pecg.ecg package

pecg.ecg.FiducialPoints

class pecg.ecg.FiducialPoints.FiducialPoints(signal: numpy.array, fs: int)

Bases: object

The purpose of the FiducialPoints class is to calculate the fiducial points.

Parameters:

• signal – the ECG signal as a ndarray, with shape (L, N) when L is the number of channels or leads and N is the number of samples.

• fs – The sampling frequency of the signal.[Hz]

import pecg
from pecg.Example import load_example
from pecg.ecg import FiducialPoints as Fp

signal, fs = load_example(ecg_type='Holter')
fp = Fp.FiducialPoints(signal, fs)

wavedet(matlab_pat: str, peaks: numpy.array = array([], dtype=float64))

The wavedet function uses the matlab algorithm wavedet which was compiled for Windows OS for its usage in python. The algorithm is described in the the work of Martinez et al. [1]. The function is calculating the fiducial points of the ECG time series using the wavelet transform.

[1]

Martínez, Juan Pablo, Rute Almeida, Salvador Olmos, Ana Paula Rocha, and Pablo Laguna. “A wavelet-based ECG delineator: evaluation on standard databases.” IEEE Transactions on biomedical engineering 51, no. 4 (2004): 570-581.

Parameters:

• matlab_pat – path to matlab runtime 2021a directory

• peaks – Optional input- Annotation of the reference peak detector (Indices of the peaks), as an ndarray of shape (L,N), when L is the number of channels or leads and N is the number of peaks. If peaks are not provided they are calculated using the jqrs detector.

Returns:

fiducials: Nested dictionary of leads - For every lead there is a dictionary that includes indexes for each one of nine fiducials points.

matlab_pat = '/usr/local/MATLAB/R2021a'
fiducials = fp.wavedet(matlab_pat)

epltd()

This function calculates the indexes of the R-peaks with epltd peak detector algorithm. This algorithm were introduced by [2].

[2] (1,2)

Pan, Jiapu, and Willis J. Tompkins. “A real-time QRS detection algorithm.” IEEE Trans. Biomed. Eng 32.3 (1985): 230-236.

Returns:

indexes of the R-peaks in the ECG signal, as an ndarray of shape (L,N), when L is the number of channels or leads and N is the number of peaks.

peaks = fp.epltd()

xqrs()

This function wraps the XQRS function of the WFDB package.

Returns:

indexes of the R-peaks in the ECG signal, as an ndarray of shape (L,N), when L is the number of channels or leads and N is the number of peaks.

peaks = fp.xqrs()

jqrs(thr: float = 0.8, rp: float = 0.25)

The function is an Implementation of an energy based qrs detector [3]. The algorithm is an adaptation of the popular Pan & Tompkins algorithm [2]. The function assumes the input ecg is already pre-filtered i.e. bandpass filtered and that the power-line interference was removed. Of note, NaN should be represented by the value -32768 in the ecg (WFDB standard).

[3]

Behar, Joachim, Alistair Johnson, Gari D. Clifford, and Julien Oster. “A comparison of single channel fetal ECG extraction methods.” Annals of biomedical engineering 42, no. 6 (2014): 1340-1353.

Parameters:

• thr – threshold, default value is 0.8.

• rp – refractory period (sec), default value is 0.25.

Returns:

indexes of the R-peaks in the ECG signal, as an ndarray of shape (L,N), when L is the number of channels or leads and N is the number of peaks.

peaks = fp.jqrs()

pecg.ecg.Biomarkers

class pecg.ecg.Biomarkers.Biomarkers(signal: numpy.array, fs: int, fiducials: dict)

Bases: object

The purpose of the Biomarkers class is to calculate the biomarkers, we divided the morphological biomarkers into two main groups: intervals and waves.

Parameters:

• signal – The ECG signal as a ndarray.

• fs – The sampling frequency of the signal [Hz].

• fiducials – Nested dictionary of leads - For every lead there is a dictionary that includes indexes for for each one of nine fiducials points. this nested dictionary can be calculated using the FiducialPoints module.

import pecg 
from pecg.ecg import Biomarkers as Bm
from pecg.ecg import FiducialPoints as Fp
from pecg.Example import load_example

signal, fs = load_example(ecg_type='12-lead')
fp = Fp.FiducialPoints(signal, fs)
matlab_pat = '/usr/local/MATLAB/R2021a'
fiducials = fp.wavedet(matlab_pat)
bm = Bm.Biomarkers(signal, fs, fiducials)
ints, stat_i = bm.intervals()
waves, stat_w = bm.waves()

intervals()

Returns:

• intervals_b: Dictionary that includes all the raw data, for the Intervals and segments biomarkers.

• intervals_statistics: Dictionary that includes the mean, median, min, max, iqr and std, for every Intervals and segments biomarker.

Intervals and segments:

Biomarker	Description
P-waveint	Time interval between P-on and P-off.
PRint	Time interval between the P-on to the QRS-on.
PRseg	Time interval between the P-off to the QRS-on.
PRint2	Time interval between P-peak and R-peak as defined by Mao et al.
QRSint	Time interval between the QRS-on to the QRS-off.
QTint	Time interval between the QRS-on to the T-off.
QTcBint	Corrected QT interval (QTc) using Bazett’s formula.
QTcFriint	QTc using the Fridericia formula.
QTcFraint	QTc using the Framingham formula.
QTcHint	QTc using the Hodges formula.
T-waveint	Time interval between T-on and T-off.
TPseg	Time interval between T-off and P-on.
RRint	Time interval between sequential R-peaks.
Rdep	Time interval betweem Q-on and R-peak.

waves()

Returns:

• waves_b: Dictionary that includes all the raw data, for every Wave characteristic biomarker.

• wave_statistics: Dictionary that includes the mean, median, min, max, iqr and std, for every Wave characteristic biomarker.

Waves:

Biomarker	Description
P-wave	Amplitude difference between P-peak and P-off.
T-wave	Amplitude difference between T-peak on and T-off.
R-wave:	R-peak amplitude.
P-waveArea	P-wave interval area defined as integral from the P-on to the P-off.
T-waveArea	T-wave interval area defined as integral from the T-on to the T-off.
QRSArea	QRS interval area defined as integral from the QRS-on to the QRS-off.
STseg	Amplitude difference between QRS-off and T-on.
J-point	Amplitude in 40ms after QRS-off as defined by Hollander et al.

Module contents