Michal M. Radai

Induced Current Impedance Technique for Monitoring Brain Cryosurgery in a Two-Dimensional Model of the Head

AbstractA fast and robust finite volume solver of the two-dimensional induced current electrical impedance forward problem was developed. The numerical solver was validated by comparison with an existing analytical solution for a symmetrical geometry case, showing an accuracy of 0.07%. The solver was used to theoretically examine the sensitivity of the induced current impedance technique for the medical procedure of monitoring brain cryosurgery. The simulation was performed using a two-dimensional approximation of otherwise realistic geometry model of the head with different ice-ball sizes, simulating the expansion of the frozen lesion. The sensitivity of the scalp potential to the ice-ball size was found to be 53×10-4 (relative scalp potential mm−2).

Induced current electrical impedance tomography system: experimental results and numerical simulations.

In electrical impedance tomography (EIT), measurements of developed surface potentials due to applied currents are used for the reconstruction of the conductivity distribution. Practical implementation of EIT systems is known to be problematic due to the high sensitivity to noise of such systems, leading to a poor imaging quality. In the present study, the performance of an induced current EIT (ICEIT) system, where eddy current is applied using magnetic induction, was studied by comparing the voltage measurements to simulated data, and examining the imaging quality with respect to simulated reconstructions for several phantom configurations. A 3-coil, 32-electrode ICEIT system was built, and an iterative modified Newton-Raphson algorithm was developed for the solution of the inverse problem. The RMS norm between the simulated and the experimental voltages was found to be 0.08 +/- 0.05 mV (<3%). Two regularization methods were implemented and compared: the Marquardt regularization and the Laplacian regularization (a bounded second-derivative regularization). While the Laplacian regularization method was found to be preferred for simulated data, it resulted in distinctive spatial artifacts for measured data. The experimental reconstructed images were found to be indicative of the angular positioning of the conductivity perturbations, though the radial sensitivity was low, especially when using the Marquardt regularization method.

A Novel Telemedicine System for Monitoring Congestive Heart Failure Patients

Congestive heart failure is a widespread cardiac disease in western countries. At present, the main measure for monitoring the level of pulmonary edema in telemedicine systems is weight, which is not a reliable indicator. The authors propose a novel bioimpedance telemedical system to monitor these patients. The system measures the resistivity of each lung using optimization methods and transmits the measurements via a modem to a call center. Preliminary results show that the measured resistivity values among healthy young patients are consistent and reproducible within 48 hours. The mean resistivity values in patients with pulmonary congestion were lower than those of the healthy patients: 887 [Omega*cm]+/-117 vs 1244 [Omega*cm]+/-87 (P<.01). The system is noninvasive, safe, and portable. It retrieves unique information correlated with the amount of fluid in the lungs and transmits the data to a medical call center in order to improve the diagnostics and treatment of congestive heart failure.

Dynamic beat-to-beat QT-RR relationship during physiotherapy effort in elderly patients without primary heart disease.

The ECGs from 18 patients hospitalized in a rehabilitation setting, following surgery for hip fracture, were examined to characterize the dynamic behavior of uncorrected QT interval in relation to changing RR interval during physiotherapy effort. ECG waveforms were analyzed to extract beat‐to‐beat QT and RR intervals using a computerized ECG Analyzer (CEA‐1100). The method of defining the QT and RR intervals is based on performing multiple cross‐correlations that enable rejection of artifacts from the analysis. The relationship between the RR and QT intervals was found using the following general formula QTi= cRRbi ‐ 1. Linear regression was performed on the logarithms of QT and RR measurements obtained to estimate the constant (a = log c) and the slope (b) values, reflecting the dynamic change of QT during physiotherapy effort. Having these two values, the dynamic QT extrapolated to a heart period of 1 second (QTcd) was calculated. The results were compared to the conventional corrected static QT according to the Bazzet formula (QTcs). The mean values of constants (a = log c) and slopes (b) over all patients were found to be 1.61 ± 0.23 and 0.33 ± 0.08, respectively, giving a QT (ms) heart‐period (ms) dynamic relation of QTi= 41 × RRi‐10.33. The correlation between the dynamic QT and the static QT intervals was not significant. The mean values of the QTcd and QTcs intervals were significantly different (392 ± 25 ms vs 434 ± 28 ms; P < 0.0001). This dynamic measurement method of QT intervals may provide additional information on normal and abnormal cardiac repolarization in health and disease, helping in the diagnosis of cardiac disorders and arrhythmia risk.

