Potential value of the low frequency/high frequency ratio of heart rate variability to guide atrial fibrillation ablation procedures

Funding Acknowledgements

Type of funding sources: None.

Background.

The  differentiation  between  vagally  and  non-vagally mediated  atrial fibrillation (AF)  could be useful  to decide  adding  ganglionated  plexi  ablation  (GPA)  to  pulmonary vein isolation (PVI)  in  patients with vagally mediated  AF  onset.

The role of the autonomic nervous system in triggering paroxysmal AF can be analyzed through heart rate variability (HRV) parameters. The low frequency/high frequency ratio (LF/HF) is used as an index of sympatho-vagal interaction. Based on clinical criteria, values of LF/HF <1.5 are considered to reflect an increase of vagal modulation. 

Purpose

Identify  patients with paroxysmal AF who may benefit from GPA in addition to PVI, based on analysis of HRV parameters

Material and methods

From our database of unselected Holter recordings containing paroxysmal AF episodes, 199 analyzable AF onsets were labelled. We analyzed HRV frequential parameters by segments every 300 RR intervals to identify adrenergic and vagal onsets of AF episodes. We categorized AF onsets, depending of their respective values of LF/HF. We also checked the relationship of these parameters to heart rate (HR). We followed the evolution of the mean normalized values of LF/HF from the beginning of the recordings to the AF onsets (see figure).

Results

We found a progressive increase of LF/HF  in normalized units  (mean [95% confidence interval]) from  2.44  [2.12-2.76]  to 3.12  [2.85-3.39]  in the adrenergic group  (n = 113)  and a  progressive decrease of LF/HF from  1.51 [1.28-1.75]  to 1.02 [0.95-1.09]  in the  vagal group (n = 86). Differences between adrenergic and vagal AF were highly significant using Mann-Whitney test (p < 0.001). A given patient shows the same type of AF in 80% of cases.  We found no correlation between frequential parameters and HR.

Conclusions

Our data show  potential  added value of  frequential  HRV  analysis  to  guide additional ganglionated  plexi ablation  to PVI. This  remains  to be demonstrated in a prospective study. The ratio LF/HF should be  considered  to optimize the decision-making process for  the ablation procedure. Abstract Figure. Frequential HRV parameters before AF

Unravelling the black box of machine learning for atrial fibrillation forecast: role of heart rate variability and of premature beats

Background

The identification of patients still in sinus rhythm who will present one month later an atrial fibrillation episode is possible using machine learning (ML) techniques. However, these new ML algorithms do not provide any relevant information about the underlying pathophysiology.

Purpose

To compare the predictive performance for forecasting AF between a machine learning algorithm and other parameters whose pathophysiological mechanisms are known to play a role in the triggering of arrhythmias (i.e. the count of premature beats (PB) and heart rate variability (HRV) parameters)

Material and methods

We conducted a retrospective study from an outpatient clinic. 10484 Holter ECG recordings were screened. 250 analysable AF onsets were labelled. We developed a deep neural network model composed of convolutional neural network layers and bidirectional gated recurrent units as recurrent neural network layers that was trained for the forecast of paroxysmal AF episodes, using RR intervals variations. This model works like a black box. For comparison purposes, we used a “random forest” (RF) model of ML to obtain forecast results using HRV parameters with and without PB. This model allows the evaluation of the relevance of HRV parameters and of PB used for the forecast. We calculated the area under the curve of the receiving operating characteristic curve for the different time windows counted in RR intervals before the AF onset.

Results

As shown in the table, the forecasting value of the deep neural network model (ML) was not superior to the random forest algorithm. Prediction value of both decreased when analyzing the RR intervals further away from the onset of AF

Conclusions

These results suggest that HRV plays a predominant role in triggering AF episodes and that premature beats could add minor information. Moreover, the closer the window from AF onset, the better the accuracy, regardless of the method used. Such detection algorithms once implemented in pacemakers, might prove useful to prevent AF onset by changing pacing sequence while patients would still be in sinus rhythm, however this remains to be demonstrated

Funding Acknowledgement

Type of funding source: None

P6568Forecasting atrial fibrillation using machine learning techniques

Background

Forecasting atrial fibrillation (AF) a few minutes before its onset has been studied, mainly based on heart rate variability parameters, derived from 24-hour ECG Holter monitorings. However, these studies have shown conflicting, non-clinically applicable results. Nowadays, machine learning algorithms have proven their ability to anticipate events. Therefore, forecasting AF before its onset should be (re)assessed using machine learning techniques. A reliable forecasting could improve results of preventive pacing in patients with cardiac electronic implanted devices (CEID).

Purpose

To forecast an oncoming AF episode in individual patients using machine learning techniques.

To evaluate the effect if the onset of an AF episode can be forecasted on longer time frames.

Methods

The totality of the raw data of a data set of 10484 ECG Holter monitorings was retrospectively analyzed and all AF episodes were annotated. Onset of each AF episode was determined with a precision of 5 msec. We only took AF events into consideration if they lasted longer than 30 seconds. Of all patients in the dataset, 140 presented paroxysmal AF (286 recorded AF episodes). We only used RR intervals to predict the presence of AF. We developed two different types of machine learning algorithms with different computational power requirements: a “dynamic” deep and recurrent neural net (RNN) and a “static” decision-tree with adaboost (boosting trees) more suitable for embedded devices. These algorithms were trained on one set of patients (around 90%) and tested on the remaining set of patients (around 10%).

