Quantifying cognition: Applications for ubiquitous data
Year: 2018
Type: dissertation
Abstract: The amount of information collected by personal health records, smartphone ecosystems, and other cloud services has increased enormously in recent years. This has, for instance, led to new ways of automated physical exercise assessment, but this also introduces the potential for novel methods and applications in the automated evaluation of various mental factors such as cognitive state and stress. Extracting such latent variables holds considerable promise, in particular in group-level analysis. Furthermore, the current initiatives and research programs collecting masses of health data from large cohorts create opportunities for developing the methodology. The lack of targeted research is partially hindering the development of the analysis of obscure factors, progress of machine learning and other algorithmic solutions, and the evolution of novel applications and technologies. As described in this introduction, various features inherent in biosignals increase the complexity in the research. In this thesis I provide an introduction to the emerging ubiquitous recording of physiological signals, its effects on the industry, opportunities for organizations and management, and data analytics and measurement techniques. The aim is to seek the future prospects of systemic scenarios and large-scale initiatives. The original publications reviewed in this thesis seek biosignals for features responsive for cognitive states such as stress and, more interestingly, second-order factors derived from inter-individual responses and activations. By introducing more complex features to psychophysiological research, group analytics can be further developed. Second-order analyses provide robust signal features and may lead to advanced applications in assessing well-being and performance. The original publications consist of three research articles and a primer review. The experiments include recordings of magnetoencephalography (MEG), heart rate variability (HRV), and electrodermal activity (EDA), and they exemplify systemic use cases of biosignals. The included primer review discusses general methods in psychophysiological research in human–computer interaction (HCI). Together with this introduction, my experimental results provide evidence that taking psychophysiological measurements from the laboratory to ecologically valid environments is plausible and effective. The literature suggests that consumer-grade devices and personal internet of things will revolutionize myriad sectors, e.g., organizational management. Together with an exponentially increasing data collection and novel applications, the industry will have large economical impacts.
Source: Työväentutkimus Vuosikirja
Author: Lauri Ahonen
Language: English
Topic: Emotion and Mood Recognition
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2883887190