Human Robot Interaction in older adults with Mild Cognitive Impairments

Eleonora Zedda ~ University of Pisa and ISTI-CNR [email protected] Outline

Motivation Cognitive Degeneration Mild Cognitive Impairments Socially Assistive Robots First study Preliminary Results Future Works Motivation-Aging of Italian Population

Average Age

istat.it Motivation-Aging of Italian Population

Average Age

Average Age 2017

istat.it Motivation

Degeneration in cognitive and physical domain

Increased number of caregivers

Gap health system

Assistive technologies Cognitive Degeneration

Preclinical Predementia Dementia

Cognitive decline normal age Mild Cognitive Impairments (MCI) Alzheimer's disease

Asymptomatic stage Middle cognitive degeneration Severe cognitive degeneration Cognitive Degeneration

Predementia

Mild Cognitive Impairments

completion of complex tasks

reduced coordination of brain activity

performance of different cognitive domains (attention, memory, ..)

maintain their independence during the activities of daily living Mild Cognitive Impairments

The MCI people are at high risk of dementia; every year about 10% of MCI people progress to dementia MCI Interventions

Cognitive and physical training

Improve cognitive performance

Reducing the incidence of dementia.

Intervene as soon as possible

Assistive technologies Assistive Technologies

Assistive technologies paper-based material Assistive technologies computer-based support Socially Assistive Robot Socially Assistive (SAR)

Providing assistance and achieving measurable progress in convalescence, rehabilitation, learning Benefit SAR

Users did not experience any anxiety

Features helps the user to facilitate the interaction with the robot

More engagement

Different type of interaction (voice, gesture, …)

Secure environments

Fill the gap between current healthcare and self-care First study

HIIS Laboratory ISTI-CNR collaboration with Institute of Neuroscience-CNR

Investigate how seniors with MCI relate and perceived a cognitive game accessed Goal through a humanoid robots, as a part a training program aimed to improve their cognitive status

12 Session at CNR Session 14 participants (average age 75.3) with MCI Low level of computer experience

Stages Familiarization with devices

Test (UES, demographic and computer experience questionnaires) First study- Familiarization

Picture taken by our group during the familiarization with the robot at the Institute of Neuroscience of CNR First study-Test

Questionnaire User Engagement Scale questionnaire Demographic questionnaire Computer experience questionnaires

Groups Tablet (control group) Robot

Picture taken during the Pepper Session Pepper Robot

Robot developed by Softbanks robotics

Interaction is one of the key features of the Pepper robot’s capabilities

Tablet

Different laser sensors, cameras, tactile and movement sensors First Study – Test set up First study

Cognitive Music Quiz Training 15 songs of 20 seconds each Recognize singer and song

Memory retrograde memory (known songs) anterograde memory (unknown songs)

autobiographical memory (known songs evoke personal memories)

Domains attention, memory, reasoning Game - App

Ionic Game

Python Connection

Receives events

Pepper controller Game - Example

Question Wrong answer Expired time

Timer

Singer Game - Ionic App Game Feedback

Adaptive Feedback Changes depending on the answer chosen Tablet Voice

Visuals

Adaptive Feedback Changes depending on the answer chosen Answer Robot Voice

Visuals Animation

Coloured led Sounds Game Feedback

Positive feedback

Negative feedback Preliminary Results Data still under analysis

Talked about the game after the training session and with persons external to the programme

Game did not perceive as a tasks but as a stimolous

Robot perceive as a friends and describe it in a human-like manner

“Pepper you are so cute”

“Ciao Pepper, you know…I woke up with back pain and I was deciding whether to come this morning ... I decided to come because I knew I would play with you” Future works

Gesture interaction

Different training games for different domain (math, logical, ..)

Adaptive behaviour (Reinforcement Learning)

Robot Operating System (ROS) is a set of software libraries and tools that help you build robot applications Summary

Motivation of the study linked with the aging of population and the increased of cognitive impairments create a gap in the health system

The benefit of the Socially Assistive robot as an Assistive technologies to fill the gap in the health system

Develop a cognitive training for elderly with MCI using a Socially Assistive Robot

Preliminary results of the project and future works References

Bechade, Lucile, et al. (2019) Towards Metrics of Evaluation of Pepper Robot as a Social Companion for the Elderly. In Advanced Social Interaction with Agents. Lecture Notes in Electrical Engineering, 89-101. Springer, Cham. DOI: https://doi.org/10.1007/978-3-319-92108-2_11

Pandey, A.K., and Gelin, R. (2018) A Mass-Produced Sociable : Pepper: The first Machine of Its kind, IEEE Robotics & Automation Magazine. 25(3), 40-48, DOI: 10.1109/MRA.2018.2833157

Calvo, R. A., D’Mello S., et al. (Eds.). (2015). The Oxford Handbook of Affective Computing. Oxford University Press. USA. DOI: 10.1093/oxfordhb/9780199942237.001.0001

Pino, Olimpia, et al. (2019) The Humanoid Robot as Trainer in a Memory Program for Elderly People with Mild Cognitive Impairment. International Journal of Social Robotics, 1-13. DOI: https://doi.org/10.1007/s12369-019-00533-y

Broekens, J., and Chetouani. (2019) “Towards Transparent Robot Learning through TDRL-based Emotional Expressions. IEEE Transactions on Affective Computing. DOI: 10.1109/TAFFC.2019.2893348