This article is part of our Summer reads series. Visit the full collection for book lists, guest essays and more seasonal distractions. Learning new things is hard.
Remembering what has already been learned is harder. Any successful learning system, be it a brain or a piece of artificial-intelligence software, must strike the right balance between stability and flexibility. It must be stable enough to remember important old things yet flexible enough to learn new ones without destroying old memory traces—preferably for as long as it exists.
Learning is a result of changes in the pattern of neural connectivity in the brain. Each connection between nerve cells, called a synapse, is a tiny gap between the ends of branches ramifying from such cells. Messages jump across these gaps in the form of molecules called neurotransmitters.
Current estimates suggest there are 600 trillion synapses in a human brain. How, then, to deal with the stability-plasticity dilemma—particularly as brains age and, as it were, fill up? Research by Dimitra Vardalaki, Kwanghun Chung and Mark Harnett at the Massachusetts Institute of Technology, just published in Nature, suggests one way is to preserve into adulthood a type of memory-forming synapse found in children. These are called silent synapses.
Silent synapses—which, as their name suggests, transmit no signal from one nerve cell to another—are often found on the ends of slender, immature protrusions from nerve cells, called filopodia. Until now, it had been thought that these disappeared as a brain matured. But Drs Vardalaki, Chung and Harnett have shown not only that they are present in adulthood, but also that they are common, at least in mice.
Read more on livemint.com