Two models of visual orienting of attention are frequently described.
Voluntary orienting is usually induced by central cues that direct sub
jects' attention to a given location in the visual space. Automatic or
ienting is provoked by presentation of peripheral cues. It is shown th
at automatic orienting induces greater attentional costs and benefits,
and is less under a subject's control (Jonides 1981). Furthermore, it
is not similarly affected by factors such as signal eccentricity (Umi
lta et al. 1991). The present experiment was undertaken to investigate
how sensitive automatic orienting produced by peripheral cues is to v
oluntary modulations of attention. In experiment 1, subjects facing si
tuations daily in which attentional requirements are high, were compar
ed to non-practiced ones. In experiment 2, other groups of practiced s
ubjects facing high or low attentional-demanding situations were teste
d. In both experiments, subjects were asked to respond to signals, pre
sented in rapid succession, in one of two possible locations in space,
on each side of central fixation point. The cue signal automatically
oriented attention to one of the two locations in which a first stimul
us was presented with 100% probability; 100 ms after the first respons
e (RT1), a second response signal (RT2) was delivered either in the sa
me location (valid condition) or in the opposite location (invalid con
dition). Four cue probabilities were manipulated for this second stimu
lus: 100%/0%, 80%/20%, 50%/50%, and 20%/80%. Two eccentricities of 3 d
egrees and 6 degrees were tested. RT2 data demonstrated that (1) there
was no eccentricity effect; (2) the higher the cue probability, the g
reater were the attentional costs; (3) the attentional effects were sm
aller in the practiced subjects who faced attention-demanding situatio
ns daily, than in the other groups. Put together, these data suggest t
hat automatic orienting of attention can be modulated by voluntary att
entional processes, according to cue probability. Furthermore, experie
nced subjects seem to be able to better distribute their attentional r
esources in space, with increased task requirements. The adoption of a
n optimal criterion might lead to the use of a cost minimizing strateg
y.