LEARNING is the process by which new knowledge is acquired about the world. MEMORY is the process by which what is learned can be retained in storage with the possibility of drawing on it later. Most of what humans know about the world is not built into the brain at the time of birth but is acquired through experience. It is learned, stored in the brain as memory, and is available later to be retrieved.
Memory is localized in the brain as physical changes produced by experience. Memory is thought to be stored as changes in synaptic connectivity within large ensembles of neurons. New synaptic connections may be formed, and there are changes as well in the strength of existing synapses. What makes a memory is not the manufacture of some chemical code, but rather increases and decreases in the strength of already existing neural connections and formation of new connections. What makes the memory specific (memory of a trip to England instead of memory of a drive to the hardware store) is not the kind of cellular and molecular event that occurs in the brain, but where in the nervous system the changes occur and along which pathways.
The brain is highly specialized and differentiated, organized so that different regions of neocortex simultaneously carry out computations on separate features of the external world (e.g., the analysis of form, color, and movement). Memory of a specific event, or even memory of something so apparently simple as a single object, is thought to be stored in a distributed fashion, essentially in component parts. These components are stored in the same neural systems in neocortex that ordinarily participate in the processing and analysis of what is to be remembered. In one sense, memory is the persistence of perception. It is stored as outcomes of perceptual operations and in the same cortical regions that are ordinarily involved in the processing of the items and events that are to be remembered.
It has long been appreciated that severe memory impairment can occur against a background of otherwise normal intellectual function. This dissociation shows that the brain has to some extent separated its intellectual and perceptual functions from its capacity for laying down in memory the records that ordinarily result from intellectual and perceptual work. Specifically, the medial temporal lobe and the midline diencephalon of the brain have specific memory functions, and bilateral damage to these regions causes an amnesic syndrome. The amnesic syndrome is characterized by profound forgetfulness for new material (anterograde amnesia), regardless of the sensory modality through which the material is presented and regardless of the kind of material that is presented (faces, names, stories, musical passages, or shapes). Immediate memory, as measured by digit span, is intact. However, a memory deficit is easily detected with conventional memory tests that ask subjects to learn and remember an amount of information that exceeds what can be held in immediate memory or with memory tests that ask subjects to learn even a small amount of information and then hold onto it for several minutes in the face of distraction. The impairment appears whether memory is tested by unaided (free) recall, by recognition, or by cued recall. As assessed by these various instruments, the deficit in amnesic patients is proportional to the sensitivity with which these tests measure memory in intact subjects. Recognition is easier than recall for all subjects, amnesic patients and normal subjects alike.
The same brain lesions that cause difficulties in new learning also cause retrograde amnesia, difficulty in recollecting events that occurred prior to the onset of amnesia. Typically, retrograde amnesia is temporally graded such that very old (remote) memory is affected less than recent memory. Retrograde amnesia can cover as much as a decade or two prior to the onset of amnesia. These observations show that the structures damaged in amnesia are not the repositories of long-term memory. Rather, these structures are essential, beginning at the time of learning, and they are thought to drive a gradual process of memory consolidation in neocortex. As the result of this process, memory storage in neocortex comes to be independent of the medial temporal lobe and diencephalic structures that are damaged in amnesia.
Information about what specific structures are important for human memory comes from carefully studied cases of amnesia, which provide both neuropsychological and neurohistological information, and from the study of an animal model of human amnesia in the monkey. The available human cases make several points. First, damage limited bilaterally to the CA1 region of the HIPPOCAMPUS is sufficient to cause a moderately severe anterograde amnesia. Second, when more damage occurs in the hippocampal formation, (e.g., damage to the CA fields, dentate gyrus, subicular complex, and some cell loss in entorhinal cortex), the anterograde amnesia becomes more severe. Third, damage limited bilaterally to the hippocampal formation is sufficient to produce temporally limited retrograde amnesia covering more than 10 years.
