Computational modelling of complex physiological systems and their interaction with medical devices forms a major area of bioengineering research worldwide. Biomedical engineers use computer simulations to provide quantitative and qualitative data for developing safe and efficacious medical devices, as well as gaining insights into normal and abnormal physiological function.
In the Graduate School of Biomedical Engineering (GSBmE), various computational modelling research projects are currently underway, including those in the areas of electrical activation of excitable tissue, neuroscience, heat treatment of tumours, and biomechanics. Most of these projects involve the development of finite-element models of devices and physiological systems incorporating realistic anatomy and function, and many involve extensive validation of models using in-vitro data and imaging obtained from various laboratory facilities in the School.
Specific computational modelling projects being undertaken are described below.
A 3D realistic model of atrial propagation is being developed using ionic mathematical models able to reconstruct electrical activity of individual cardiac cells. The cell models have been fitted to action potentials (APs) recorded intracellularly in our laboratory from intact tissue preparations. Parameter optimisation techniques developed in GSBmE allows direct incorporation of experimental data into anatomically realistic geometries, and is an important step towards developing patient-specific models for the treatment of atrial arrhythmias.
Retinal Electric Stimulation
A major modelling effort is currently underway in GSBmE to develop finite-element models of the retina and its electrical activation by a neuroprosthesis implant. The models will inform stimulation strategies for a vision prosthesis currently being developed by the Bionic Vision Australia (BVA) consortium, which includes investigators from GSBmE. BVA was awarded a four year ARC Special Research Initiative in Bionic Vision Technologies to develop a neuroprosthesis to restore sight to individuals blinded by degenerative retinal disease. The computational model will incorporate data acquired from in-vitro and in-vivo experiments, and will require extensive implementation of non-linear parameter optimisation methods developed within GSBmE.
Brain Activation During Electroconvulsive Therapy
Electroconvulsive therapy (ECT) is the most effective treatment for severe depressive disorder, yet the mechanisms underlying its therapeutic effect remain largely unknown. A novel computational model is being developed in GSBmE to simulate and investigate direct brain excitation by ECT, using an anatomically-accurate finite element model of the human head. The model includes the skull with anisotropic electrical conductivity, fatty tissue layers, as well as an active neural model of brain tissue. The model will provide insights into brain activation during ECT, informing appropriate stimulus parameters and electrode positioning on the scalp to maximise therapeutic efficacy.
Large-Scale Parameter Optimisation of Ionic Cell Models
Biological systems models typically contain many parameters, many of which have not been determined experimentally. Parameters of various subsystems are often based on values reported in separate experiments conducted on disparate cells, tissues or species. In GSBmE, we have developed custom methods for large-scale parameter optimisation of ionic cell models based on the concept of curvilinear gradient search in multidimensional parameter space, as well as clamping model outputs to experimental data and fitting to the feedback currents. The methods are being used to simultaneously fit multiple action potential data recorded from paced cardiac tissue for developing computational models of cardiac arrhythmias, as well as fitting models to neural spiking patterns recorded from retinal ganglion cells.
Cardiovascular–Rotary Blood Pump Interactions
Implantable rotary blood pumps (iRBPs) have potential as bridge-to-transplantation and destination therapy devices for end-stage heart failure patients. We are developing a computational model to simulate the response of the heart to an iRBP, in order to investigate the dynamics of assisted circulation. Such a model will assist in the development of robust physiological pump control algorithms by allowing reproducible numerical experiments under identical conditions.
Key Contact: Assoc Prof Socrates Dokos
Top – down and bottom – up study of organs, tissues, and cell mechanophysiology
Tissues such as bone exhibit spatial and time dependent behavior as well as remarkable adaptive properties in response to biophysical forces inherent to living in Earth’s gravitational environment. To understand physiology and pathophysiology of the complex organismal system of the human body, we build virtual models. Some of these models include virtual representations of organs comprising tissues, cells and their extracellular matrix (top-down models).
Using these models, we can predict how exercise or bedrest affects tissue structure and function. For instance, computational poroelastic (finite element) models predict the role of mechanical loading on transport of molecules through the complex tissues of the human joint. Furthermore, nano-micron scale computational fluid dynamics models of percellular fluid flow demonstrate how natural narrowings in canalicular spaces amplify shear stresses at regular intervals along the osteocyte process (bone cell projections implicated in mechanosensation). Similarly, using bottom – up models, we can predict how stem cells adapt to subtle biophysical signals engendered through proliferation, growth and cell-signaling cascades. These models help us to predict generation of tissue in development and healing scenarios. We use this information to optimize tissue engineering template, implant and device designs.
Engineering of novel, biologically-inspired materials and devices
The smart, or stimuli responsive properties, of tissues and other natural materials provide ample inspiration for development of novel materials. Using virtual computational models of physiological systems of interest, it is possible to unravel the key system parameters underlying such smart properties. Such models have led to the invention of flow-directing and mechanoresponsive materials.