San Diego Scientists Probe The Machinery Of Memory
What actually happens inside the human brain when it forms a memory? Monday, San Diego researchers report seeing that process up close, cell by cell, inside a living human brain.
That marks a step forward for memory research, since most studies are performed on mice. Or they rely on brain scans, a method that can be great at painting a broad overview of brain activity, but isn't so great at picking up on fine details.
UC San Diego'sJohn Wixted wanted to go deeper. But he says, "To do that, you have to actually have electrodes in the human brain. And that's not something that you often have."
This time, Wixted did have that chance. His colleagues at the Barrow Neurological Institute in Phoenix, Arizona, were treating nine epilepsy patients. Due to the severity of their condition, they all had electrodes implanted in their brains.
With the patients' consent, Wixted used this rare opportunity to put them through a simple memory test. They read a list of words. Later, they read through a longer list that included some of the original words, and they were asked to recall which words they'd seen before.
The experiment allowed Wixted to train his sights on the hippocampus, a small part of the brain known to be crucial for memory formation. It's one of the first regions damaged in early stages of Alzheimer's. And after the famous amnesiac patient Henry Molaison had his hippocampus surgically removed, he spent the rest of his life unable to form new memories.
But how exactly the hippocampus takes raw experience and cements it in memory remained a bit of a mystery. Wixted expected to see large networks of neurons throughout the hippocampus activate when the subjects in his experiment remembered a word.
But what Wixted actually saw was this: "Each experience was encoded by a small percentage of neurons," he said. "And the next experience was encoded by a different small percentage of neurons — non-overlapping sets of neurons"
Only 2 percent of neurons in the hippocampus were active with each memory. And instead of seeing one large web of neurons producing all these memories, Wixted saw small, varied constellations light up for each individual memory.
The results initially surprised Wixted, but he soon discovered that many neurocomputational theorists have long thought that's exactly how our brain makes memory. Now, Wixted says, there's data to back up their theories.