What the orbitofrontal cortex does not do
Thomas Stalnaker, Nisha Cooch & Geoffrey Schoenbaum, US National Institutes of Health
Nature Neuroscience, Vol. 18, No. 5, May 2015
Summary and review of the above paper
The paper focuses mainly on the lateral orbitofrontal. Orbitofrontal damage can lead to difficulties in switching from behaviours that were previously advantageous, but have ceased to be so. However, the ability to make an initial discrimination, as opposed to altering an existing one, is usually unaffected. Nevertheless, changes in response are not seen as a core function of the orbitofrontal, and damage to the orbitofrontal is seen to effect a range of other tasks.
Final account of value
The authors instead stress the importance of the orbitofrontal in providing a final account of value. This involves converting information about the probability, size, timing and costs of particular rewards into a common neural currency. Single neurons in the orbitofrontal have been found to fire, not in response to any specific feature of a reward, but only in response to subjective value. Subjective value and neural firing rate have a linear relationship, and are found to be predictive of idiosyncratic changes in preference. This result is taken to rule out more simplistic reasons for orbitofrontal responses, such as the suggestion that subjects reacted to the unchanging ingredients of fruit juices that were being consumed. Subsequent studies have replicated these results, and human brain scanning has also confirmed the result. Results in humans show blood oxygen levels correlated to a subject’s willingness to pay during a decision-making task.
It is suggested that preferences among familiar rewards may be made independently of the orbitofrontal, leaving the this brain region do deal with more novel choices. The orbitofrontal is not seen as being necessary for choices where it has been possible to pre-compute the choice. The authors also stress that other parts of the brain produce the same sorts of value signals as the orbitofrontal, notably fronto-parietal areas and especially the anterior cingulate, which is involved with the dimensions of possible rewards. The orbitofrontal is not apparently needed to discriminate between different sized rewards, but it is needed to discriminate between the outcome reward.
Orbitofrontal activity is seen as being related to the processing of a model of the possible outcomes. It is thought to be needed when brain processing has to link two independently acquired pieces of information. The orbitofrontal is seen as related to processing ‘on the fly’, using information acquired since earlier learning processes. This brain region is argued to be able to experience hypothetical, imagined or predicted outcomes. In particular it can handle comparison of things that have not been experienced together.