Evaluation and optimization of novel extraction algorithms for the automatic detection of atrial activations recorded within the pulmonary veins during atrial fibrillation

Background and objective

The automated detection of atrial activations (AAs) recorded from intracardiac electrograms (IEGMs) during atrial fibrillation (AF) is challenging considering their various amplitudes, morphologies and cycle length. Activation time estimation is further complicated by the constant changes in the IEGM active zones in complex and/or fractionated signals. We propose a new method which provides reliable automatic extraction of intracardiac AAs recorded within the pulmonary veins during AF and an accurate estimation of their local activation times.

Methods

First, two recently developed algorithms were evaluated and optimized on 118 recordings of pulmonary vein IEGM taken from 35 patients undergoing ablation of persistent AF. The adaptive mathematical morphology algorithm (AMM) uses an adaptive structuring element to extract AAs based on their morphological features. The relative-energy algorithm (Rel-En) uses short- and long-term energies to enhance and detect the AAs in the IEGM signals. Second, following the AA extraction, the signal amplitude was weighted using statistics of the AA sequences in order to reduce over- and undersensing of the algorithms. The detection capacity of our algorithms was compared with manually annotated activations and with two previously developed algorithms based on the Teager–Kaiser energy operator and the AF cycle length iteration, respectively. Finally, a method based on the barycenter was developed to reduce artificial variations in the activation annotations of complex IEGM signals.

Results

The best detection was achieved using Rel-En, yielding a false negative rate of 0.76% and a false positive rate of only 0.12% (total error rate 0.88%) against expert annotation. The post-processing further reduced the total error rate of the Rel-En algorithm by 70% (yielding to a final total error rate of 0.28%).

Conclusion

The proposed method shows reliable detection and robust temporal annotation of AAs recorded within pulmonary veins in AF. The method has low computational cost and high robustness for automatic detection of AAs, which makes it a suitable approach for online use in a procedural context.

A new method for analysis of atrial activation during chronic atrial fibrillation in man

To further clarify the mechanisms maintaining chronic atrial fibrillation (CAF), a method identifying preferable activation patterns of the atria during fibrillation, by time averaging of multiple discrete excitation vectors, was developed. Repeated recordings, each of 56 atrial bipolar electrograms simultaneously acquired during 8 s, were made at multiple sites in the right atrial free wall and the left atrial appendage in 16 patients with CAF using a 2.17 x 3.54 cm electrode array. The local activation times (LAT's) in each recording were estimated as the median activation time at the respective measurement point. By calculating the time difference between the LAT's at adjacent measurement points in two spatial dimensions, a direction vector was created for each activation wave passing each set of measurement points, a total of 42 sets. By time averaging of the individual direction vectors (typically n = 55) at each set of measurement points, preferable activation patterns were determined. Three types of activation patterns were found: 1) inconsistent activation (n = 5), 2) consistent activation with preferential propagation directions (n = 7) and 3) consistent activation with impulses originating from a localizable site within the recording area (n = 4). All activation patterns were reproducible and the two latter patterns were proven significant using statistical tests. We conclude that this new method is useful in further clarification of the mechanisms involved in the maintenance of atrial fibrillation.

Activation time determination by high-resolution unipolar and bipolar extracellular electrograms in the canine heart

INTRODUCTION

To identify the optimal criteria for activation time (AT) determination of bipolar electrograms from normal hearts, a high-resolution cross electrode array comprising 128 unipolar electrodes of 500-microns spacing was used to record extracellular potentials from the left ventricular epicardium of 12 dog hearts.

METHODS AND RESULTS

Recordings were made during broad wavefront propagation (B wave) and local elliptical wavefront propagation (E wave). Characteristics of 863 bipolar electrograms (1-mm spacing) were constructed from unipolar data standardized for differences in polarity, then classified morphologically. Features for bipolar AT determination were compared to the time of the negative peak of the first temporal derivative of a unipolar electrogram situated mid-way between the bipoles. During B wave, three distinct morphologies were observed: uniphasic (61%), biphasic (23%), and triphasic (16%). Peak voltage of uniphasic and triphasic signals was the best predictor of AT (error: 0.6 +/- 0.6 msec and 0.6 +/- 0.8 msec, respectively). During E wave, parallel orientation of the bipoles with respect to the direction of impulse propagation wavefront resulted in uniphasic signals (> 99%), while for perpendicular orientation of the bipoles, electrogram morphology was variable. For parallel orientation of the bipoles, peak negative voltage was the best predictor of AT for both longitudinal and transverse propagation, while for perpendicular bipole orientation, peak negative voltage was a less reliable predictor for propagation along both fiber axes. Increasing interpolar distance resulted in a degradation in AT accuracy for B wave (from 0.6 +/- 0.6 msec at 1 mm to 1.1 +/- 1.2 msec at 7 mm) and for E wave (from 0.4 +/- 0.3 msec at 1 mm to 3.1 +/- 2.9 msec at 7 mm).