Results

The performance figures are summarized in the table. Both algorithms can be tuned to increase their specificity (at a loss of sensitivity) or vice versa, depending on the objective.

Performance of forecasting algorithms RR-distance Boosting trees AUC RNN AUC 30–1 RR-Interval(s) before an AF event 97.1% 98.77% 60–31 RR-Intervals before an AF event 97.5% 99.1% 90–61 RR-Intervals before an AF event 96.9% 99.1% 120–91 RR-Inervals before an AF event 98.2% 98.9% AUC for Area Under ROC Curves.

Conclusion

Based upon this retrospective study, we show that AF can be forecasted on an individual level with high predictive power using machine learning algorithm, with little drop-off of predictive value within the studied distances (1–120 RR intervals before a potential AF episode). We believe that the embedding of our new algorithm(s) in CEID's could open the way to innovative therapies that significantly decrease AF burden in selected implanted patients.

Separation of samples into their constituents using gene expression data

Gene expression measurements are a powerful tool in molecular biology, but when applied to heterogeneous samples containing more than one cellular type the results are difficult to interpret. We present here a new approach to this problem allowing to deduce the gene expression profile of the various cellular types contained in a set of samples directly from the measurements taken on the whole sample.

Frustrated chaos in biological networks

The behaviour of three biologically inspired networks is investigated: Immune Idiotypic Network (IIN), Hopfield Network (HN) and Coupled Map Lattice (CML), mainly the effect on the dynamics induced by connecting the network in a frustrated way. Frustration occurs when the global structure is such that local connectivity patterns responsible for stable behaviour are intertwined, leading to mutually competing attractors and chaotic itinerancy among brief appearance of these attractors. Frustration destabilizes the network and provokes an unpredictable "wavering" among the stable dynamic regimes which characterize the same network when it is interconnected in a non-frustrated way. As the main contribution of this paper, an immune idiotypic network in which the prevailing behaviour is oscillatory is studied in detail. It is shown how connecting an elementary three-clone network in a frustrating way transforms the oscillatory regime into a chaotic one. This chaotic regime is further analysed and several interesting aspects are discussed such as the variable homogeneity, the intrinsic chaotic itinerancy among brief oscillatory regimes and the strong unpredictability. In addition, dynamical regimes obtained by frustrating the connectivity of HN and CML are presented and the similarities as well as the differences with the IIN dynamics are emphasized. Common to all these networks is the description of the frustrated chaos as a succession of attempts to relax the network into one of the oscillatory regimes given by a weaker and non-frustrated connectivity, an impossible achievement making the dynamics rambling over brief but repelling orbits.

A simplification of the backpropagation-through-time algorithm for optimal neurocontrol

Backpropagation-through-time (BPTT) is the temporal extension of backpropagation which allows a multilayer neural network to approximate an optimal state-feedback control law provided some prior knowledge (Jacobian matrices) of the process is available. In this paper, a simplified version of the BPTT algorithm is proposed which more closely respects the principle of optimality of dynamic programming. Besides being simpler, the new algorithm is less time-consuming and allows in some cases the discovery of better control laws. A formal justification of this simplification is attempted by mixing the Lagrangian calculus underlying BPTT with Bellman-Hamilton-Jacobi equations. The improvements due to this simplification are illustrated by two optimal control problems: the rendezvous and the bioreactor.

Natural tolerance in a simple immune network

The following basic question is studied here: In the relatively stable molecular environment of a vertebrate body, can a dynamic idiotypic immune network develop a natural tolerance to endogenous components? The approach is based on stability analyses and computer simulation using a model that takes into account the dynamics of two agents of the immune system, namely B-lymphocytes and antibodies. The study investigates the behavior of simple immune networks in interaction with an antigen whose concentration is held constant as a function of the symmetry properties of the connectivity matrix of the network. Current idiotypic network models typically become unstable in the presence of this type of antigen. It is shown that idiotypic networks of a particular connectivity show tolerance towards auto-antigen without the need for ad hoc mechanisms that prevent an immune response. These tolerant network structures are characterized by aperiodic behavior in the absence of auto-antigen. When coupled to an auto-antigen, the chaotic attractor degenerates into one of several periodic ones, and at least one of them is stable. The connectivity structure needed for this behavior allows the system to adopt particular dynamic concentration patterns which do not lead to an unbounded immune response. Possible implications for the understanding of autoimmune disease and its treatment are discussed.

Development of an idiotypic network in shape space

Based upon the shape-space formalism, a model of an idiotypic network including both bound and free immunoglobulins is simulated. Our point of interest is the network development in the context of self antigens. The investigations are organized around simulations initiated by various spatial configurations of antigens; the behavior of the system with respect to antigens is analyzed in terms of morphogenetic processes occurring in the shape space. For certain values of the parameters, the network expands by traveling waves. The resulting spatial pattern is a partition of the shape space into zones where introduction of an antigen entails an infinite growth of the clones binding to it, and into zones where, on the contrary, the anti-antigen idiotypes decrease. Among the parameter combinations tested, some produce a partition that remains static whereas others produce a partition that changes in time. For other values of the parameters, the patterns generated do not partition shape space into zones; in these cases, it is observed that the system systematically explodes when an antigen is present.