to influence the development of subsequent cognitive cycles. If that is the case,
attention and central executive functions such as sustainment of goals, task
switching and inhibition can be understood as the boosting and suppression
(Gazzal ey et al. 2005) of localized processors that is triggered by a global
broadcast, meaning that there is no little man shouting orders from the rooftop,
but only emergent behavior arising from the complex interaction of multiple
systems that have been fine-tuned by evolution.
1.3 A neurocognitive approach to processing and storage
The reduction of attention to global broadcast-dependent boosting and sup-
pression does not mean, nonetheless, that some regions of the brain cannot be
thought of as responsible for cognitive control: they would have to be those
that are well situated to inhibit or boost across long distances, and to receive
bottom-up feedback that will regulate their functioning. The most studied
‘control region’ regarding WM is the prefrontal cortex, which fulfills those
requirements by being highly interconnected with subcortical regions, and also
part of the network that engages during the performance of WM-related tasks,
the Multiple Demand (MD) network (Duncan 2 010). The MD network is the
physical realization of the process described in the global workspace model, as
it connects with parietal regions of the cortex dedicated to sensory, cognitive
and memory processing. Such long-distance interconnectivity should be the
product of synchronization of neural oscillations (Buszáki 2006; Kitzbich ler
et al. 2011). A key finding by Duncan ( 2010) is the presence of independently
evolving patterns of activation in the MD network (orthogonality), which would
be useful in sustaining identity and content of stimuli across cognitive cycles.
From the perspective of the MD network, prefrontal cortex can be understood
as an enhancer of capacities that are performed somewhere else in the brain,
and only in this interconnected scenario can be said to support higher-order
cognition, including learning and memory processes, but also the implementa-
tion of goals and abstract representations (cf. Fedorenk o and Thompson-Schill
2014). This enhancement can be understood as an increase in the time window
in which a localized processor remains active (a quantitative measure that can
correspond to the operation of refreshing, cf. Rowe et al. 2000), and also as a
qualitative increase in the granularity of represented content (Mercado 2008).
As an instance of the latter, Miller et al. (2011) discovered that disrupting
prefrontal cortex activity blurs the distinctiveness with which occipital visual
areas activate to face versus scene stimuli.
Since Goldman- Rakic (1987), however, a theoretical assumption about the
standard WM model was that the two perceptual buffers could be localized in
the prefrontal cortex, correlating with sustained activity of some of its neurons
after the presentation of a stimulus. Indeed, the literature on areas of the prefrontal
cortex supporting short-term memory functions is large. Postle ( 2006) lists some
of the buffers that have been proposed: egocentric and allocentric spatial working
memory, including hand-centered vs. eye-centered vs. foot-centered spatial
108 Gonzalo Castillo