Systematic and cumulative work in monkeys has further demonstrated that the full medial temporal lobe memory system consists of the hippocampus and adjacent, anatomically related structures: entorhinal cortex, perirhinal cortex, and parahippocampal cortex. The critical regions of the medial diencephalon important for memory appear to be the mediodorsal thalamic nucleus, the anterior nucleus, the mammillary nuclei, and the structures within and interconnected by the internal medullary lamina.
One fundamental distinction in the neuropsychology of memory separates immediate memory from long-term memory. Indeed, this is the distinction that is revealed by the facts of human amnesia. In addition, a number of distinctions can be made within long-term memory. Memory is not a unitary mental faculty but depends on the operation of several separate systems that operate in parallel to record the effects of experience. The major distinction is between the capacity for conscious recollection about facts and events (so-called declarative or explicit memory) and a collection of nonconscious memory abilities (so-called nondeclarative or implicit memory), whereby memory is expressed through performance without any necessary conscious memory content or even the experience that memory is being used.
Declarative memory is the kind of memory that is impaired in amnesia. Declarative memory is a brain- systems construct. It is the kind of memory that depends on the integrity of the medial temporal lobe - diencephalic brain structures damaged in amnesia. Declarative memory is involved in modeling the external world, in storing representations of objects, episodes, and facts. It is fast, specialized for one-trial learning, and for making arbitrary associations or conjunctions between stimuli. The acquired representations are flexible and available to multiple response systems. Nondeclarative memory is not itself a brain-systems construct, but rather an umbrella term for several kinds of memory, each of which has its own brain organization. Nondeclarative memory underlies changes in skilled behavior, the development through repetition of appropriate ways to respond to stimuli, and it underlies the phenomenon of priming -- a temporary change in the ability to identify or detect perceptual objects. In these cases, performance changes as the result of experience and therefore deserves the name memory, but like drug tolerance or immunological memory, performance changes without providing a record of the particular episodes that led to the change in performance. What is learned tends to be encapsulated and inflexible, available most readily to the same response systems that were involved in the original learning.
Among the prominent kinds of nondeclarative memory are procedural memory (memory for skills and habits), simple classical CONDITIONING, and the phenomenon of priming. Skill and habit memory depends importantly on the dorsal striatum, even when motor activity is not an important part of the task. Thus, nondemented patients with Parkinson's disease, who have dorsal striatal damage, are impaired at learning a two-choice discrimination task where the correct answer on each trial is determined probabilistically. In this task, normal subjects learn gradually, not by memorizing the cues and their outcomes, but by gradually developing a disposition to respond differentially to the cues that are presented. Classical conditioning of skeletal musculature (e.g., eyeblink conditioning) depends on cerebellar and brain stem pathways. Emotional learning, including fear conditioning, depends on the amygdaloid complex. In the case of fear conditioning, subjects will often remember the unpleasant, aversive event. This component of memory is declarative and depends on the medial temporal lobe and diencephalon. But subjects may also develop a negative feeling about the stimulus object, perhaps even a phobia, and this component of remembering depends on the amygdala. The AMYGDALA also appears to be an important modulator of both declarative and nondeclarative forms of memory. For example, activity originating in the amygdala appears to underlie the observation that emotional events are typically remembered better than neutral events. Finally, the phenomenon of priming appears to depend on the neocortical pathways that are involved in processing the material that is primed. Neuroimaging studies have described reductions in activity in posterior neocortex in correspondence with perceptual priming.
Information is still accumulating about how memory is organized, what structures and connections are involved, and what jobs they do. The disciplines of both psychology and neuroscience contribute to this enterprise.
Knowlton, B. J., J. Mangels, and L. R. Squire. (1996). A neostriatal habit learning system in humans. Science 273:1399-1402.
Rempel-Clower, N., S. M. Zola, L. R. Squire, and D. G. Amaral. (1996). Three cases of enduring memory impairment following bilateral damage limited to the hippocampal formation. Journal of Neuroscience 16:5233-5255.
Squire, L. R., B. J. Knowlton, and G. Musen. (1993). The structure and organization of memory. Annual Review of Psychology 44:453-495.
Squire, L. R., and S. Zola-Morgan. (1991). The medial temporal lobe memory system. Science 253:1380-1386.