Digital Phenotype Mood Episode Prediction Using Wearable AI
Solution Overview
Problem
Current methods for predicting mood episodes in patients with mood disorders, such as depression and bipolar disorder, rely heavily on subjective self-reporting and lack precise monitoring of circadian rhythms, which are crucial for early detection and management of relapses.
Innovation Solution
An apparatus and method utilizing digital phenotypes from wearable devices and user terminals to collect log data, extract relevant features, and input them into AI-based prediction models to forecast the occurrence of mood episodes, including major depressive, manic, and hypomanic episodes, by analyzing circadian rhythms, sleep patterns, and other biometric data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If subjective self-reporting methods are used to predict mood episodes, then the method is simple to implement, but the prediction accuracy and reliability are insufficient
Solution Approach 1:
The patent introduces digital phenotype data from wearable devices as an intermediary between the patient and the prediction system. This intermediary objective measurement (sleep patterns, circadian rhythms, activity levels) replaces subjective self-reporting, providing reliable prediction data without requiring complex clinical assessments or patient self-evaluation capabilities.
Solution Approach 2:
The patent replaces the mechanical/subjective self-reporting system with an automated digital sensing system. Wearable devices continuously collect objective physiological data (sleep, circadian rhythms, activity) that feeds into AI prediction models, eliminating the need for patients to manually report their mood states while maintaining high prediction accuracy.
2Reliability
If circadian rhythm monitoring is added to predict mood episodes, then the prediction reliability improves, but the device complexity and data collection requirements increase
Solution Approach 1:
The patent leverages multi-functional wearable devices that already collect various health metrics (sleep, activity, heart rate) and repurposes this existing data for circadian rhythm monitoring and mood episode prediction. By making the wearable device universal for both general health tracking and psychiatric monitoring, the system avoids adding separate complex dedicated circadian monitoring equipment.
Solution Approach 2:
The patent transforms raw wearable device data (sleep duration, activity patterns, circadian rhythm variations) into clinically meaningful parameters that predict mood episodes. By changing the representation of data from basic sensor readings to derived physiological parameters, the system maintains reliable prediction while simplifying the data collection interface to standard wearable sensors.
3Loss of time
If real-time digital phenotype data is collected continuously, then the prediction timeliness and early detection capability improve, but the energy consumption and data processing load increase
Solution Approach 1:
The patent performs preliminary data processing and feature extraction on wearable device data in the background, continuously monitoring for patterns that predict mood episodes. By preparing and analyzing data in advance before actual relapse occurs, the system achieves rapid detection response without requiring intensive real-time processing at the moment of crisis, thereby reducing peak energy consumption.
Solution Approach 2:
The patent selectively processes only the most relevant digital phenotype parameters (sleep patterns, circadian rhythms, activity levels) rather than continuously analyzing all possible sensor data. By focusing computational resources on the most predictive parameters, the system maintains timely relapse detection while minimizing unnecessary energy consumption from processing irrelevant data streams.
Data Source
AI summary
A method for predicting an occurrence of a mood episode using a digital phenotype according to the exemplary embodiment may include acquiring log data associated with a circadian rhythm of a target user from at least one of a user terminal of the target user and a wearable device which is attached to or worn on a body of the target user, extracting predetermined main feature information from the log data, and deducing prediction information about the occurrence of the mood episode of the target user by inputting the main feature information to a previously trained artificial-intelligence based prediction model.