Evaluation of impedance technique for detecting breast carcinoma using a 2-D numerical model of the torso.

Abstract: Previous experimental studies showed that significant changes occur in the electrical properties of breast cancer tissue compared to the surrounding normal tissue. This phenomenon motivated studies on cancer detection using electrical impedance techniques. In the present study, a two‐dimensional model of the torso and a numerical method were used to investigate the changes in the potential distribution as a result of a malignant tissue present in the breast. A transverse MRI image of the womans torso was scanned. Noise reduction and contour‐following algorithms were applied to differentiate between eight compartments in the torso. The extracted tissue types were lungs, blood, ribs, bone marrow of the cord, breast fat, skin, skeletal muscle, and heart muscle. Isotropic homogeneous conductivity was assigned to each one of these compartments. The volume conductor problem was solved numerically using the finite volume method to determine the potential distribution developed due to the dipole source. Cases without and with artificially inserted malignant region with realistic sizes were examined to investigate the sensitivity of impedance techniques to detect breast cancer. Significant changes were detected in the potential distribution inside the volume conductor as a result of the realistic size of breast tumors. A linear relation was found between the surface potential in the vicinity of the tumor region and the size of the tumor. For a small malignant area of 0.22 cm2, the surface potential near the tumor region decreased only slightly from a value of 13.81 mV in the normal case to 13.67 mV (0.14 mV change; 1.0%). For a larger malignant area of 5.43 cm2, the potential decrease was more pronounced, 11.29 mV (2.52 mV change; 18.3%), indicating that realistic sizes of breast tumor result in significant changes in the surface potential. Thus, impedance techniques employed in the present study show very good promise in detecting breast cancer.

Effect of brain damage and source location on left-right asymmetry of visual evoked potentials in a realistic model of the head.

The left-right asymmetry of visual evoked potentials (VEP) was studied using a realistic three-dimensional model of the head, constructed from CT scans. The integral equation describing the potential distribution due to an inner current-dipole (simulating the VEP source), inside the biological volume conductor (the head) was solved numerically using the finite volume method. The effect of several structural parameters, such as the natural head geometry, the location and orientation of the current source, and conductivity changes (simulating a brain damage) on the VEP asymmetry was examined. The results revealed that the natural anatomical asymmetry presented in a normal head induces some degree of VEP asymmetry (up to 0.42% at the occipital electrodes), yet the major cause of left-right asymmetry is due to asymmetrical location of the source: up to 6.53% in the O(2)-O(1) electrodes for an angular shift of 3 degrees to the left. It was also found that conductivity changes inside the head have a smaller effect on the VEP asymmetry (up to 3.35% in hemorrhaged brain compared to 1.96% in the normal brain at the C(4)-C(3) electrodes). These findings may help in better understanding VEP sources of asymmetry, originating not only due to cerebral functionality, but from structural parameters as well.

Dynamic cardiophysiologic variables correlate with lesion location and FIMTM in patients with ischemic stroke.

