Centralised interpretation of electrocardiograms.

A system was devised so that a peripheral hospital could transmit electrocardiograms (ECGs) to a central computer for interpretation. The link that transmits both ECGs and reports is provided by the telephone network. Initial results showed that telephone transmission did not significantly affect the accuracy of the ECG interpretation. The centralised computer programme could be much more widely used to provide ECG interpretations. A telephone link would not be justified in health centres, where the demand for ECGs is fairly small, but ECGs recorded at a health centre can be sent to the computer for interpretation and returned the next day. The most cost-effective method of providing computer interpretation for several health centres in a large city would be to have a portable electrocardiograph and transmission facilities, which could be moved from centre to centre.     

Ventricular beat detection in single channel electrocardiograms

Background  

Detection of QRS complexes and other types of ventricular beats is a basic component of ECG analysis. Many algorithms have been proposed and used because of the waves' shape diversity. Detection in a single channel ECG is important for several applications, such as in defibrillators and specialized monitors.     

Frequency modulation within electrocardiograms during ventricular fibrillation

Periods of reentrant activation and effective refractory periods are correlated with dominant frequency or reciprocal of cycle periods during ventricular fibrillation (VF). In the present study, we used an analysis technique based on Wigner transforms to quantify time-varying dominant frequencies in electrocardiograms (ECGs) during VF. We estimated dominant frequencies within orthogonal ECGs recorded in 10 dogs during trials of 10 s of VF and in 9 dogs during trials of 30 s of VF. In four additional dogs, we compared dominant frequencies during 10 s of VF before and after administration of amiodarone. Our results showed the following. 1) There was substantial frequency variation or modulation within the ECGs during 10 and 30 s of VF, the average variation being +/-15% from the mean frequency. Amiodarone decreased mean frequencies (P < 0.05) as expected; however, amiodarone also decreased the variation in frequencies (P < 0.05). 2) During 30 s of VF, the dominant frequencies increased continuously from 7.3 to 8.1 Hz (P < 0.05). The increase in frequency was almost linear with a rate of 0.022 Hz/s (r(2) = 0.93, P < 0.0005). 3) Modulation of frequencies during the first and the last one-half of 30 s of VF was not different. Average (in time) mean frequencies and modulation of frequencies were similar in all three ECGs. 4) Although the averages were similar, during any VF episode, dominant frequencies in ECGs recorded from different locations on the body surface were similar to each other at some times and markedly different from each other at other times. We conclude that during VF, 1) frequencies in ECGs vary considerably and continuously, and amiodarone decreases this variation; 2) mean frequencies increase linearly during first 30 s; 3) the variability in frequency does not change during 30 s; and 4) at any given time, the frequencies within spatially different body surface ECGs can be either similar or markedly different.     

Assessment of the technical quality of electrocardiograms.

The technical quality of 600 electrocardiograms (ECG''s) was assessed for missing leads and clipping, and graded from 1 to 5 for each of noise, lead drift and beat-to-beat drift. Three subgroups of 200 ECGs each were studied: group A, those obtained by emergency department staff (non-technicians); group B, records obtained by ECG technicians; and group C, telephone-transmitted records obtained by technicians performing all the laboratory work at a smaller, outlying hospital. Records with missing leads, clipping, grade 4 or 5 noise, grade 5 lead drift or grade 5 beat-to-beat drift were classified as unsatisfactory or rejected. With these stringent criteria the rejection rate was 71.0% for group A records, 58.5% for group B and 44.5% for group C. The proportions of records with peak quality (no missing leads or clipping, and grade 1 noise, lead drift or beat-to-beat drift) were 4.5% for group A, 5.5% for group B and 23.0% for group C. Suggested revisions in the grading of technical quality of ECGs are presented.     

Detection of somatic muscle fasciculation on electrocardiograms.

Small discrete spikes appearing on the electrocardiogram (ECG) of a man with a pacemaker turned out to be fasciculation potentials picked up from leg muscles. To find out how common these spikes were all routine ECGs made in one month were reviewed. Spike potentials were found in six patients, and in each case they originated from somatic musculature. Spikes could nearly always be detected on ECGs that were recorded subsequently in patients known to have lower motor neurone lesions that produced fasciculation. Thus the presence of these fasciculation spikes may suggest underlying neuromuscular disease; this possibility should be recognised by those who record ECGs so that the spikes are not suppressed by overuse of the filter.     

Method for non-invasively recording electrocardiograms in conscious mice

Background  

The rapid increase in the development of mouse models is resulting in a growing demand for non-invasive physiological monitoring of large quantities of mice. Accordingly, we developed a new system for recording electrocardiograms (ECGs) in conscious mice without anesthesia or implants, and created Internet-accessible software for analyzing murine ECG signals. The system includes paw-sized conductive electrodes embedded in a platform configured to record ECGs when 3 single electrodes contact 3 paws.     

A beat-by-beat analysis of electrocardiograms from cardiac transplant recipients

A software system is described for producing a beat-by-beat analysis of the electrocardiograms from patients after heart transplantation. Pacemaker spikes are automatically detected and eliminated from the signals. R waves are located by a robust and accurate two-step algorithm. Based on the variable length of single heart beats, the Fourier coefficients of three orthogonal surface leads and of two intracardiac leads are calculated beat-by-beat. Power spectra are then obtained by combining contributions from the variable fundamental frequency and its multiples (harmonics) into fixed frequency classes of 1 Hz width and averaging over 60–120 cardiac cycles. Additionally, averaged beat-by-beat power spectra are calculated for windowed QRS complexes and T waves. As by-products, single beat quantities, such as R-R and R-T intervals, averaged signals for all leads, and the orientational autocorrelation function of the electrical vector of the heart, are obtained. Following beat-by-beat evaluation, mean values and standard deviations are obtained for all quantities.