Creativity is a mental journey between ideas or concepts that involves either a novel route or a novel destination. Many underlying causes have been suggested, including divine intervention and serendipity, but an adequate account must consider its neurobiological bases (Zeki, 2001). Brain theories of creativity might be divided in the broadest terms into global theories (based on universal aspects of brain function or universal influences on it) and local theories that emphasize certain components. Links between creativity and intelligence (Hofstadter, 1995; Sternberg, 2001), transmitter function (Beversdorf et al., 1999; Schrag and Trimble, 2001) or mood (Ludwig, 1994) all arguably reflect global mechanisms. In terms of local theories, there is ongoing debate about the relative importance of frontal and posterior association cortices (Heilman et al., 2003) and the relative importance of the right and left hemisphere (Amaducci et al., 2002). Other structures that have been emphasized include the cerebellum, albeit as part of a broader network (Chavez-Eakle et al., 2007).
Particular problems with the study of natural creativity are the fact that a given creative event is unique to any individual and the restriction of creativity to particular circumstances; the creative process that led Einstein to develop a special theory of relativity only happened once (further rehearsal of the concepts does not constitute creativity) and we would not dispute his creative genius based on a consideration of his violin playing. The creative process is therefore not well suited to the use of traditional psychological methods based on the observation of group performance over multiple trials in stereotyped circumstances. Specific forms of creativity can be studied using imaging in constrained environments (Geake and Hansen, 2005; Chavez-Eakle et al., 2007; Mashal et al., 2007), but such studies beg the question whether natural creativity can be captured by tests that depend on particular interpretations of its generic basis. In this issue of Brain, Seeley and colleagues take a more ethological approach by observing the effect of focal degenerative brain disease on the natural expression of creativity. They describe a fascinating example of brain degeneration not only sparing creativity but being preceded by increased creative behaviour.
The creative output of a scientist and artist with primary progressive aphasia (PPA) is described and related to the distribution of brain atrophy measured prospectively by voxel-based morphometry. This was made possible by the existence of MRI scans before the onset of aphasia that were carried out to follow an acoustic neuroma. The patient started painting at the age of forty-six and the paintings evolved into complex cross-modal representations by the age of fifty-two, eight years before any speech symptoms. Her painting is assessed from before the onset of aphasia to a stage when she was still painting despite being mute. What is particularly interesting in this case is the focus that the patient developed on the work of another progressive aphasia sufferer, Maurice Ravel (Amaducci et al., 2002), before the onset of her illness: a painting based on Bolero executed before the aphasia converts the repetitive musical form into a repetitive visual form. Another painting produced before the aphasia demonstrates a different conversion, from mathematics, in which the number pi is represented in visual form. During the illness itself the painting style became increasingly close to life. An evolving pattern of left frontal and temporal cortical degeneration associated with her aphasia is demonstrated in the accompanying voxel-based morphometry, with associated increases in grey matter in the right parietal cortex. The behavioural data are interpreted by the authors in terms of an increase in ‘transmodal creativity’ realized by intact right posterior mechanisms that are released from inhibition by the frontal cortex.
The work raises a number of interesting questions, the most central of which is whether the study really demonstrates increased creativity corresponding to the disease process? The early paintings demonstrate a form of conversion of acoustic or mathematical structure to visual form beyond simple translation and evidence true creativity. However, peak creativity defined in terms of the cross-modal work occurred eight years before the speech symptoms and structural brain changes. When the disease was established she exhibited what is arguably a less creative style. That might still represent a gain of function compared to when she did not paint, although measuring the creativity of her previous scientific work is even more difficult than judging the creativity of her artistic work. However, the other cases of increased creativity in early PPA reported by this group argue against coincidence, and it is possible that the sensitivity of voxel-based morphometry is not sufficient to show early structural correlates of disease at a time when the suggested gain of function was most marked.
How can we define creativity in a robust way? The methodology of arts criticism leads to crude metrics such as the ranking of creative endeavours, likened by one observer to ‘World Federation Wrestling for middle-class people’ (Costello, 2003). That is not a basis for useful metrics that might correlate with brain morphology. In the laboratory, quantification of creativity in the form of ‘fluid analogies’ is possible when the input to subjects and output from them is constrained (Geake and Hansen, 2005), but how can we quantify spontaneous creativity? Quantification of creativity based on information theory is conceivable, based on characterizing the novelty and complexity of either concepts or the relationship between them. Boléro would not actually come out highly in an assessment of creativity based on information theory, due to the repetitive structure, whilst the silent piano piece 4’33’’ by John Cage corresponds to a baud rate that can only ever be beaten. In visual art the use of information theory to define creativity would probably suggest that photorealism was more creative than cubism, whilst in literature the use of related measures like automated textual analysis (Garrard et al., 2005) might place Hemingway at the bottom of the creative heap. But the ultimate futility of possible information theoretic approaches is highlighted by a consideration of the richest possible creative outputs based on metrics derived from it: acoustic noise, visual noise or random words. Information theory can never capture the holy grail of creativity: beautiful concepts, ideas or images, which can be very simple. A reverse view of creativity is an ‘ability to introduce order into the randomness of nature’ (E. Hoffer, unpublished). But, at the limit, that takes us back to 4’33’’ as the ultimate expression of creativity. We are left, then, with an imbalance between the anatomical precision of voxel-based morphometry and the absence of any robust measure of creativity to which this might be related.
If the association between this pattern of brain structural change and creativity is accepted, it raises further questions about the adequacy of local theories of creativity. These all emphasize areas of association cortex that mediate connections between cortical areas (or the cerebellum as a subcortical structure with extensive cortical connections). From first principles, posterior association cortex might be adequate for the abstraction between symbolic forms communicated by the senses, whilst abstraction between ideas could involve frontal association cortex. The current study suggests that isolated posterior mechanisms are not an adequate basis for creativity expressed through abstractions between sense data.