One of the main functions of the prefrontal cortex
that has been alluded to in a great deal of the literature
is in working memory
. The role of working memory seems to be in guiding and adapting behaviour
through its ability to hold information “on-line
”. This has been demonstrated many times through the impaired performance of monkeys
with prefrontal damage on delayed matching to sample
, and delayed non-matching to sample tasks
(e.g., Passingham, 1975
). Such inability to adapt and change behaviour is seen in humans
with frontally located lesions. Such patients tend to be absent-minded
and very easily distracted. They do not seem to organise and plan actions very well. For instance, patients have been known to be completely unable to do shopping efficiently. They might go into the shop, purchase just one item, return to their car, re-enter the shop, buy one more item, return to their car, and so forth (Shallice and Burgess, 1991
). Difficulties in such “multiple sub-goal tasks” are fairly common among frontal patients.
Patients with frontal lesions also tend to display behavioural responses inappropriate to the situation and this can be clearly seen in the Wisconsin Card Sort Test (where cards have to be sorted according to colour, shape, or number, depending on whether the experimenter says “right” or “wrong”). Using this apparatus, Milner (1963) observed that patients with frontal lesions had difficulty in figuring out the first sorting principle, and furthermore, were impaired in the ability to shift to a second sorting principle when the so required.
Can mnemonic deficits explain the problems of patients with prefrontal damage?
Memory does seem to be a likely candidate for explaining the aforementioned problems of frontal patients, particularly as they tend to explain their actions (such as going on a simple errand and being later found on the local golf course) by saying that they had “completely forgotten” what they were previously intending to do. Also, Hécaén (1964) has previously identified that frontal patients can have memory problems. However, although such problems might initially strike one as caused by deficits in working memory, this does not necessarily seem to be the case. For instance, Shallice and Burgess (1991) account for problems in the performance of multiple sub-goal tasks by the lack of a “supervisory system” (Shallice, 1982), controlling processes such as goal articulation, plan formulation, and the creation and triggering of markers (a “marker” can be thought of as a sort of message that some future behaviour or event is significant to current goals). Rolls (1998) attributes the problems in performing the Wisconsin Card Sort Test, not to working memory problems, but rather to problems in altering behaviour to changes in stimulus-reinforcement associations. Rolls suggests that the “orbitofrontal cortex is involved in emotional responses by correcting stimulus-reinforcement responses when they become inappropriate”.
Roll’s stimulus-reinforcement view of the orbitofrontal cortex seems to contradict any “working memory” processes, at least in this frontal area. The fact that there is a significant projections from primary taste cortex (Rolls et al., 1990) and olfactory areas (Rolls and Baylis, 1994) seems to be good supporting evidence of this view, since taste and olfactory stimuli can act as primary reinforcers, that is, they can reinforce without any learning that they are a reward or a punishment. Rolls further suggests that visual inputs to the orbitofrontal cortex can be linked with reward associations (Thorpe et al., 1983; Rolls et al., 1996). This is quite possible as the visual inputs are projecting from the temporal lobes, which are part of the ventral visual processing stream and are able to detect the form of objects. Face information appears to be another significantly represented type of associative processing in the orbitofrontal cortex, with many neurons responding in similar ways to “face neurons” in the temporal cortical visual areas (Rolls, 1984a, 1992a, 1994a, 1995b, 1996). Hasselmo et al. (1989) have made the link here to social reinforcement conveyed by facial expression being processed in the orbitofrontal cortex.
Attentive processes in prefrontal cortex
The prefrontal cortex has also been implicated in attention. Passingham (1998) highlights the role of the dorsal prefrontal cortex in subjects attending to their own actions. Using PET (Positron Emission Tomography), Jenkins et al. (1994) and Juepter et al. (1996) have shown that the dorsal prefrontal cortex is activated during the learning of new tasks, but not significantly activated above baseline levels when performing prelearned tasks (or tasks which have been “learned by heart”). Passingham concludes that this data agrees with Shallice’s (1982) idea that the prefrontal cortex acts as a supervisory attentional system, influencing the actions subjects perform according to whether they are routine or not. The attentive role of the prefrontal cortex seems to be specific to action. For instance, when attending to spatial location, imaging studies have shown that it is parietal cortical areas that are activated, whereas attending to action does not activate these areas (Posner and Petersen, 1990; Corbetta et al., 1993). Doral prefrontal cortex also has been shown to be activated when subjects are required to make their own decisions about which movements to make, rather than being given a series of movements to learn by an experimenter (Dieber et al., 1991; Frith et al., 1991).
