Signal averaging

• US 9,504,427 B2
• Filed: 05/04/2012
• Issued: 11/29/2016
• Est. Priority Date: 05/04/2011
• Status: Active Grant

• Current Assignee: CardioInsight Technologies Inc. (Medtronic Plc)
• Original Assignee: CardioInsight Technologies Inc. (Medtronic Plc)
• Inventors: George, Brian P., Dubois, Remi, Ramanathan, Charulatha, Wodlinger, Harold
• Primary Examiner(s): Dejong, Eric S
• Application Number:
  US14/115,235
• Publication Number:
  US 20140067279A1
• Time in Patent Office: 1,670 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: A61B 5/282 Holders for multiple electrodes for introduction into the body A61B5/287
• CPC Class Codes: A61B 5/287 Holders for multiple electrodes, e.g. electrode catheters for electrophysiological study [EPS]
• CPC Class Codes: A61B 5/327 Generation of artificial ECG signals based on measured signals, e.g. to compensate for missing leads
• CPC Class Codes: A61B 5/35 by template matching
• CPC Class Codes: A61B 5/7203 for noise prevention, reduction or removal
• CPC Class Codes: A61B 5/7221 Determining signal validity, reliability or quality A61B5/7207 and A61B5/7217 take precedence
• CPC Class Codes: A61B 5/7235 Details of waveform analysis A61B5/349 takes precedence
• CPC Class Codes: A61B 5/7278 Artificial waveform generation or derivation, e.g. synthesising signals from measured signals A61B5/327 takes precedence
• CPC Class Codes: G06F 2218/04 Denoising
• CPC Class Codes: G06F 2218/14 by matching peak patterns

Abstract

A method (500) can comprise performing principal component analysis (PCA) on data corresponding to a subset of a plurality of signals and a selected template to generate a virtual lead and an optimized template (530). The method can also comprise calculating a cross correlation on the virtual lead and the optimized template to determine a strength of linear dependence between the virtual lead and the optimized template to determine regions of interest (ROIs) of the virtual lead (540). The method can further comprise detecting peak correlation coefficients in the virtual lead (550). The method can still further comprise comparing the amplitude of each of the ROIs of the virtual lead with the selected template to determine an error between the template and each ROI of the virtual lead (560). The method can yet further comprise averaging the ROIs to generate averaged data (570).

Claims

1. A method comprising:

receiving measurement data characterizing a subset of a plurality of electrical biological signals detected by sensors;

selecting a template region of the signals characterized by the measurement data;

performing principal component analysis (PCA) on the measurement data to generate a virtual lead corresponding to a single channel representation of the subset of signals characterized by the measurement data and performing PCA on the selected template region to generate an optimized template corresponding to a single channel representation of the selected template region;

calculating a cross correlation on the virtual lead and the optimized template to determine a strength of linear dependence between the virtual lead and the optimized template to determine regions of interest (ROIs) of the virtual lead;

detecting peak correlation coefficients in the virtual lead based on the cross-correlation to identify potential ROIs;

comparing the amplitude of each of the ROIs of the virtual lead with the optimized template to determine an error between the optimized template and each ROI of the virtual lead for each of the potential ROIs; and

generating averaged data for the each of the plurality of signals by determining an average for each respective ROI of the virtual lead based on the error.

2. The method of claim 1, further comprising:

detecting regions of poor signal quality for each of the plurality of signals; and

determining which of the plurality of signals is associated with a bad channel based on the detected regions of poor signal quality, each bad channels being excluded from the PCA.

3. The method of claim 1, further comprising mapping the averaged data to a surface of an organ to reconstruct electrical activity of the organ.

4. The method of claim 3, further comprising generating one of potential and isochrone maps based on the mapped averaged data.

5. The method of claim 1, wherein the plurality of signals comprises at least 200 signals acquired concurrently from a body surface of a patient, the averaged data including signal averaged data computed for a subset of the at least 200 signals.

