Neuroimaging predictors of creativity in healthy adults

Journal Pre-proof Neuroimaging predictors of creativity in healthy adults Adam Sunavsky, Jordan Poppenk PII:

S1053-8119(19)30883-3

DOI:

https://doi.org/10.1016/j.neuroimage.2019.116292

Reference:

YNIMG 116292

To appear in:

NeuroImage

Received Date: 2 June 2019 Revised Date:

29 September 2019

Accepted Date: 16 October 2019

Please cite this article as: Sunavsky, A., Poppenk, J., Neuroimaging predictors of creativity in healthy adults, NeuroImage (2019), doi: https://doi.org/10.1016/j.neuroimage.2019.116292. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc.

Neuroimaging predictors of creativity in healthy adults Sunavsky, Adama,†, & Poppenk, Jordana,b,c,†,*

a. Department of Psychology, Queen’s University, Kingston, ON, Canada, 62 Arch St., K7L 3N6 b. Centre for Neuroscience Studies, Queen’s University, Kingston, ON, Canada, 18 Stuart St., K7L 3N6 c. School of Computing, Queen’s University, Kingston, ON, Canada, 557 Goodwin Hall, K7L 2N8 † Co-first authors

* Corresponding author: Dr. Jordan Poppenk, 62 Arch St., Kingston, ON K7L 3N6. 613533-6009. [email protected].

Abstract Neuroimaging has revealed numerous neural predictors of individual differences in creativity; however, with most of these identified in only one study, sometimes involving very small samples, their reliability is uncertain. To contribute to a convergent cognitive neuroscience of creativity, we conducted a pre-registered conceptual replication and extension study in which we assessed previously reported predictors of creativity using a multimodal approach, incorporating volumetric, white matter, and functional connectivity neuroimaging data. We assessed sets of pre-registered predictors against prevailing measures of creativity, including visual and verbal tests of divergent thinking, everyday creative behaviour, and creative achievement. We then conducted wholebrain exploratory analyses. Greater creativity was broadly predicted by features of the inferior frontal gyrus (IFG) and inferior parietal lobe (IPL), including both local grey matter and white matter predictors in the IFG, the superior longitudinal fasciculus that connects them, and IFG-IPL functional connectivity. As IFG and IPL are important nodes within executive control and default mode networks (DMN), respectively, this result supports the view that executive modulation of DMN activity optimizes creative ideation. Furthermore, white matter integrity of the basal ganglia was also a generalizable creativity predictor, and exploratory analyses revealed the anterior lobe of the cerebellum and the parahippocampal gyrus to both be reliable predictors of creativity across neuroimaging modalities. This pattern aligns with proposals ascribing roles of working and long-term memory to problem-solving and imagination. Overall, our findings help to consolidate some, but not all neural correlates of individual differences that have been discussed in the cognitive neuroimaging of creativity, yielding a subset that appear particularly promising for focused future investigation.

Keywords Creativity; Individual differences; Neuroimaging; Executive function; Memory

