Nano-car OK, Nano-robot OK, Nano-neuron?

As soon as we hear the word neuron, we think about a creature having a spherical head and a long tail. When we say that we wish to mimic neuron in a molecule, we do not mean something that would look like this shape, rather, would carry out two most fundamental peculiarities of neurons, first, it would be a multilevel switch and second it should communicate with many others at a time.

See movie: A journey through our brain while large number of neurons are processing signal

Till date several neural network models have been developed, and all of them (to the best of our knowledge no multilevel neuron is used in any models) have used neuron flipping between binary states.

A neuron bursts beyond a threshold voltage, therefore normally it is considered that neuron is binary. However, beyond threshold bias connectivity increases continuously, to take that factor into account, we consider multilevel aspect of neuron. Multiple-valued logic has long history read wikipedia and follow current trends here.

If number of digital levels an analogue neuron can process and number of other neurons it can communicate at a time is equal, we would say that we have ideally realised a neuron.

Duroquinone based molecular neuron reported for the first time in the year 2006 which finally has led to the invention of nano brain reported in the year 2008.

Till now we have tested more than 60 molecules for the multi-level data processing applications. Though we have published only one molecule for multilevel switch application as shown in the above diagram.

Multiple negative differential resistance is the physical phenomenon which suggests first criterion of a molecular neuron that is multilevel data processing. However, for the second criterion we need to study all of them in their organic monolayer. Thats the challenge. The molecule that demonstrates reversible conductance switching between multiple NDR states have many different configurations and forms distinctly different monolayer at different conditions.

This would be one fo the major breakthorughs in this field, therefore we are zooming the phenomenon much more reproducibly before disclosing it to ope public.

Contact: Anirban Bandyopadhyay, National Institute for Materials Science, Tsukuba, Japan-305-0037