The Anatomy of Memory

One of the first questions about memory to be addressed using animals was
its relationship to the underlying structure of the nervous system. American
psychologist Karl Lashley (1890-1958) was an early pioneer in this field. His
main interest was in finding what was then referred to as the engram, the
physical location in the brain where memories are stored. Lashley trained
rats on a variety of tasks, such as the ability to learn mazes or perform simple
discriminations, and then lesioned various parts of the cerebral cortex
(the convoluted outer covering of the brain) in an attempt to erase the
memory trace. Despite years of effort, he found that he could not completely
abolish a memory, no matter what part of the cortex he lesioned.
Lashley summed up his puzzlement and frustration at these findings in this
now well-known quote: “I sometimes feel, in reviewing the evidence on the
localization of the engram, that the necessary conclusion is that learning
just is not possible.”

While the specific location of the brain lesion did not appear important,
Lashley found that the total amount of brain tissue removed was critical.
When large lesions were produced, as compared to smaller ones, he found
that memories could be abolished, regardless of the location in the cortex
where they were made. This led Lashley to propose the concepts of mass action
and equipotentiality, which state that the cortex works as a whole and
that all parts contribute equally to complex behaviors.

Further research has generally supported Lashley’s original conclusions
about the localization of the engram. However, better memory tests and
more sophisticated techniques for inducing brain damage have revealed
that certain brain regions are more involved in memory than others and
that different brain regions are actually responsible for different types of
memory. For example, classical conditioning, which is the modification of a
reflex through learning, appears primarily to involve the brain stem or cerebellum,
which are two evolutionarily old brain structures. Specific circuitry
within these structures that underlies a number of forms of classical conditioning
has been identified.

In the rabbit, a puff of air blown into the eye produces a reflexive blinking
response. When researchers repeatedly pair the air puff with a tone, the
tone itself will eventually elicit the response. The memory for this response
involves a very specific circuit of neurons, primarily in the cerebellum. Once
the response is well learned, it can be abolished by lesions in this circuit. Importantly,
these lesions do not affect other forms of memory. Similarly, taste
aversion learning, a process by which animals learn not to consume a food
or liquid that has previously made them ill, has been shown to be mediated
by a very specific circuit in the brain stem, specifically the pons and medulla.
Animals with lesions to the nucleus of the solitary tract, a portion of this circuit
in the medulla where taste, olfactory, and illness-related information
converge, will not readily learn taste aversions.

More complex forms of learning and memory have been shown to involve
more recently evolved brain structures. Many of these are located in either
the cortex or the limbic system, an area of the brain located between
the newer cortex and the older brain stem. One component of the limbic
system believed to be heavily involved in memory is the hippocampus. One
of its primary functions appears to be spatial memory. Rats and monkeys
with damage limited to the hippocampus are impaired in maze learning and
locating objects in space but have normal memory for nonspatial tasks. Animals
that require spatial navigation for their survival, such as homing pigeons
and food-storing rodents (which must remember the location of the
food that they have stored) have disproportionately large hippocampi.
Moreover, damage to the hippocampus in these species leads to a disruption
in their ability to navigate and find stored food.

One area of the cortex that has been shown to be involved in memory is
the prefrontal cortex. This area has been implicated in short-term memory,
which is the ability to hold temporarily a mental representation of an object or event. Monkeys and rats that received lesions to the prefrontal cortex
were impaired in learning tasks that required them to remember briefly the
location of an object or to learn tasks that require them to switch back and
forth between strategies for solving the task. Studies involving the measurement
of brain function have also demonstrated that this area of the brain is
active during periods when animals are thought to be holding information
in short-term memory.

While experimental brain damage has been one of the predominant
techniques used to study structure/function relationships in the nervous
system, difficulty in interpretation, an increased concern for animal welfare,
and the advent of more sophisticated physiological and molecular techniques
have led to an overall decline in their use.