Brainy robot predicts danger

Researcher Nik Kasabov (left) and student Nathan Scott. Photo / Michael Craig

Kiwi researchers say their invention will save lives and help paralysed people to move

New Zealand researchers have created a machine with the potential to accurately predict strokes, help the paralysed to move and to perform Jedi mind tricks.

It could also be used to help predict earthquakes.

The NeuCube was developed by a team led by Dr Nik Kasabov, director of the Auckland University of Technology's knowledge, engineering and discovery research institute.

"We looked at the brain for inspiration - how the brain deals with this enormous amount of information we learn every day," he told the Herald on Sunday.

NeuCube is based on "spiking neural networks" and uses information-processing principles observed in the brain. The machine can deal with complex data including ecological and environmental, seismic, radio astronomy, health and brain signals.

NeuCube had the potential to be a life-changer, and even a life-saver.

Working with AUT's National Institute for Stroke and Applied Neurosciences, the team compared brain scans of people who had a stroke with those of healthy brains.

The machine was able to predict retrospectively with 95 per cent accuracy whether someone was going to have a stroke in the next 24 hours.

This may see the machine being used in hospitals to predict the risk of stroke.

Rehabilitation robot developed at the China Academy of Sciences Institute of Automation.

And working in collaboration with the Chinese Academy of Sciences, Kasabov's team is planning to use NeuCube to develop exoskeletons that may enable paralysed people to move using the power of thought by transforming brain signals into movements of robotic arms or legs.

One of Kasabov's PhD students, Nathan Scott, is also working with robots and master's student Akshay Gollahalli is trying to create a drone aircraft which he can operate using signals from his brain.

Kasabov and his team have also been trialling NeuCube with seismic information, using historic earthquake data collected in New Zealand.

The machine retrospectively predicted all 14 of the last major earthquakes in New Zealand to within 24 hours.

Kasabov - who recently returned from India where he was the only invited New Zealand speaker at the first Commonwealth Science Conference - said there was potential the machine could be used to evaluate real-time seismic activity in a bid to predict a pending major quake.

The NeuCube

• NeuCube is able to interpret complex data beyond the capabilities of the most powerful computers.

• Health implications include predicting strokes and establishing treatments.

• It could more accurately predict natural disasters like earthquakes and tsunamis.

• It can interpret complex brain signals, potentially enabling paralysed people to walk in robotic exoskeletons controlled by their thoughts.

• Other uses include interpreting business and financial data; cyber-security data; social media data; and audio/visual data.