Award Abstract # 1955590
CHS: Medium: Collaborative Research: Teachable Activity Trackers for Older Adults

NSF Org: IIS
Division of Information & Intelligent Systems
Recipient: THE PENNSYLVANIA STATE UNIVERSITY
Initial Amendment Date: August 19, 2020
Latest Amendment Date: October 15, 2020
Award Number: 1955590
Award Instrument: Standard Grant
Program Manager: Dan Cosley
dcosley@nsf.gov
 (703)292-8832
IIS
 Division of Information & Intelligent Systems
CSE
 Directorate for Computer and Information Science and Engineering
Start Date: October 1, 2020
End Date: September 30, 2024 (Estimated)
Total Intended Award Amount: $119,992.00
Total Awarded Amount to Date: $119,992.00
Funds Obligated to Date: FY 2020 = $119,992.00
History of Investigator:
  • David Conroy (Principal Investigator)
Recipient Sponsored Research Office: Pennsylvania State Univ University Park
201 OLD MAIN
UNIVERSITY PARK
PA  US  16802-1503
(814)865-1372
Sponsor Congressional District: 15
Primary Place of Performance: Pennsylvania State Univ University Park
110 Technology Center Building
University Park
PA  US  16802-1503
Primary Place of Performance
Congressional District:
15
Unique Entity Identifier (UEI): NPM2J7MSCF61
Parent UEI:
NSF Program(s): HCC-Human-Centered Computing
Primary Program Source: 01002021DB NSF RESEARCH & RELATED ACTIVIT
Program Reference Code(s): 7367, 7924
Program Element Code(s): 736700
Award Agency Code: 4900
Fund Agency Code: 4900
Assistance Listing Number(s): 47.070

ABSTRACT

Self-tracking of physical activities can support people of all ages in understanding their lifestyle behaviors and making healthy choices, reducing chronic disease risks. For older adults, movement behaviors are especially critical. They help people maintain functional abilities and live independently. Smart watches and other activity tracking technologies have become available, making self-tracking easier than before, but older adults have adopted them less. One barrier is that current physical activity trackers do not effectively identify and track older adults? activities. This project aims to understand (1) what kind of data are needed from older adults to make activity tracking work for them; and (2) how to engage older adults to collect the needed data. This project will develop a new approach to personalizing older adults? activity tracking. It will open up new research avenues on personalized and multimodal self-tracking that affect healthcare, quality of life, and privacy. This project is expected to make broader impacts for older adults in enhancing their motivation to engage in physical activities, as well as societal impacts in nurturing a culture of diversity and inclusion that benefits the lives of older adults and people with and without disabilities or health conditions.

This project uses ?teachable interfaces? to facilitate personalized, self-tracking for older adults? physical activities, while considering their changes in mobility and diverse physical characteristics. The teachable interfaces are intended to help people provide personalized activity labels, which will be used to recognize their unique movements. They will also enable self-tracking of meaningful and modifiable movement and non-movement activities, supporting older adults to displace inactivity with physical activity, which can provide significant health benefits. The research team will investigate: (1) older adults? movement and non-movement activities that they wish to change; (2) new personalized, multimodal activity trackers that provide opportunities for self-reflection through teachable interfaces; and (3) commonalities and differences in efficacies for subgroups of older adults (e.g., people with mild dementia) and what adjustments are needed to accommodate them. Combining expertise from human-computer interaction, interactive machine learning, accessibility, aging, and kinesiology, the project will employ a mixed-methods research approach: co-design with older adults, technology design and development, and evaluations both in the lab and in people?s natural environments.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

PUBLICATIONS PRODUCED AS A RESULT OF THIS RESEARCH

Note:  When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external site maintained by the publisher. Some full text articles may not yet be available without a charge during the embargo (administrative interval).

Some links on this page may take you to non-federal websites. Their policies may differ from this site.

