Bioengineering

ActSole: The Property Adjusting Shoe Sole

Supervisor

Denys J.C. Matthies
Auckland Bioengineering Institute
Project code: ABI001

In future, a smart sports shoe would be able to sense the user's context and to change the property of the sole, such as the rigidity, profile, and cushioning. As the sensing part (e.g., recognizing the user's activity and the terrain the user is walking on) has been explored in previous projects already, it remains an open question about how to change the properties of the shoe's sole. This project has a great exploratory component, in which the student creates several flexible designs of shoe sole prototypes. The expected results are several iterations of 3D-printed designs of a shoe sole offering mechanisms to adjust its rigidity, profile, and cushioning.

Modelling mechanics of the stomach

Supervisor

Dr Kumar Mithraratne, Dist. Prof. Peter Hunter
Auckland Bioengineering Institute
Project code: ABI002

The stomach is the key organ in the digestive system. Although it has a relatively simple topology, its wall tissue structure is far more complex consisting of several layers. The smooth muscle, which also has several sub-layers, is found in the middle of the wall with muscle fascicles (fibre bundles) organised in different directions in different sub-layers. The vagus nerve controls the stomach activity. It contains 80% sensory fibres known as afferents and 20% secreto-motor fibres known as efferents. When food enters the stomach, the stretch receptors via afferent pathways inform the brain, which in turn through efferents activate the secretory glands and peristaltic action of the stomach by contracting the smooth muscle. Our overseas collaborators are currently acquiring high resolution imaging of the whole rat stomach at different fill levels and mapping nerve endings on the stomach wall. We can use this data to construct the stomach geometry at different fill levels accurately using finite element (piece-wise parameterised) meshes. This information can then be used to determine the tissue stretch via a suitable strain measure throughout the stomach and investigate whether there is a correlation between the afferent nerve ending distribution and stretch.

Modelling the interaction between Lactoferrin and Staphylococcus aureus biofilms

Supervisor

Dr Jagir R. Hussan, Dist. Prof. Peter Hunter
Auckland Bioengineering Institute
Project code: ABI003

Staphylococcus aureus is a major cause of nosocomial infections and forms treatment resistant biofilms. S. aureus can adhere to the surface of indwelling medical devices and is a major cause of infection in implant procedures. Further, the diversity of S. aureus virulence factors along with its ability to form disease-associated biofilm reduces the ability of the host immune defence to fully eradicate the bacteria. Consequently, the bacteria is highly tolerant to antibiotics, leading to persistent chronic infections.
Lactoferrin, a naturally occurring protein, exhibits antimicrobial activity and is a regulated component of the immune system. Some of the reported mechanism of its antibacterial include sequestration of free iron and destroying bacterial membranes through oxidization. It has also been observed to be effective against S. aureus biofilms, however there is not much direct evidence of how lactoferrin affects the pathogen. Side effects such as amyloidosis, developing iron-resistance and promoting horizontal gene transfer are serious concerns that need to addressed prior to its wider use. In this project, an agent based computer model will be developed to investigate the interactions between lactoferrin and S. aureus biofilms. In addition to representing the microbial ecosystem, lactoferrin concentrations, additional aspects such as iron-concentration, immune modulators, etc. will be modelled using detailed systems biology models. The model will be validated against in-vitro experimental data from our collaboration partners. An ideal candidate will have a background in microbiology, interest in systems biology, computer coding and analysis skills.

Integrated Physiological Data Collection – Design & Development of a Software Tool for Simultaneous Recording of Physiological Data from Multiple Sources

Supervisor

Mark Billinghurst
Auckland Bioengineering Institute
Project code: ABI004

The project aims to develop a ‘plug & play’ software platform that will allow researchers to connect multiple physiological sensing devices such as EEG, ECG, GSR, and eye trackers etc. to enable synchronised collection of such data. Current platforms such as the Lab Streaming Layer (LSL) require the experimenter to have some programming knowledge in order to integrate the tool into their workflow. The Lab Recorder that is available with LSL also requires similar knowledge in programming to be able to build and deploy in an appropriate manner. Such pre-requisites tend to hinder the ease with which multiple physiological sensing devices can be deployed. The software tool that we envision will allow for any person, irrespective of programming literacy, to easily collect physiological data from multiple devices with a single click.