1. Introduction Creativity is celebrated for its whimsical application (Richards, 2007), therapeutic utility (Schouten et al., 2015), and practical importance in domains such as academic performance (Furnham & Bachtiar, 2008), entrepreneurship aptitude (Lee et al., 2004), and artistic prowess (Carson et al., 2005). However, many current empirical questions about creativity lie in the domain of cognitive neuroscience: for example, what neurocognitive systems facilitate creativity, and can it be predicted based on observable characteristics of participants’ brains? Answering this question can be challenging because creativity itself is a multifaceted construct with no standard operational measure. Because it is spontaneous, it is difficult to capture in event-related designs and instead is frequently measured as a between-subjects variable (Christoff et al., 2016). Additionally, the construct of creativity is believed to be multi-dimensional (Cropley, 2000), incorporating dimensions including divergent thinking, everyday creative behaviour, and creative achievement. Below, we briefly review what neuroimaging research on this topic has revealed. 1.1. Predictors of creativity The default mode network (DMN), known for being more active during rest than during tasks, is frequently associated with creativity (e.g., Beaty et al., 2014; Kühn et al., 2014). The DMN has been associated with spontaneous mental processes and mindwandering (Mason et al., 2007; Christoff et al., 2009), but has also been linked to goaldirected tasks, such as future thinking and episodic memory retrieval (Spreng et al., 2009). It includes the medial prefrontal cortex (mPFC), posterior cingulate cortex (PCC), inferior parietal lobes (IPL), and medial temporal lobes (MTL; Spreng et al., 2010; Andrews-Hanna et al., 2010). Connectivity among these regions has been found to predict divergent thinking, with resting-state functional connectivity (rs-fc) between the inferior frontal gyrus (IFG), IPL, mPFC, and PCC predicting creativity in one study (Beaty et al., 2014), and between the middle temporal gyrus (MTG) and mPFC in another (Wei et al., 2014). Takeuchi et al. (2012) also found positive associations between creativity and mPFC-PCC rs-fc in a larger sample (n = 159). Research on anatomical correlates of creativity has also implicated the DMN. For example, greater fractional anisotropy (FA) values (corresponding to greater white matter integrity) predicted greater divergent thinking in the bilateral subgyral frontal lobe, right medial frontal gyrus (MFG), temporal-parietal junction (TPJ) and right IPL (Takeuchi et al., 2010b). Similarly, grey matter volume (GMV) correlates have included the MFG (Takeuchi et al., 2010a) and TPJ, as well as ventromedial prefrontal cortex (vmPFC; Kühn et al., 2014). Creative achievement, in turn, has been linked to lower cortical thickness in left vmPFC and higher thickness in right angular gyrus (Jung et al., 2010); as well as to greater GMV and left superior frontal gyrus within the right vmPFC (Chen et al., 2014). Chen et al. (2015) found that GMV and thickness correlates of verbal creativity included the right precuneus, which some argue is part of the DMN (Buckner et al., 2008). Finally, a study using both regional grey and white matter volume as predictors revealed positive correlations between creativity and both grey and white matter volumes in the bilateral IFG (Zhu et al., 2013).

There is also some evidence surrounding the possible involvement of an executive control network (ECN) in creativity. One view holds that the DMN and ECN act in opposition to each other, with one at rest while the other is active (Fox et al., 2005), and with creative output principally occurring during unconstrained operation of the DMN. An increasingly well-supported alternative view is that cognitively demanding thought processes also underlie creative tasks, with the ability to creatively problemsolve arising from coupling of the DMN and ECN (for review, see Beaty et al., 2016). Beaty et al. (2015) present relevant evidence by showing regions involved in attention and executive function to be active during divergent thinking tasks. Li et al. (2016) provide further support, showing a positive association in rs-fc between the left dorsolateral prefrontal cortex (DLPFC) and right DLPFC. Also, a meta-analysis by Gonen-Yaacovi et al. (2013) revealed that activation of parts of the ECN, such as the lateral PFC, tended to occur during divergent thinking tasks. Dopaminergic systems, which mediate reward-seeking behaviours, also previously have been implicated in creativity. One proposal is that, from an evolutionary perspective, participants with the greatest motivation will generate a greater number of novel and productive ideas (Flaherty, 2005). One study found creative participants to have higher baseline levels of arousal and greater responses to sensory stimulation, consistent with attributes of the dopaminergic system (Martindale, 1999). Some frontal and temporal dopaminergic network regions overlap with DMN regions, and the striatum and PFC both utilize dopamine (Alexander et al., 1986). Dopaminergic modulation of these regions contributes to constructs that promote creativity (Cools et al., 2007), such as flexibility and persistence (Nijstad et al., 2010). Takeuchi and colleagues (2010a, 2010b) observed GMV correlations with divergent thinking in dopaminergic system structures that included the striata, substantia nigra (SN), ventral tegmental area (VTA), and periaqueductal grey, while also observing an FA-based predictor within the basal ganglia. As well, parts of these subcortical dopaminergic structures (the ventral striatum, SN, and the VTA) are key parts of the larger salience network (Menon, 2015) that is also thought to facilitate creative cognition (Beaty et al., 2015). Consolidating these previous findings, a multitude of regions within the default, executive control, dopaminergic, and salience networks seem to facilitate creative cognition. Supporting this, Beaty et al. (2018) found a pattern of dense functional connections consisting of default (e.g., posterior cingulate cortex), executive (e.g., DLPFC), and salience (e.g., anterior insula) neural networks in highly creative individuals in a large (n = 163) sample. Finally, a meta-analysis by Wu and colleagues (2015) analyzed functional and structural creativity studies to report on which brain regions were associated with creative processes. Across the studies, they found regions within the DMN (IPL, MTG, fusiform gyrus, precuneus), ECN (DLPFC, lateral PFC cortex), and the salience/dopamine network (anterior cingulate cortex, caudate nucleus) to be involved in divergent thinking tasks (for further review on the role that these interacting networks may play in creative cognition, see Beaty et al., 2016). Finally, it is worth considering that what cognitive neuroscience predictors of creativity do exist could simply be the same ones associated with cognitive ability more broadly. Guilford (1967) argued against this possibility, suggesting that convergent thinking (processes that lead to narrow thought and correct answers) underpin intelligence, whereas creativity is more assosciated with divergent thinking. However,

