7.1 Types of Memory
Psychologists and neuroscientists have divided memory systems into two broad categories, declarative and nondeclarative (Figure 7.1).
The declarative memory system is the system of memory that is perhaps the most familiar. It is the memory system that has a conscious component and it includes the memories of facts and events. A fact like ‘Paris is the capital of France’, or an event like a prior vacation to Paris.
Nondeclarative memory, also called implicit memory, includes the types of memory systems that do not have a conscious component but are nevertheless extremely important. They include the memories for skills and habits (e.g., riding a bicycle, driving a car, playing golf or tennis or a piano), a phenomenon called priming, simple forms of associative learning [e.g., classical conditioning (Pavlovian conditioning)], and finally simple forms of nonassociative learning such as habituation and sensitization. Sensitization will be discussed in detail later in the Chapter. Declarative memory is “knowing what” and nondeclarative memory is “knowing how”.
7.2 Testing Memory
[…] We like to think that memory is similar to taking a photograph and placing that photograph into a filing cabinet drawer to be withdrawn later (recalled) as the “memory” exactly the way it was placed there originally (stored). But memory is more like taking a picture and tearing it up into small pieces and putting the pieces in different drawers. The memory is then recalled by reconstructing the memory from the individual fragments of the memory.
The word list [memory test] gives insights into memory processing and retrieval, but it is not a really good test of “raw” memory ability because it can be affected by distortions and biases. To avoid these problems, psychologists have developed other memory tests. One is the object recognition test (Figure 7.3) to test declarative memory. [It] involves presenting a subject with two different objects and they are asked to remember those objects. There is a pause and then two objects are shown again, one of which is new and the other having been shown previously. Subjects are asked to identify the novel object, and to do so, they need to remember which one was shown previously. A somewhat related test is the object location test in which subjects are asked to remember the location of an object on a two-dimensional surface.
Examples of nondeclarative memory, such as associative learning, can be tested by pairing one stimulus with another and later testing whether a subject has learned to make the association between the two stimuli. The classical example is the paradigm developed by the Russian physiologist Ivan Pavlov, which is now called classical or Pavlovian conditioning. In classical conditioning (Figure 7.4), a novel or weak stimulus (conditioned stimulus, CS) like a sound is paired with a stimulus like food that generally elicits a reflexive response (unconditioned response, UR; unconditioned stimulus, US) such as salivation. After sufficient training with contingent CS-US presentations (which may be a single trial), the CS is capable of eliciting a response (conditioned response, CR), which often resembles the UR (or some aspect of it).
7.3 Localization of Memory
Now let us turn to this issue about where is memory located. There are three basic approaches.
- Imaging. Modern imaging techniques like fMRI (functional magnetic resonance imaging) or PET (positron emission tomography) allows one to “see” areas of the brain that are active during specific brain tasks. If a subject is placed in an fMRI scanner and given a memory test, one can determine what areas of the brain are active, and that activity presumably is related to where in the brain the memory is processed and/or stored.
- Brain lesions. In this experimental procedure, small parts of the brains of mice or rats are surgically removed or chemically inactivated and the animals are systematically examined to determine whether the lesion affected any memory system.
- Brain disease and injury. Here scientists take advantage of individuals who have had unfortunate brain injuries, for example, through stroke or through a brain tumor in a specific area of the brain. If one finds a memory deficit in the patient, it is likely that the region of the brain that was injured is involved in that memory.
A classic study on localization of memory was the result of surgery performed on Henry Molaison, a patient who was only known to the scientific community as “H.M.” until his death in 2008. H. M. is famous in the neuroscience literature because his brain provided major insights into the localization of memory function. In the 1950’s, H.M. was diagnosed with intractable epilepsy, and while there are pharmacologic treatments, in some cases the only treatment is to remove the portion of the brain that is causing the seizures. Consequently, H.M.’s hippocampus was removed bilaterally.
Before the operation, H.M. had a fine memory, but after the operation, H.M. had a very severe memory deficit. Specifically, after the operation H.M.’s ability to form any new memories for facts and events was severely impaired; he had great difficulty learning any new vocabulary words; he could not remember what happened the day before. So if H.M. had an interview the day following a previous interview, he would have little or no memory about the interview or events during it.
