Research into memory processes has progressed in recent years through the combined efforts of neuroscientists and cognitive scientists, aided by positron emission tomography and functional magnetic resonance imaging (Schacter, 1997). Most of the research advances in memory that help scientists understand learning come from two major groups of studies: studies that show that memory is not a unitary construct and studies that relate features of learning to later effectiveness in recall.

Memory is neither a single entity nor a phenomenon that occurs in a single area of the brain. There are two basic memory processes: declarative memory, or memory for facts and events which occurs primarily in brain systems involving the hippocampus; and procedural or nondeclarative memory, which is memory for skills and other cognitive operations, or memory that cannot be represented in declarative sentences, which occurs principally in the brain systems involving the neostriatum (Squire, 1997).

Different features of learning contribute to the durability or fragility of memory. For example, comparisons of people’s memories for words with their memories for pictures of the same objects show a superiority effect for pictures. The superiority effect of pictures is also true if words and pictures are combined during learning (Roediger, 1997). Obviously, this finding has direct relevance for improving the long-term learning of certain kinds of information.

Research has also indicated that the mind is not just a passive recorder of events, rather, it is actively at work both in storing and in recalling information. There is research demonstrating that when a series of events are presented in a random sequence, people reorder them into sequences that make sense when they try to recall them (Lichtenstein and Brewer, 1980). The phenomenon of the active brain is dramatically illustrated further by the fact that the mind can “remember” things that actually did not happen. In one example (Roediger, 1997), people are first given lists of words: sourcandy-sugar-bitter-good-taste-tooth-nice-honey-soda-chocolate-heart-caketart-pie. During the later recognition phase, subjects are asked to respond “yes” or “no” to questions of whether a particular word was on the list. With high frequency and high reliability, subjects report that the word “sweet” was on the list. That is, they “remember” something that is not correct. The finding illustrates the active mind at work using inferencing processes to relate events. People “remember” words that are implied but not stated with the same probability as learned words. In an act of efficiency and “cognitive economy” (Gibson, 1969), the mind creates categories for processing information. Thus, it is a feature of learning that memory processes make relational links to other information.

In view of the fact that experience alters brain structures and that spe-