researchers have consistently reported a positive association between these constructs (Batey & Furnham, 2006; Kim et al., 2010), and a meta-analysis drawing on 21 studies with a total of 45,880 participants revealed a positive, albeit small (r = .174) association between creativity and intelligence (Kim, 2005). As well, Silvia (2008) found that lowerorder factors of intelligence (fluid intelligence, verbal fluency, and strategy generation) are modest predictors of creativity, but that general intelligence is a better predictor. 1.2. Current approach Nearly all relevant neuroimaging studies have drawn upon only one of the dimensions of creativity believed to be important to the construct, the most common target being divergent thinking (and specifically verbal divergent thinking; see Supplemental Table 1 and Piffer, 2012). This raises uncertainty about how well what is known generalizes to the construct of creativity as a whole. In addition, although much has been learned from the above research on the DMN, ECN, dopaminergic, and salience networks, progress has been slowed by many contradictory results (Dietrich & Kanso, 2010). Accordingly, consolidation of existing results is an important directive for establishing a better foundation for future research on creativity. Here, we pre-registered predictions arising from the above findings, aiming through our predictions to confirm involvement of the DMN, ECN, dopaminergic and salience networks. Because we hoped to integrate findings from a variety of reports that varied in implementation, and because we sought effects that were robust to minor deviations in protocol, we employed measures and techniques that we felt to be representative of the constructs under evaluation rather than exact re-implementations of past experiments. In other words, our effort concerned conceptual replications aimed at obtaining previously identified neural predictors of creativity, rather than attempting to single out the reproducibility of the pattern of results found in any specific study. We employed a sample size near the median of recent studies in this area (Supplemental Table 1). This sample size fell short of the scale of several hundred participants or more that would be useful for suggesting that past results are attributable to type II error; accordingly, our work should not be taken as an authoritative guide to which effects cannot be reproduced. Rather, our experiment is best viewed as an initial exploration of which are robust enough to replicate using a sample attainable by smaller labs, and that are therefore particularly practical for carrying forward into a future cognitive neuroscience of individual differences in creativity. By drawing upon multiple neuroimaging modalities to predict a spectrum of creativity measures, we also aimed to generate a dataset unique in addressing the multimodality and generalizability of neuroimaging predictors to a spectrum of instruments targeting a common underlying construct of creativity.

