MS Pattern Explorer: interactive visual exploration of temporal activity patterns for multiple sclerosis

Area of Science:

• Biomedical Informatics
• Human-Computer Interaction
• Data Visualization

Background:

• Fitness wearables generate vast amounts of activity data, posing challenges for clinical interpretation.
• Analyzing this data is crucial for understanding disease progression and patient symptomatology, particularly in conditions like multiple sclerosis (MS).
• Existing methods often struggle with signal overload and rapid insight generation from complex sensor data.

Purpose of the Study:

• To design and evaluate MS Pattern Explorer, a novel visual analytics tool.
• To leverage interactive machine learning for analyzing fitness wearable data in MS patients.
• To address challenges in managing activity signals, accelerating insight generation, and contextualizing patterns.

Main Methods:

• User-centered design approach prioritizing clinician needs for pattern exploration and contextualization.
• Computation of meaningful patient activity and sleep sequences using clustering and proximity search.
• Development of an interactive visual interface with coordinated views.
• Evaluation involving 15 participants (clinicians, data scientists, non-experts) using usability and insight generation scoring.

Main Results:

• MS Pattern Explorer facilitates understanding of activity patterns in temporal data.
• The tool enables rapid insight generation and contextualization of data within and between patient cohorts.
• Consistent performance was observed across diverse participant groups.
• Effective support for generating insights from MS patient fitness tracker data was demonstrated.

Conclusions:

• MS Pattern Explorer effectively reduces signal overload for clinicians analyzing activity data.
• The tool offers novel opportunities for data exploration, sense-making, and hypothesis generation in clinical research.
• The system has broad applicability for analyzing sensor data in chronic condition studies and cohort comparisons.