... Salvaris[*]
The authors are with the School of Computer Science and Electronic Engineering, University of Essex. CO4 3SQ, Colchester, UK, Email: {rpoli,ccinel,lciti,missal}@essex.ac.uk
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...[*]
This work was supported by the UK Engineering and Physical SciencesResearch Council (grant EP/F033818/1).
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...[*]
As a by product, this also makes it possible to get new information from the many EEG datasets which have been recorded without EOG.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...[*]
Naturally, we performed a much larger number of smaller runs while developing the fitness function.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...[*]
The high frequency of eye movements in the Perception experiment is due to the particular protocol used for visual stimulation. Subjects are asked to focus on the fixation cross in the middle of the screen. However, because the stimuli appear in peripheral locations, it is not always easy for subjects to avoid saccading towards the stimuli.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...fig:best_Perception.[*]
For easier display the program in the figure has been first syntactically simplified replacing expressions involving only numerical constants with their value and expressions of the form (if True ExpA ExpB) or (if False ExpA ExpB) with ExpA and ExpB, respectively. Then, starting from the leaves, we replaced sub-trees with their median output if the replacement influenced performance by less than 2%.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.