The effect of perceptual noise on the Eriksen flanker effect: insights from Bayesian parameter estimation & model comparison

Deakin, Jordan Theresa ORCID: 0000-0003-2796-5622 (2024). The effect of perceptual noise on the Eriksen flanker effect: insights from Bayesian parameter estimation & model comparison. University of Birmingham. Ph.D.

Preview

Deakin2024PhD.pdf
Text - Accepted Version
Available under License All rights reserved.
Download (6MB) | Preview

Abstract

Decades of research using the Eriksen Flanker Task have provided valuable insight into the mechanisms of visual attention. However, this body of work has largely overlooked a critical aspect of our visual experience – the dynamic and noisy nature of our visual environments. This thesis aims to address this gap in the literature by exploring how adding noise to the stimuli influences attentional focusing in the flanker task. To reach this aim, we present three novel flanker tasks, two of which explore noise in the motion domain while the final study explores noise in the color domain. Furthermore, by comparing these results to a standard task, we explore differences in flanker suppression across noisy and standard, noise-free conditions. Within this field, there is also debate as to whether increases in attentional selectivity, or the suppression of flankers, proceeds as a gradual or discrete process. To test these differing views, two computational models have been proposed which realize the two viewpoints, namely the Shrinking Spotlight Model (SSP) of White et al. (2011) implementing the former and the Dual-Stage Two-Phase Model of Hübner et al. (2010) implementing the latter. Using the rarely used Bayesian approach, we fit both models to the data, and through a combination of average performance analyses, distributional analyses and model simulation, we can explain how noise influences the flanker effect and how this is accounted for under each theoretical framework. Firstly, we find that noise does indeed influence the flanker effect, in that adding noise to the stimuli impedes flanker suppression. Secondly, by independently manipulating the noise level in targets and flankers, we find solid evidence that the adverse effect of noise stems mainly from adding noise to the target. Finally, and in contrast to previous research using the standard paradigm, we find decisive evidence that a discrete improvement model (DSTP) best accounts for flanker suppression under noisy conditions. Our distributional analyses allow us to ascertain that the superiority of DSTP can be localized mainly to its ability to better model data patterns indicative of considerably poor flanker suppression, particularly observed in high noise conditions. Combining our data analyses with the predictions of both models, we propose a theory of flanker suppression under noisy conditions, in which the extent of flanker interference is primarily determined by the saliency of the target: the less salient the target, the greater the flanker interference.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):