Population-scale hand tremor analysis via anonymized mouse cursor signals

Area of Science:

• Neurology
• Epidemiology
• Human-Computer Interaction

Background:

• Tremors are common movement disorders with diverse causes.
• Clinical studies of tremors are often limited by small sample sizes and short observation periods.
• Understanding tremor prevalence across large populations is challenging with traditional methods.

Purpose of the Study:

• To assess the feasibility of using anonymized computer mouse cursor movement data to estimate hand tremor prevalence.
• To validate mouse cursor movement as a population-scale tremor sensor for epidemiological research.
• To explore potential applications in tremor screening and illness progression monitoring.

Main Methods:

• Collected anonymized mouse cursor movement data from millions of web search engine users.
• Developed methods to estimate hand tremors based on tremor-like cursor movements.
• Compared estimated tremor characteristics against user self-reports, published clinical findings, and a targeted study.

Main Results:

• Significant alignment was found between estimated tremors and user self-reports and clinical findings.
• Users with cursor tremor events showed higher likelihood of tremor-related search interests.
• Demographic variations in cursor tremor quantity and frequency mirrored established clinical study patterns.

Conclusions:

• Anonymized mouse cursor motion is a valid population-scale tremor sensor for epidemiological studies.
• This method offers a feasible approach for large-scale tremor detection and analysis.
• Future applications may include opt-in services for tremor screening and disease progression monitoring.