI researchers are forced to use less obtrusive and less invasive mobile EDA devices to conduct their studies which record lower quality EDA signals. Therefore, before designing such studies, considering the activities in our study, we must make sure if EDA is a fit for measuring psychological constructs in our study. EDA caused by the activation of the sympathetic nervous system which is responsible for fight or flight actions and mostly causes negative feeling such as stress and arousal. Therefore, the ability of EDA in measuring physiological signals is limited to negative physiological constructs. For instance, EDA has not shown any evidence of insight about the positive experience of valence [10].