ADAPTIVE CORRELATION COEFFICIENT THRESHOLD
Publication Date: 2005-Jul-20
The IP.com Prior Art Database
The present method allows for automaticity with real-time morphology variation monitoring and automatic correlation coefficient threshold adjustment. Correlation coefficient threshold adjustment increases specificity for a particular patient and can improve methods for diagnosing and treating cardiac conditions. This feature is desirable for implantable cardiac rhythm management (CRM) devices used in patients with rate dependent morphology variation. Here, correlation coefficient threshold adjustment may improve the specificity. CRM devices used for heart failure patients that show unstable morphology may benefit from correlation coefficient threshold adjustment because, when the correlation coefficient values at low rate are already compromised, automatically adjusted correlation coefficient threshold may improve CRM performance.
Discriminating ventricular from supraventricular cardiac rhythms may involve comparing an electrogram signal of one or more cardiac beats to a template characterizing a supraventricular rhythm. If the cardiac beats are sufficiently correlated to the template then the rhythm is determined to be a supraventricular rhythm (SVR). A beat may be determined to be sufficiently correlated if the correlation coefficient calculated using sample points or feature points of the beat and corresponding sample point or feature points of the template meets or exceeds a correlation coefficient threshold.
When the correlation coefficient for supra-ventricular tachycardia (SVT) rhythms exhibit a different distribution from the correlation coefficient of ventricular tachycardia and ventricular fibrillation (VT/VF) rhythms, the correlation coefficient threshold used for discriminating SVT from VT/VF may be adjusted.
In one method, a correlation coefficient threshold may be adjusted based on detected morphology variations during a sustained sinus tachycardia (ST). When a cardiac rhythm is confirmed as a sinus tachycardia (ST) rhythm, the correlation coefficient variations can be monitored as the rate is accelerated. If the correlation coefficient is lowered, but higher than the correlation coefficient values of previously detected VT/VF, then the correlation coefficient threshold may be lowered to improve the specificity without lowering sensitivity. A cardiac rhythm may be confirmed as sinus tachycardia (ST) by gradual rate change, Vrate=Arate, and stable heart rate consistent with the sensor rate. During sustaining ST, morphology variations may be monitored, trended, and used to adjust a correlation coefficient threshold.
The correlation between cardiac beats and the SVR template is continuously monitored to detect a morphological variation. The correlation can be evaluated at every pre-determined beat per minute (bpm) increment, e.g., a 5 bpm increment, at pre-determined rates, e.g., 120 bpm increment 5 bpm, or periodically, which may be determined as a function of the heart rate. Morphological variation may detect a conduction variation such as bundle branch block (BBB). If the morphological variation is detected at a rate less than the VT rate, or at a rate greater than the VT rate if ST is confirmed, then the correlation coefficient for VT detection may be adjusted, or other rhythm discrimination methods may be used.
In another implementation, the correlation coefficient values of stored tachycardia electrograms (EGM) are monitored and compared. Correlation coefficient values of stored VT/VF episodes and SVT episodes are compared, and based on the comparison, the correlation coefficient threshold may be adjusted to improve the specificity without lowering the sensitivity.