Adventures in Memory. Hilde østby
as they should have. In other words, you must treat your test animals nicely if you want to learn from them. The same goes for humans: when we are stressed, we don’t retain memories as easily as when we are happy and relaxed.
At about the same time as Lømo’s discovery, there were other breakthroughs in the hunt for the memory trace. In 1971, John O’Keefe at University College London found cells in the hippocampus that remember certain locations. For example, there are some cells in the hippocampus that are active only when we sit on a certain chair, and not on another chair—even in the same room. It is evidently up to some cells (place cells) to remember where we have been at all times. But to remember a place in and of itself—is that a memory? The Norwegian neuropsychologists May-Britt Moser and Edvard Moser—together with John O’Keefe—were awarded the Nobel Prize in Physiology or Medicine in 2014 for their work on that very question. The two Norwegians received the prize because they decided to develop O’Keefe’s research further and look beyond the hippocampus. Their work examined the entorhinal cortex, which connects the hippocampus and the rest of the brain. The Mosers experimented with rats, which, when they were free to explore their environment, showed cells firing in exactly that part of the brain.
With tiny metal electrodes surgically inserted into their brains, these rats wandered around their cages. A single neuron in the entorhinal cortex didn’t react to just one place the rat had scurried to, like place cells, but to several places. Amazing, that what they were expecting to be place cells didn’t remember only one location but several locations in the same area! But when the Mosers marked the points in the cage where the cells had fired, they formed a perfect hexagon on their computer screen. The more the rats ran around in their cages and mazes, the more obvious it became; on the Mosers’ computer, a clear honeycomb pattern emerged. One cell, one hexagonal grid pattern. It was a coordinate system of the environment.
“At first, we thought there was something wrong with our equipment,” Edvard Moser says. “The pattern that emerged was too perfect to come out of something real.”
Each of these neurons makes its own grid, each slightly offset from that of the neighboring cells, so that all points in the environment are covered. Some grids are fine-meshed, while others react to points far away from each other, even farther than it is physically possible for the researchers to measure indoors. Without these grid cells, we are not able to understand or remember locations and where we are in relation to where we have been. We make these patterns wherever we go—wherever we stand, lie, or drive.
“We sent the rats into a ten-armed maze, and it turned out that they continued to make the grid pattern but also started a new one for each ‘arm.’ We believe that these patterns are patched together, so that the rats remember how to get through the maze,” Edvard Moser says.
Since then, other researchers have found the same result in patients undergoing epilepsy surgery. It was as anticipated: in humans, as in rats, all locations are stored in a hexagonal pattern. We are all bees! We all organize the world around us as a hexagonal grid.
“We believe that this was developed very early in the evolution of mammals,” Edvard Moser says. “And we believe that what we have discovered about grid cells is central for episodic memory. It is, after all, impossible to create memories without tying them to a place.”
Other researchers agree that place and grid cells play a special role in episodic memory. Some go so far as to say that this system in the hippocampus and the entorhinal cortex has become specialized to assign each memory its unique memory trace, as part of a unique memory network. Perhaps, at first, the sense of place was the primary task for the hippocampus and the entorhinal cortex. But as evolution proceeded, our memory maps were given a new function: to take our individual experiences and tie them together in a grid. Hexagonal maps of the environment became hexagonal-patterned fishnets of memories.
Recently, researchers in California have been able to demonstrate, in the hippocampi of mice, how memory networks link themselves to context-dependent memory. Like Terje Lømo, they made a window into the hippocampus, to the tiny little piece of it called cornu ammonis 1, Ammon’s horn. Looking at the hippocampus in cross-section, it looks like a goat’s horn, bent inward, into a spiral. Here, through the tiny window into the cradle of memory, the California researchers could see, under a slightly fancy microscope, how the neurons lit up when the mice were placed in different environments. They made three different cages, which would give rise to three different memories: a round cage, a triangular cage, and a square cage. The smell, texture, and other conditions also varied between the three cages. The crucial factor was how close in time the various experiences took place. Two groups of mice were compared. Half of the mice had a go at the triangular cage, and then directly afterward they were placed in the square cage. These mice got to experience two different cages in quick succession. The rest of the mice were placed in the round cage and then, seven days later, in the square cage. The second group of mice had two experiences—episodic memories—distinctly separated in time. When the researchers watched through the microscope while the mice were exploring the cages, they could see activity in the neurons in a defined area. Each of the three cages created a signature pattern of neuronal activity in the hippocampus, meaning distinct memories. The exciting part was that the experiences that took place close together in time led to activity in groups of neurons that overlapped. These two experiences hooked on to each other, not only in time but also in place, in the hippocampus of the mice. Meanwhile, when mice visited two cages a week apart, it was accompanied by activity in two separate groups of neurons in the hippocampus.
The researchers believe this happens because the activation of the one group of neurons causes other nearby neurons to become easier to activate. Everything links together in a network. The main point of Godden and Baddeley’s context-dependent memory experiment has, in this way, been demonstrated in the brain—not in diving mice, but by diving into the cortex of the mice.
WHEN WE EXPERIENCE something—as we find ourselves in a specific situation at a specific location—and it becomes a memory in the brain, it spreads out across the cortex until we recall it. A memory is composed of thousands of connections between neurons; it is not one connection that makes a memory. A memory is more than Terje Lømo’s long-term potentiation.
But what does a memory look like? Can we see a complex memory the way we can see a simple memory trace? To be able to do this, we must exit the rabbit and mouse brains and enter the human brain. And we must watch the brain while memories are recalled. Fortunately, we don’t need to sedate humans and open their heads to get a glimpse of their memories. As we learned in chapter 1, Eleanor Maguire at University College London has used an MRI scanner and some reminiscing volunteers to observe the traces of their memories as they relive their past experiences.
An MRI machine uses a strong magnetic field to take pictures of the body. Different body tissues react differently to the magnetic field, which results in detailed images. The MRI machine can be programmed in a certain way to read the oxygen level in the blood flowing through the brain. Since neurons use oxygen to function, we can tell from the images where there’s a lot of activity. We then know where in the brain nerve cells are most active while the test subjects remember things. This is called functional magnetic resonance imaging (fMRI)—images of the brain while it is working—as opposed to structural MRI, which shows us only what the brain looks like. The memories light up like tiny flashlights underwater, flashes that light up the sea in little spurts.
But is it really possible to see what memory a person is recalling? In Eleanor Maguire’s laboratory, participants allowed themselves to be scanned by an MRI machine at the same time as they were asked to remember their own experiences. The professor actually managed to figure out what they remembered by studying the fMRI images. Maguire watched the activity in the hippocampus while the test subjects were thinking of episodes from their past, and she could see that each memory had a unique pattern of activity. She had a computer program that learned which of the test subjects’ memories were tied to which patterns of activity. From that, the computer program could pick out which fMRI images hung together with which memories.
Is this simply a mind-reading machine?
“These