| |
Abstract:
Many neurons in LIP maintain spatially-tuned memory activity
prior to a delayed saccadic eye movement. Furthermore, these
neurons appear to compensate for changes in gaze position made
during the memory period, updating the inferred location of an
extinguished world-fixed stimulus by combining retinal and gaze
position signals. However, as gaze changes, the target of a future
eye movement may instead move with the subject's gaze. Thus, the
computation of invisible target location ought to incorporate gaze
position signals when targets are world-fixed but ignore them when
targets are subject-fixed. To compare LIP activity in these
situations, we recorded from 79 neurons in two monkeys trained to
perform memory-guided saccades. An initial instruction indicated
whether a remembered target would be world-fixed or subject-fixed.
During the memory period, a whole-body rotation, smooth pursuit, or
saccadic eye movement shifted gaze to an intermediate location.
Subsequently, the subject was directed to look to the remembered
location. Following slow gaze shifts (rotation and smooth pursuit
trials), both neuronal spatial sensitivity and saccade precision
were reduced for world-fixed relative to subject-fixed trials.
However, spatial accuracy during world-fixed trials was reduced in
the neuronal population but not in the behavior, suggesting that
areas outside of LIP must help guide saccades to world-fixed
targets when slow, but not fast, gaze changes are incorporated into
oculomotor plans.
|
