ArXiv Preprint
The effectiveness of digital treatments can be measured by requiring patients
to self-report their mental and physical state through mobile applications.
However, self-reporting can be overwhelming and may cause patients to disengage
from the intervention. In order to address this issue, we conduct a feasibility
study to explore the impact of gamification on the cognitive burden of
self-reporting. Our approach involves the creation of a system to assess
cognitive burden through the analysis of photoplethysmography (PPG) signals
obtained from a smartwatch. The system is built by collecting PPG data during
both cognitively demanding tasks and periods of rest. The obtained data is
utilized to train a machine learning model to detect cognitive load (CL).
Subsequently, we create two versions of health surveys: a gamified version and
a traditional version. Our aim is to estimate the cognitive load experienced by
participants while completing these surveys using their mobile devices. We find
that CL detector performance can be enhanced via pre-training on stress
detection tasks and requires capturing of a minimum 30 seconds of PPG signal to
work adequately. For 10 out of 13 participants, a personalized cognitive load
detector can achieve an F1 score above 0.7. We find no difference between the
gamified and non-gamified mobile surveys in terms of time spent in the state of
high cognitive load but participants prefer the gamified version. The average
time spent on each question is 5.5 for gamified survey vs 6 seconds for the
non-gamified version.
Michal K. Grzeszczyk, Paulina Adamczyk, Sylwia Marek, Ryszard Pręcikowski, Maciej Kuś, M. Patrycja Lelujko, Rosmary Blanco, Tomasz Trzciński, Arkadiusz Sitek, Maciej Malawski, Aneta Lisowska
2023-02-07
