The human brain is comprised of more than 100 billion neurons with north of 1,000 trillion synapses. Each neuron is capable of both processing and storing data, making them far more efficient than the architecture found in traditional computers. The proof-of-concept memristor was built on a micro-scale and is about the size of a human hair. Lead researcher Feng Zhao tested the honey memristors’ switching speeds and found them to be comparable to human synapses (100ns on and 500ns off). They were also able to emulate functions called spike-timing dependent plasticity and spike-rate dependent plasticity, which assist with learning and retaining new information.
Honey was chosen due in part to its biodegradable and renewable nature. “Honey does not spoil,” Zhao said. “It has a very low moisture concentration, so bacteria cannot survive in it. This means these computer chips will be very stable and reliable for a very long time.” Zhao is also looking into the viability of proteins and other sugars, like those found in Aloe vera leaves, for similar applications. Future iterations will be developed on a nanoscale, or about 1/1000 the size of a human hair. By bundling millions or even billions together, the researchers eventually hope to construct a fully functional neuromorphic computing system. Their work has been published in the Journal of Physics D: Applied Physics. Image credit Three-shots
