Scientists build an artificial neuron chip that recognizes biological signals in real time

A team of researchers in Zurich has developed a compact, energy-efficient device made from artificial neurons that can decode brain waves. The chip uses data recorded from the brain waves of people with epilepsy to identify which areas of the brain are responsible for seizures. This opens up new applications for treatment.

Current neural network algorithms produce impressive results and help solve a surprising number of problems. However, the electronics used to run these algorithms still require a lot of processing power. When it comes to processing sensory information or interacting with the environment in real time, these artificial intelligence (AI) systems simply can't compete with the actual brain. And neuromorphic engineering is a promising new approach that Bridges the gap between artificial and natural intelligence.

Using this approach, an interdisciplinary research team from the University of Zurich, ETH Zurich and University Hospital Zurich has developed a chip based on neuromorphic technology that can reliably and accurately identify complex biological signals. The scientists were able to use this technique to successfully detect previously recorded high-frequency oscillations (HFOs). These specific waves, measured using an intracranial electroencephalogram (IEEG), have proved to be promising biomarkers for identifying the brain tissue that causes seizures.

The researchers first designed an algorithm to detect HFOs by mimicking the brain's natural neural network: a tiny so-called spike neural network (SNN). The second step is to implement SNN in a fingernail-sized piece of hardware that receives nerve signals through electrodes and, unlike traditional computers, has huge energy efficiency. This makes computing with very high temporal resolution possible without relying on the Internet or cloud computing.

"Our design enables us to identify spatiotemporal patterns in biological signals in real time," said Giacomo Indiveri, a professor at the Institute of Neuroinformatics at the University of Zurich and ETH Zurich.

The researchers are now planning to use their findings to create an electronic system to reliably identify and monitor HFOS in real time. When used as an additional diagnostic tool in the operating room, the system can improve the outcome of neurosurgical interventions.

However, this is not the only area where HFO identification can play an important role. The team's long-term goal is to develop a device that can be used outside of hospitals to monitor seizures, which would make it possible to analyse signals from a large number of electrodes over weeks or months.

Johannes Sarnthein, a neurophysiologist at the University Hospital Zurich, explains: "We wanted to incorporate low-energy wireless data communications into the design -- for example, to connect it to a mobile phone. Portable or implantable chips like this can identify periods when seizure rates are higher or lower, which will allow us to deliver personalised drugs."