A primer on executive function in the prefrontal cortex

Gilbert and Burgess offer a brief review of our prefrontal brain cortex structures that enable our flexible responses to situations with alternative choices. I think it provides a good look-up reference, and so I want to excerpt a few of the summary figures and text here:

At the heart of most (but not all) theories of executive function is a distinction, or gradation, between routine (or ‘automatic’) and non-routine (or ‘controlled’) processing. Routine processing refers to mental operations that are well rehearsed or overlearned, for example reading out a word. By contrast, non-routine processing most commonly refers to mental operations that are used in situations when there is not a well-established stimulus-response association, or where a behavioural impasse has occurred (for example one notices an error, or realises that one is behaving in a sub-optimal fashion). The term ‘executive functions’ has become synonymous with those behaviours and abilities.

Determining the relative contributions of different frontal subregions to different executive functions is a highly complex matter, both theoretically and methodologically. On current evidence, however, one can make some preliminary suggestions. The figure illustrates some of the major subdivisions of the human PFC, which may be divided into lateral and medial surfaces. On the lateral surface, the PFC may be further subdivided into ventrolateral, dorsolateral, and rostral regions. Although the medial PFC is depicted as a single area in the figure, there is now strong evidence that this part of PFC can also be subdivided both on cytoarchitectonic and functional grounds. The figure shows the lateral surface split into a ventrolateral region (VLPFC), dorsolateral region (DLPFC) and rostral region (RPFC). The medial surface (MPFC) is illustrated as a single region, but recent studies indicate considerable anatomical and functional variation within this region as well. (Click to enlarge)


Ventrolateral PFC (VLPFC) is thought to be involved in comparatively simple tasks, such as short-term maintenance of information that cannot currently be perceived in ‘working memory’ (for example, memorising a phone number you have just been told, before keying the numbers into a telephone). It has also been proposed — although this is controversial — that different parts of the VLPFC are used to store different types of information (for example, the sound of a word versus its meaning). By contrast, dorsolateral PFC (DLPFC) has been most commonly implicated not so much in maintaining information that is no longer available in our environment, but in manipulating that information. For example, although DLPFC is probably not involved in processes such as remembering a telephone number, it does seem to play a role in more difficult tasks, such as dialling the number in reverse order (rearranging the digits that we have just been told). DLPFC has also been suggested to be involved in complex functions such as making plans for the future.

A brain region with strong projections to and from the DLPFC is the anterior cingulate cortex (ACC), part of the medial PFC. One influential theory proposes that this brain region detects the need for control, for example where there is competition between two or more ways of behaving in a certain situation, both of which may be triggered by events in our environment, requiring top-down input to resolve the conflict. It is suggested that the ACC does not itself provide higher-level modulation of lower-level processes, but instead signals to DLPFC when such higher-level modulation is required.

The largest, but most mysterious, sub-region of prefrontal cortex is the rostral PFC (RPFC). As a proportion of whole-brain volume, some have estimated the human RPFC to be twice as large as the corresponding region in the chimpanzee brain. Yet curiously, patients with damage restricted to the RPFC often perform well on standard neuropsychological tests, including ‘classical’ tests of executive function such as the Wisconsin card sorting test. Instead, patients with damage to this region seem to have particular difficulty in real-world ‘multitasking’ situations, such as organising a shopping trip when there are few strict constraints — participants are relatively free to organise their behaviour however they like — but there are also multiple instructions to be remembered, rules to be followed, and potential distractions in the environment. Recent accounts have focused on the role of RPFC in the most high-level human abilities, such as combining two distinct cognitive operations in order to perform a single task, trying to work out what other people are thinking (‘mentalising’), and reflecting on information we retrieve from long-term memory (‘source memory’, for example trying to work out when we last saw a person familiar to us). We recently put forward the unifying hypothesis that this brain region serves as a ‘gateway’ between cognitive processes directed towards current incoming perceptual information, versus information that we generate ourselves. We have also shown (see following figure) by a meta-analysis of functional neuroimaging results that there are distinct functions associated with different parts of the RPFC, with segregation especially between lateral versus medial regions, and between rostral versus caudal regions. (Click to enlarge)