Name: Cody A. Mashburn 

Master's Thesis Defense Meeting

Date: Friday, February 3, 2023

Time: 4:00 p.m.

Location: J.S. Coon bldg. room 150 and on Zoom (click here)


Advisor: Randall W. Engle, Ph.D. (Georgia Tech)


Thesis Committee Members:

Randall W. Engle, Ph.D. (Georgia Tech)

James S. Roberts, Ph.D. (Georgia Tech)

Rick P. Thomas, Ph.D. (Georgia Tech)


Title: Are prosaccades always automatic?: Validating the antisaccade task as a measure of controlled attention


Abstract: Recently, mainstream cognitive psychology has become aware of difficulties in measuring individual differences in the ability to direct attention in a goal-direct manner. Such difficulties may suggest that attention control is not a measurable general cognitive ability but may instead be highly task-specific. Accuracy rates from the antisaccade task are a notable exception to the measurement difficulties often seen in other tasks, but the measure’s construct validity has been questioned. Some researchers have argued that antisaccade accuracy is a function of individual differences in general processing speed (e.g., Rey-Mermet et al., 2019). The present study evaluated this position in a combined differential-experimental study. I assessed whether the adaptive procedures adopted by previous studies in non-attention-demanding tasks increased attention control demands, leading to inaccurate estimates of criterion-related validity. I compared two versions of the prosaccade task (a non-attention-demanding variant of the antisaccade task), a non-adaptive version and an adaptive version which adjusted the presentation duration of a target stimulus on a trial-by-trial basis. I also attempted to eliminate the relationship between antisaccade accuracy and working memory capacity/fluid intelligence by accounting for speed measures from both prosaccade tasks. Mean pupil size was larger in the pre-target period of the adaptive prosaccade task than in the non-adaptive prosaccade task, suggesting the adaptive procedure made the task more effortful. Crucially, however, no matter how I attempted to control for processing speed, I could not eliminate the relationship between antisaccade accuracy and cognitive abilities, implying that antisaccade accuracy is not merely a proxy measure for general speed.