Decision making in prefrontal cortex
Decision making is indeed another function that has been implicated in the prefrontal cortex. A good example of how this ability can be impaired in humans with prefrontal damage is a patient described by Eslinger and Damasio (1985). The patient was a man who had an IQ of 130, but just could not organise his life. Six years after the removal of a bilateral orbitofrontal meningioma, they found that he had been dismissed from a series of jobs despite having the required skills and an appropriate manner (some patents are often impulsive, apathetic, and display inappropriate affect (Blumer and Benson, 1975)). He had also gone bankrupt and had two divorces in two years. He would take literally hours to decide even simple matters. For instance, when going out to dinner he would need to investigate the menu, seating plan, atmosphere, and management for each restaurant. He would also drive to each establishment to see how busy each was. Even then, he still could not decide.
A study by Bechara, Damasio, Tranel, and Anderson (1998) is effective in confirming that this process is located in the prefrontal region, and in addition, that decision making is a distinct process from the working memory processes that are known to exist in prefrontal cortex. The experimenters tested patients with various frontal lesions with delay tasks that assessed working memory and with a gambling task that assessed decision making. It was found that among the subjects with ventromedial prefrontal lesions, those with more anterior lesions had an impaired performance on the gambling, but not the delay task, whereas those with more posterior lesions were impaired on both tasks. Among the patients with dorsolateral/high mesial lesions, those with right hemisphere lesions were impaired on the delay task, but not the gambling one, whereas those with left hemispherelesions displayed no impairment of either task. This experiment successfully demonstrated a cognitive and anatomical double dissociation between decision making deficits (localised in anterior ventromedial prefrontal cortex) and working memory (localised in right dorsolateral/high mesial prefrontal cortex). The authors suggest that the explanation for this is in accordance with preliminary evidence from Anderson et al. (1996) that the decision making impairment, caused by anterior vetromedial lesions, is linked to “insensitivity to future consequences”. In other words, such patients seem oblivious to the future and instead are guided by immediate prospects, whether they are positive or not. They specifically do not link their results to an explanation by failure to inhibit previously rewarded responses (perhaps suggesting that Rolls’ stimulus-reinforcement ideas are confined to orbitofrontal cortex). Nor do they agree with the view that there is an impairment in selective attention in that subjects are unable to shift their attention to a new and beneficial objects in the gambling task.
Integration of visual information
The prefrontal cortex has been further implicated by Rao, Rainer, and Miller (1997) in integrating the visual information from the ventral and dorsal visual processing streams. This function is therefore concerned with integrating the form and colour of objects (“what” – ventral stream) with spatial location (“where” – dorsal stream) in order to direct action to objects. This study found that 48% of neurons recorded from were either object-tuned or location-tuned. Also there was a split in that object-tuned neurons tended to be in ventrolateral prefrontal cortex, whereas location-tuned neurons tended to be in dorsolateral prefrontal cortex. However, they also found that 52% of the neurons recorded from where tuned to both object and location, and furthermore could mirror the demands of a task (by first conveying object information, then location information).
