Limits...
Optimisation of ionic models to fit tissue action potentials: application to 3D atrial modelling.

Al Abed A, Guo T, Lovell NH, Dokos S - Comput Math Methods Med (2013)

Bottom Line: A 3D model of atrial electrical activity has been developed with spatially heterogeneous electrophysiological properties.Membrane potentials of myocytes from spontaneously active or electrically paced in vitro rabbit cardiac tissue preparations were recorded using intracellular glass microelectrodes.The tissue-based optimisation of ionic models and the modelling process outlined are generic and applicable to image-based computer reconstruction and simulation of excitable tissue.

View Article: PubMed Central - PubMed

Affiliation: Graduate School of Biomedical Engineering, The University of New South Wales, Sydney, NSW 2052, Australia. amra@unsw.edu.au

ABSTRACT
A 3D model of atrial electrical activity has been developed with spatially heterogeneous electrophysiological properties. The atrial geometry, reconstructed from the male Visible Human dataset, included gross anatomical features such as the central and peripheral sinoatrial node (SAN), intra-atrial connections, pulmonary veins, inferior and superior vena cava, and the coronary sinus. Membrane potentials of myocytes from spontaneously active or electrically paced in vitro rabbit cardiac tissue preparations were recorded using intracellular glass microelectrodes. Action potentials of central and peripheral SAN, right and left atrial, and pulmonary vein myocytes were each fitted using a generic ionic model having three phenomenological ionic current components: one time-dependent inward, one time-dependent outward, and one leakage current. To bridge the gap between the single-cell ionic models and the gross electrical behaviour of the 3D whole-atrial model, a simplified 2D tissue disc with heterogeneous regions was optimised to arrive at parameters for each cell type under electrotonic load. Parameters were then incorporated into the 3D atrial model, which as a result exhibited a spontaneously active SAN able to rhythmically excite the atria. The tissue-based optimisation of ionic models and the modelling process outlined are generic and applicable to image-based computer reconstruction and simulation of excitable tissue.

Show MeSH

Related in: MedlinePlus

Optimised cSAN and pSAN models using a cSAN-pSAN 2D tissue axisymmetric disc model with σSAN = 800 μS·cm−1. (a) The optimised cSAN and pSAN AP waveforms (Em) are overlaid on top of AP traces recorded experimentally (Vm) from cSAN, pSAN, and RA myocytes, respectively, from a rabbit sino-atrial tissue preparation. (b) Inward, outward, and leakage ionic currents reconstructed from the optimised cSAN, pSAN, and RA models, respectively. Note the different scales used for currents in cSAN, pSAN, and RA.
© Copyright Policy - open-access
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3713319&req=5

fig6: Optimised cSAN and pSAN models using a cSAN-pSAN 2D tissue axisymmetric disc model with σSAN = 800 μS·cm−1. (a) The optimised cSAN and pSAN AP waveforms (Em) are overlaid on top of AP traces recorded experimentally (Vm) from cSAN, pSAN, and RA myocytes, respectively, from a rabbit sino-atrial tissue preparation. (b) Inward, outward, and leakage ionic currents reconstructed from the optimised cSAN, pSAN, and RA models, respectively. Note the different scales used for currents in cSAN, pSAN, and RA.

Mentions: A generic ionic model with two time-dependent (inward and outward) currents and one background current was implemented for all cell types in the 2D axisymmetric disc. Figure 6 illustrates the optimised model and experimental AP waveforms for cSAN, pSAN, and RA cells for SAN tissue conductivity (σSAN) of 800 μS·cm−1 and atrial tissue conductivity (σRA) of 104 μS·cm−1. Table 3 lists the optimised model parameters for all cell types. The root mean square error (RMSE) between experiment and model APs was 2.22 mV, 2.45 mV, and 4.12 mV for cSAN, pSAN, and RA cells, respectively. There was a transition in the reconstructed currents from cSAN to RA models, with the peak magnitude of each time-dependent current increasing from cSAN to pSAN to RA (Figure 6). The SAN activated spontaneously and was able to entrain the RA region. AP propagation velocity was 22 and 96 cm·s−1 in the SAN and RA segments, respectively. The maximum rates of change of the membrane potential during phase 0 depolarisation and phase 1 repolarisation of the tissue-optimised RA AP were 46.03 and 1.689 V·s−1, respectively, compared to 26.55 and 0.7114 V·s−1 for the single-cell optimised RA AP.


