| |
Abstract:
The working-with-memory model (e.g., Moscovitch & Umilta,
1990, 1991) states that episodic retrieval from medial-temporal
brain regions is manditory once the appropriate retrieval cues are
made available. In a series of divided attention (DA) experiments,
healthy younger and older adults engaged in a memory task and a
reaction time (RT) secondary task. Time-accuracy functions were
derived from trials in which the memory task was performed under
full attention (FA) conditions and the time for encoding or for
retrieval was manipulated across trials. In addition, the RT task
was performed alone, or concurrently either during encoding or
during retrieval.
The time available for encoding or retrieval in DA conditions was
calculated from RT costs (RTs in DA conditions minus RTs in FA
conditions) plus the time devoted to motor aspects of the RT task.
For both age groups, time-accuracy functions overestimated memory
in DA at encoding conditions, but accurately predicted memory in DA
at retrieval conditions. The older adults had more time available
for encoding and retrieval (larger RT costs and slower motor
speed), but their encoding and retrieval also proceeded more
slowly. These results are consistent with the working-with-memory
model; however, the additional assumption must be made that there
exists a retrieval bottleneck (c.f. Pashler, 1984), in which
response selection on the secondary task cannot proceed until
ecphoric processes are completed.
|
