Exploiting GPUs to investigate an inversion method that retrieves cardiac conductivities from potential measurements

Johnston, Barbara M.; Barnes, Josef Peter

doi:10.21914/anziamj.v55i0.7790

Cited by 3 publications

(3 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…New subsets of electrodes, corresponding to 'closely-spaced' sets of electrodes for the first pass to find the extracellular conductivities and 'widely-spaced' sets of electrodes for the second pass to find the intracellular conductivities and fibre rotation angle, were investigated. The accuracy of retrieval for various simulated noise levels was presented [45,40,46] and it was found that the three extracellular conductivities could be retrieved extremely accurately; even with added noise of up to 40%, relative errors are around 2% on average. The fibre rotation angle and g il can also be retrieved quite accurately, with around 8% error for 40% noise.…”

Section: Johnston and Johnstonmentioning

confidence: 99%

Approaches for determining cardiac bidomain conductivity values: progress and challenges

Johnston

2020

Med Biol Eng Comput

View full text Add to dashboard Cite

Modelling the electrical activity of the heart is an important tool for understanding electrical function in various diseases and conduction disorders. Clearly, for model results to be useful, it is necessary to have accurate inputs for the models, in particular the commonly used bidomain model. However, there are only three sets of experimentally determined four conductivity values for cardiac ventricular tissue and these are inconsistent, were measured around forty years ago, often produce different results in simulations, and do not fully represent the three dimensional anisotropic nature of cardiac tissue. Despite efforts in the intervening years, difficulties associated with making the measurements and also determining the conductivities from the experimental data have not yet been overcome. In this review, we summarise what is known about the conductivity values, as well as progress to date in meeting the challenges associated with both the mathematical modelling and the experimental techniques.

show abstract

Section: Johnston and Johnstonmentioning

confidence: 99%

Approaches for determining cardiac bidomain conductivity values: progress and challenges

Johnston

2020

Med Biol Eng Comput

View full text Add to dashboard Cite

show abstract

“…This is the angle through which the sheets of cardiac fibres rotate relative to one another between the inner and outer heart surfaces. A recent implementation of the inversion routine on GPUs [23] has resulted in a speedup that has allowed a more realistic investigation into the accuracy of these retrievals to be undertaken.…”

Section: Introductionmentioning

confidence: 99%

Determining six cardiac conductivities from realistically large datasets

Johnston

2015

Mathematical Biosciences

View full text Add to dashboard Cite

Simulation studies of cardiac electrophysiological behaviour that use the bidomain model require accurate values for the bidomain extracellular and intracellular conductivities to produce useful results. This work considers an inversion algorithm, which has previously been shown, using simulated data, to be capable of retrieving six bidomain conductivities and the fibre rotation angle from measurements of electric potential made in the heart. The aim here is to see whether it is possible to improve the accuracy of the retrieved parameters. The scenario of retrieving only conductivities and not fibre rotation is examined but this does not lead to a worthwhile improvement in retrieval accuracy. It is also found that it is possible to retrieve the bidomain conductivities using not two but just one pass of the algorithm, made on a 'widely-spaced' electrode set. This appears to work because the algorithm is still very sensitive to the extracellular conductivities. However, the single-pass method is not recommended because the intracellular conductivities that are retrieved are not as accurate as those that are retrieved in the usual two-pass method, particularly for higher values of added noise. The second part of this work considers retrieving the six conductivities and fibre rotation from realistically large sets of potential measurements and identifies the best data analysis method. It is found that, even with added noise of up to 40%, the extracellular conductivities can still be retrieved extremely accurately (relative errors of around 2% on average) and so can the intracellular longitudinal conductivities and fibre rotation (errors less than 8% on average). The remaining intracellular conductivities have errors that are generally less than twice the added noise, particularly for the higher noise values.

show abstract

“…Two Intel Xeon(R) E5-2630 v3 containing eight cores, each with a clock speed of 2.40GHz, were used to run these inversions. The inversion algorithm was written in C++ and parallelised to run on multiple CPU threads using a message passing interface (MPICH2) and techniques given by [47]. Under this scenario, a first pass modified protocol inversion with 100 measurement sets typically had a runtime of approximately 3-5 minutes, while the second pass modified protocol inversion with 100 measurement sets typically had a run-time of 2-3 minutes.…”

Section: Computational Costs: the Original Protocol Vs The Modified P...mentioning

confidence: 99%

A modified approach to determine the six cardiac bidomain conductivities

Kamalakkannan

Johnston

2021

Computers in Biology and Medicine

View full text Add to dashboard Cite

Exploiting GPUs to investigate an inversion method that retrieves cardiac conductivities from potential measurements

Cited by 3 publications

References 20 publications

Approaches for determining cardiac bidomain conductivity values: progress and challenges

Approaches for determining cardiac bidomain conductivity values: progress and challenges

Determining six cardiac conductivities from realistically large datasets

A modified approach to determine the six cardiac bidomain conductivities

Contact Info

Product

Resources

About