Logo Logo
Switch Language to German

Harmelen, Anna van ORCID logoORCID: https://orcid.org/0000-0003-3935-2209 (August 2023): Investigating stimulus-driven selective attention inside visual working memory. Master Thesis, Faculty of Biology, Ludwig-Maximilians-Universität München. [PDF, 5MB]

[thumbnail of Thesis by Anna van Harmelen for completion of the Master of Neuroscience at the Graduate School of Systemic Neurosciences]
Published Version
Download (5MB)


Multiple objects can exist within (visual) working memory simultaneously, and internal selective attention is required to select amongst those objects. It is known this internal selection can occur in a goal-driven manner (Griffin & Nobre, 2003; van Ede et al., 2019), but the question remains whether this can also occur in a stimulus-driven manner. To answer this question, van Ede et al. (2020) developed a retrocue visual working memory task. They were able to induce attentional shifts with a completely uninformative retrocue, an effect that is best explained by the occurrence of stimulus-driven internal attention. Using eye-tracking, van Ede et al. showed that these attentional shifts were accompanied by a shift in gaze position towards the original location of the currently active memory item. Liu et al. (2022) showed that an internal attentional shift is often (also) accompanied by a bias in microsaccade direction.

However, the retrocue employed by van Ede et al. was only completely uninformative in one third of all trials, creating the possibility that the retrocue retained some implied informational value for participants. Our main research question therefore is: is a completely uninformative cue capable of and sufficient for creating an involuntary attentional-capture effect? If so, is this again reflected in an ocular signature consisting of a bias in gaze position and microsaccade direction? And is the behavioural effect that results from the attentional capture due to a benefit to the cued memory item, a cost to the uncued memory item, or a combination of both?

To answer these questions, the visual working memory task from van Ede et al. (2020) was employed after changing it in two critical ways. Based on the previous research we expected to find a behavioural attentional-capture effect, by observing less reproduction error and/or shorter decision times when the capture cue colour-matched the eventually probed item. We also expected that this behavioural effect would be accompanied by both a bias in gaze position and microsaccade direction towards the cued memory item.

Two separate experiments were performed for this research, which only differed in the shape of the capture cue. Experiment 1 produced no significant effect of capture cue congruency on decision times or accuracy, and this was accompanied by an extremely small, but significant, gaze bias. No significant microsaccade bias was found. Experiment 2 produced a significant effect of congruency on decision times, but not on accuracy: participants responded faster in congruent trials than in both neutral and incongruent trials. On group-level, this behavioural effect was again accompanied by a small, but significant, gaze bias, and no significant microsaccade bias was found.

The relationship between behavioural effect and ocular bias was further investigated by calculating a decision time effect size and a microsaccade bias effect size. A median-split of decision time effect size showed a significant microsaccade bias was present in the above-median decision time effect group. This microsaccade bias occurred within the expected timeframe, as based on previous research by Liu et al. (2022).

We concluded that a completely uninformative capture cue is capable of and sufficient for creating an involuntary attentional-capture effect within a visual working memory task. In a subset of participants this was also reflected in a significant bias in microsaccade direction, but this did not occur at a group-level. Finally, it seems the behavioural effect is due to a benefit for cued items, at no cost to uncued items. The current experiment complements earlier work by providing further proof that stimulus-driven internal attention does indeed exist and is sometimes accompanied by an ocular signature, consisting of at least a microsaccade direction bias. This research emphasises that the external world plays an important role in shaping the contents of our mind.

Actions (login required)

View Item View Item