The hypothesis was tested that the field of a premature (S2) stimulus, interacting with relatively refractory tissue, can create unidirectional block and reentry in the absence of nonuniform dispersion of recovery. Simultaneous recordings from a small region of normal right ventricular (RV) myocardium were made from 117 to 120 transmural or epicardial electrodes in 14 dogs. SI pacing from a row of electrodes on one side of the mapped area generated parallel activation isochrones followed by uniform parallel isorecovery lines. Cathodal S2 shocks of 25 to 250 V lasting 3 ms were delivered from a mesh electrode along one side of the mapped area to scan the recovery period, creating isogradient electric field lines perpendicular to the isorecovery lines. Circus reentry was created following S2 stimulation; initial conduction was distant from the S2 site and spread towards more refractory tissue. Reentry was clockwise for right S1 (near the septum) with top S2 (near the pulmonary valve) and for left SI with bottom S2; and counterclockwise for right S1 with bottom S2 and left SI with top S2. The center of the reentrant circuit for all S2 voltages and coupling intervals occurred at potential gradients of 5.1±0.6 V/cm (mean±standard deviation) and at preshock intervals 1±3 ms longer than refractory periods determined locally for a 2 mA stimulus. Thus, when S2 field strengths and tissue refractoriness are uniformally dispersed at an angle to each other, circus reentry occurs around a "critical point" where an S2 field of -

Real time spike detection is an important requirement for developing brain machine interfaces (BMIs). We examined three classes of spike-detection algorithms to determine which is best suited for a wireless BMI with a limited transmission bandwidth and computational capabilities. The algorithms were analyzed by tabulating true and false detections when applied to a set of realistic artificial neural signals with known spike times and varying signal to noise ratios. A design-specific cost function was developed to score the relative merits of each detector; correct detections increased the score, while false detections and computational burden reduced it. Test signals both with and without overlapping action potentials were considered. We also investigated the utility of rejecting spikes that violate a minimum refractory period by occurring within a fixed time window after the preceding threshold crossing. Our results indicate that the cost-function scores for the absolute value operator were comparable to those for more elaborate nonlinear energy operator based detectors. The absolute value operator scores were enhanced when the refractory period check was used. Matched-filter-based detectors scored poorly due to their relatively large computational requirements that would be difficult to implement in a real-time system.

To determine the mechanism of ventricular vulnerability to electrical stimulation, we simultaneously recorded from 120 transmural electrodes in a 35 X 20 X 5-mm portion of right ventricular infundibulum in seven dogs. Baseline pacing (S1) was performed from outside the mapped region followed by single premature stimulation (S2) of increasing strength at the center of the mapped region. In five of six episodes of ventricular fibrillation and 26 of 30 episodes of repetitive responses, complete reentrant pathways were observed. Earliest activation following S2 was not at the site of S2 stimulation but was at a point between the S1 and S2 sites of stimulation. Activation spread away from the early site toward the opposite side of the mapped region around the sides of an arc of block near the S2 site to form a "figure-of-eight." The activation fronts coalesced to activate the region around the S2 site last and, if the difference in times between activation at the early site and near the S2 site was large, reentered the tissue toward the S1 site. Ventricular refractory periods were determined in four dogs following S1 pacing; the regions with the greatest nonuniformity in the dispersion of refractoriness were not the regions of unidirectional block after S2 stimulation. Thus, 1) ventricular fibrillation and repetitive responses induced electrically with S1 and S2 stimuli at different ventricular sites arise by figure-of-eight reentry, 2) this reentry is caused by the ability of S2 stimulation both to prolong refractoriness near the S2 site and to initiate a propagated response in the region between the S1 and S2 sites, and 3) a nonuniform dispersion of refractoriness is not crucial for the electrical induction of reentry leading to ventricular fibrillation or repetitive responses when S1 and S2 stimuli are given at different locations on the right ventricular outflow tract.

Acoustic radiation force impulse (ARFI) imaging has been demonstrated to be capable of visualizing variations in local stiffness within soft tissue. Recent advances in ARFI beam sequencing and parallel imaging have shortened acquisition times and lessened transducer heating to a point where ARFI acquisitions can be executed at high frame rates on commercially available diagnostic scanners. In vivo ARFI images were acquired with a linear array placed on an exposed canine heart. The electrocardiogram (ECG) was also recorded. When co-registered with the ECG, ARFI displacement images of the heart reflect the expected myocardial stiffness changes during the cardiac cycle. A radiofrequency ablation was performed on the epicardial surface of the left ventricular free wall, creating a small lesion that did not vary in stiffness during a heartbeat, though continued to move with the rest of the heart. ARFI images showed a hemispherical, stiffer region at the ablation site whose displacement magnitude and temporal variation through the cardiac cycle were less than the surrounding untreated myocardium. Sequences with radiation force pulse amplitudes set to zero were acquired to measure potential cardiac motion artifacts within the ARFI images. The results show promise for real-time cardiac ARFI imaging.