Semantic Alignment of Chinese and New Zealand Acute Coronary Syndrome Registries and Development of a Data Quality Assessment Framework

Supervisor

Dr Koray Atalag (UoA), Xudong Lu (Zhejiang University)
Auckland Bioengineering Institute
Project code: ABI005

Disease registries are essential for driving clinical quality improvement and supporting care delivery, research, and health policy making. Interoperability between registries in different regions and countries is important for being able to aggregate a much larger pool of data and allow for comparative studies. We have been collaborating to apply the openEHR health informatics standard and clinical terminology systems (SNOMED CT and LOINC) to align the structure and semantics of the datasets of the two acute coronary syndrome (ACS) registries from China and New Zealand. The next step will be to further extend and validate the data quality assessment (DQA) framework developed by the Chinese team based on extensive experience gained in New Zealand with the mature ANZACS-QI ACS registry in the past 5 years. This framework will then serve to inform future development of the Chinese ACS Registry with the aim to conduct comparative benchmarking studies between the two countries in the area of CVD risk assessment and prediction. This includes a planned feasibility study to apply a deep learning based prediction algorithm developed by the Chinese team on NZ Registry data (conditional on ethics approval). The student will take an active role in helping with preliminary exploratory work for this study.

Specific goals and tasks for this project may include:

  1. Defining the data quality requirements for the ACS registries for collecting information about CVD for clinical research and decision-making.
  2. Extending and validating the existing DQA Framework
  3. Using openEHR to describe the data quality requirements of the two registries and defining corresponding data quality rules.
  4. Trialling and validating the DQA approach to assess the data quality of the NZ ACS registry.
  5. Comparison of semantic alignment and DQA results between the NZ and Chinese ACS registries
  6. Exploratory work to assess the feasibility and applicability of a Deep Learning based valuating an algorithm of prediction model on NZ ACS Registry.

Virtual Agent Emotional Study Data Analysis

Supervisor

Mark Billinghurst, Amit Barde
Auckland Bioengineering Institute
Project code: ABI006

The goal of this project is to explore people's emotional response to talking with a virtual character. A considerable amount of data has been collected from an experiment with people talking with a virtual character. This includes video and audio recordings and physiological data (heartrate/GSR), as well as questionnaire responses. Facial expression annotation and transcription has been completed as well as questionnaire analysis but there is still a considerable amount to do, in particular analysing the physiological data and aligning it against the other measures of emotion. The goal is to isolate key features that can be used to identify emotional cues and so help the character better recognize users emotions. This project would involve processing of the collected data (especially the physiological data) and generating a feature set suitable for deep learning training or other uses. The student should have skills/interests in data processing, Matlab (or similar), emotion recognition. It is a joint project wth the Department of Psychology.

3D Scene Sharing for Remote Collaboration

Supervisor

Mark Billinghurst, Huidong Bai
Auckland Bioengineering Institute
Project code: ABI007

The goal of this project is to explore how a 3D virtual model can be made of a person's real surroundings and shared with a remote person to improve remote collaboration. Depth sensor technology allows a 3D virtual model to be created in real time of a person's surroundings. We have developed technology that combines several sensors together to enable a full 360 scan to be completed at once, enabling live 360 3D data streaming. This project will develop software that will allow this 3D model to be streamed to a remote user who can experience it as an immersive Virtual Reality (VR), and collaborate back with the person sending the data. The goal is to enable remote collaboration as easily as if the people were in the same environment. It will also involve developing an Augemtned Reality (AR) interface for a head mounted display that will allow the local person see virtual collaboration cues from their remote partners. The student should have skills/interests in computer vision, AR/VR and ideally experience with the Unity game engine.

Experimental designs for identifying mechanical properties of skin

Supervisor

Dr Prasad Babarenda Gamage, Prof. Poul Nielsen, Prof. Martyn Nash
Auckland Bioengineering Institute
Project code: ABI008

Simulating the mechanical behaviour of skin using biomechanical models is useful for a wide range of applications from simulating surgical interventions to helping to create more realistic models of the face for the animation industry. Existing biomechanical models of skin require many parameters to describe the complex mechanical behavior observed in real skin. Attempting to identify these parameters using ad-hoc experimentation protocols often results in non-unique parameter estimates. This situation arises because of difficulties in predicting whether the experimental protocols being applied can provide sufficiently rich information for robustly identifying the model parameters. A model-based design of experiments framework has recently been developed at ABI to help address these issue by determining experimental protocols that maximise the identifiability of the mechanical parameters.

The main aim of this project is to conduct experiments to validate our model-based design of experiments framework and apply it to computational models of skin. The framework will first be validated by performing experiments on silicone gel phantoms. The framework will then be applied to determine optimal indentation protocols for identifying mechanical properties of skin using existing computational models and a novel micro-robot indenter developed at ABI for applying controlled deformation to skin. The student would ideally be keen on computer modelling and will develop skills in finite element modelling, design of experiments techniques, nonlinear parameter optimisation, and code development using Python and Matlab.

