Non-invasive estimation of atrial fibrillation driver position using long-short term memory neural networks and body surface potentials.

BACKGROUND AND OBJECTIVE: Atrial Fibrillation (AF) is a supraventricular tachyarrhythmia that can lead to thromboembolism, hearlt failure, ischemic stroke, and a decreased quality of life. Characterizing the locations where the mechanisms of AF are initialized and maintained is key to accomplishing an effective ablation of the targets, hence restoring sinus rhythm. Many methods have been investigated to locate such targets in a non-invasive way, such as Electrocardiographic Imaging, which enables an on-invasive and panoramic characterization of cardiac electrical activity using recording Body Surface Potentials (BSP) and a torso model of the patient. Nonetheless, this technique entails some major issues stemming from solving the inverse problem, which is known to be severely ill-posed. In this context, many machine learning and deep learning approaches aim to tackle the characterization and classification of AF targets to improve AF diagnosis and treatment. METHODS: In this work, we propose a method to locate AF drivers as a supervised classification problem. We employed a hybrid form of the convolutional-recurrent network which enables feature extraction and sequential data modeling utilizing labeled realistic computerized AF models. Thus, we used 16 AF electrograms, 1 atrium, and 10 torso geometries to compute the forward problem. Previously, the AF models were labeled by assigning each sample of the signals a region from the atria from 0 (no driver) to 7, according to the spatial location of the AF driver. The resulting 160 BSP signals, which resemble a 64-lead vest recording, are preprocessed and then introduced into the network following a 4-fold cross-validation in batches of 50 samples. RESULTS: The results show a mean accuracy of 74.75% among the 4 folds, with a better performance in detecting sinus rhythm, and drivers near the left superior pulmonary vein (R1), and right superior pulmonary vein (R3) whose mean sensitivity bounds around 84%-87%. Significantly good results are obtained in mean sensitivity (87%) and specificity (83%) in R1. CONCLUSIONS: Good results in R1 are highly convenient since AF drivers are commonly found in this area: the left atrial appendage, as suggested in some previous studies. These promising results indicate that using CNN-LSTM networks could lead to new strategies exploiting temporal correlations to address this challenge effectively.

Non-invasive Estimation of Atrial Fibrillation Driver Position With Convolutional Neural Networks and Body Surface Potentials.

Atrial fibrillation (AF) is characterized by complex and irregular propagation patterns, and AF onset locations and drivers responsible for its perpetuation are the main targets for ablation procedures. ECG imaging (ECGI) has been demonstrated as a promising tool to identify AF drivers and guide ablation procedures, being able to reconstruct the electrophysiological activity on the heart surface by using a non-invasive recording of body surface potentials (BSP). However, the inverse problem of ECGI is ill-posed, and it requires accurate mathematical modeling of both atria and torso, mainly from CT or MR images. Several deep learning-based methods have been proposed to detect AF, but most of the AF-based studies do not include the estimation of ablation targets. In this study, we propose to model the location of AF drivers from BSP as a supervised classification problem using convolutional neural networks (CNN). Accuracy in the test set ranged between 0.75 (SNR = 5 dB) and 0.93 (SNR = 20 dB upward) when assuming time independence, but it worsened to 0.52 or lower when dividing AF models into blocks. Therefore, CNN could be a robust method that could help to non-invasively identify target regions for ablation in AF by using body surface potential mapping, avoiding the use of ECGI.

Autonomic Nervous System and Recall Modeling in Audiovisual Emotion-Mediated Advertising Using Partial Least Squares-Path Modeling.

Interest in improving advertisement impact on potential consumers has increased recently. One well-known strategy is to use emotion-based advertisement. In this approach, an emotional link with consumers is created, aiming to enhance the memorization process. In recent years, Neuromarketing techniques have allowed us to obtain more objective information on this process. However, the role of the autonomic nervous system (ANS) in the memorization process using emotional advertisement still needs further research. In this work, we propose the use of two physiological signals, namely, an electrocardiogram (heart rate variability, HRV) and electrodermal activity (EDA), to obtain indices assessing the ANS. We measured these signals in 43 subjects during the observation of six different spots, each conveying a different emotion (rational, disgust, anger, surprise, and sadness). After observing the spots, subjects were asked to answer a questionnaire to measure the spontaneous and induced recall. We propose the use of a statistical data-driven model based on Partial Least Squares-Path Modeling (PSL-PM), which allows us to incorporate contextual knowledge by defining a relational graph of unobservable variables (latent variables, LV), which are, in turn, estimated by measured variables (indices of the ANS). We defined four LVs, namely, sympathetic, vagal, ANS, and recall. Sympathetic and vagal are connected to the ANS, the latter being a measure of recall, estimated from a questionnaire. The model is then fitted to the data. Results showed that vagal activity (described by HRV indices) is the most critical factor to describe ANS activity; they are inversely related except for the spot, which is mainly rational. The model captured a moderate-to-high variability of ANS behavior, ranging from 38% up to 64% of the explained variance of the ANS. However, it can explain at most 11% of the recall score of the subjects. The proposed approach allows for the easy inclusion of more physiological measurements and provides an easy-to-interpret model of the ANS response to emotional advertisement.