Page 44 - Artificial Intelligence for Computational Modeling of the Heart

14 Chapter 1 Multi-scale models of the heart for patient-specific simulations

Modeling the electrical conduction system
An accurate EP model needs to include the description of the
effect of high-speed conducting tissues in the heart [60], which
determine the pattern of ventricular excitation and contraction
both in normal sequences and cardiac arrhythmias [84]. Com-
putational models have been developed to try to reproduce the
tree-like structure of the Purkinje network, and to capture pecu-
liar features of the high-speed conducting system such as retro-
grade propagation. This may play a role in complex scenarios such
as cardiac resynchronization therapy, in which pacing electrodes
may be placed in areas of the myocardium that are reached by the
distal termination of the Purkinje networks. Current imaging tech-
niques do not allow the patient-specific measurement of the heart
conduction system in vivo, therefore this kind of models cannot
be directly validated and personalized. They are typically based
on anatomical information from histological studies and tuned
to correctly reproduce sites of earliest activation and normal ac-
tivation sequence from electrical mapping studies. An extensive
literature review of this modeling approach is provided by ten
Tusscher and Panfilov [84], who also propose a model in which the
location of the Purkinje fiber-ventricular muscle junction (PVJ) is
derived from electrical activation maps. More recent examples of
anatomically realistic models of the Purkinje network include [85,
86]. An alternative, phenomenological approach can be followed,
to reproduce the depolarization pattern without describing the
detailed anatomy of the high-speed conducting tissue. Some stud-
ies focused on the estimation of a time delay function to apply
to regions in the endocardium [87,88], or proposed an increase
of the conduction velocity in the sub-endocardial tissue to repre-
sent the influence of the Purkinje system [89]. Despite its inability
to describe secondary and retrograde wavefronts, this approach
is adequate under the assumption that the distribution of PVJ is
dense in the endocardium, so that a single wavefront is generated.

1.2.3 Body surface potential modeling
The electrical activity of the heart induces spatial and tem-
poral changes of the electrical potential in the whole body. As
the heart tissue depolarizes, a wave of positive electrical signal
diffuses through the surrounding tissue and reaches the outer
body surface. Electrocardiography aims at non-invasively record-
ing such activity by measuring signals through electrodes placed
on the skin. Each electrode reports the instantaneous difference
between the local electrical potential and that of a reference elec-
trode. By defining multiple leads, connections between a refer-