Feature-based attentional weighting and spreading in visual working memory

Abstract

Attention can be directed at features and feature dimensions to facilitate perception. Here, we investigated whether feature-based-attention (FBA) can also dynamically weight feature-specific representations within multi-feature objects held in visual working memory (VWM). Across three experiments, participants retained coloured arrows in working memory and, during the delay, were cued to either the colour or the orientation dimension. We show that directing attention towards a feature dimension (1) improves the performance in the cued feature dimension at the expense of the uncued dimension, (2) is more efficient if directed to the same rather than to different dimensions for different objects, and (3) at least for colour, automatically spreads to the colour representation of non-attended objects in VWM. We conclude that FBA also continues to operate on VWM representations (with similar principles that govern FBA in the perceptual domain) and challenge the classical view that VWM representations are stored solely as integrated objects.

Mnemonic and attentional roles for states of attenuated alpha oscillations in perceptual working memory: a review

“…Finally, it is noteworthy that the vast majority of studies linking attention to WM (as well as alpha to attention) regard spatial‐based prioritisation mechanisms. It has recently been shown that visual feature dimensions (such as orientation and colour of all items in the array) can also be dynamically up‐ and downregulated during visual WM retention (Ye et al ., ; Niklaus et al ., ). Whether alpha modulations also support such feature dimension‐based attentional prioritisation (as reported for the perceptual domain; see Snyder & Foxe, ), remains an interesting target for future research.…”

Age Differences in the Efficiency of Filtering and Ignoring Distraction in Visual Working Memory

“…Information regarding the relevance of some features over others may become available, however, only after the objects were already encoded to WM. A handful of recent studies have also investigated whether people can selectively weight features already encoded to WM, or in other words, whether they can "ignore" irrelevant features in WM [10,29,30]. In these studies, participants encoded a set of multifeatured objects (e.g., colored and oriented gratings).…”

Two distinct mechanisms of selection in working memory: Additive last-item and retro-cue benefits

“…In a typical retro-cue experiment, after presentation of a memory array, a cue identifies the location of one item that is most likely to be tested in a subsequently following recognition test. A validly retro-cued item can be accessed faster and more accurately in comparison to conditions where no cue, or an uninformative cue, is provided (Griffin & Nobre, 2003;Niklaus, Nobre, & van Ede, 2017;Rerko & Oberauer, 2013;Souza, Rerko, Lin, & Oberauer, 2014;Souza, Rerko, & Oberauer, 2016, 2014van Ede, Niklaus, & Nobre, 2016).…”