Investigating Rhythmicity in App Usage to Predict Depressive Symptoms: Protocol for Personalized Framework...

Area of Science:

• Digital Phenotyping
• Computational Psychiatry
• Behavioral Informatics

Background:

• Depressive symptom prediction in students is crucial for timely diagnosis and treatment.
• Existing unobtrusive prediction systems face limitations in sample size, performance, and resource requirements.
• The potential of app usage behavioral rhythms for predicting depressive symptoms remains underexplored.

Purpose of the Study:

• To investigate the existence of statistically significant rhythms in resource-insensitive app usage behavioral markers.
• To predict depressive symptoms using these marker-based rhythmic features.
• To understand the association between rhythmic features and depressive symptoms.

Main Methods:

• Collected countrywide app usage behavioral data from 2952 students.
• Employed zero-amplitude tests, cosinor models, and nonparametric rhythmic feature analysis.
• Developed a personalized multitask learning (MTL) framework for symptom prediction.

Main Results:

• Analyzed data from 2902 students, encompassing over 24 million app usage events.
• Data represented a diverse student population across Bangladesh.
• Findings on rhythmic features and their association with depressive symptoms are forthcoming.

Conclusions:

• Understanding app usage rhythms can aid healthcare professionals in better supporting depressed students.
• App usage-based rhythms may offer novel possibilities for mental health interventions.
• The MTL framework shows potential for accurate depressive symptom prediction through rhythmic features.