Diffusion Tensor Imaging (DTI) offers increased resolution compared to conventional structural MRI regarding white matter microstructure by measurement of water diffusion through cellular compartments in vivo [7]. Compared to more isotropic movement of water in gray matter, water diffusion in white matter moves anisotropically, meaning that water diffuses preferentially along the length of the axon compared to perpendicular to the axon. This anisotropic diffusion of water appears to be due to the highly structured axonal membranes and their associated myelin sheaths [8]. By tracking the diffusion of water in the brain a measure of fractional anisotropy (FA) can be derived. Higher FA suggests greater axonal coherence and myelination [7], increasing in a roughly linear manner that conforms to normal developmental brain processes [9], [10]. Measures of fractional anisotropy (FA) are considered to be overall measures of axonal integrity, reflecting either increased axonal caliber, increased myelin thickness, increased fiber coherence in a given direction, or some combination of these factors [11].

Both schizophrenic and bipolar patients have been shown to have reduced fractional anisotropy (FA) in the anterior thalamic radiation [12], [13] and uncinate fasciculus [14] within frontal brain regions. Similarly, reduced FA was observed within the uncinate fasciculus of a cohort with schizotypal personality disorder, providing strong support for the hypothesis that similar neural phenotypes may not result in full-blown clinical symptoms [15]. Finally, in normal subjects, the Neuroregulin-1 (NRG1) single nucleotide polymorphisms (SNP's) SNP8NRG243177 and SNP8NRG221533 were found to predict lower FA in the left anterior thalamic radiation [16]. As NRG1 has been found to predict higher risk of schizophrenia [17], [18] and bipolar disorder [19], and is linked with axonal myelination and migration [20], these authors hypothesize a mechanistic link between NRG1 within the anterior thalamic radiation and risk for psychotic disorders [16].

Figure 1 (upper panels) shows regions in which FA was inversely related to CCI (p<.05), with arrows pointing to MNI coordinate (in coronal, axial, and sagittal planes respectively) of the maximally significant cluster within the left anterior thalamic radiation.

Table 3 (upper) shows regions in which CCI was significantly inversely related to FA, RD, and AD within numerous regions within the left hemisphere, especially the anterior thalamic radiation. There was a small region within the right anterior thalamic radiation where the CCI was inversely related to FA. Openness was significantly inversely related to FA within the right anterior thalamic radiation. (p<.05), with arrows pointing to the MNI coordinate (in coronal, axial, and sagittal planes respectively) of the maximally significant cluster within the right inferior frontal white matter/anterior thalamic radiation (Figure 1 lower panels). Table 3 (lower) shows regions in which Openness was significantly inversely related to FA, RD, and AD within a region localized to right anterior thalamic radiation. There were no clusters where positive correlations were found.

Figure 2 demonstrates scatterplots within the maximally significant clusters obtained from the anterior thalamic radiation for CCI (left) and Openness (right). The upper panel shows individual subject data for FA; the middle shows RD, and the lower panel shows AD. Post hoc analyses (Table 3) revealed that the major contribution, in terms of significant cluster size, of the FA-CCI and FA-Openness relationships was driven by RD as compared to AD contribution.

The anterior thalamic radiations include association fibers linking the thalamus to projection zones within the prefrontal cortices, in one of many functionally segregated thalamo-fronto-striatal loops [32]. Interestingly, researchers have found that lower FA was “unrelated to age, duration of illness, current measures of psychopathology, prescribed medication, or the presence of psychotic symptoms” in bipolar and schizophrenia disorders, with implications that “white matter connectivity represents a disease or vulnerability related feature common to both disorders” [14]. In contrast to the current frontal and predominantly left lateralized CCI-FA findings, our previous morphometry study of creativity found generally posterior and right lateralized CCI-thickness relationships [22]. Researchers studying morphometric correlates of personality found an inverse correlation between thickness and Openness limited to the right inferior parietal cortex (i.e., angular gyrus) in an elderly cohort [33]. With so few studies in the extant literature utilizing morphometric measures and measures of Openness or DT, it is difficult to link systematically our DTI findings with these previous findings other than to note the general inverse brain-behavior nature of the relationships across research groups. Future studies in large normal cohorts undertaken with both DTI and cortical thickness measures, would further elucidate potential brain networks involved in creative cognition and personality. Similarly, measures of creative cognition and personality (e.g., DT and Openness) should be administered in tandem with DTI measures, both to patients diagnosed with psychotic disorders (e.g., schizophrenia, bipolar) and their unaffected relatives, to better understand the causal mechanisms underlying lower FA, creativity, and psychosis within frontal white matter.