Arad M, Ring H, Abboud S, Radai MM, Tamir A, Adunsky A: Dynamic cardiophysiologic variables correlate with lesion location and FIMTM in patients with ischemic stroke. Am J Phys Med Rehabil 2002;81:590–596. Objective To examine the correlation between a clinical measure of function in patients undergoing rehabilitation who had recently had an ischemic stroke and to examine cardiophysiologic measures registered during effort. Design A cohort study comprising a sample of consecutive patients, without a history of cardiac disease or rhythm disturbances, admitted for rehabilitation after an ischemic supratentorial stroke. All patients were examined for the FIM™ score and dynamic cardiophysiologic variables. Results were analyzed in relation to stroke location. Thirty-eight patients participated in the study. Ten patients had a superficial lesion, 20 had a deep brain lesion, and eight had no noticeable lesion by computed tomographic imaging. Function was measured with the FIM instrument 48–72 hr after admission and repeated at 1 wk before discharge. Electrocardiographic activity was recorded during physiotherapy treatment. The relationship between the RR and QT intervals in the electrocardiographic waveforms was found to estimate two cardiophysiologic variables, the constant (a) and the slope (b) values, reflecting the dynamic change of QT during physiotherapy effort. Results Only for the subgroup of patients who sustained a deep brain lesion did the motor items in the FIM instrument on admission and discharge scores have a statistically significant relationship with the slope variable (b) and an inverse statistically significant relationship with the constant variable (a). Conclusions Deep brain infarction seems to result in a significant dysfunction of the autonomic nervous system, manifesting itself as distorted dynamic behavior of the QT interval and with impaired functional performance.

Effect of ischemic stroke on the dynamic beat-to-beat QT-RR relationship.

ARAD, M., et al.: Effect of Ischemic Stroke on the Dynamic Beat‐to‐Beat QT‐RR Relationship. The ECGs of 26 patients following ischemic stroke and 18 control patients with various orthopedic problems, all without primary heart disease, were examined to characterize the dynamic behavior of uncorrected QT interval in relation to changing RR interval during physiotherapy effort. ECG waveforms were analyzed to extract beat‐to‐beat QT and RR intervals. Based on performing multiple cross‐correlations, the relationship between the RR and QT intervals was calculated using the following general formula QTi = c RRbi‐1. Linear regression was performed on the logarithms of QT and RR measurements to estimate the constant (a = log c) and the slope (b) values, reflecting the dynamic change of QT during physiotherapy effort. Having these two values, the dynamic QT extrapolated to a heart period of 1 second (QTcd) was calculated. It was found that the mean slope (b) of the linear regression line in the ischemic stroke group was significantly lower than in the control group (0.26 ± 0.08 vs 0.33 ± 0.08, P < 0.02), and the constant (a) was significantly higher (1.83 ± 0.21 for the ischemic stroke vs 1.61 ± 0.23 for the controls, P < 0.002). No significant difference was found in QTcd values between the two groups (386 ± 27 ms for the ischemic stroke vs 392 ± 25 ms for the controls, P > 0.05). In conclusion, hemispheric brain infarction seems to result in alteration in the autonomic activity during exercise manifesting itself as distorted dynamic behavior of the QT interval.

Left-right asymmetry of visual evoked potentials in a realistic 3-dimensional numerical model of the head

The distribution of electric potentials was studied by using a realistic 3-D numerical model of the head. Computerized Tomography (CT) imaging scans were used to construct the three dimensional model of the head. Several compartments of the head were defined and a different conductivity was assigned for each compartment. The integral conservation equation in biological volume conductor was solved numerically by a finite volume method for the potential distribution created by dipole sources in the occipital region. The influence of several parameters, such as the geometry of the head, the location of the current source, and the presence of damaged tissue, on the left-right asymmetry of the surface Visual Evoked Potentials (VEP) were examined. The numerical model revealed that the major source of scalp potential asymmetry (Right-Left) is due to asymmetric location of the source (a change of 1.30 Arbitrary Units in O/sub 1/-O/sub 2/ pair of electrodes for falx deviation of 1/spl deg/ between the occiput and the nasion-inion line). It was also found that the left-right asymmetry in the scalp VEP due to changes in the conductivity of the volume conductor (a damaged region) between the source and the electrodes is insignificant (0.001 Arrbitry Units in P/sub 3/-P/sub 4/ pair of electrodes).