The Effect of Visual Attention on Eye Movements



Journal Title

Journal ISSN

Volume Title



Purpose: Visual attention along with eye movements help scan the visual world by processing only the relevant information and filtering the extraneous information. Attention could enhance eye movement responses in more than one way. This dissertation aims to understand how endogenous visual attention affects eye movement responses.

Methods: Experiments were designed to study the interaction between attention and saccades during fixation as well as pursuit. The dual Purkinje image eye tracking system sampling at either 120Hz or 358Hz was used to track saccades and pursuit respectively. To understand how saccadic reaction times (SRTs) are shortened with endogenous attention, two dual-task experiments were performed to assess psychophysical and oculomotor responses to non-foveal targets (3° or 6° eccentricity) of various contrasts for different spatial-attention-cueing conditions. Cues were either: valid, an arrow at fixation pointing in the direction of the upcoming target; invalid, an arrow pointing in a different direction from the target; or neutral, a small circle instead of an arrow. In both experiments, subjects were instructed to make a saccade to the location of a subsequent, briefly flashed target and make psychophysical judgements (2 alternative force choice) on contrast detection and contrast matching. It has been suggested that spatial attention leads pursuit as inferred from SRTs to targets presented at multiple locations during pursuit. These findings were revisited to investigate if saccadic amplitude influenced the difference in SRTs between forward and backward saccades during ongoing pursuit. The pursuit target was a letter from the set, C,D,H,K,N,O,R,S,V,Z,E, that traversed the screen horizontally in sinusoidal motion. Each time the letter crossed the straight-ahead position, it jumped either forward or backward with different amplitudes (2.25o to 4o). After confirming that it did not, another experiment was designed to determine the amount of equivalent contrast to explain the forward-backward difference in saccadic latency during pursuit. In the last experiment, scleral search coil eye tracking system at a sampling rate of 500Hz was used to track torsional eye rotations (cycloversion and cyclovergence). Subjects wore red–green anaglyph glasses and were asked to hold their head and gaze steady on the central fixation dot and pay attention to targets rotating in a sinusoidal fashion with different temporal frequencies (0.25Hz and 0.50Hz) at different spatial locations. Fourier analysis was used to determine the torsional tracking amplitude at each frequency for each condition.

Results and Conclusion :Saccades exhibit a robust, ~40-50 milliseconds (ms) reduction of SRTs for attended compared to unattended targets. Cueing effects on contrast detection (0.08 log units) and matching were small and inconsistent across subjects. Hence, the observed decrease in SRTs could not be accounted for fully by an enhancement in the target’s effective contrast due to attention, as attended and unattended targets that were equally detectable or were perceived to have the same suprathreshold contrast showed substantial differences in SRT. Saccades elicited during pursuit had an average latency 30-40 ms less for target jumps in the direction of pursuit compared to targets that jumped in the opposite direction. This was observed at all contrast levels. This difference could not be explained by a difference in the saccade amplitudes of forward vs. backward saccades. The magnitude of equivalent contrast that accounts for the asymmetry in SRTs during pursuit is greater than a log unit. These results suggest that attention influences SRTs in more than one way. In the last experiment scleral search coil method of eye tracking was used to track torsional eye responses at a sampling rate of 500 Hz. Consistent responses to both cyclovergence and cycloversion stimuli were observed in all subjects for both frequencies, with tracking gain ranging from 0.05 to 0.1. Enhanced cycloversion responses were observed with attention. None of the subjects showed a change in cyclovergence responses with attention suggesting that the mechanisms controlling cyclovergence are outside the influence of attentional enhancement. Attention appears to primarily impact voluntary eye responses with advanced oculo-motor planning.



Visual attention, Eye movements, Saccades, Smooth pursuit, Torsion