Does Short-Term Memory Load Influence Visual Search? An Oculomotor Study

ABSTRACT SUBMISSION DETAILS
Type of Submission:	Poster
Submission Date:	January 31, 2008
Review Status:	PENDING
Title:	Does Short-Term Memory Load Influence Visual Search? An Oculomotor Study
Subject Area:	Cognitive
Keyword:	Attention
Presenters:	Jibo He, University of Illinois, Urbana/Champaign Jason S McCarley, University of Illinois at Urbana-Champaign
Abstract:
A dual-task experiment examined the influence of STM load on visual processing and saccade targeting in visual search. Increased load altered saccade amplitudes, but did not appear to slow visual search nor to compromise foveal stimulus analysis. Results imply independence between STM maintenance and visual search.
Supporting Summary:
INTRODUCTION: A dual-task experiment examined the influence of short-term memory (STM) load on visual processing and saccade targeting in visual search. Performance of visual search under high and low memory load was compared through analysis of oculomotor data. The search task was designed to allow insight into the quality of the participants’ foveal analysis and saccade control. METHOD: Participants performed a visual search task while concurrently maintaining either a low or high memory load. The search task required participants to locate a circle (O) among a set of 35 gapped-circle (C) distractors. The memory task required participants to hold either a single alphanumeric character or six characters in STM. Each trial began with a fixation cross, followed by a memory set of either one character (low memory load) or six characters (high memory load). The visual search task was classified as coded or uncoded. In the coded condition, the Cs were oriented to face in the direction of the target, such that the participant could use distractors to guide search toward the target’s location. In the uncoded condition, distractors were oriented randomly. Comparison of performance in the coded and uncoded conditions thus provides a measure of the participants’ ability to utilize foveal analysis of distractor orientation to facilitate saccade targeting during search (Hooge & Erkelens, 1998). Memory performance was measured with a recognition test after the visual search each trial. RESULTS: Visual search was significantly more efficient under coded than uncoded conditions, as evident in changes in response times and saccade frequencies. Mathematical analysis (Wagenmakers et al., 2007) of saccade latency and accuracy data also revealed an increase in information accumulation rate for saccade target selection under coded search conditions. High memory load increased first-saccade latency each trial and led to larger saccade amplitudes, but otherwise did not hinder performance. No significant differences in memory recognition or saccade targeting accuracy across the four conditions were found. DISCUSSION: For coded search task, the benefits of cues outweighed the cost of additional cognitive processing of direction of the cues than uncoded search task. Higher STM load did not appear to slow down visual search, nor to compromise participants’ ability to guide search on the basis of distractor analysis. Results imply independence between STM maintenance and visual search.