PRACTICAL GUIDANCE ON THE SELECTION OF PHYSIOLOGICAL SENSORS FOR GERONTOLOGICAL RESEARCH. (11th November 2018)

• Record Type: Journal Article
• Authors: Carmichael, A; Raichur, V; Burnside, L; Nielsen, K; Mejia, S; Gonzalez, R
• Is Part Of: Innovation in aging. Volume 2(2018)Supplement 1
• Page Start: 677
• Page End: 677
• Publication Date: 2018-11-11
• Subjects: Aging -- Periodicals; Gerontology -- Periodicals; 612.67
• DOI: 10.1093/geroni/igy023.2521
• Languages: English
• ISSNs: 2399-5300

Abstract

Abstract: Physiological measures have been traditionally employed in gerontological research to study changes in health and functional abilities associated with aging. More recently, these measures have also been used in behavioral and social research to study decision making and user experience. Selection of appropriate sensors to record a pertinent physiological measures is a critical choice in the design of such studies. Poor sensor selection can alter study conclusions and have far-reaching, if insidious, consequences for participant retention, data quality, and data completeness. We argue that sensor selection is an essential process in research design deserving of deep consideration. When choosing an appropriate sensor, one should consider a broad range of factors. For purposes of discussion, these factors can be grouped into four broad domains: (1) sensor specifications, (2) characteristics of the research population, (3) context of application, and (4) analysis goals. While researchers seek to optimize their selection across all factors, these domains are often in tension with one another. Sensor selection thus requires careful prioritization and management of tradeoffs. Many resources are available to help navigate this decision. In our experience, however, no single method has proven more effective at resolving sensor selection dilemmas than thoughtful testing and piloting. We present practical guidance on sensor selection using examples drawn from our own testing, piloting, and research recordings including electrodermal activity, electromyography, and photoplesmography data in older adult populations. Unique design considerations based on age, in-lab vs. at-home recording, and single-sensor vs. multi-sensor integration are addressed.