6. The method of claim 5, wherein each of the plurality of signals represents an electrocardiograph (ECG) signal.

7. The method of claim 5, wherein the template defines a P wave.

8. The method of claim 1, wherein the plurality of signals are acquired non-invasively from a patient.

9. The method of claim 1, wherein the plurality of signals are acquired invasively from a patient.

10. The method of claim 1, wherein PCA is performed on data corresponding to a sensing zone that comprises a proper subset of the plurality of signals, the averaging being performed on the plurality of signals.

11. The method of claim 1, further comprising detecting and removing correlated noise from the subset of the plurality of signals.

12. A non-transitory machine readable medium for storing machine readable instructions, the machine readable instructions comprising:

a signal averager that receives measurement data that characterizes a plurality of concurrently detected biological signals, the signal averager comprising;

a principal component analysis (PCA) function that performs PCA on the measurement data for a subset of the biological signals to generate a virtual lead corresponding to a single channel representation of the subset of the biological signals and performs PCA on a selected template region of the plurality of biological signals to generate an optimized template corresponding to a single channel representation of the selected template region;

a cross correlation (CC) calculator that determines a strength of linear dependence between the virtual lead and the optimized template to determine potential regions of interest (ROIs) of the virtual lead;

an amplitude comparator that compares the amplitude of each of the potential ROIs of the virtual lead with the optimized template to determine an error between the optimized template and each ROI of the virtual lead; and

a segment averaging function that averages each of the ROIs based on the error between the optimized template and each ROI to calculate averaged data for each of the biological signals.

13. The non-transitory machine readable medium of claim 12, wherein the machine readable instructions further comprise a mapping system comprising:

a reconstruction component to convert the averaged data to corresponding reconstructed electrical activity of an organ of a patient; and

a map generator to produce map data based on the reconstructed electrical activity of the organ of the patient.

14. The non-transitory machine readable medium of claim 12, wherein the signal averager further comprises a zone selector to select a sensing zone comprising a proper subset of the biological signals, wherein the PCA function performs the PCA on the sensing zone.

15. The non-transitory machine readable medium of claim 12, wherein the signal averager further comprises a peak correlation coefficient detector to determine maximum correlation coefficients for the virtual lead.

16. The non-transitory machine readable medium of claim 12, wherein the machine readable instructions further comprise a bad channel detector to:

detect regions of poor signal quality for each of the plurality of biological signals; and

determine which of the plurality of biological signals is associated with a bad channel based on the detected regions of poor signal quality, each bad channel being identified for exclusion from the PCA function.

17. A system comprising:

a memory for storing machine readable instructions; and

a processing unit to access the memory and execute the machine readable instructions, the machine readable instructions comprising;

a signal averager to;

detect regions of poor signal quality for each of a plurality of signals, wherein each of the plurality of signals corresponds to an electrical biological signal;

determine which of the plurality of signals is associated with a bad channel based on the detected regions of poor signal quality; and

provide averaged data corresponding to an average of regions of interest (ROIs) that are determined based on a strength of linear dependence between a virtual lead and an optimized template,

wherein the virtual lead and the optimized template are determined based on the plurality of signals that are not associated with the bad channel and based on a selected template region of the plurality of signals, and

wherein the virtual lead corresponds to a single channel representation of the subset of the biological signals and the optimized template corresponds to a single channel representation of the selected template region.

18. The system of claim 17, further comprising a sensor array mounted on a substrate, the sensor array comprising a plurality of sensors, the sensor array providing the plurality of signals.

19. The system of claim 18, further comprising a graphical user interface to provide study data to the signal averager in response to user input, the study data identifying a sensing zone that comprises a proper subset of the plurality of sensors in the senor array, wherein the signal averager provides the averaged data based on signals of the plurality of signals that are provided from the proper subset of the plurality of sensors in the sensing zone.

20. The system of claim 17, further comprising a graphical user interface that provides an interface for selecting the template in response to user input.