1. Explain how you calculated the highest frequency component of your subject’s
ID: 3507787 • Letter: 1
Question
1. Explain how you calculated the highest frequency component of your subject’s ECG. What was the frequency and, using this information, what sampling rate would you use to adequately capture a good representation of the ECG when recording?
2. The ECG waveform is not symmetrical unlike the differential recording example provided in lecture (with a bipolar waveform - equal deflections in positive and negative directions). What reasons can you think of for this observation?
3. When you switched to recording the EMG signals you still have two electrodes and a ground just as you did with the ECG. Explain why you think you do not see an ECG signal.
4. The position of electrodes often influences the appearance of the recorded signal (it does not change the actual signal). In the EMG experiment, explain why the signals are different when the electrodes are positioned along the muscle versus across the muscle.
5. If a signal that you were interested in recording had a frequency of 60Hz and background noise was at a similar frequency (50-60Hz) would it be possible to eliminate the noise by using filters? Explain your answer.
Explanation / Answer
1. The QRS complex of the ECG is important information in heart-rate monitoring and cardiac disease diagnosis. Before applying ECG signals, all ECG signals are filtered to produce the baseline corrected ECG signal. The R-waves are detected by a peak detection algorithm, which begins by scanning for local maxima in the absolute value of ECG data. For certain window durations, the searching continues to look for a larger value. If this search finishes without finding a larger maximum, the current maximum is assigned as the R peak. Centered on the detected R peak, the QRS complex portion is extracted by applying a window of 280 ms, and P-wave and T-wave are removed by this window duration. Based on a 360 sampling rate, 100 samples can be acquired around the R peak (Sampling point P=100, 50 points before and 50 points after). After sampling and analogue-to-digital conversion, individual QRS complex is extracted. Then, frequency spectrum of each QRS complex is computed by Eqs. (1) and (2).
The spectrum varies with different cardiac arrhythmias, and power spectra are observed in the frequency range from 1 to 20 Hz. The spectra are plotted and analysed, and all amplitudes are normalized with maximum amplitude. The amplitudes decrease as the frequency increases, and rapidly vanishes above 12 Hz. Frequency components from 1 to 12 Hz (n=12) are selected for multiple ECG beat recognition. These spectra are not disturbed by high-frequency components above 20 Hz such as power-line interference (50 Hz/60 Hz) and muscle noise, and very low-frequency components (<1 Hz) such as baseline drift and breath. Therefore, power line noise, very low-frequency and high-frequency components are excluded without affecting the frequency-domain features.
For certain window duration, each QRS complex is extracted as VQRS=[v1,v2,v3,…,vp,…,vP], P is the number of sampled points, p=1,2,3,…,P. Frequency spectra are computed by the function fft
(1) X=[x1x2xixn]=fft(VQRS).
The DFT is found by taking the n-point FFT. The FFT returns a two-sided spectrum in complex form (Real and Imaginary Parts), which can scale and convert to polar form to obtain amplitude and phase. The amplitude of FFT is related to the number of points in the time-domain signal. If X is complex, compute the amplitude of the FFT of a sequence by the function abs(•)
(2) A=[a1a2aian]=abs(X)max[abs(X)].
In this study, the dataset of QRS complexes for six typical heartbeat classes are taken from the MIT–BIH arrhythmias database. The database contains 48 records, and each record is slightly over 30 min long. In most records, the upper signal is a modified limb lead II (ML II) and the lower signal is a modified lead V1 (VI). Six classes have been included in the investigations, involving Nor, VEB, SEB, BBEB, UB, and FB. From these records (ML II Signal), a total of 50 QRS complexes (K=50) are selected including patient numbers 103, 107, 109, 111, 118, 119, 124, 200, 202, 208, 209, 212, 214, 217, 221, 231, 232, and 233, and classified into six types:
Nor: Normal Beat (•), weighted factors could be encoded as [1 0 0 0 0 0];
VEB: Premature Ventricular Contraction (V), weighted factors could be encoded as [0 1 0 0 0 0];
SEB: Atrium Premature Beat (A), weighted factors could be encoded as [0 0 1 0 0 0];
BBEB: Right and Left Bundle Branch Block Beat (R/L), weighted factors could be encoded as [0 0 0 1 0 0];
UB: paced beat (P), weighted factors could be encoded as [0 0 0 0 1 0];
FB: Fusion of Paced/Ventricular and Normal Beat (f/F), weighted factors could be encoded as [0 0 0 0 0 1];
The weighted factors wkj, k=1,2,3,…,K, j=1,2,3,…,m, are encoded as binary values with signal “1” for belonging to Classj. FFT are applied to ECG signals for power spectrum estimation to construct various patterns. The frequency-domain features of six classes are produced for further analysis. To quantify the differences among various classes, the comparative sequences for each class are created as A(k)=[a1(k),a2(k),…,ai(k),…,an(k)], i=1,2,3,…,n. Frequency-based features will be quantified and used to classify cardiac arrhythmias. The numbers of averaged patterns from the same class are 8-, 13-, 2-, 15-, 6-, and 6-set data respectively.
Related Questions
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.