On the use of discrimination to assess memory

A number of studies have made use of sensory discrimination to assess memory. These experiments have typically employed a temporal two-alternative forced choice paradigm in which the interstimulus interval (ISI) is varied. Many of these tests have shown that ISI has little or no effect on discrimination performance. This has been taken as evidence for perfect or very good memory. An alternative interpretation would be that discrimination tests lack sufficient sensitivity to reveal memory effects. In the present study, computer simulations based on signal detection theory were used to estimate the relative effects of memory noise and direct noise (i.e., noise inherent in the discrimination task) on discrimination performance. These simulations indicate that the amount of memory noise has to be quite large to cause a marked elevation in discrimination thresholds and that sensory discrimination is a relatively insensitive test of memory.