Publications between 2021 and 2025
2025
2024
Shang, L.*, Yeh, L.-C., Zhao, Y., Wiegand, I., & Peelen, M. V.* (2024). Category-based attention facilitates memory search. eNeuro, 11(2), ENEURO.0012-24.2024.https://doi.org/10.1523/ENEURO.0012-24.2024’
Highlight: 1) multiple object templates can be activated at the same time in manner of one categorical single-chunk template; 2) people can categorically reject distractors before memory search starts.
We frequently perform memory search to decide whether the object in front of us is one of several objects we look for; an example is deciding whether a specific product in the supermarket is on our shopping list. Previous research showed that this decision slows down as the number of objects we look for increases. However, if the viewed object (e.g., a milk bottle) is of a different category than the memorized set (e.g., various types of fruit), memory search is very efficient. Why?
One possibility is that attention biases visual processing, resulting in the quick rejection of objects belonging to different categories, largely avoiding memory search. Alternatively, all objects may enter the memory search stage, and the search itself is more efficient. To test this, in Exp 1, participants first memorized sets of either animate or inanimate objects and then decided whether individual objects were part of the memory set. Our analyses focused on the EEG response to non-target objects, which were all unique in the experiment. Results supported the attentional selection account: the categorical match between the non-target object and the memory set influenced the evoked response (N1) and representational strength (decoding) before 200 ms, i.e., during visual processing.
If participants indeed directed attention to the category of the memory set, we would expect that category-matching objects attract spatial attention. To test this, in Exp 2, we presented two images side by side and measured the lateralized N2pc EEG component indexing attention. Results confirmed this prediction: when both objects were non-targets, we found that attention was quickly directed to the non-target object that matched the category of the memory set.
Together, these results show that participants spontaneously used the memorized objects’ shared categorical features to direct attention at that level, leading to the efficient rejection of non-target objects belonging to other categories.
2023
2022
2021
Shang, L., Little, D. R., Webb, M. E., Eidels, A., & Yang, C.-T.* (2021). The workload capacity of semantic search in convergent thinking. Journal of Experimental Psychology: General, 150(11), 2230-2245. https://doi.org/10.1037/xge0001045
What is it about?
Convergent thinking tasks require individuals to generate a solution that depends on combining multiple cue components (i.e., that are multiply-constrained). For example, a commonly used convergent thinking task, the remote associates task, requires participants to find a fourth word related to three cue words (e.g., potato, tooth, and heart can be combined with sweet to form three compound words). Most previous studies focused on exploring the relationship between candidate solutions, whereby the search strategy and mechanisms of convergent thinking could be investigated. The present study attempts to explore this issue from another point of view. We are concentrating on understanding how multiple cues interact with each other via workload capacity, and finally reveal individual differences in workload capacity, suggesting individuals are much more likely to accept different search strategies to solve multiply-constrained problems. We also observe a negative correlation between workload capacity and convergent thinking.
Why is it important?
Our study provides a new possible interpretation to understand the mechanisms of convergent thinking. Traditional wisdom holds that working memory (WM) may play an important role in convergent thinking, but discrepant findings do not support directly associating convergent thinking with WM. Hence, Chuderski (2014) proposed that reasoning mediates the effects of WM on convergent thinking which was still based on WM. However, the present study bypasses WM and switches to workload capacity, a crucial but easily overlooked factor, by which a negative relationship is directly observed between workload capacity and convergent thinking. This contributes to the current understanding of convergent thinking, despite limitations to be solved by future study (e.g. measure context invariance, directly comparing the number of candidate solutions loaded).