Effects of mood and aging on keystroke dynamics metadata and their diurnal patterns in a large open-science sample: A...

Area of Science:

• Digital health
• Psychological assessment
• Human-computer interaction

Background:

• Traditional psychological assessments can be complemented by ecologically relevant metrics from ubiquitous technologies.
• Smartphone-derived keyboard metadata presents a novel avenue for digital biomarkers of mood.

Purpose of the Study:

• To determine the feasibility of using smartphone-derived real-world keyboard metadata as digital biomarkers of mood.
• To investigate the relationship between typing metadata and depression severity.

Main Methods:

• A real-world observation study (BiAffect) utilized a custom iPhone keyboard app to unobtrusively collect typing metadata.
• Data from over 14 million keypresses from 250 users, including demographics and Patient Health Questionnaire-8 (PHQ-8) for depression severity, were analyzed.
• Hierarchical growth curve mixed-effects models were employed to analyze the effects of mood, demographics, and time of day on keyboard metadata.

Main Results:

• Increased depression severity correlated with more variable typing speed, shorter session duration, and lower accuracy.
• Typing speed and variability displayed diurnal patterns, peaking at midday.
• Age influenced typing speed and variability, with older users typing slower and more variably, especially in the evening.

Conclusions:

• Unobtrusively collected keystroke dynamics are significantly associated with mood.
• Real-world typing metadata can serve as a foundation for developing digital biomarkers for mood.
• These digital biomarkers are robust despite diurnal patterns and age-related effects.