Estimation of the omitted-lead potentials with the temporal and spatial correlations.

概要

論文の詳細を見る
By making use of the correlations among the potentials measured at different sites on the body surface, the number of electrodes required to reproduce the body surface potential map (BSPM) can be reduced as Lux et al. pointed out. In their method, the potentials at some instant are estimated only from those measured at the same time. However, the omitted-lead potentials are correlated with those in the past and the future, so that the estimation error can be reduced by utilizing these correlations.The omitted-lead potentials at time T are estimated from those measured at T, T±τ, T±2τ, and so on. The more terms being included, the more accurate the estimation, but the longer the computation time. In the present paper, four combinations, (T, T-τ), (T, T-τ, T-2τ), (T, T±τ) and (T, T±τ, T±2τ) are applied to the BSPM measured from 75 subjects with various heart diseases including normal ones, and the relations between the time-lag r and the estimation errors are investigated. For example, if the number of electrodes is 40 and the combination (T, T±τ, T±2τ) is ued, the estimation error becomes 81% of that estimated with the Lux method. The same error is obtained in the Lux method if 45-46 electrodes are used. In other words, the new method we propose here saves 5-6 electrodes. When the electrodes are restricted on the chest. our method is more effective and saves no less than 12 electrodes.
一般社団法人日本生体医工学会の論文