The neural processes involving creative cognition can be included in the realm of optimal brain functioning. The present findings suggest a complex picture of “optimal” white matter microstructure in health—especially when the white matter correlates of DT and intelligence appear in direct opposition [34], [35]. Similarly disparate results were evident in our previous studies showing that DT ability was associated with both increased and decreased cortical thickness at various loci [22], and results from MR spectroscopy showing that N-acetyl-aspartate and DT relationships may be direct or inverse—depending on whether IQ falls above or below 120 [23]. These findings of “less is sometimes more,” broadly consistent with the notion of “neural efficiency,” were evident in the earliest functional studies of intelligence [36], and have subsequently been supported in both intelligence [37] and creativity neuroimaging studies [38]. A picture begins to emerge regarding the manifestation of complex cognitive processes in the brain, including the interplay of intelligence and creativity. Whereas more neural resources are often associated with higher intellectual capacity in a parieto-frontal network of brain regions [39], studies in DT appear to suggest that less is often better in a different network of brain regions, particularly fronto-cingulate-subcortical networks linked via white matter loops [40].

There are several important limitations for interpreting results. First, there does not exist one “creativity”; rather, this construct is hypothesized to reside upon a continuum between cognitive (i.e., scientific) and emotional (i.e., artistic) behavioral domains [41], [42]. Thus, when comparing scientists and artists directly, researchers have found lower lifetime rates of psychopathology for: 1) scientists compared to artists, 2) natural scientists compared to social scientists, 3) nonfiction writers compared to fiction writers and poets, and 4) formal artists compared to “expressive” artists [3], [4], [43]. These findings have led researchers to hypothesize a hierarchical structure of creativity across disciplines [42], which echoes the notions of “paradigmatic” (i.e., a fundamental model of events) versus “revolutionary” (i.e., rejection of doctrines) approaches as applied to the sciences [44]. The benefits of working within the lines of a given field appear to be lower levels of psychopathology; alternately, individuals with lower levels of psychopathology may be attracted to such endeavors. Similarly, there is increasing evidence that the cost of “revolutionary” approaches to creative endeavors, whether it is in the arts or sciences, may be associated with increased levels of psychopathology although, again, causative links are weak at best.

Second, measures of DT are only proxy measure of one aspect of creativity, as we and numerous other researchers have noted [5], [22], [29], with this measure “having to do with fluency of thinking and flexibility of thinking, abilities concerned with the ready flow of ideas and with readiness to change direction or to modify information” [21]. If the definition of creativity, as widely accepted, involves a tension between novelty and usefulness, then measures of DT appear to pull for aspects of novel cognitive production amenable to laboratory settings. Viewed within the context of the hierarchical construct of creative endeavors reviewed above, this tension plays out among creative actors by means of their chosen profession (more or less disciplined) and their role within that profession (more or less conforming) [42]. Thus, DT measures would generally appear to bias measurement of novelty over usefulness, would value measurement of less disciplined and conforming responses (as scored by our judges), and would be expected to have a higher overlap with psychopathology than creativity measures biasing “usefulness” (e.g., Creative Achievement Questionnaire) based on the incidence of psychopathology across creative professions [43].

Finally, DTI measures themselves lack the level of specificity desired to determine subtle overlap between patient and normal cohorts. Indeed, measures of FA represent a conglomeration of ratio measures reflecting diffusion of water along the axon and perpendicular to the axon, the combination of which is considered to be a proxy measure for axonal health. However, while lower FA is commonly seen in diseases where both cognition and white matter integrity are impaired (e.g., Traumatic Brain Injury, Schizophrenia, Alzheimer's disease) [45], [46], [47], evidence is accumulating that higher FA in particular brain regions may also be associated with clinical disorders including post-traumatic stress disorder [48], obsessive-compulsive disorder [49], panic disorder [50], synaesthesia [51], and Williams syndrome [52]. We have attempted to provide additional specificity to our findings by including measures of RD and AD, but the general point remains that optimal white matter integrity, as quantified by measures of FA, RD, and AD may underpin both cognitive functioning and clinical disease status. Recent cross-sectional evidence showing a curvilinear relationship between age and measures of FA, AD, and RD would tend to support the notion of optimal regional white matter integrity [53], [54].