Q.How is information processed through sensory, short-term and long-term memory systems?

The Stage Model of memory, proposed by Atkinson and Shiffrin in 1968, describes memory as a system of three separate stores through which information passes in sequence, borrowing its basic logic from how a computer registers, stores, and manipulates information. The three stores are sensory memory, short-term memory (STM), and long-term memory (LTM), and each has its own characteristic capacity and duration.

Information first enters sensory memory, which has a large capacity — able to register input from every sense — but an extremely short duration, with information decaying automatically in well under a second unless something is done with it. Sensory memory holds this information as a fairly exact replica of the original stimulus; the visual version is called iconic memory and the auditory version is called echoic memory, and everyday experiences like a lingering visual after-image or a fading echo of a sound illustrate these registers directly.

Only information that is actually attended to moves on from sensory memory into short-term memory. This is governed by the first control process, selective attention, which filters out the great majority of sensed information (which simply fades away unattended) and lets through only what is being actively focused on. STM holds a comparatively small amount of information — traditionally described as around seven items, plus or minus two — for a brief period, usually thirty seconds or less, and Atkinson and Shiffrin proposed that this information is mainly encoded acoustically, in terms of sound. Two further control processes operate within STM: maintenance rehearsal, which is simply repeating information (silently or aloud) to keep it available for as long as needed, though it is lost the moment the repetition stops; and chunking, which expands STM's effective capacity by combining several smaller units into larger, meaningful chunks — for example, breaking an overly long string of digits into recognisable groups, such as memorable years, so that the whole string becomes easier to hold in mind.

Information that is to be retained beyond STM's tight limits must move into long-term memory, and this transfer happens through a fourth control process, elaborative rehearsal. Unlike maintenance rehearsal's simple repetition, elaborative rehearsal works by connecting the new information to material that already exists in long-term memory, building meaningful associations with concepts and life experience; the richer this network of associations, the more likely the new material is to become a permanent part of long-term memory. LTM itself has a vast, essentially unlimited capacity and can hold information for up to a lifetime; it is thought to rely primarily on semantic encoding, meaning the information is represented in terms of the meaning it carries rather than its surface form.

While the Stage Model gives a clear, structured picture of how information supposedly flows from sensory memory through STM into LTM, experimental evidence has been mixed on whether STM and LTM really are two entirely separate stores — case studies such as KF (whose STM was damaged but LTM remained intact) and evidence that both stores can use more than one type of encoding have pushed some psychologists toward alternative views, such as the Levels of Processing approach, discussed later in the chapter.