Summer research experience program

Biomedical Engineering Summer Research Program (SREP)

The BME Summer Research Experience Program (SREP) provides students a meaningful experience in a Biomedical Engineering research laboratory. These placements help inform future career choices, including Research or Research & Development pathways via PhD study or working in research laboratory positions as a Masters graduate.

Students participating in SREP will undertake work experience with academics and research staff in the Department of Biomedical Engineering, gaining valuable exposure to the world of research. A limited number of placements are offered and are allocated on a competitive basis.  Projects are sourced either from the list below, or by students approaching individual academics requesting supervision in a topic in which they are interested.

For the first time, the SREP is available either as a 12.5 point Approved Elective in the Master of Engineering (Biomedical) or as a Volunteer Placement. Time commitment, expectations and outcomes for each stream vary accordingly.

  • Approved Elective (12.5 point) subject: The SREP subject will run over the summer semester, from 4th January to 27th February 2022. The handbook entry for this subject will be available from the 15th November 2021.
  • Eligibility
    1. Current enrolment in UoM Master of Engineering (Biomedical or Biomedical with business) and study plan space for an Approved Elective.
    2. 75% WAM or better.
  • Volunteer Placements: These are unpaid, voluntary work experience opportunities, and can occur between November 2021 - February 2022 as agreed upon by the student and supervisor. Students can volunteer up to 10 hours per week.
  • Eligibility
    1. Current enrolment in a UoM Master of Engineering degree or commencing 3rd year of a UoM Bachelor of Science or Bachelor of Biomedicine degree in 2022
    2. H2A average or better.

The Summer Research Experience Program runs subject to University of Melbourne restrictions related to the COVID-19 pandemic.

How to apply

If you meet the eligibility criteria, please send Natasha Baxter (nbaxter@unimelb.edu.au) an email from your UoM student account containing the following information:

  1. Name
  2. Student number
  3. Current postal address
  4. Current phone number
  5. Whether you would like to take SREP as an Approved Elective or a Volunteer Placement
  6. Your top three projects in order of preference (if selected from the list below), or attach proof of correspondence with a supervisor agreeing to supervise you in a different project
  7. Why you are interested in undertaking the program and your plans for further studies
  8. Attach your UoM Statement of Results to-date, and your undergraduate transcript if it was completed at a non-UoM institution.

Applications should be submitted by 17 November 2021 to guarantee consideration. It is expected that decisions will be announced by 26 November.

Projects List

  • Visual Neuroscience Summer Research Experience (Multiple Students)

    Supervisor: A/Prof. Hamish Meffin

    Visual neuroscience studies how the brain processes visual information to perceive the world around us. It has applications in neural implants and artificial intelligence. This experience will introduce you to a range of techniques and methods used in visual neuroscience, including some of the following (as permitted by COVID restrictions): optical imaging and multielectrode recording of cortical activity, data analysis, including topographic map estimation and spike sorting, and mathematical modelling and simulation of visual processing.  For those who may be interested in doing a PhD in neural engineering or neuroscience, this experience will provide a taste of what it is like to do research in a neuroscience lab.

  • Multi-scale modelling of the heart for cardiology applications (1 or 2 students)

    Supervisor: Dr. Vijay Rajagopal

    Multi-scale modelling of the heart for cardiology applications (1 or 2 students)

    We develop patient-specific digital twin models of patient hearts that can be used to test treatment strategies for a variety of significant heart diseases including arrhythmias and heart failure. In this project the student will learn finite element modelling and application of computer modelling to cardiology applications.

    Identifying new drug targets for heart disease that harness exercise signaling pathways (1, or 2 students)

    In collaboration with colleagues at the Baker Institute, we are apply computational systems biology and high-throughput single-cell measurement techniques to investigate ways to identify new drug targets to treat heart disease. The heart is a muscular organ that adapts to long-term changes in blood supply demand by changing its size. In many heart diseases this adaptation does not work well and is classified as pathological heart growth. But there is another type of heart growth that is beneficial to our health. This physiological heart growth occurs when we do endurance training. In this project students will be working with our biomedical engineering PhD students and physiologists at the Baker to develop a computational model of signalling pathways related to pathological and physiological heart growth . We will use the model to identify strategies that activate exercise-related hear growth to treat heart disease. Students will be exposed to a range of techniques including live microscopy, image processing and data analysis, as well as mathematical and computational modelling.

