Cautionary response strategy and impairment of post-conflict response selection underlie age-related differences in a location-based Stroop task

Abstract

Introduction: Research suggests that older adults have deficits in selective attention, a cognitive process often queried through the Stroop task. To tease apart whether this is due to failures to inhibit distracting information or to upregulate attention towards target information, younger and older adults completed a task called the Shape Stroop.

Methods: In this task, participants had to name the color of a shape that was occluded by another shape. Critically, congruent or incongruent Stroop words were placed in either the target shape, the occluding (distractor) shape or in the background. We first modeled performance as a function of age-group, Stroop word congruency, and location.

Results: The results indicate that older adults were more accurate but slower than younger adults to choose the correct shape color. For both younger and older adults, incongruent words induced slower reaction times when words were in the target location, indicating age-invariance in the Stroop effect. To further probe how early and/or late attentional processes contribute to performance and to interrogate the decision strategies adopted across different conditions, we also fit the dual-stage two-phase model of selective attention to our data.

Discussion: Our results indicate that older adults tend to be more cautious and require more information before choosing a stimulus to attend to or making a decision. Although older adults' ability to inhibit irrelevant information seems intact, they show signs of slower information processing in the later stages of attentional processing.

Keywords: aging; attention; decision; inhibition; perception.

Figure 1

The DSTP model of selective attention. In each graph, the top panel represents two phases of response selection, and the lower panel represents the late stages of stimulus selection. The x-axis of each panel represents time, with zero representing stimulus onset. The four solid black horizontal lines within each graph represent a response or a stimulus selection boundary. The red vertical line represents the time point on the x-axis, where the accumulation of evidence in the first phase of response selection (top panel of each graph) or in the late stage of response selection (top panel of each graph) reaches a boundary. Note that the first phase of response selection and the late stimulus selection processes compete to reach a boundary. Depending on the outcome of this competition, six different patterns can occur. If the evidence accumulation responsible for the first phase of response selection reaches the boundary first, the participant initiates a response that can be correct (A) or incorrect (B). On the other hand, if the evidence accumulation responsible for the late stimulus selection reaches the boundary first, the first phase of response selection is terminated, and the second phase of response selection begins, which is based on the input of the late stimulus selection stage. In this case, four possible outcomes can occur. The late stimulus selection processing may choose the target (C,D) or the distractor (E,F). Although the second phase of stimulus selection most likely chooses the same stimulus as the late stimulus selection processing (C,F), it may instead choose the other stimulus (E,D).

Figure 2

Schematic representation of different conditions in the Shape Stroop Task. Participants had to either indicate the first letter of the color of the shape that was on top, or underneath.

Figure 3

Overall performance results. Panel (A) shows the accuracy results and panel (B) shows the reaction time results. The mean of reaction time and accuracy within each condition is represented with solid black and green circles for older and younger adults, respectively.

Figure 4

Parameters of the DSTP model fit for younger and older adults as a function of word location. Congruency is incorporated into the model by including congruent vs. incongruent trials as a factor in the fitted DSTP models.