Testing the Efficacy of Social Robot-led Yoga

Research demonstrates that yoga is beneficial for the older adults with or nearing mild cognitive impairment. Social robots are increasingly being used as companions and caregivers for older adults with or nearing MCI. The purpose of this study is to test the efficacy of social robot-led yoga for older adults with or nearing MCI. The study hypothesizes that social robot-led yoga is similar in effectiveness to human-led yoga across a range of measures.

A group of approximately 40-50 older adults (65+) will follow a yoga routine twice per week for 12 weeks taught by either a social robot or human instructor. Details of the yoga instruction can be found below. Yoga sessions will be conducted at an assisted living facility in Denver, Colorado.

Prior to, at the midway point, and after the yoga intervention, researchers will measure participants for the following:

• Range of motion, overhead flexion: Ryan, the social robot, can scan participants using infrared cameras and create a digital remembrance of their end range of overhead flexion. n = 40 participants will complete the range of motion assessment prior to intervention, midway through the intervention (week 6), post intervention (week 12), and at follow up (week 18).
• Single-leg balance: Ryan, the social robot, can measure the duration of participant single-leg balance. participants will be asked to balance on their nondominant leg for up to 60 seconds. n = 40 participants will complete the single-leg balance assessment prior to intervention, midway through the intervention (week 6), post intervention (week 12), and at follow up (week 18).
• The Saint Louis University Mental Status examination (SLUMS): the SLUMS questionnaire will be administered to measure cognitive performance. n = 40 participants will complete the slums prior to intervention, midway through the intervention (week 6), post intervention (week 12), and at follow up (week 18). If at any point the participant meets the exclusion criteria based on SLUMS score, they will be removed from the study.
• The patient health questionnaire (PHQ-9): researchers will administer the PHQ-9 to measure depression. n = 40 participants will complete the phq-9 prior to intervention, midway through the intervention (week 6), post intervention (week 12), and at follow up (week 18).
• Kentucky inventory of mindfulness skills (KIMS): researchers will administer the KIMS to measure mindful attention capacity. n= 40 participants will complete the KIMS prior to intervention, midway through the intervention (week 6), post intervention (week 12), and at follow up (week 18).
• Task engagement: researchers will administer a task engagement questionnaire to measure engagement in the yoga experience. This modified scale will provide a reliable measure of user engagement with both instructors. n = 40 participants will complete the user engagement questionnaire after each yoga session (week 1-12).

At the end of each yoga session, participants will fill out a brief qualitative questionnaire. the questionnaire will ask the following: 1) Please described your experience. 2) What did your instructor do well, where did they struggle? To increase questionnaire brevity, the task engagement questionnaire and qualitative questionnaire are displayed on the same single page of paper.

After six weeks have passed from the date of the last yoga session, participants will be measured one final time as a form of post-participation follow up. Measures will include range of motion, single-leg balance, SLUMS examination, PHQ-9, and KIMS.

Yoga Instruction: Social robot-led yoga will be instructed by Ryan, the social robot. Ryan will be programmed to instruct 10-14 yoga postures in addition to an opening and closing meditation. Each yoga posture will be selected for appropriateness to the developmental age of participants. Several postures will be seated. Others will be standing. Participants will be given the option to stand if they wish and encouraged to move at their own pace. Each posture focuses on creating requisite strength for older adults and correlates to the study's physical measures of single-leg balance and overhead flexion, range of motion. All postures re-enforce each other. Postures will be randomized to ensure variety and updated periodically throughout the study to ensure proper level of challenge and engagement. Updates may include new postures, new verbal or video-based instruction in conjunction with physical instruction changes. Human-led yoga will follow the same format as the social robot-led yoga sessions. Instruction will be 45-minutes of yoga designed with an integrative focus. Human-led yoga will mirror social robot-led yoga in intent, duration, and posture selection.

Inclusion criteria

• Age 65 years or older.
• SLUMS score of 20-26 will include persons with moderate cognitive impairment
• SLUMS score of >26 will include persons without moderate cognitive impairment
• Verbal skill to interact
• Available for a period of 12 weeks to participate in yoga
• Ability to stand without supports

Exclusion criteria

• Age younger than 65 years
• Aggressive behaviors
• Medical diagnosis of dementia or memory loss
• An inability to balance on a single leg without supports
• Acute physical illness that impairs or precludes physical activity
• SLUMS score equal to or below 20, the cutoff for dementia
• A drop in SLUMS score greater than 3 points from the beginning of the study to the midpoint of the study