Lessons learned in the Apple Heart Study and implications for the data management of future digital clinical trials.

The digital clinical trial is fast emerging as a pragmatic trial that can improve a trial's design including recruitment and retention, data collection and analytics. To that end, digital platforms such as electronic health records or wearable technologies that enable passive data collection can be leveraged, alleviating burden from the participant and study coordinator. However, there are challenges. For example, many of these data sources not originally intended for research may be noisier than traditionally obtained measures. Further, the secure flow of passively collected data and their integration for analysis is non-trivial. The Apple Heart Study was a prospective, single-arm, site-less digital trial designed to evaluate the ability of an app to detect atrial fibrillation. The study was designed with pragmatic features, such as an app for enrollment, a wearable device (the Apple Watch) for data collection, and electronic surveys for participant-reported outcomes that enabled a high volume of patient enrollment and accompanying data. These elements led to challenges including identifying the number of unique participants, maintaining participant-level linkage of multiple complex data streams, and participant adherence and engagement. Novel solutions were derived that inform future designs with an emphasis on data management. We build upon the excellent framework of the Clinical Trials Transformation Initiative to provide a comprehensive set of guidelines for data management of the digital clinical trial that include an increased role of collaborative data scientists in the design and conduct of the modern digital trial.

The development of a mobile app-focused deduplication strategy for the Apple Heart Study that informs recommendations for future digital trials.

An app-based clinical trial enrolment process can contribute to duplicated records, carrying data management implications. Our objective was to identify duplicated records in real time in the Apple Heart Study (AHS). We leveraged personal identifiable information (PII) to develop a dissimilarity score (DS) using the Damerau-Levenshtein distance. For computational efficiency, we focused on four types of records at the highest risk of duplication. We used the receiver operating curve (ROC) and resampling methods to derive and validate a decision rule to classify duplicated records. We identified 16,398 (4%) duplicated participants, resulting in 419,297 unique participants out of a total of 438,435 possible. Our decision rule yielded a high positive predictive value (96%) with negligible impact on the trial's original findings. Our findings provide principled solutions for future digital trials. When establishing deduplication procedures for digital trials, we recommend collecting device identifiers in addition to participant identifiers; collecting and ensuring secure access to PII; conducting a pilot study to identify reasons for duplicated records; establishing an initial deduplication algorithm that can be refined; creating a data quality plan that informs refinement; and embedding the initial deduplication algorithm in the enrolment platform to ensure unique enrolment and linkage to previous records.

Arrhythmias Other Than Atrial Fibrillation in Those With an Irregular Pulse Detected With a Smartwatch: Findings From the Apple Heart Study.

[Figure: see text].

From Machine Learning to Artificial Intelligence Applications in Cardiac Care.

Artificial intelligence offers the potential for transformational advancement in cardiovascular care delivery, yet practical applications of this technology have yet to be embedded in clinical workflows and systems. Recent advances in machine learning algorithms and accessibility to big data sources have created the ability for software to solve highly specialized problems outside of health care, such as autonomous driving, speech recognition, and game playing (chess and Go), at superhuman efficiency previously not thought possible. To date, high-order cognitive problems in cardiovascular research such as differential diagnosis, treatment options, and clinical risk stratification have been difficult to address at scale with artificial intelligence. The practical application of artificial intelligence in the underlying operational processes in the delivery of cardiac care may be more amenable where adoption has great potential to fundamentally transform care delivery while maintaining the core quality and service that our patients demand. In this article, we provide an overview on how these artificial intelligence platforms can be implemented to improve the operational delivery of care for patients with cardiovascular disease.

Automated classification of optical coherence tomography images of human atrial tissue.

Tissue composition of the atria plays a critical role in the pathology of cardiovascular disease, tissue remodeling, and arrhythmogenic substrates. Optical coherence tomography (OCT) has the ability to capture the tissue composition information of the human atria. In this study, we developed a region-based automated method to classify tissue compositions within human atria samples within OCT images. We segmented regional information without prior information about the tissue architecture and subsequently extracted features within each segmented region. A relevance vector machine model was used to perform automated classification. Segmentation of human atrial ex vivo datasets was correlated with trichrome histology and our classification algorithm had an average accuracy of 80.41% for identifying adipose, myocardium, fibrotic myocardium, and collagen tissue compositions.

A role for synaptic inputs at distal dendrites: instructive signals for hippocampal long-term plasticity.