Kim, Young-Ho and Chou, Diana and Lee, Bongshin and Danilovich, Margaret and Lazar, Amanda and Conroy, David E. and Kacorri, Hernisa and Choe, Eun Kyoung "MyMove: Facilitating Older Adults to Collect In-Situ Activity Labels on a Smartwatch with Speech" 2022 CHI Conference on Human Factors in Computing Systems , 2022 https://doi.org/10.1145/3491102.3517457 Citation Details

PROJECT OUTCOMES REPORT

Disclaimer

This Project Outcomes Report for the General Public is displayed verbatim as submitted by the Principal Investigator (PI) for this award. Any opinions, findings, and conclusions or recommendations expressed in this Report are those of the PI and do not necessarily reflect the views of the National Science Foundation; NSF has not approved or endorsed its content.

The main objective of this grant was to design and build personalized activity tracking systems that better match older adults’ lifestyle and physiological characteristics, with a long-term goal of supporting older adults to actively engage in their physical activity. Specifically, we devised “teachable interfaces,” a mechanism that enables older adults to capture data and tailor a machine learning algorithm with their training examples. The novel approach lies in using activity annotations with a teachable interface to serve dual purposes: (1) increasing older adults’ awareness of their own movement and non-movement activities, and (2) improving system accuracy by allowing individuals to tailor it to their personal patterns and idiosyncrasies. To facilitate this, we developed a smartwatch app that older adults used to label their behavior with verbal annotations. To obtain ground truth data, these annotations were then aligned with data from a research grade accelerometer that participants wore. Older adults also completed extensive interviews about their experiences using the smartwatch app and what activities they consider meaningful and worth capturing.

Smartwatches proved to be a feasible medium for verbal annotation by older adults. In the deployment study, older adults wore the watches for an average of 11.6 hours/day and provided a balanced mix of prompted and user-initiated reports on their behavior. From these data, we observed that older adults rarely labeled their activity as a form of structured exercise (e.g., running) but did report variability in the experienced intensity of activities undertaken. Based on this finding, we used data from a research-grade accelerometer to develop markers of intensity for older adults. We also explored a variety of “few-shot learning” methods that would use sparse labels for an individual to generate better models for classifying their data. Thematic analysis of interviews with older adults also identified a set of values that were competing and influencing the meaning of different activities. For example, some older adults obtained meaning from improving physical health where others obtained meaning from enhancing mental engagement or social connections. Some activities were more acceptable as tracking targets and older adults preferred not to track activities when the goal was immersion due to the cognitive demands of tracking. In other words, there was a distinction between meaningful activities for older adults and the activities that were meaningful to track. A key implication of this finding is that aging research should attend more to some low intensity activities that are common throughout an older adult’s day and not just focus on moderate-to-vigorous intensity physical activities.

This project converged with work undertaken by the Science Board for the President’s Council on Sports, Fitness & Nutrition which was developing their 2023 Midcourse Report on Implementation Strategies for Promoting Physical Activity in Older Adults. One of our team members (Conroy) was a member of that board and reviewed evidence on physical activity interventions for older adults based on both device-based and self-report measures of physical activity. That report highlighted some of the gaps that our work attempted to address with respect to understanding the scope of activities in which older adults engage and the impact of disability and health conditions on activity levels.

Ongoing work by the collaborating research team/site is focused on co-designing teachable activity trackers for and with older adults. Through co-design, the research team identified older adults’ preferences for multimodal data input to support more accurate and precise activity annotations. The participant pool was also expanded to include older adults with dementia or mild cognitive impairments, yielding insights into how interfaces could be adjusted to better support memory-related challenges during annotation. Activity annotation can provide both a measure of behavior and a stimulus for initiating behavior change; however, it remains unclear what types of activities are most effective to track for promoting positive behavior change, and how much annotation is needed to achieve both acceptable model performance and meaningful impact on behavior. This work is especially important for personalized activity tracking across the full spectrum of intensities.

In sum, this project shed light on the unique needs of older adults and how self-tracking strategies can be designed to enhance older adults’ experience engaging with the strategy. The findings have implications for future monitoring strategies and behavioral interventions aimed at promoting physical activity in older adults via self-tracking.

 


Last Modified: 05/17/2025
Modified by: David E Conroy

Please report errors in award information by writing to: awardsearch@nsf.gov.

Print this page

Back to Top of page