Symmetric Projection Attractor Reconstruction analysis of murine electrocardiograms: Retrospective prediction of Scn5a+/- genetic mutation attributable to Brugada syndrome

TY - JOUR

T1 - Symmetric Projection Attractor Reconstruction analysis of murine electrocardiograms: Retrospective prediction of Scn5a+/- genetic mutation attributable to Brugada syndrome

AU - Bonet-Luz, Esther

AU - Lyle, Jane

AU - Huang, Chris

AU - Nandi, Manasi

AU - Jeevaratnam, Kamalan

AU - Aston, Philip

PY - 2020/9/17

Y1 - 2020/9/17

N2 - BACKGROUND Life-threatening arrhythmias resulting from genetic mutations are often missed in current electrocardiogram(ECG) analysis. We combined a new method for ECG analysis thatuses all the waveform data with machine learning to improve detection of such mutations from short ECG signals in a mouse model.OBJECTIVE We sought to detect consequences of Na1 channeldeficiencies known to compromise action potential conduction incomparisons of Scn5a1/- mutant and wild-type mice using shortECG signals, examining novel and standard features derived fromlead I and II ECG recordings by machine learning algorithms.METHODS Lead I and II ECG signals from anesthetized wild-typeand Scn5a1/- mutant mice of length 130 seconds were analyzedby extracting various groups of features, which were used bymachine learning to classify the mice as wild-type or mutant. Thefeatures used were standard ECG intervals and amplitudes, as wellas features derived from attractors generated using the novel Symmetric Projection Attractor Reconstruction method, which reformulates the whole signal as a bounded, symmetric 2-dimensionalattractor. All the features were also combined as a single featuregroup.RESULTS Classification of genotype using the attractor featuresgave higher accuracy than using either the ECG intervals or theintervals and amplitudes. However, the highest accuracy (96%)was obtained using all the features. Accuracies for differentsubgroups of the data were obtained and compared.CONCLUSION Detection of the Scn5a1/- mutation from shortmouse ECG signals with high accuracy is possible using our Symmetric Projection Attractor Reconstruction method.

AB - BACKGROUND Life-threatening arrhythmias resulting from genetic mutations are often missed in current electrocardiogram(ECG) analysis. We combined a new method for ECG analysis thatuses all the waveform data with machine learning to improve detection of such mutations from short ECG signals in a mouse model.OBJECTIVE We sought to detect consequences of Na1 channeldeficiencies known to compromise action potential conduction incomparisons of Scn5a1/- mutant and wild-type mice using shortECG signals, examining novel and standard features derived fromlead I and II ECG recordings by machine learning algorithms.METHODS Lead I and II ECG signals from anesthetized wild-typeand Scn5a1/- mutant mice of length 130 seconds were analyzedby extracting various groups of features, which were used bymachine learning to classify the mice as wild-type or mutant. Thefeatures used were standard ECG intervals and amplitudes, as wellas features derived from attractors generated using the novel Symmetric Projection Attractor Reconstruction method, which reformulates the whole signal as a bounded, symmetric 2-dimensionalattractor. All the features were also combined as a single featuregroup.RESULTS Classification of genotype using the attractor featuresgave higher accuracy than using either the ECG intervals or theintervals and amplitudes. However, the highest accuracy (96%)was obtained using all the features. Accuracies for differentsubgroups of the data were obtained and compared.CONCLUSION Detection of the Scn5a1/- mutation from shortmouse ECG signals with high accuracy is possible using our Symmetric Projection Attractor Reconstruction method.

U2 - 10.1016/j.hroo.2020.08.007

DO - 10.1016/j.hroo.2020.08.007

M3 - Article

SP - 368

EP - 375

JO - Heart Rhythm O2

JF - Heart Rhythm O2

ER -