To what extent does the multi-store model offer a reasonable account of human memory?

Research studies have tended to support the distinction proposed by Atkinson and Shiffrin regarding differences between the stores with respect to capacity, encoding and duration. The existence of separate stores in memory is also supported by the use of modern brain-scanning techniques (such as MRI and PET). These have shown, for example, that the prefrontal cortex is active when individuals are working on a task in immediate(i.e. short-term) memory, whereas the hippocampus is active when LTM is engaged. The notion of a different physiological basis for short and long term memory is further supported by the study of a man called HM (Milner, 1966) who had to have both hippocampi removed due to a tumour. The effect was that he was unable to form any new long term memories, though he could perform STM tasks.

Despite this research support, the multi-store model is probably an oversimplification of memory processes. The multi-store model just proposes one LTM store, whereas research suggests several different forms of LTM such has episodic, semantic and procedural memory. This model also proposes one STM store. Finally, the multi-store model proposes one mechanism for hoe data are stored in LTM by rehearsal. Despite of Atkinson and Shiffrin claims rehearsal is not the only way that information finds it way into LTM.