How Are Memories Stored and Retrieved?

See allHide authors and affiliations

Science  01 Jul 2005:
Vol. 309, Issue 5731, pp. 92
DOI: 10.1126/science.309.5731.92

Packed into the kilogram or so of neural wetware between the ears is everything we know: a compendium of useful and trivial facts about the world, the history of our lives, plus every skill we've ever learned, from riding a bike to persuading a loved one to take out the trash. Memories make each of us unique, and they give continuity to our lives. Understanding how memories are stored in the brain is an essential step toward understanding ourselves.

Neuroscientists have already made great strides, identifying key brain regions and potential molecular mechanisms. Still, many important questions remain unanswered, and a chasm gapes between the molecular and whole-brain research.

The birth of the modern era of memory research is often pegged to the publication, in 1957, of an account of the neurological patient H.M. At age 27, H.M. had large chunks of the temporal lobes of his brain surgically removed in a last-ditch effort to relieve chronic epilepsy. The surgery worked, but it left H.M. unable to remember anything that happened—or anyone he met—after his surgery. The case showed that the medial temporal lobes (MTL), which include the hippocampus, are crucial for making new memories. H.M.'s case also revealed, on closer examination, that memory is not a monolith: Given a tricky mirror drawing task, H.M.'s performance improved steadily over 3 days even though he had no memory of his previous practice. Remembering how is not the same as remembering what, as far as the brain is concerned.

Thanks to experiments on animals and the advent of human brain imaging, scientists now have a working knowledge of the various kinds of memory as well as which parts of the brain are involved in each. But persistent gaps remain. Although the MTL has indeed proved critical for declarative memory—the recollection of facts and events—the region remains something of a black box. How its various components interact during memory encoding and retrieval is unresolved. Moreover, the MTL is not the final repository of declarative memories. Such memories are apparently filed to the cerebral cortex for long-term storage, but how this happens, and how memories are represented in the cortex, remains unclear.

More than a century ago, the great Spanish neuro-anatomist Santiago Ramòn y Cajal proposed that making memories must require neurons to strengthen their connections with one another. Dogma at the time held that no new neurons are born in the adult brain, so Ramòn y Cajal made the reasonable assumption that the key changes must occur between existing neurons. Until recently, scientists had few clues about how this might happen.

Memorable diagram.

Santiago Ramòn y Cajal's drawing of the hippocampus. He proposed that memories involve strengthened neural connections.

Since the 1970s, however, work on isolated chunks of nervous-system tissue has identified a host of molecular players in memory formation. Many of the same molecules have been implicated in both declarative and nondeclarative memory and in species as varied as sea slugs, fruit flies, and rodents, suggesting that the molecular machinery for memory has been widely conserved. A key insight from this work has been that short-term memory (lasting minutes) involves chemical modifications that strengthen existing connections, called synapses, between neurons, whereas long-term memory (lasting days or weeks) requires protein synthesis and probably the construction of new synapses.

Tying this work to the whole-brain research is a major challenge. A potential bridge is a process called long-term potentiation (LTP), a type of synaptic strengthening that has been scrutinized in slices of rodent hippocampus and is widely considered a likely physiological basis for memory. A conclusive demonstration that LTP really does underlie memory formation in vivo would be a big breakthrough.

Meanwhile, more questions keep popping up. Recent studies have found that patterns of neural activity seen when an animal is learning a new task are replayed later during sleep. Could this play a role in solidifying memories? Other work shows that our memories are not as trustworthy as we generally assume. Why is memory so labile? A hint may come from recent studies that revive the controversial notion that memories are briefly vulnerable to manipulation each time they're recalled. Finally, the no-new-neurons dogma went down in flames in the 1990s, with the demonstration that the hippocampus, of all places, is a virtual neuron nursery throughout life. The extent to which these newborn cells support learning and memory remains to be seen.

View Abstract

Stay Connected to Science


Navigate This Article