Synaptic potentials originating at distal dendritic locations are severely attenuated when they reach the soma and, thus, are poor at driving somatic spikes. Nonetheless, distal inputs convey essential information, suggesting that such inputs may be important for compartmentalized dendritic signaling. Here we report a new plasticity rule in which stimulation of distal perforant path inputs to hippocampal CA1 pyramidal neurons induces long-term potentiation at the CA1 proximal Schaffer collateral synapses when the two inputs are paired at a precise interval. This subthreshold form of heterosynaptic plasticity occurs in the absence of somatic spiking but requires activation of both NMDA receptors and IP(3) receptor-dependent release of Ca(2+) from internal stores. Our results suggest that direct sensory information arriving at distal CA1 synapses through the perforant path provide compartmentalized, instructive signals that assess the saliency of mnemonic information propagated through the hippocampal circuit to proximal synapses.

HCN1 channels constrain synaptically evoked Ca2+ spikes in distal dendrites of CA1 pyramidal neurons.

HCN1 hyperpolarization-activated cation channels act as an inhibitory constraint of both spatial learning and synaptic integration and long-term plasticity in the distal dendrites of hippocampal CA1 pyramidal neurons. However, as HCN1 channels provide an excitatory current, the mechanism of their inhibitory action remains unclear. Here we report that HCN1 channels also constrain CA1 distal dendritic Ca2+ spikes, which have been implicated in the induction of LTP at distal excitatory synapses. Our experimental and computational results indicate that HCN1 channels provide both an active shunt conductance that decreases the temporal integration of distal EPSPs and a tonic depolarizing current that increases resting inactivation of T-type and N-type voltage-gated Ca2+ channels, which contribute to the Ca2+ spikes. This dual mechanism may provide a general means by which HCN channels regulate dendritic excitability.

On the electrical function of dendritic spines.

Dendritic spines mediate most excitatory inputs in the brain, yet their function is still unclear. Imaging experiments have demonstrated their role in biochemical compartmentalization at individual synapses, yet theoretical studies have suggested that they could serve an electrical function in transforming synaptic inputs and transmitting dendritic spikes. Recent data indicate that spines possess voltage-dependent conductances and that these channels can be spine-specific. Although direct experimental investigations of the electrical properties of spines have not yet taken place, spines could play a significant electrical role, greatly influencing dendritic integration and the function of neural circuits.

Role of dendritic spines in action potential backpropagation: a numerical simulation study.

Two remarkable aspects of pyramidal neurons are their complex dendritic morphologies and the abundant presence of spines, small structures that are the sites of excitatory input. Although the channel properties of the dendritic shaft membrane have been experimentally probed, the influence of spine properties in dendritic signaling and action potential propagation remains unclear. To explore this we have performed multi-compartmental numerical simulations investigating the degree of consistency between experimental data on dendritic channel densities and backpropagation behavior, as well as the necessity and degree of influence of excitable spines. Our results indicate that measured densities of Na(+) channels in dendritic shafts cannot support effective backpropagation observed in apical dendrites due to suprathreshold inactivation. We demonstrate as a potential solution that Na(+) channels in spines at higher densities than those measured in the dendritic shaft can support extensive backpropagation. In addition, clustering of Na(+) channels in spines appears to enhance their effect due to their unique morphology. Finally, we show that changes in spine morphology significantly influence backpropagation efficacy. These results suggest that, by clustering sodium channels, spines may serve to control backpropagation.

Calcium dynamics of spines depend on their dendritic location.

Dendritic spines are morphologically and functionally heterogeneous. To understand this diversity, we use two-photon imaging of layer 5 neocortical pyramidal cells and measure action potential-evoked [Ca(2+)]i transients in spines. Spine calcium kinetics are controlled by (i) the diameter of the parent dendrite, (ii) the length of the spine neck, and (iii) the strength of spine calcium pumps. These factors produce different calcium dynamics in spines from basal, proximal apical, and distal apical dendrites, differences that are more pronounced without exogenous buffers. In proximal and distal apical dendrites, different calcium dynamics correlate with different susceptibility to synaptic depression, and modifying calcium kinetics in spines changes the expression of long-term depression. Thus, the spine location apparently determines its calcium dynamics and synaptic plasticity. Our results highlight the precision in design of neocortical neurons.

Calcium dynamics in spines: link to synaptic plasticity.

Dendritic spines are morphologically and functionally heterogeneous. Here, we use two-photon imaging of layer V pyramidal neurons in slices from mouse visual cortex to characterize differences in spine calcium dynamics between individual spines. By measuring action potential-evoked [Cell transients in spines, we find different calcium dynamics in spines from proximal apical and distal apical dendrites. Using a mathematical multi-compartmental model, we demonstrate that these differences are even more pronounced in the absence of exogenous calcium buffers. We also find that these different calcium dynamics cause different susceptibility to synaptic depression in proximal and distal apical synapses, and that modifying calcium decay kinetics in spines changes the expression of long-term depression. We conclude that the location of the spine determines its time window of calcium compartmentalization and degree of calcium-dependent synaptic plasty. Our results highlight the precision of the design of neocortical neurons.