Cognitive Performance Measurements and the Impact of Sleep Quality Using Wearable and Mobile Sensors

TL;DR

This work addresses how sleep quality shapes cognitive performance in daily life by combining wearable sleep metrics, passive smartphone sensing, and active cognitive testing in a two-month field study. The authors model the impact of detailed sleep variables on alertness measured by the Psychomotor Vigilance Task (PVT) and demonstrate that smartphone typing dynamics can serve as a viable proxy for cognitive performance, enabling unobtrusive, long-term monitoring. Key findings show that night-time heart rate, sleep latency, sleep timing, sleep restfulness, and total sleep duration significantly influence next-day alertness, with distinct patterns for single-night versus multi-night averages; typing speed patterns closely track PVT outcomes. These results advance ubiquitous computing by supporting cognition-aware systems and enabling continuous cognitive assessment through passively collected typing data, reducing the need for frequent active testing in real-world settings.

Abstract

Human cognitive performance is an underlying factor in most of our daily lives, and numerous factors influence cognitive performance. In this work, we investigate how changes in sleep quality influence cognitive performance, measured from a dataset collected during a 2-month field study. We collected cognitive performance data (alertness) with the Psychomotor Vigilance Task (PVT), mobile keyboard typing metrics from participants' smartphones, and sleep quality metrics through a wearable sleep tracking ring. Our findings highlight that specific sleep metrics like night-time heart rate, sleep latency, sleep timing, sleep restfulness, and overall sleep quantity significantly influence cognitive performance. To strengthen the current research on cognitive measurements, we introduce smartphone typing metrics as a proxy or a complementary method for continuous passive measurement of cognitive performance. Together, our findings contribute to ubiquitous computing via a longitudinal case study with a novel wearable device, the resulting findings on the association between sleep and cognitive function, and the introduction of smartphone keyboard typing as a proxy of cognitive function.

Figures (6)

• Figure 1: The Oura sleep tracking ring. The sensors - infrared photoplethysmography (PPG) for heart rate, negative temperature coefficient (NTC) for temperature, and 3D accelerometer for movement - are placed on the inner surface of the ring.
• Figure 2: The reaction test application ("Circog") used in our study. Once the user clicks start (left figure), at random intervals the counter (right figure) starts running after 3-10s. Clicking anywhere on the surrounding box stops the counter, stores the reaction time result and starts a new counter after a random 1-5s interval. This is repeated 8-12 times. Details of the application can be found in dingler2017building.
• Figure 3: Study periods and participants of both participant groups in the study. For participant group 2 the end interview was replaced with an online survey with same topics being addressed, due to the COVID-19 social distancing restrictions. Group 2 could re-use sleep trackers from group 1, thus for most their sleep tracking period begun right away.
• Figure 4: Hourly sum of our typing event dataset and reaction test dataset. Samples between midnight and 7AM (24-06) are omitted from further analysis.
• Figure 5: Regression model results between reaction test performance (difference to mean, lower value is better) and individual sleep metrics. Blue line denotes a rolling average from previous five nights for each sleep metric. Red line denotes previous night's result. Grey area denotes the 95% confidence interval of the model. Table \ref{['table:models']} includes details on model selection for each metric and one-day vs. 5-day results. 'Total time in bed' equals the duration variable and 'Total sleep duration' equals the total sleep variable.