The visuospatial sketchpad binds visual features to objects fi les, indexes that
mark potential objects to which visual properties are attached (Kah neman et al.
1992). Their separation is motivated by the fact that it is possible to fi nd
modality-dependent performance when simultaneously carrying out a visual and
a language comprehension task (Bad deley and Hitch 1974).
The complexity and noise of the environment are automatically fi ltered through
these perceptual buffers, providing the material on which processing mechanisms
can act. Within the model, these processing mechanisms are summarized under
the notion of the central executive, a domain-general capacity to implement
goals. In order to achieve those goals, the central executive can switch between
tasks and integrate perceptual information into coherent multimodal episodes,
effectively supporting what we call cognition. This integration and coordination
is assumed to be based on the allocation of top-down attention (Tre isman and
Gelade 1980), which can be considered as the fuel of the central executive.
Top-down attention results from the existence of volition, and is opposed to
bottom-up attention, which refers to the automatic shifts of attention that are
triggered by the sudden presence of unexpected stimuli in our fi eld of percep-
tion. From a WM perspective, triggers of bottom-up attention need to be
avoided to effectively carry out the resolution of a specifi c cognitive task, so
the central executive also has the function of shielding itself from interferences
(inhibition) (Miy ake et al. 2000).
Figure 7.1 The multicomponent model of WM (adapted from B addeley 2012).
“Visual semantics”, “Episodic long-term memory” and “Language” cor-
respond to ‘crystallized’ information stored in long-term memory, while
WM itself comprises the rest of the systems.
104 Gonzalo Castillo