Memory

Important New Theory Explains Where Old Memories Go

Why some memories disappear, some remain, and others blend with fiction

October 15, 2013

Emilie Reas  http://www.scientificamerican.com

www.nasa.gov

Think back to your first childhood beach vacation. Can you recall the color of your bathing suit, the softness of the sand, or the excitement of your first swim in the ocean? Early memories such as this often arise as faded snapshots, remarkably distinct from newer memories that can feel as real as the present moment. With time, memories not only lose their rich vividness, but they can also become distorted, as our true experiences tango with a fictional past.

The brain’s ability to preserve or alter memories lies at the heart of our basic human experience. The you of today is molded not only by your personal history, but also by your mental visits to that past, prompting you to laugh over a joke heard yesterday, reminisce about an old friend or cringe at the thought of your awkward adolescence. When we lose those pieces of the past we lose pieces of our identity. But just where in the brain do those old memories go? Despite decades studying how the brain transforms memories over time, neuroscientists remain surprisingly divided over the answer.

Some of the best clues as to how the brain processes memories have come from patients who can’t remember. If damage to a particular brain area results in memory loss, researchers can be confident that the region is important for making or recalling memories. Such studies have reliably shown that damage to the hippocampus, a region nestled deep inside the brain, prevents people from creating new memories. But a key question, still open to debate, is what happens to a memory after it’s made. Does it stay in the hippocampus or move out to other areas of the brain? To answer this, scientists have studied old memories formed before brain damage, only to discover a mix of inconsistent findings that have given rise to competing theories.

One popular theory proposes that the hippocampus is critical only for recent memories, but not older ones. Over time, the hippocampus teaches the surrounding brain – the cortex – how to represent a memory. As the memory matures, the hippocampus kicks it out to reside independently in the cortex. If you lost your hippocampus today, you could still remember your childhood vacation to Florida, but the memory from last weekend’s dinner party would be lost. This is the exact pattern scientists have observed in many patients, including the celebrated amnesic H.M. After surgery to remove a large chunk of his hippocampus, H.M. could recall some very old memories, but could no longer make new memories or remember the years leading up to his surgery.

But researchers have seen other patients with hippocampal damage who have memory deficits extending back through most of their life. An alternative theory accounts for these discrepancies by proposing that the hippocampus selectively stores one type of memory – “episodic” – while the surrounding cortex stores another – “semantic.” Episodic memories are usually rich in details about our past experiences, whereas semantic memories are based on impersonal, factual knowledge. As a memory ages, the model proposes, it is copied many times in both the hippocampus and the cortex. All of those cortical copies generate a new semantic memory, representing only the gist or key facts about the experience, without all of its elaborate episodic details. Intriguingly, this theory is also backed by patients who exhibit memory problems depending on the type of memory, rather than its age. Without a hippocampus, such patients couldn’t remember the experience of their childhood Florida trip – an episodic memory – but could remember the fact that they visited Florida – a semantic memory.

Both theories have camps of staunch advocates, drawing support from particular amnesia cases that only their model can explain. But as neither theory perfectly fits together all pieces of the puzzle, the field has entered a stalemate.

Researchers from Johns Hopkins University have come up with a new theory that just might settle some of this controversy. Their explanation rests on the premise that memories are transformed each time we revisit them. According to this theory, a memory is first encoded by the coordinated activity of neurons in the hippocampus and cortex. The hippocampus acts as the brain’s director, telling the cortex which particular neurons to activate. Each time we recall that memory, a similar, but not identical set of neurons are activated. Neurons that are frequently activated become part of the permanent memory trace in the cortex, while the rarely activated ones are lost. Every reactivation re-encodes the memory, and depending on what cortical neurons are engaged, can strengthen, weaken or update particular memory features.

On the surface, this new model sounds a lot like the earlier ones. But it breaks the longstanding stalemate by proposing that what we do with a memory, rather than its age or type, determines where it’s stored in the brain. While the competing theories debated whether the hippocampus is only needed for recent memories or episodic memories, the new model suggests that what really matters is how often you revisit the memory. When a memory is recalled often, it will more rapidly become stored in the cortex, become less episodic and become independent of the hippocampus. But a memory that’s rarely revisited will remain dependent on the hippocampus. Older memories might be recalled more often, but the relationship isn’t perfect. This would explain why one amnesic’s memory impairment extends back forty years, while another’s extends only ten years.

The theory also nicely accounts for our subjective sense for how our memories change over time; namely, how the hippocampus and cortex collaborate to gradually fade or distort our memories. Say you’re reminded of that beach vacation every summer. With each memory reactivation, some features are reinforced while others disappear, explaining why the memory seems to get fuzzy over time. And the more details that are lost, the less “episodic” and the more “semantic” the memory becomes, accounting for the sense of personal detachment often associated with very old memories. Each time you think back to your Florida vacation, you re-encode fewer details, making the memory feel less vivid now than it did decades ago. Today, you might still be able to recall your striped blue bathing suit even though the smell of the ocean air is lost.

Each mental trip back to Florida is not only an opportunity to strengthen or weaken the memory, but also to incorporate fictional tidbits. You used to be sure that vacation was in Fort Lauderdale, but your sister always talks about the fun family trip to Miami. Every time you reminisce together, the memory of Fort Lauderdale is reactivated, but so is a competing representation of Miami. Next time you think of the vacation, the Fort Lauderdale and Miami representations conflict, causing uncertainty over where you actually went. Recall the beach in Miami enough times and voilà, a false memory is born!

Memories fade and transform as they age. This intriguing new theory suggests that these changes have to do less with the age or content of a memory, and more with what we do with that memory. Changing the past just might be easier than we thought. Chances are, you do it every time you remember.

Are you a scientist who specializes in neuroscience, cognitive science, or psychology? And have you read a recent peer-reviewed paper that you would like to write about? Please send suggestions to Mind Matters editor Gareth Cook, a Pulitzer prize-winning journalist and regular contributor to NewYorker.com. He can be reached at garethideas@gmail.com or Twitter @garethideas.

ABOUT THE AUTHOR
Emilie Reas is a Neuroscience doctoral student in the Brewer lab at the University of California, San Diego. She uses fMRI to study how we create and recall memories. You can follow her on Twitter at @etreas.