Multi-Window Heart Rate Analysis for Seizure Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current seizure detection algorithms using heart data struggle to accurately and rapidly identify the onset and end of seizures, often confusing pathological changes with non-pathological exertional or positional changes, leading to delayed intervention and inaccurate assessment of patient condition.
Innovation Solution
A method and system that determine heart rate (HR) parameters by comparing HR measures across multiple windows, using thresholds to identify the onset and end of seizures, thereby distinguishing between seizure-related HR changes and normal physiological changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If current seizure detection algorithms use heart data to identify seizures, then seizure detection capability is provided, but accuracy is reduced due to confusion between pathological and non-pathological heart rate changes
Solution Approach 1:
The patent segments the heart rate analysis into multiple time windows (first window, second window, third window) with different purposes. The first window captures pre-seizure baseline, the second window captures seizure onset, and the third window captures seizure continuation. This segmentation allows the algorithm to distinguish between transient exertional changes and sustained seizure-related changes by comparing patterns across windows.
Solution Approach 2:
The patent performs preliminary analysis by establishing a baseline heart rate pattern in the first window before seizure onset. This baseline is stored and used for comparison in subsequent windows, enabling the algorithm to detect deviations from normal physiology that indicate seizure onset while filtering out expected variations.
2Productivity
If single-window heart rate analysis is used, then computational simplicity is maintained, but detection speed and accuracy of seizure onset identification deteriorates
Solution Approach 1:
The patent implements periodic analysis by continuously evaluating heart rate data in sequential time windows. The algorithm periodically compares the current window (second window) with the baseline window (first window) and follows up with a continuation window (third window) if seizure onset is detected. This periodic multi-window approach enables rapid identification of seizure onset while maintaining computational efficiency through structured, repetitive analysis patterns.
3Ease of operation
If heart rate monitoring is used to detect seizures, then non-invasive detection is achieved, but ability to distinguish seizure-related changes from exertional changes deteriorates
Solution Approach 1:
The patent applies dynamics by analyzing the temporal evolution of heart rate patterns rather than static snapshots. The algorithm examines how heart rate changes unfold across multiple time windows, looking for the dynamic pattern characteristic of seizures (gradual onset, sustained elevation) versus the dynamic pattern of exertion (rapid onset, quick recovery). This dynamic analysis maintains non-invasive monitoring while improving discrimination capability.
Data Source
AI summary
Methods and systems for characterizing a seizure event in a patient, including determining a time of beat sequence of the patient's heart, determining a first HR measure for a first window, determining a second HR measure for a second window, wherein at least a portion of the first window occurs after the second window, determining at least one HR parameter based upon said first HR measure and said second HR measure, identifying an onset of the seizure event in response to determining that at least one HR parameter crosses an onset threshold, identifying an end of the seizure event in response to determining that at least one HR parameter crosses an offset threshold.