Electronics engineers can play a greater role in medical breakthroughs, according to a researcher at the University of South Australia.
Mark McDonnell, associate professor at the University of South Australia, has been researching computational neuroscience since 2009. While the term is relatively new in Australia, he said engineers play an important role in the field.
The term computational neuroscience was first coined in the 1980s, and is loosely defined as the study of the brain using computational techniques, such as analysing and understanding cell and circuit behavior.
“There are billions of cells in the brain and the nervous system that have the property of conducting electricity. Given this, there’s a strong connection with electronic engineering, which is about designing artificial systems which function using electricity,” McDonnell said.
McDonnell said neuroscience is about understanding something that already exists. For example, looking at how a brain operates on a computational level and what goes wrong when there is a brain disorder.
“The goal of neuroscience is understanding the building blocks themselves, which are the neurons – the brain cells – and how they work together in large networks to result in … the ability to take in information from the world to control our emotions, for example,” McDonnell said.
On the other side, engineering is about designing systems that can be built from scratch. This is where engineers come in – they are integral to designing the sensors and diagnostic equipment that help neuroscientists. This equipment can then be used to examine disorders such as epilepsy, schizophrenia and Alzheimer’s disease.
Engineers have previously been involved in developing bionic devices, such as the cochlear implant to help people hear. McDonnell said there is also a push to develop bionic eyes, an area that is being heavily funded in Australia.
McDonnell is currently carrying out research around neuromorphic engineering and building electronic systems that function as closely as possible like real neurons in the brain.
“This means designing custom electronic chips and circuits which function a bit differently to our normal electronics,” McDonnell said. “Neuromorphic engineering is one approach to try to do better with our electronics and being able to do that requires substantial input from neuroscience.”
That involves studying real neurons in the brain and its networks using simulation.
McDonnell said the most recent breakthrough in neuromorphic engineering has been vision sensors that function like real eyes – information flows from the eyes and down optic nerves to the brain.
A practical area of his research is cochlear implants and trying to predict how much better they could be if there were more electrodes and they were better positioned. It also includes factoring in improved electrode array technology in the future.
“(In artificial intelligence), one of the goals is to design algorithms which enable computers to automatically recognise patterns or extract information from large datasets,” McDonnell said.
“In my own research, I aim to design algorithms based on the knowledge of how neurons in the brain work together in networks to perform that task in biological organisms.
“Ultimately, the application is in better algorithms for machine learning and artificial intelligence inspired by actual biology.”
Working in computational neuroscience can have an added layer of complexity – McDonnell said it’s one thing to take inspiration from nature for robotics, but it’s a very different story to actually work with biology. In this case, he said it’s vital for engineers to know how biology works.
“That takes time to get to that level of knowledge, or at least a very close working relationship with someone who has that knowledge,” he said.
This can pose obstacles for engineers looking to enter the field. However, McDonnell said having the mindset of a robotics designer can be a valuable asset in understanding basic science.
“Essentially it’s a reverse engineering problem to understand how neurons in the brain can function computationally. It’s a difficult problem, but I think one of the best ways to attack it is to think like a designer – how can we put together a system with building blocks that have these properties that achieve some goal?” McDonnell said.
McDonnell is adamant that putting these building blocks together requires a shared understanding and vision, and is not just about one profession informing the other.
“Often it’s a very interdisciplinary team that’s needed,” McDonnell said. “The cochlear implant is a classic example where you needed involvement from surgeons who can do the actual surgery, and the electronic engineers who design the sound processing and electrode arrays.”
In the future, he believes the relationship between engineers and neuroscientists will continue to develop and there will be a global trend towards more research and interest in neuromorphic electronics.
But McDonnell said Australia is lagging behind other countries when it comes to computational neuroscience. For example, Germany has a large presence in the field and has centres for computational neuroscience. But he is hopeful things will change in Australia.
“When I visit those places, it’s amazing what they’ve managed to bring together in the same place in terms of people from different fields working towards common goals,” he said.