This project is investigating the effects of ischemia on cardiac mechano-electrical feedback and its role in sustained cardiac arrhythmias. Experimentally, we apply controlled, non-traumatic local deformation, while monitoring the electrical responses using voltage- and ion-sensitive optical mapping and surface electrocardiogram, during both acute ischaemia and in control, and with and without pharmacological modulation of stretch activated channel activity. Results are combined computationally, using anatomically-based representations of the rabbit ventricle. Thus far, we have shown that local mechanically-induced excitation is the initiator of electrophysiological responses with local deformation and that the spatio-temporal nature of applied strain is critical to outcome. Specifically, strain in diastole causes focal activation from the site of stimulation and ectopic excitation of the ventricles. Strain during repolarisation causes focal excitation followed by ventricular fibrillation if applied during the early T-wave of the electrocardiogram (vulnerable window), when there is spatio-temporal overlap of the repolarisation wave-front and mechanically stimulated tissue (creating an ectopic focus and region of functional block). This supports previous computational predictions of the mechanism for re-entry.  Furthermore, preliminary results show that block of cation non-selective stretch-activated channels with Grammostola spatulata mechanotoxin-4 (GsMTx-4) can eliminate the observed ectopy.Furthermore, preliminary results show that block of cation non-selective stretch-activated channels with Grammostola spatulata mechanotoxin-4 (GsMTx-4) can eliminate the observed ectopy. With the continuation of these studies, we hope to elucidate causally-linked events involved in mechanically-induced changes in cardiac rhythm important in pathophysiological states, while linking whole organ and sub-cellular mechanisms.

The aim of Sara's project is to investigate mechano-electric feedback mechanisms involved in the initiation and sustenance of cardiac arrhythmias during myocardial ischaemia, with the hope of elucidating new diagnostic and therapeutic options. Whole ventricular rabbit and human computational models of ischaemia-related changes in electro-mechanical activity are being developed using the cardiac simulation environment Chaste, informed and validated by optical mapping experiments in isolated hearts and clinical data from patients. Thus far, Sara has performed a combined experimental and computational investigation to quantify how photon scattering distorts optical mapping recordings in the presence of ischemia-induced transmural heterogeneities. Dual-wavelength optical mapping experiments, using camera frame-synchronized LEDs with alternating 470 nm (shallow tissue penetration) and 640 nm (deep tissue penetration) light, were performed on globally ischemic rabbit hearts. Control recordings showed no discernable difference in action potential morphology between the two excitation wavelengths. However, after 5 min of ischemia, clear differences appeared, with increased upstroke duration and decreased action potential duration at 640 nm compared to 470 nm excitation. Three-dimensional tissue simulations, representing both electrophysiological changes during no-flow ischemia and photon scattering effects, showed similar differences, demonstrating they are a result of enhanced wavefront curvature and variation in action potential duration due to increased transmural heterogeneities in ischemia. Furthermore, ischemic transmural heterogeneities increase the distortion of the optical signal, such that there is an increase in upstroke velocity and action potential duration compared to homogeneously ischemic tissue, an effect which is amplified with an increase in the penetration depth of excitation. This reveals that transmural heterogeneities in ischemia have a significant effect on optical mapping recordings and that dual wavelength excitation can be used to reveal these heterogeneities, which may be useful for investigating their role in the initiation and sustenance of arrhythmias.

The aim of Phil's project is to address uncertaintity and degeneracy in computational cardiac cell model parameters by examining the effects that model inputs have on the resulting output. Techniques for multi-dimensional parameter sweeping, visualisation, and comparison with experimental data are being developed, which will allow the determination of a physiologically appropriate parameters space for a given computational model, and allow investigations into how (patho-)physiological spatial variation in underlying tissue electrophysiology may contribute to arrhythmogenesis. Thus far, Phil has examined the effects of simultaneously varying six ion channel conductances in a rabbit ventricular action potential model. This has demonstrated a highly non-linear relationship between model parameters and output, with clear degeneracy present for some measured variables. Furthermore, it has revealed that common measurements, such as action potential duration, are not necessarily sufficient to characterise model validity. This is currently being compared to experimental data, with the hope of producing a fully characterised computational model for use in investigations of sources of ventricular arrhythmias.