This study clearly indicated that the hippocampus was critical for memory formation. But whereas H.M. had great difficulty forming new memories for facts and events, he still had all of his old memories for facts and events. Specifically, he had all his childhood memories, and all of his memories prior to the operation. This type of memory deficit is called anterograde amnesia. (In contrast, retrograde amnesia refers to loss of old memories.) The studies on H.M. clearly indicated that whereas the hippocampus is critical for the formation of new memories, it is not where the old memories are stored.
It is now known that those old memories are stored in other parts of the brain, such as in the frontal cortex. The process by which an initially labile memory is transformed into a more enduring form is called consolidation. This process involves the memory being stored in a different part of the brain than the initial site of its encoding.
H.M. was also interesting in that while his ability to form new memories for facts and events was severely impaired, he could form new memories for skills and habits. While he could form new memories for skills and habits, he did not know that he had the skills! He had no awareness of the memory; he couldn’t declare that he had it. This finding clearly indicated that the memory for skills and habits are not formed in the hippocampus. Collectively, we learned from these studies on H.M. and other patients that memory is distributed throughout the nervous system, and different brain regions are involved in mediating different types of memory.
Figure 7.7 summarizes many decades of research on the anatomical locus of memory systems. The medial temporal lobe and structures like the hippocampus are involved with memories for facts and events; the striatum is involved with memories for skills and habits; the neocortex is involved with priming; the amygdala is involved with emotional memories; and the cerebellum with simple forms of associative learning. Lower brain regions and the spinal cord contain even simpler forms of learning. In summary, memory is not stored in a single place in the brain. It is distributed in different parts of the brain.
7.4 Mechanisms of Memory
Much of what has been learned about the neural and molecular mechanisms of learning and memory have come from the use of so called “model systems” that are amenable to cellular analyses. One of those model systems is illustrated in Figure 7.8A. Aplysia californica is found in the tidal pools along the coast of Southern California. It is about six inches long and weighs about 150 grams. At first glance it is an unpromising looking creature, but neuroscientists have exploited the technical advantages of this animal to gain fundamental insights into the molecular mechanisms of memory. Indeed, the pioneering discoveries of Eric Kandel using this animal were recognized by his receipt of the Nobel Prize in Physiology or Medicine in 2000. […]
One principle about learning and memory derived from studies of this simple animal, and this principle holds true in our brains as well, is that learning involves changes in the strength of synaptic connections between neurons. Learning is not due to a reorganization of the nervous system or the growth of new neurons. What has changed is that the strength of a previously existing connection is modified.
Now we can take this analysis one step further and ask what are the biochemical mechanisms that underlie learning and memory. We will divide the discussion into two temporal domains of memory; short-term memory and long-term memory. We have already discussed different types of memory such as declarative and nondeclarative memory. There are also different temporal domains of memory. Short-term memories are like the memory for a telephone number that last several minutes, and long-term memory are memories that last days, weeks or a lifetime.
[…] We have discussed a mechanism for a short-term memory. It is “short-term” because the memory is transient and that is so because the underlying biochemical changes are transient. The duration of the memory is dependent on how long the various substrate proteins (e.g., membrane channels) are phosphorylated. PKA will only be activated for a short time after a brief stimulus because cyclic AMP will be degraded and PKA levels will decrease. Protein phosphatases will remove the phosphate groups on the substrate proteins that are “storing” the memory.
Mechanisms of long-term sensitization. There are two major differences between short-term and long-term memories. Long-term memories involve changes in protein synthesis and gene regulation, whereas short-term memories do not. And, long-term memories in many cases involve structural modifications. The neuron from the trained animal has a greater number of branches and a greater number of synaptic varicosities than the neuron from the untrained animal. Therefore, long-term memory involves changes in the structure of neurons including growth of new processes and synapses. So, to the extent that you remember anything about this material on memory tomorrow, or next week, or next year, it will be because structural changes in synapses are beginning in your brains!
Given that long term memory involves changes in gene expression, a major goal of neuroscientists is to identify the specific genes and proteins that are involved in long-term memory. […]
Note that there is not a single “magic memory gene” – rather, the induction and maintenance of memory, even in a single neuron, involves the engagement of multiple genes and proteins that act synergistically to change the properties of the neurons and regulate the properties of the neuron and the strength of the synapse. Also note that changes in gene expression do not occur all at once – there are different phases. Some changes in gene expression occur early, some even 24 hours after the learning occurs.