CONCLUSIONS

(1) The accuracy of bipolar electrograms is sensitive to wavefront direction, bipole orientation, and interpolar distance; (2) peak negative voltage of uniphasic and triphasic signals is a reliable predictor of AT, but only for B wave; (3) a maximum interpolar distance of 2 mm and bipole orientation parallel to the direction of the impulse wavefront are minimally required for accurate determination of AT during impulse propagation initiated near the recording electrodes; and (4) for impulses initiated near the recording site in normal tissue, a biphasic or triphasic morphology almost certainly indicates that the bipolar electrode is oriented perpendicular to the wavefront direction, irrespective of fiber orientation.

Improved spatial resolution for cardiac mapping using current source density-based electrode arrays

The time of the minimum slope (i.e., the fastest negative deflection) in monopolar (MP) electrograms from normal hearts compares closely with time of phase 0 of the action potential in cells underlying the electrode, but poor rejection of far-field activity may limit the utility of MP electrode technology in dense arrays used for the study of ventricular tachycardia and fibrillation. The purpose of this study is to evaluate more myopic discrete bipolar (BP) and nondirectional, two-dimensional current source density (CSD) based arrays for rejection of far-field potentials and precision of activation time determination. Simultaneous recordings of the CSD, MP, and multiple BP electrograms were performed on normal dog epicardium. The time of the minimum slope in MP electrograms was compared to activation times in CSD and BP derivations using: (1) peak; (2) steepest slope; (3) zero crossing of the steepest sloping segment in either direction; and (4) waveform morphology. In vivo, CSD amplitude was reduced significantly more than MP and BP amplitudes by insertion of inert media between the heart and the electrodes. The time of the steepest slope in CSD electrograms designated activation times closest to the time of the minimum slope in MP electrograms (0.9 +/- 1.3 msec). We conclude that CSD provides a nondirectional electrode system that accurately defines the time of local activation and possesses better spatial specificity than MP electrode systems and BP electrode systems having the same interelectrode distances.

Spectral analysis of activation time sequences

Adequate spatial resolution of local activation is fundamental for the correct depiction of myocardial activation during ablative treatment of ventricular tachycardia. The widest allowable distances between recording sites that provide an accurate description of the field potential distribution is dictated by the Nyquist sampling theorem. Activation times are derived from the field potentials. However, because of noise intrinsic in activation detection algorithms, closer recording sites may be required than those theoretically computed. The purpose of this study is to examine the spatial frequency spectrum of epicardial activation time sequences computed by common activation detection algorithms, determine which algorithm is least noisy, and derive the recording site density necessary to avoid distortion of the epicardial activation map. Using 40 to 80 electrode linear arrays, monopolar and bipolar electrograms from the epicardium were recorded vertically (base to apex) and horizontally in 11 dogs. Activation times for bipolar electrograms were estimated using Peak, Fastest Zero Crossing, and Morphological algorithms. Activation times for monopolar electrograms were set equal to the time of the fastest negative deflection. Activation time sequences were analyzed using conventional Fourier techniques. Anomalous activation times from serially adjacent bipolar electrograms, which constitute algorithm noise, were studied. Horizontal and vertical interelectrode distances are 3.2 mm and 2.4 mm, respectively. Of the bipolar algorithms, the Morphological algorithm produced the fewest anomalous activation times. Mapping systems having more than 256 channels are required for accurate representation of epicardial activation in a typical 20-kg dog. The endocardial electrode spacing is unknown, but is expected to be at least as dense. Large global mapping systems or regionally dense arrays may offer advantages during catheter ablations for ventricular tachycardia and for studies into the mechanisms of ventricular tachycardia by accurately defining the cardiac activation pattern.

High-density mapping of electrically induced atrial fibrillation in humans

BACKGROUND

Mapping studies in animals have suggested that atrial fibrillation (AF) is based on multiple reentering wavelets. Little information is available about the patterns of activation during AF in humans. The objective of the present study was to reconstruct and classify the patterns of human right atrial (RA) activation during electrically induced AF.