Project Aims:

  1. Validate the design of experiments framework by performing a series of indentation and gravity loading experiments on phantoms made of silicone gel.
  2. Apply the design of experiments framework to existing biomechanical models of skin to determine the optimal indentation trajectories that maximise the identifiability of the mechanical parameters in the model.
  3. Perform the optimal indentation protocol on the forearm skin of an individual using the micro-robot and identify their individual-specific mechanical parameters using an existing numerical optimisation algorithm.

Biomechanics for breast cancer imaging

Supervisor

Dr Prasad Babarenda Gamage, Prof. Poul Nielsen, Prof. Martyn Nash
Auckland Bioengineering Institute
Project code: ABI009

Breast cancer affects 1 in 9 New Zealand women. Early detection is key to improving the likelihood of survival. The ABI’s Biomechanics for Breast Imaging group is developing an automated, biomechanics-based image processing pipeline in collaboration with clinicians at Auckland City Hospital to address challenges in detecting and treating breast cancer.

This pipeline automatically builds computational models of the breast from clinical magnetic resonance images (MRI) acquired in the prone position, and uses these models to predict the locations of suspected tumours in the supine position (in which treatment procedures are performed). This is helpful during diagnostic and treatment investigations, because supine MRI are generally not available. To accurately predict breast deformations, individual-specific estimates of the mechanical properties and the load-free shape of the breast are necessary. The accuracy with which existing biomechanical models can predict breast deformation needs to be analysed to identify sources of error, and to improve our ability to track and locate suspicious tumours in the breast.

Interested students will have the opportunity to work with our clinical collaborators at Auckland City Hospital. Students would ideally be keen on biomechanical modelling, and will (further) develop skills in image processing, nonlinear parameter estimation, statistical regression, and some code development using Python and Matlab.  

Project Aims
This project consists of two main parts: (A) development and validation of a robust approach for simultaneously estimating the mechanical properties and load-free shape of the breast; and (B) quantifying and improving the accuracy of the clinical biomechanics pipeline using available images from healthy participants and patients.

Part A:

  • Develop a combined biomechanical/statistical approach for simultaneously identifying the mechanical properties and load-free shape of the breast, and verify the methodology/implementation using synthetically generated data. This will build on a novel, previously prototyped approach, which uses principal component analysis to substantially reduce the number of parameters required to represent breast shape and properties.
  • Validate this approach using an existing database of prone and supine MRIs from a population of healthy participants.

Part B:

  • Quantify the accuracy of an existing breast biomechanics pipeline for predicting breast tissue deformation. This will involve comparing the model predicted motion of landmarks within the breast from the prone to supine positions, with deformations determined manually by expert clinical collaborators.
  • Investigate strategies for improving regional accuracy of the biomechanical models.

PBPK modelling of hepatobiliary transporters and enzyme-mediated metabolism processes

Supervisor

Dr Harvey Ho, Zeng Su (Zhejiang University)
Auckland Bioengineering Institute
Project code: ABI010

Elucidation of the rate-determining process in the overall hepatic elimination of drugs is critical for predicting their hepatic clearance and their systemic and regional exposures. A physiologically based pharmacokinetic (PBPK) model that includes the transporter-mediated membrane transport and enzyme-mediated metabolism processes is a powerful tool to investigate the effect of changes in transporter (influx, efflux) function and metabolizing enzyme function on the pharmacokinetics of drugs in the blood and the liver. The prediction efficacy of such an in silico tool however is affected by the limited data for determining parameters in the model. Hence numerical estimation/optimisation methods are required to evaluate these parameters and to validate a PBPK model.

The aim of this project is to learn a novel PBPK approach, whereby the parameter sensitivity analysis is performed through a Cluster Newton Method and/or a 'hypercube' algorithm, so as to elucidate the quantitative relationship between the transport activities and drug response. We will collaborate with Prof. Yuichi Sugiyama, RIKEN Cluster for Science, RIKEN, Yokohama, Japan to analyse the plasma clearance data of drugs in PET imaging and to evaluate the transporter function in vivo including PET probes for hepatic uptake transporters (OATP1B1, OATP1B3) and biliary excretion transporters (MRP2, BCRP) both in experimental animals and in human. Through this project the student is expected to gain a good understanding of the basal and apical transporters, metabolism pathways modelling techniques.