What are Memristors?
Memristors are two terminal devices that can be programmed to take a number of memory states at very low power. Their nanometre-scale size and capacity for storing multiple bits of information per device means electronics industries are interested in exploiting these devices to enhance their memory applications as they could store more data on smaller hard disks at much less power consumption.
How can these devices be applied to biological platforms?
The dynamics of memristors actually resemble the functioning of chemical synapses that form the links between neurons in our brain. In the electronics industry, all the efforts go into minimising any mechanisms that prohibit the storage of memory states in the long-term, but we can advantageously apply this to neural networks as biological synapses are equally volatile. Their temporal characteristics means memristors can be programmed to emulate the formations of not just long-term, but short-term memory events so we can look at exploiting these devices to create more realistic neuromorphic systems; advanced electronic systems that mimic the functioning of biological neural networks.
What are the benefits?
In gaining a better understanding of neural networks we can boost research and development in cross-disciplinary areas such as neurobiology and develop improved diagnostic tools and treatments for patients suffering from neurological disorders. This research represents a fundamental step towards the creation of ‘autonomous intelligent systems’ for robotics applications and neuroprostheses, with a tremendous societal impact.