The 'anti-shift': Shifting attention opposite to a saccade

The eyes usually shift with attention, but is this connection necessary? We tested whether saccades can be de-coupled from endogenous attention shifts by comparing pro-shifts, in which they move together, to anti-shifts in which attention is tasked to shift in one direction while the eyes saccade in the opposite direction. Saccade latency was measured by an eye tracker. The attention shift latency was derived using the RSVP 'attention shift paradigm' of Reeves and Sperling (1986), in which attention shifts from a target embedded in a stream of letters to a stream of numerals, of which the first four are to be noted. Initially, anti-shifts were impossible; either the eye or attention shifted first. But after 10 h of practice, five of the six adult subjects learnt how to execute saccades, shift attention, and retain numerals, almost equally well whether performing anti-shifts or pro-shifts. The results imply that brain structures such as the frontal eye fields do not inevitably yoke fixational and attentional movements together.