Optimisation of ionic models to fit tissue action potentials: application to 3D atrial modelling.

Al Abed A, Guo T, Lovell NH, Dokos S - Comput Math Methods Med (2013)

Optimised cSAN and pSAN models using a cSAN-pSAN 2D tissue axisymmetric disc model with σSAN = 800 μS·cm−1. (a) The optimised cSAN and pSAN AP waveforms (Em) are overlaid on top of AP traces recorded experimentally (Vm) from cSAN, pSAN, and RA myocytes, respectively, from a rabbit sino-atrial tissue preparation. (b) Inward, outward, and leakage ionic currents reconstructed from the optimised cSAN, pSAN, and RA models, respectively. Note the different scales used for currents in cSAN, pSAN, and RA.
© Copyright Policy - open-access
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3713319&req=5

fig6: Optimised cSAN and pSAN models using a cSAN-pSAN 2D tissue axisymmetric disc model with σSAN = 800 μS·cm−1. (a) The optimised cSAN and pSAN AP waveforms (Em) are overlaid on top of AP traces recorded experimentally (Vm) from cSAN, pSAN, and RA myocytes, respectively, from a rabbit sino-atrial tissue preparation. (b) Inward, outward, and leakage ionic currents reconstructed from the optimised cSAN, pSAN, and RA models, respectively. Note the different scales used for currents in cSAN, pSAN, and RA.
Mentions: A generic ionic model with two time-dependent (inward and outward) currents and one background current was implemented for all cell types in the 2D axisymmetric disc. Figure 6 illustrates the optimised model and experimental AP waveforms for cSAN, pSAN, and RA cells for SAN tissue conductivity (σSAN) of 800 μS·cm−1 and atrial tissue conductivity (σRA) of 104 μS·cm−1. Table 3 lists the optimised model parameters for all cell types. The root mean square error (RMSE) between experiment and model APs was 2.22 mV, 2.45 mV, and 4.12 mV for cSAN, pSAN, and RA cells, respectively. There was a transition in the reconstructed currents from cSAN to RA models, with the peak magnitude of each time-dependent current increasing from cSAN to pSAN to RA (Figure 6). The SAN activated spontaneously and was able to entrain the RA region. AP propagation velocity was 22 and 96 cm·s−1 in the SAN and RA segments, respectively. The maximum rates of change of the membrane potential during phase 0 depolarisation and phase 1 repolarisation of the tissue-optimised RA AP were 46.03 and 1.689 V·s−1, respectively, compared to 26.55 and 0.7114 V·s−1 for the single-cell optimised RA AP.

Bottom Line: A 3D model of atrial electrical activity has been developed with spatially heterogeneous electrophysiological properties.Membrane potentials of myocytes from spontaneously active or electrically paced in vitro rabbit cardiac tissue preparations were recorded using intracellular glass microelectrodes.The tissue-based optimisation of ionic models and the modelling process outlined are generic and applicable to image-based computer reconstruction and simulation of excitable tissue.

View Article: PubMed Central - PubMed

Affiliation: Graduate School of Biomedical Engineering, The University of New South Wales, Sydney, NSW 2052, Australia. amra@unsw.edu.au

ABSTRACT
A 3D model of atrial electrical activity has been developed with spatially heterogeneous electrophysiological properties. The atrial geometry, reconstructed from the male Visible Human dataset, included gross anatomical features such as the central and peripheral sinoatrial node (SAN), intra-atrial connections, pulmonary veins, inferior and superior vena cava, and the coronary sinus. Membrane potentials of myocytes from spontaneously active or electrically paced in vitro rabbit cardiac tissue preparations were recorded using intracellular glass microelectrodes. Action potentials of central and peripheral SAN, right and left atrial, and pulmonary vein myocytes were each fitted using a generic ionic model having three phenomenological ionic current components: one time-dependent inward, one time-dependent outward, and one leakage current. To bridge the gap between the single-cell ionic models and the gross electrical behaviour of the 3D whole-atrial model, a simplified 2D tissue disc with heterogeneous regions was optimised to arrive at parameters for each cell type under electrotonic load. Parameters were then incorporated into the 3D atrial model, which as a result exhibited a spontaneously active SAN able to rhythmically excite the atria. The tissue-based optimisation of ionic models and the modelling process outlined are generic and applicable to image-based computer reconstruction and simulation of excitable tissue.

Show MeSH
Related in: MedlinePlus