Studying psychological affective states through physiological signals
The proliferation of smartwatches and wearable devices, equipped with a variety of sensors to obtain physiological readings, enables the study of psychological aspects of human behaviour ‘in-the-wild’ and establish physiological computing as an emerging research area. Along these lines, I will share a summary and some thoughts from a recent talk that I attended about this topic at the UCLIC’s (UCL’s Interaction Centre) seminar in London. More about UCLIC’s seminars and events: https://uclic.ucl.ac.uk/news-events-seminars
The talk was titled “Physiological Computing towards Mental Disability Innovation” and was delivered by Youngjun Cho, an Assistant Professor in the Department of Computer Science at UCL. The central topic revolved around physiological computing — as nicely framed by Youngjun — “technology that listens to our bodily functions and psychological needs and adapts its functionality”. This technology is capable of tracking our biological signals in our daily lives and offers an alternative source of implicit data to study psychological aspects of our behaviour.
The talk was structured around three key components: (a) sensing: The use of different sensors to capture physiological signals (e.g., cardiac, respiratory), (b) recognition: The application of machine-learning algorithms to extract meaningful information from physiological signals, which in turn are being used as input to classification algorithms, and (c) feedback: How the feedback is communicated with the end user (e.g., visual, haptic, auditory).
Youngjun presented a series of studies in which he utilized thermal imaging cameras for sensing physiological signals and developed models that recognize people’s psychological stress levels and respiration tracking. Thermal imaging camera is a type of thermographic camera that is used to study heat patterns of materials and organisms. The different use cases presented in the talk utilized low-cost thermal imaging cameras to associate thermal signatures with a person’s psychological affective state.
In particular, I liked the inference of perceived mental stress using smartphone-based photoplethysmography (PPG) and thermal imaging. PPG is a simple and low-cost optical technique to detect blood volume changes by shedding light, which in turn the increased pulse pressure causes measurable difference in the amount of light reflected back, resulting the light absorption. Combining all these technologies, he presented a system that continuously measures a person’s blood volume pulse, the time variability between successive heartbeats (R-R intervals) and 1D thermal signature of the nose tip. By exposing participants in a series of mathematical tasks that induced different levels of perceived stress, they were able to develop models for instant mental stress recognition. He discussed a new preprocessing approach that enhances signal’s quality obtained from thermal imaging cameras, and presented a neural network model trained using an array of R-R intervals and thermal directionality features to learn high-level mental stress levels. More information about his studies here: http://youngjuncho.com/
To sum up, data obtained from mobile- and wearable-devices represent a rich source of implicit data about people’s daily lives. That is, it will enable the study of people’s psychological affective states such as stress, emotions, experiences, perception through the lens of people’s biological responses, away from the lab, in order to facilitate the creation of new tools and forms of augmented human interactions.
