Actimetric Symptom Onset Prediction Using Time-Frequency Spike Trains
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting the onset of psychiatric illnesses or medical conditions, such as depressive episodes, are hindered by performance heterogeneity, data quality concerns, and the inability to capture subtle behavioral and physiological changes, limiting early detection and intervention.
Innovation Solution
A sensor-based method utilizing multi-resolution time-frequency analysis of actimetric data, including empirical mode decomposition and Hilbert transformation, to identify time-frequency spike trains and anomalies, predicting symptom onset through anomaly scoring and issuing alerts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional time-domain methods are used for anomaly detection, then the analysis is simple and computationally efficient, but subtle changes associated with symptom onset cannot be captured
Solution Approach 1:
The patent transforms the analysis from traditional time-domain methods to time-frequency domain analysis. By applying multi-resolution time-frequency analysis, the system adds a frequency dimension to the temporal data, enabling detection of subtle changes in behavioral and physiological signals that are not visible in the time domain alone. This dimensional transformation allows capture of transient patterns and anomalies associated with symptom onset while maintaining computational feasibility through efficient signal processing techniques.
2Productivity
If summary statistics of windowed and under-sampled data are used, then the computation is simplified, but granular data analysis required for mood regulation detection is lost
Solution Approach 1:
The patent applies segmentation by dividing the continuous actimetric data into multiple temporal scales through multi-resolution analysis. Instead of using a single windowed approach that loses granularity, the system segments the data at different resolution levels, allowing both detailed granular analysis and efficient summary statistics to coexist. This enables detection of subtle mood regulation changes while maintaining computational efficiency through hierarchical processing.
3Reliability
If current prediction methods are applied, then some symptom onset can be detected, but performance heterogeneity across participants and data quality issues limit reliability
Solution Approach 1:
The patent changes the parameters of analysis by moving from traditional time-domain features to time-frequency domain features. This parameter transformation enables more reliable detection of symptom onset by capturing patterns that are invariant to individual variations in behavior. The multi-resolution approach adapts to different participants' data characteristics, reducing performance heterogeneity while the systematic methodology maintains reliability without excessive complexity.
Data Source
AI summary
A sensor-based system, method and computer program product for individualized prediction of onset of a symptom of a psychiatric illness or medical condition. Actimetric patient data is obtained from at least one actimetric sensor associated with a patient. The actimetric patient data comprises at least one of time series sleep variable data and time series activity variable data. Multi-resolution time-frequency analysis is applied to the actimetric patient data to obtain time-frequency spike train data for the patient, and onset of a symptom is predicted from the time-frequency spike train data. An alert is issued when the onset of the symptom is predicted from the time-frequency spike train data.