  • Synthetic biology: encoding cellular recognition circuits to improve targeted medicine

    Supervisor: Dr. Matt Faria

    Truly “targeted” treatments, which only affect chosen diseased cells while ignoring healthy tissue could be regarded as the holy grail of medicine. One strategy to accomplish this is to functionalize drugs with antibodies to diseased cells. Ideally, antibody functionalization will cause drugs to preferentially bind and be taken up by diseased cells. But in practice, this approach often fails. Functionalized antibodies can be shielded by biomolecules in the blood prior to reaching their target; and many drugs are recognized, taken up, and/or filtered by elements of the immune system.

    We are interested in developing a complementary approach to antibody targeting, which we term “post-uptake cellular recognition”. This approach is to use the techniques of synthetic biology to encode a simple “program” using DNA. This program will recognize the type of cell the cargo has been delivered to (i.e., diseased or healthy), and do different things within diseased cells vs healthy ones (i.e., kill diseased cells and do nothing in healthy cells).  When a cell takes up the DNA cargo, existing cellular transcription machinery will “execute” this program, leading to a new frontier for targeting.

  • Deep Learning for Neuroscience – Electroencephalography (EEG) Analysis (Multiple Students)

    Supervisor: Dr. Joe West

    EEG measures electrical activity of the brain and an important line of research is mapping these EEG signals to human functions. For people with neurological disorders EEG can be used to assist with diagnosis and characterisation of an illness but an existing challenge is to detect onset of symptoms before they occur. If a signal can be found within the EEG which indicates the impending onset of debilitating symptoms then therapeutic measures can be taken to alleviate those systems. The key challenge is that these signals of interest are very small, if they exist at all, and reside within a large noisy background. Depending on the student’s interest I am open to investigating the application of any deep learning method to analysing EEG. Open source EEG signals are available as are a number of open source deep learning systems which could be used as the starting point for the project. Suggested topics include:

    EEG representation analysis. Build a simple EEG deep learning system and report on the impact of different EEG pre-processing methods/representations.

    Deep Fake EEG. Using generative adversarial networks, build a system which can generate realistic fake EEG.

    Build EEG Brain-Stimulation Game. Overlay AI-Gym control dynamics with EEG signals and build a reinforcement learning agent that can play the game.

    Transformer EEG Processing. Build a transformer based neural network that predicts the next time slice of EEG given a prior sequence.

    Other EEG Signal Processing with Deep Learning. Open to student ideas if they have one.

  • Improving patient-specific seizure forecasting with population data 

    Supervisor: Dr. Pip Karoly

    This project will utilise a large, existing database of mobile seizure diaries from people with epilepsy. The objective is to use statistics and machine learning to explore how population-level trends (i.e., common daily and weekly patterns) can be used as a prior likelihood to improve individual's forecasts. The research student will work closely with the team at Seer Medical as part of our pilot study of a mobile seizure forecasting app. The project can be undertaken via remote work. The student will be able to learn and develop skills in Python programming and gain insight into software development.

    Incorporating sleep data into personalised forecasts of seizure likelihood 

    This project will utilise a large, existing database of long-term sleep monitoring recorded from a wearable smartwatch (Fitbit) in conjunction with mobile seizure diaries from people with epilepsy. The objective is to use machine learning to explore how various sleep features (such as wake time, bedtime, sleep stages) relate to individuals’ seizure likelihood. The research student will work closely with the team at Seer Medical as part of our pilot study of a wearable and mobile seizure forecasting app. The project can be undertaken via remote work. The student will be able to learn and develop skills in Python programming and gain insight into software development.

  • Changes in knee joint kinematics with increased lateral tibial slope (Multiple Students)

    Supervisor: A/Prof. David Ackland

    The slope of the tibial plateau is known to increase risk of anterior cruciate ligament (ACL) tear and re-tear of ACL reconstructions; however, the role of the tibial slope on knee joint motion remains poorly understood. The objective of this study will be to evaluate the influence of varying degrees of knee joint slope on knee kinematics by employing a cadaveric knee model. This project will involve close collaboration with orthopaedic surgeons, and the results are expected to provide guidance for tibia re-alignment surgery.