Copyright © 2007-2018 Russ Dewey
How are Memories Stored?
We have seen that formation of one type of memory, episodic memory, involves a specific part of the brain: the hippocampus. However, other areas of the brain are involved in other forms of memory.
"Memory is modular," says neuroscientist Patricia Goldman-Rakic (Service, 1993). Wilson, Scalaidhe, and Goldman-Rakic (1993) found that one set of neurons was active when monkeys remembered the identity of a stimulus; another was active when the monkeys remembered the location of a stimulus.
Probably each part or module of the brain remembers its own role in each distinct type of brain activity. If so, each part involved in an experience should show activity if that experience is remembered later.
What is probably true of memory storage in the brain?
Dingledine (1995) noted that "long-lasting changes in synaptic strength are not limited to the hippocampus." Instead, such changes "are encountered in nearly every layered structure one examines from the cerebellum to olfactory and neocortices, in keeping with the expectation that memory storage areas are distributed throughout the brain."
Brennan, Kaba, and Keverne (1990) showed that a rat's memory for sexual attractant odor was stored at the first synapse after the odor receptor. That means the memory was stored (among other places) in the first neurons that responded to an odor. Again, memory is located in structures that represent or respond to an event.
What "old principle" may apply to memory storage?
If memory is a construction, then it makes sense that memories would be stored in the same neurons that originally constructed an experience. They are likely to be the neurons called upon later to re-construct it.
This is like an old principle of efficient workplace organization: store at the point of first use. In other words, store something at the location where you expect to need it later.
In the case of memory representation, brain regions involved in an experience would be the first place to look for storage of a memory. The big exception, as earlier noted, is event memory of humans.
The distinctive thing about event memories is that they bind together a variety of sensory, cognitive, and emotional elements into a single gestalt that is stored as memory for a particular space-time event. The hippocampus and neighboring tissue apparently draws together information from widespread parts of the brain, for this type of memory.
If each part of the nervous system stores memories of its own role in neural activity, this may help explain a famous series of experiments that baffled psychologists in the 1930s and 1940s. Karl Lashley, one of the world's foremost brain researchers, tried to locate the area in the brain where engrams or memory traces were stored.
Lashley sliced or removed sections of rat brains after teaching the rats to run mazes. None of the brain injuries abolished the "maze-running habit," although Lashley tried removing tissue in almost every area that allowed the rat to remain alive. Lashley concluded that memories had to be spread all over the brain, throughout all the tissue.
What led Karl Lashley to his conclusion?
In retrospect, Lashley probably picked the worst possible laboratory task to study, if he was trying to find a specific location that stored a memory. Maze running involves many parts of the brain.
At minimum, maze-running involves vision (remembering the sight of correct pathways), spatial sense (remembering the direction to turn), olfaction (smelling the cheese and moving toward the more powerful odor), and kinesthesis (the feeling of arms and legs running a certain direction). If one type of clue is eliminated, there are many others remaining, allowing the rat to guide itself to the end of the maze.
In what way did Lashley "pick the worst possible laboratory task to study"?
Lashley may have been half right is saying memory is widely distributed. He was also half wrong, because he assumed memory was unitary and there was one type of memory trace stored all over the brain.
The closest thing to "unitary" memory traces are event memories, and they are localized near the hippocampus. Other forms of memory are located where they might be needed later, in areas where experiences are converted into neural codes.
Lashley's research was very influential and led to a persistent anti-localization bias among psychologists. For decades, psychologists were inclined to believe that complex mental functions were spread all over the brain in equal measure.
The assumption was that one area of brain tissue must be able to substitute for another after brain injury. How else could one explain Lashley's results?
What bias was created by Lashley's research? How was the idea "put to rest" by brain scanning technologies?
Lashley's conclusions implied that a skill would never depend upon one tiny area of neurons. However, as it turned out, almost the exact opposite is true.
Specialized circuits exist in every part of the brain. The reason the "maze-running habit" seemed to be so widely spread was that many different brain areas were involved with it. Eliminating one part of the brain was not enough to disrupt the entire act.
Posner (1993) noted the "popularly held belief in psychology that the cognitive functions of the brain are widely distributed among different brain areas." However, he concluded, "imaging studies reveal a startling degree of region-specific activity."
In other words, most brain tissue is highly specialized. A typical cognitive act does indeed activate many places in the brain, but each area is doing something specific. Loss of any particular area should lead to missing or slowed information processing, if a researcher knows exactly what to look for.
Brennan, P., Kaba, H., & Keverne, E. B. (1990). Olfactory recognition: A simple memory system. Science, 250, 1223-1226.
Dingledine, R. (1995) Machinery of learning. Science, 267, 265-266.
Posner, M. I. (1993). Seeing the mind. Science, 262, 673-676.
Wilson, F. A., O'Scalaidhe, S. P., & Goldman-Rakic, P. S. (1993). Dissociation of object and spatial processing domains in primate prefrontal cortex. Science, 260, 1955-1958.
Service, R. F. (1993). Making modular memories. Science, 261, 1876.
Write to Dr. Dewey at email@example.com.