Kostas Kostarelos, Professor of Nanomedicine at The University of Manchester and the Catalan Institute of Nanoscience and Nanotechnology (ICN2) in Barcelona, Spain, says developments at the intersection between smart materials and artificial intelligence mean we are on the verge of a breakthrough that could help us better learn the workings of the human brain.

As the development of artificial intelligence (AI) accelerates, there is a fascination about the potential for AI to become ‘human’, and whether it can accurately reflect human thought processes.

While this question – for the time being at least – remains unanswered, our research teams are exploring how we can use AI to accurately help understand the brain’s function from a physiological perspective.

As advancements in both AI and smart materials like graphene continue, we are starting to develop new and powerful tools for exploration of the cerebral space and function, which can lead to significant medical advancements. More than this, it opens up the potential to better understand the workings of the brain – of memory, the way we think and our capacity to interact with our environment.

Opening the door to medical advancements

At The University of Manchester and ICN2, we have worked with many different research institutions from across Europe to create a flexible graphene depth neural probe (gDNP), which consists of a millimetre-long linear array of micro-transistors imbedded in a micrometre-thin polymeric flexible substrate.

As a result, for the first time we can observe brainwaves at the lowest end of the frequency spectrum (infra-low) with the greatest precision ever achieved. Infra-low brainwaves (also known as Slow Cortical Potentials) are believed to be implicated in basic cortical rhythms, the electrical oscillations taking place in the brain – and these underline various brain functions.

Very little is known about infra-low brainwaves because it is difficult to detect them accurately with current equipment. However, they appear to play an important role in brain timing and network function[1].

This is incredibly exciting for the neuroscience community. It opens up a whole new frontier, particularly from a medical perspective. One important such application will be to better understand and help cure the various pathologies that affect the brain, such as epilepsy. Research programmes across the UK and Europe aim to discover how new graphene-based neural probes can improve detection of epileptic brain signals, even before seizures occur.

This is just the first step – sophisticated graphene brain sensors could help us to better investigate other brain disorders. Previously we couldn’t properly record, for example, the electrophysiological phenomena associated with painful migraine headaches, because this occurs too low in the brain’s bandwidth for current electrodes to accurately detect. But now we can look to better observe this pathology too.

Exploring beyond the physical with AI

If this wasn’t exciting enough, there is potential for graphene-enhanced probes to go beyond pathology and to discover much more about the workings of the brain that have been invisible up until now.

As researchers gradually gain much better access and recording capabilities of signals being emitted at the lower and higher brain frequencies, artificial intelligence will be essential in our ability to interpret this plethora of new data.

AI analytics will provide the necessary tools and techniques to discover insights, identify patterns and discover relationships in this new neurological data – and in collaboration with neuroscientists, electrophysiologists and neurologists unlock some of the complex processes involved in our higher functions.

Steve Furber, ICL Professor of Computer Engineering and a senior research leader at The University of Manchester, has developed a machine called SpiNNaker (short for Spiking Neural Network Architecture). SpiNNaker supports computer models of systems that work in ways that are similar to the brain, enabling researchers to simulate areas of the brain and to test new hypotheses about how it might work.

SpiNNaker and gDNP and are complementary to one another – gDNP takes a “top down” approach, aiming to constantly dive deeper into the human brain. SpiNNaker enables a “bottom up” approach by building large-scale models of the brain and attempting to emulate its functions.

By combining this mutual knowledge, it may be possible to create an AI-led “decoder” which will eventually open a new door to understanding human pathology and go beyond just medical applications, to exploration of perception, capacity and conscience.

Society is not quite entering the realm of science fiction just yet, but ongoing developments are taking us closer to being able to understand considerably better the functions of the human brain with AI and smart materials. It is helping to get inside the human brain like never before, and the potential is almost boundless.

[1] https://www.ncbi.nlm.nih.gov/sites/ppmc/articles/PMC4362698/