| |
Abstract:
A verbal recurrent perseveration is the inappropriate
repetition of a recent utterance when a different response is
expected (Sandson & Albert, 1984). Several accounts have
suggested that such perseverations result when processing of the
current stimulus is too weak or slow to dominate a recently primed
response (e.g. Cohen & Dehaene, 1998; Martin, et al, 1998;
Plaut & Shallice, 1993). We present data from several
picture-naming experiments conducted with an anomic patient (E.B.)
which require further specification of current proposals. Small
sets of pictures were repeatedly presented for naming at a fixed
response-stimulus interval (RSI), either fast (1 sec) or slow
(10-15 sec). Consistent with existing priming accounts,
perseverations were more likely after one within-block repetition
and when the current target name was of low frequency. However,
there was no difference in the number of perseverations produced
under a fast vs. slow RSI, and the rate of decay was far slower
with a slow RSI. We present a simple computational model of these
results which appeals to recent theories of the neural basis of
priming (Wiggs & Martin, 1998) and a similar spacing effect
observed in the presynaptic depression of neurons (Byrne,
1982).
|