To test the hypothesis that a defibrillation shock is unsuccessful because it fails to annihilate activation fronts within a critical mass of myocardium, we recorded epicardial and transmural activation in 11 open-chest dogs during electrically induced ventricular fibrillation (VF). Shocks of 1-30 J were delivered through defibrillation electrodes on the left ventricular apex and right atrium. Simultaneous recordings were made from septal, intramural, and epicardial electrodes in various combinations. Immediately after all 104 unsuccessful and 116 successful defibrillation shocks, an isoelectric interval much longer than that observed during preshock VF occurred. During this time no epicardial, septal, or intramural activations were observed. This isoelectric window averaged 64±22 ms after unsuccessful defibrillation and 339±292 ms after successful defibrillation (P < 0.02). After the isoelectric window of unsuccessful shocks, earliest activation was recorded from the base of the ventricles, which was the area farthest from the apical defibrillation electrode. Activation was synchronized for one or two cycles following unsuccessful shocks, after which VF regenerated.Thus, (a) after both successful and unsuccessful defibrillation with epicardial shocks of 21 J, an isoelectric window occurs during which no activation fronts are present; (b) the postshock isoelectric window is shorter for unsuccessful than for successful defibrillation; (c) unsuccessful shocks transiently synchronize activation before fibrillation regenerates; (d) activation leading to the regeneration of VF after the isoelectric window for unsuccessful shocks originates in areas away from the defibrillation electrodes. The isoelectric window does not support the hypothesis that defibrillation fails solely because activation fronts are not halted within a critical mass of myocardium. Rather, unsuccessful epicardial shocks of 21 J halt all activation fronts after which VF regenerates.

A novel three-dimensional finite element model for the study of radiofrequency ablation is presented. The model was used to perform an analysis of the temperature distribution in a tissue block heated by RF energy and cooled by blood (fluid) flow. This work extends earlier models by including true flow in place of a convective boundary condition to simulate realistic experimental conditions and to improve the prediction of blood temperatures. The effect of fluid flow on the temperature distribution, the lesion dimensions, and the ablation efficiency was studied. Three flow velocities were simulated: (i) 30, (ii) 55, and (iii) 85 mm/s. The modeling results were validated qualitatively and quantitatively with in vitro data. The correlation coefficients between the modeling and the experimental temperature measurements were 0.98, 0.97, and 0.95 for flows (i)-(iii), respectively. The slopes were 0.89, 0.95, and 1.06, and the mean root mean square differences between modeling and experimental temperature measurements were 17.3% +/- 11.6%, 15.8% +/- 13.4%, and 18.8% +/- 14.9% for flows (i)-(iii), respectively. A comparison of temperature distribution obtained with a convective boundary versus inclusion of fluid motion showed that the convective boundary resulted in a similar tissue temperature distribution, but overestimated fluid temperatures and lacked the flow asymmetry seen in the true flow model.

No abstract

Epicardial and endocardial pacing are widely used, yet little is known about the three-dimensional distribution of potentials generated by the pacing stimulus or the spread of activation from these pacing sites. In six open-chest dogs, simultaneous recordings were made from 120 transmural electrodes in 40 plunge electrodes within a 35 X 20 X 5-mm portion of the right ventricular outflow tract during epicardial and endocardial pacing at a strength of twice diastolic threshold and at 1 mA. The magnitude of extracellular potentials generated by the stimulus and the activation times were compared in regions proximal (less than 10-12 mm) and distal to the pacing site. Local fiber orientation was histologically determined at each recording electrode. For endocardial pacing, endocardial potentials were larger than epicardial potentials only in the proximal region (p less than 0.001); while in the distal region, epicardial potentials were larger (p less than 0.001), and endocardial activation occurred earlier than epicardial activation for both regions (p less than 0.001). For epicardial pacing, epicardial potentials were larger than endocardial potentials in both regions (p less than 0.001), and epicardial activation occurred earlier only in the proximal region (p less than 0.02), while endocardial activation occurred before epicardial activation in the distal region (p less than 0.01). In planes of recording electrodes parallel to the epicardium and endocardium, the initial isochrones were elliptical with the major axes of the ellipses along the mean fiber orientation between the pacing site and recording plane rather than along the local fiber orientation in the recording plane. Thus, the ellipses in each plane rotated with respect to each other so that in three dimensions the activation front was helicoid, yet the twist of the helix was less than that of the corresponding transmural rotation of fibers. For pacing from the right ventricular outflow tract, we conclude that beyond 10-12 mm from endocardial and epicardial pacing sites epicardial stimulus potentials in both cases are larger than endocardial potentials because of resistivity differences inside and outside the heart wall and activation in both cases is primarily endocardial to epicardial because of rapid endocardial conduction, and we conclude that the initial spread of activation is helicoid and determined by transmural fiber direction.