Artificial neurons behave like real ones

The authors of a bio-engineering study just published in the journal Nature Communications exhibit the kind of calm rationality that makes one believe.

They point out that while a range of neuromorphic silicon devices replicating biological nerve functions have been proposed, a number of problems have hampered the up-to-now theoretical attempts to develop them. 

Devices proposed include silicon neurons, synapses and brain inspired networks, but their designs, say the authors, were not meant to copy the behaviour of biological cells, but to search for the organising principles of biology that can be applied to practical devices.

However, an increasing focus on implantable bioelectronics to treat chronic disease is changing this paradigm, they say, and “instilling new urgency in the need for low-power analogue solid-state devices that accurately mimic biocircuits”.

The joint British/Swiss/New Zealand team’s paper describes a way of making silicon chips that are much smaller than a fingertip but reproduce the electrical behaviour of biological neurons.

The approach, they say, could lead to the development of bionic chips to repair biological circuits in the nervous system when functions are damaged or lost to disease.

They designed microcircuits modelling ion channels that integrate raw nervous stimuli and respond in a similar way to biological neurons.

They then recreated the activity of individual hippocampal and respiratory neurons in silicon chips. In a series of 60 electrical stimulation protocols, they found that the solid-state neurons produced nearly identical electrical responses when compared to biological neurons.

“Our approach combines several breakthroughs,” says lead author Alain Nogaret, from the UK’s University of Bath.

“We can very accurately estimate the precise parameters that control any neurons behaviour with high certainty. We have created physical models of the hardware and demonstrated its ability to successfully mimic the behaviour of real living neurons.

“Our third breakthrough is the versatility of our model which allows for the inclusion of different types and functions of a range of complex mammalian neurons.”

Nogaret and colleagues note that respiratory neurons, such as those they have modelled, couple respiratory and cardiac rhythms and are responsible for respiratory sinus arrhythmia.

Loss of this coupling through age or disease is a prognosis for sleep apnoea and heart failure. They suggest that a device that adapts biofeedback in the same way as respiratory neurons may offer a potential therapy in the future.