It is desirable to combine optical mapping data (high temporal resolution, but generally relatively low spatial resolution) with structural information obtained at high spatial resolution, to investigate the interrelation of cardiac structure and function. However, current methods suffer from similar technical and economic difficulties as multi-parametric optical imaging (see above). To this end, we are extending Temporal Pixel Multiplexing technology (developed by Dr. Gil Bub) to use in whole heart optical mapping. By offsetting pixel-exposure times during capture of a single image frame, this novel imaging modality embeds high speed temporal information into still images, allowing simultaneous acquisition of multiple spatial and temporal resolutions (e.g. full-resolution images and high-speed image sequences at reduced resolution) with a single detector, allowing extraction of a range of spatio-temporal events from different regions of the heart.

One of the major disadvantages of optical mapping is loss of signal integrity with cardiac motion (either naturally occurring due to contraction or from extrinsic factors, such as mechanical stimulation). Various methods are used to minimise these ‘motion artefacts’, including physical restraint and pharmacological excitation-contraction uncoupling. However, these techniques invariably interfere with normal cardiac physiology, and performing investigations under conditions of normal contraction is preferable. Ratiometric approaches are effective in significantly reducing motion artefacts arising from motion perpendicular to the imaging plane; however they cannot compensate for motion within the imaging plane, where there is movement of material points in relation to the optical sensor. We are investigating the use of intensity-based registration algorithms (namely normalised cross-correlation of localised patches and advanced optical flow) for removing motion artefacts from optical mapping images of the moving heart. Thus far, we have shown that the optical flow approach compensates for the change in fluorescence intensity arising from changes in transmembrane potential, adversely affecting voltage signals, and thus requiring additional measures for these intensity-based registration algorithms to be used. However, a hybrid feature- and intensity-based algorithm which extracts surface landmarks (either naturally occurring or artificial markers), aligns them using optical flow, and interpolates the remaining regions using thin-plate splines, appears promising for reducing the amplitude of motion artefacts.

The human coronary tree is commonly assumed to have two main aortic roots, giving rise to one coronary artery on the left and one on the right. However, this generalisation may not be adequate, as a surprisingly large proportion of humans have two ostia in the right aortic sinus of Valsalva, giving rise to the right coronary artery and a separately emerging conus arteriosus artery. By high-resolution MRI scanning of explanted hearts, we found this 'third coronary artery' in 80% of humans and 100% of rabbits. With a flow-bed that feeds the right ventricular outflow tract and part of the upper septum, the potential presence of this vessel in a significant fraction of patients should be considered in heart rhythm research, diagnostics, and therapy. We are now investigating the functional relevance of this vessel by targeted occlusion and electrophysiological measurements by optical mapping in isolated rabbit hearts.

Increased resolution of cardiac MRI imaging and a growing interest in the effect of cardiac micro-structure on electrophysiology poses new challenges for the generation of cardiac atlases. This requires establishment of a standard coordinate system for the heart independent of the commonly used macro-structure. We are working towards developing methods for defining venticular descriptors that can be used to build a standardized reference frame, specifically exploring the use of the central curve of the left ventricular cavity and the smoothed internal envelope of the right ventricular crest (the curve of the endocardial surface marking the junction between the right ventricular free wall and septum). Using manual segmentation tools developed by Dr. Ramón Casero for the open source software platform Seg3D, that improves spatial coherence by providing visual feedback of the segmentation in real time, we have been deliniating cardiac anatomy, to generate rat- and rabbit-specific cardiac atlases, useful for computational modelling studies.