Re-emphasis of the mnemonic and "on-line" processing in prefrontal cortex
But still, the main function of the prefrontal cortex seems to be in mnemonic functioning. Petrides (1996) has identified the mid-dorsolateral prefrontal cortex being in a position to monitor and manipulate information within working memory due to its connections with the limbic region of the medial temporal lobe (which has mnemonic functioning) and orbitofrontal cortex. Petrides suggests that there is a distinction between the mid-dorsolateral and mid-ventrolateral frontal cortex insofar as the modality specificity of the information it processes. Petrides goes on to suggest a two-level hypothesis in that the vetrolateral frontal cortex has a role in strategic encoding and retrieval of specific information from posterior cortical regions, and comparing, deciding, and selecting information in short-term and long-term memory. The second level of his hypotheses is localised in the mid-dorsolateral frontal cortex and is involved with monitoring and manipulating working memory on the basis of task requirements and current plans.
However, a later review by Rushworth and Owen (1998) disagrees with Petrides. They suggest that there is not modality difference between mid-dorsolateral and mid-ventrolateral frontal cortex, and the dorsolateral areas instead process the manipulation of information in working memory, and the ventrolateral frontal cortex selects currently relevant information.
Organisation of prefrontal cortex
What is clear is that the prefrontal cortex plays a role in several different functions that could be thought of under the umbrella term of “executive processes”. However, it has long bee thought that there must therefore be some kind of modularity to the prefrontal areas, according to the information processing occurring within. The earliest ideas were in terms of mapping due to function (e.g., Fulton, 1950; Mishkin, 1964, Pribram, 1987). There was also another school of thought (Owen et al., 1996) that suggested a hierarchical processing relationship between superior and inferior dorsolateral cortex. However, Goldman-Rakic (1987) has postulated an organisational idea based on subregions according to informational domain, where different domains share a common specialisation. In other words, content and not function is mapped according to cytoarchtitectonic fields.
According to the Goldman-Rakic idea, each autonomous subdivision integrates attentional, memorial, and possibly affective dimensions of behaviour by virtue of network connectivity with relevant sensory, motor, and limbic areas of the brain. This theory can account for the diversity of specific behavioural impairments that are suffered by frontal patients and animals alike. Goldman-Rakic describes this as a “well designed parallel processing architecture for the brain’s highest level of cognition” (Goldman-Rakic, 1998).
The Goldman-Rakic organisational hypothesis is effective for accounting for the wide range of different processes that occur within the prefrontal cortex. However, its is no longer the favoured view of the organisation of the prefrontal cortex. Now a more hierarchical view has been adopted, with more basic processes happening in ventrolateral prefrontal cortex, and more complex ones in the dorsolateral prefrontal cortex. This seems plausible when looking at visual information processing. The ventrolateral areas receive a wealth of input from the temporal “what” processing areas, while the dorsolateral receives “where” (or more accurately, “how”) information from parietal areas. The dorsolateral processing is more complex and uses the ventrolateral processes to integrate the object-related information with the spatial information in visual processing.
What is clear is that the prefrontal cortex is the location of the most complex integrative processes in the brain, taking inputs from almost every area of the brain, and also being highly dependant on tasks and the context in which specific behaviour is required.
Bechara et al. (1998), Dissociation of working memory from decision making within the human prefrontal cortex, Journal of Neuroscience 18, 428-37
Goldman-Rakic, P. (1998), The prefrontal landscape, in; The Prefrontal Cortex, Eds. Roberts, Robbins, Weiskrantz, OUP, Oxford.
Passingham, R.E. (1998), Attention to action, in; The Prefrontal Cortex, Eds. Roberts, Robbins, Weiskrantz, OUP, Oxford.
Petrides, M. (1996), Lateral frontal cortical contribution to memory, Seminars in the Neurosciences 8, 57-63
Rao, S.C., Rainer, G., Miller, E.K., (1997), Integration of what and where in the primate prefrontal cortex, Science 1997 May 2: 276, 821-4
Rolls, E.T. (1998), The orbitofrontal cortex, in; The Prefrontal Cortex, Eds. Roberts, Robbins, Weiskrantz, OUP, Oxford.
Rushworth & Owen (1998), The functional organization of the lateral frontal cortex: conjecture or conjuncture in the electrophysiological literature, Trends in Cognitive Sciences 2, 46-53
Shallice, T., and Burgess, P.W. (1991), Deficits in strategy application following frontal-lobe damage in man, Brain 114, 727-41