METHODS AND RESULTS

AF was induced by rapid atrial pacing in 25 patients with Wolff-Parkinson-White syndrome undergoing surgery for interruption of their accessory pathway(s). The free wall of the RA was mapped using a spoon-shaped electrode containing 244 unipolar electrodes. The activation of the RA during AF showed large interindividual differences. Based on the complexity of atrial activation, three types of AF were defined. In type I (40% of patients), single broad wave fronts propagated uniformly across the RA. Type II (32%) was characterized by one or two nonuniformly conducting wavelets, whereas in type III (28%), activation of the RA was highly fragmented and showed three or more different wavelets that frequently changed their direction of propagation as a result of numerous arcs of functional conduction block. There were significant differences (P < .05) among the three types of AF in median intervals (174 +/- 28, 150 +/- 14, and 136 +/- 16 milliseconds), variation in AF intervals (P5-95) (54 +/- 25, 94 +/- 21, and 104 +/- 22 milliseconds), incidence of electrical inactivity (42 +/- 11%, 21 +/- 4%, and 8 +/- 4%) and reentry (3 +/- 7%, 36 +/- 28%, and 99 +/- 36%), and average conduction velocity during AF (61 +/- 6, 54 +/- 4, and 38 +/- 10 cm/s).

CONCLUSIONS

During pacing-induced AF in humans, the RA is activated by one or multiple wavelets propagating in different directions. Three types of RA activation during AF were identified. From type I to type III, the frequency and irregularity of AF increased, and the incidence of continuous electrical activity and reentry became higher. These various types of AF in humans appear to be characterized by different numbers and dimensions of the intra-atrial reentrant circuits.

The epicardial field potential in dog: implications for recording site density during epicardial mapping

Investigations into mechanisms and successful surgical therapy of ventricular tachycardia (VT) depend upon accurate endocardial/epicardial mapping. Deduction of local activation is based upon parameters derived from the field potential (FP) (monopolar recording) or its first spatial derivative (bipolar recording). Adequate electrode spacing is an assumption fundamental to the mapping process, but the electrode spacing required for accurate representation of the FP is unknown. The purpose of this work is to derive the electrode spacing necessary to accurately describe the FP on the epicardium. In 11 dogs, electrograms from vertical (V) (base to apex) bands having 40 electrodes and horizontal (H) bands having 40 to 80 electrodes were sampled at 1 kHz. The spatial bandwidths (BW) were computed according to two criteria: (1) the frequency yielding 2% mean squared error (MSE) computed at the time of the greatest integrated magnitudes of the Fourier transform; and (2) the highest frequency bounding 95% power computed at each msec throughout the beat. Implied electrode spacings were defined according to the sampling theorem. The 5th percentiles of the implied electrode spacing distributions were used to define the widest interelectrode distance required to prevent spatial aliasing. H-5th percentile and V-5th percentile were, respectively: 2% MSE (3.5 mm, 2.3 mm); 95% power (3.6 mm, 2.3 mm). Thus, a typical 20-kg dog requires more than 250 recording sites for accurate epicardial mapping. Extrapolating to man, these results suggest inadequate electrode density may partially be responsible for incomplete and ambiguous reentry patterns often observed during intraoperative mapping.

Observations on the epicardial activation of the normal human heart

Serial hand mapping techniques in man have identified 3 to 5 sites of epicardial breaktrough (EBT). However, transmural epicardial excitation from the widely distributed His/Purkinje system suggests a more complicated pattern may exist. Multielectrode arrays used with large mapping systems during surgery often present complicated and sometimes inconsistent activation patterns. The purpose of this work is to reconcile epicardial activation in the normal human heart with anatomical and endocardial/intramural physiological recordings using multichannel computer mapping requiring only a single beat, and rigorously defined and applied activation time detection algorithms. Eighteen subjects undergoing surgery for Wolff-Parkinson-White syndrome were recorded with a 119 site sock array during nonpreexcited sinus rhythm. None had evidence of coronary artery disease and all exhibited a normal 12-lead ECG except during periods of preexcitation or tachycardia. Each was recorded bipolarly and four also were recorded monopolarly. Recordings revealed 8.0 +/- 1.6 EBTs (range 5 to 12). Closely spaced, multiple EBTs often were observed and usually confirmed using different activation time detection algorithms. The earliest EBT always occurred over the anterior right ventricle at 14.3 +/- 6.5 msec (range -1 to 29 msec) after QRS onset. Subsequent EBTs could occur at any ventricular site with variable latencies. In contrast to previous reports describing epicardial spread of activation from a few foci, a mosaic of epicardial activation emerges. These data are consistent with endocardially initiated transmural activation of the epicardium suggested by the anatomy of the His/Purkinje system and intramural recordings.