Visuospatial perceptual sequence learning and eye movements

We examined perceptual sequence learning by means of an adapted serial reaction time task in
which eye movements were unnecessary for performing the sequence learning task. Participants
had to respond to the identity of a target letter pair ("OX" or "XO") appearing in one of four
locations. On the other locations, similar distractor letter pairs ("QY" or "YQ") were shown.
While target identity changed randomly, target location was structured according to a
deterministic sequence. To render eye movements superfluous, (1) stimulus letter pairs appeared
around a fixation cross with a visual angle of 0.63°, which means that they appeared within the
foveal visual area and (2) the letter pairs were presented for only 100 ms, a period too short to
allow proper eye movements. Reliable sequence knowledge was acquired under these conditions,
as responses were both slower and less accurate when the trained sequence was replaced by an
untrained sequence. These results support the notion that perceptual sequence learning can be
based on shifts of attention without overt oculomotor movements.