Medical systems and methods for detecting changes in electrophysiological evoked potentials
Inventors
JOHNS, Gregg • O'Brien, Richard Arthur • Snow, Robert
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-11963775-B2
Publication Date
2024-04-23
Expiration Date
Abstract
A medical method of automatically improving signals received from a patient's physiological system can include delivering stimulation signals to a nerve pathway of a patient with electrical pulses via electrodes placed over the nerve pathway to generate a plurality of resultant electrical waveforms (EPs) based on a plurality of electrophysiological responses (ERs); recording the plurality of resultant EPs; generating an ensemble average waveform (EA), the generating comprising averaging a subset of the plurality of ERs; and denoising the EA, the denoising comprising applying a wavelet transform to the EA.
Core Innovation
The invention relates to improved evoked potential (EP) monitoring by automatically improving signals received from a patient's physiological system. Stimulation signals are delivered to a nerve pathway via electrodes placed over the nerve pathway using electrical pulses to generate resultant evoked potentials (EPs) based on electrophysiological responses (ERs). A plurality of resultant EPs is recorded and an ensemble average waveform (EA) is generated by averaging a subset of the ERs.
The EA is denoised to generate a denoised EA using wavelet-transform processing. The denoising includes hierarchically decomposing the EA using a series of filter banks derived from a mother wavelet derived from an electrophysiological response of interest, applying a first wavelet transform and iterating the hierarchical decomposition to filter high-frequency noise from the EA. A dynamic coefficient threshold is applied to the filter coefficients, and a second wavelet transform includes an inverse wavelet transform using the same mother wavelet to reconstruct a denoised signal.
The denoised EA is compared to a previously denoised EA to determine whether a change has occurred in the denoised EA relative to the previously denoised EA. Based on the determination that a change has occurred, the system generates an alert. The document also describes that noise can cause false alerts and that the disclosed denoising improves waveform smoothness while preserving morphology, supporting automated detection of waveform changes by reducing confounding noise.
Claims Coverage
The document provides two independent claims. Across these claims, there are 3 core inventive features: (1) EP acquisition with ensemble averaging, (2) hierarchical wavelet denoising using a mother-wavelet-derived filter-bank decomposition and dynamic thresholding with inverse reconstruction, and (3) change detection by comparing the denoised EA to a previously denoised EA and generating an alert when a change is determined.
Ensemble average evoked potential monitoring from stimulation and recorded responses
Delivering stimulation signals with electrical pulses via electrodes placed over a nerve pathway to generate a plurality of resultant evoked potentials (EPs) based on electrophysiological responses (ERs), recording the plurality of resultant EPs, and generating an ensemble average waveform (EA) by averaging a subset of the ERs.
Hierarchical wavelet-transform denoising of the ensemble average using mother-wavelet-derived filter banks and dynamic thresholding
Denoising the EA to generate a denoised EA by hierarchically decomposing the EA using a series of filter banks derived from a mother wavelet derived from an electrophysiological response of interest, applying a first wavelet transform, iterating the hierarchical decomposition to filter high-frequency noise from the EA, applying a dynamic coefficient threshold to the filter coefficients determined with the EA, and applying a second wavelet transform comprising an inverse wavelet transform using the mother wavelet derived from the electrophysiological response of interest.
Change detection on denoised ensemble averages and alert generation
Comparing the denoised EA to a previously denoised EA, determining whether a change has occurred in the denoised EA relative to the previously denoised EA, and generating, based on the determination that the change has occurred, an alert.
The independent claims cover both a medical method and an automated EP analysis system that generate an ensemble average waveform from stimulation-evoked electrophysiological responses, denoise that EA through hierarchical wavelet-transform processing with filter banks derived from a mother wavelet and dynamic coefficient thresholding followed by inverse reconstruction, and detect changes by comparing the denoised EA to a previously denoised EA and generating an alert when change is determined.
Stated Advantages
Improves automated detection of waveform changes by reducing confounding noise.
Reduces noise-induced false alerts.
Improves waveform smoothness while preserving morphology.
Documented Applications
Evoked potential (EP) analysis apparatus/architecture for monitoring and automated detection, including graphical alerting and alert generation based on detected changes.
Integration with alerting/analysis workflows, including sliding-window trend analysis [procedural detail omitted for safety].
Attribution support by obtaining information from an anesthesia or blood pressure machine and determining whether changes in EPs result from anesthesia or blood pressure changes.
Surgical-environment workflow integration by transmitting information so other devices can identify changes between the denoised EA and the previously denoised EA.
Interested in licensing this patent?