| Summary: | <p>Flexible cognition necessary for everyday behaviour depends on working memory (WM)–the ability to keep information in mind for a later use. Not all items in WM are
represented equally. Attention can be allocated to select and prioritise task-relevant content for a forthcoming action.</p>
<p><p>Chapter 2-3 used behaviour, eye-tracking and electroencephalography (EEG) to investigate the consequences of selecting a single feature in WM for the representation
of other features belonging to the same item. Orientation judgements were faster and more precise when another feature of the same object was cued for an orthogonal task
during the delay. Colour selection in WM engaged spatial attention, indexed by contralateral alpha suppression and gaze bias, and was associated with a boost in the representation of the cued object’s memorised location (Chapter 2) and orientation (Chapter 3) as decoded from EEG voltages. The results provide evidence for obligatory attentional spread between an item’s features in WM and support a special status of location in selecting and binding features. Finally, EEG decoding revealed that the cued item was automatically compared with the probe at recall, suggesting selected WM
content may be in a prioritised state for interacting with new sensory input and guiding decision-making.</p>
<p>WM is more than a passive store of recent experiences and is fundamentally future-oriented. Chapter 4 tested whether WM contents are configured as prospective task-sets for guiding behaviour. In one experiment, a prioritised item drove responses in a task-specific manner even when irrelevant in the context of a nested task. In a second experiment, task-foreknowledge improved WM-guided behaviour and enhanced retro-cued prioritisation on reaction times. Together, the results indicate that prioritised WM items are configured as prospective templates optimised for context-dependent action. Neuroimaging studies are needed to elucidate the precise mechanisms of task-specific coding in WM. Chapter 5 proposes an EEG experiment, in the form of a Registered
Report, to test the hypothesis that prioritised content is stored in a functional state that is intrinsically linked to its anticipated use. I predict that patterns of neural activity coding for remembered items will be task-specific. The suitability of the analysis pipeline was demonstrated on simulated neural data.</p>
<p>The results of this thesis highlight a key role of attentional selection in preparing WM representations for future action and motivate further research into the ways information in WM may serve as functional states for decision-making.</p>
|
|---|