2.0 Methods We deployed the Abbreviated Torrance Test for Adults (ATTA, a divergent thinking task as well as shortened version of the Torrance Test of Creative Thinking;

Goff, 2002), the Creative Behaviour Inventory (CBI, a measure of everyday creativity; Dollinger, 2003) and the Creative Achievement Questionnaire (CAQ, a measure of creative achievement; Carson et. al, 2005). To obtain a better creativity index that was less influenced by state effects, we sampled the ATTA on a separate computer testing session from the CBI and CAQ. Administration of the WAIS-IV and neuroimaging scans took place on subsequent visits, with neuroimaging incorporating volumetric, structural and rs-fc scans. These were part of a series of visits involving experimentation on topics to be discussed elsewhere. Because the neuroimaging data have also been used to investigate other aspects of memory, aspects of our neuroimaging methods here are shared with prior reports (e.g., Matorina & Poppenk, 2019). Due to the high cost of neuroimaging as well as the demanding nature of our experiment, prior to testing, participants also completed a screening session before being invited to continue. Those who did not meet our demographic criteria or demonstrated possible issues with keeping still in a simulated fMRI scanner were not invited for subsequent sessions. 2.1. Pre-registration. To target specific neural predictors for testing, we referenced recent metaanalyses and review articles to investigate possible emergent themes in the prediction of creativity using neurological, cognitive, and behavioural markers (i.e., individual differences). From this search, we extracted hypotheses and studies relevant to the formulation of those hypotheses, with particular weight given to large studies and influential findings. Our resulting pre-registered hypotheses and procedures may be found online on the Open Science Framework website (https://osf.io/uzhvm). We complemented our selective replication of a small number of findings with whole-brain analyses aimed at uncovering effects that were prominent in our own data. Several deviations from our pre-registration were made to accommodate helpful reviewer suggestions, including across-the-board control for intelligence, separation of verbal and visual sub-scores of divergent thinking, transformation of CAQ scores, and inclusion of several new ROIs suitable for attempted replication. In addition, we conducted exploratory evaluation of contrasts using all creativity tests other than just the one used to justify inclusion of each predictor. For example, where a brain region was found to predict divergent thinking, we assessed its relationship with ATTA, but also CBI and CAQ scores. The post-hoc nature of these tests is indicated throughout our tables and results. Finally, we excluded a planned multiple-regression analysis, as the large number of plausible predictors for entry into this analysis required a selection step, and we did not wish to “double dip” into our data for both selection and evaluation of predictors. For clarity of focus in our report, pre-registered tests involving sleep EEG predictors and hippocampal subfields have been described separately in Supplemental Results. 2.2. Participants 104 participants were recruited in Kingston, Canada using posters, Facebook advertisements and web posts on Reddit, Kijiji and Craigslist. Informed consent was provided by all the participants for participating in our experiment. Participants were

required to be: between 22-35 years of age (to avoid potential developmental effects); right-handed, an English native-speaker, have normal or corrected-to-normal vision and hearing, regularly sleep at least 5 hours a night, have no contraindications for MRI scanning, have no history of neurological disorders, sleep disorders, or recurrent mental illness that included medication, and not currently be taking psychotropic medications. They were invited to undergo a two-hour in-person screening session in which demographic eligibility was confirmed, and a simulated MRI scanner was used to rule out claustrophobia, inability to keep still (operationalized as falling below the 95th percentile of low-frequency motion as measured in a reference sample), and inability to stay awake for a 20-minute eyes-open resting session. In addition, we assessed participants' ability to respond to at least 33% of encoding trials in a recognition memory task, to obtain a d-prime of at least 0.1 in a subsequent memory test, and to demonstrate acceptable reading comprehension and speed (a TOWRE score of at least 26.4 and a Nelson-Denny score of 2). Finally, participants were excluded for multiple no-shows, failing to follow instructions, rudeness, and excessive drug use. Participants were advised that they would be compensated CAD$10 per hour for their time (or a prorated amount in the case of early withdrawal). Of the original sample, 69 participants met our criteria and returned to take part in our experiment; of these, 66 completed the neuroimaging session. All participants who completed the scan also completed the CBI and CAQ, but for logistical reasons, one neuroimaging participant did not complete the ATTA. Of the 66 participants contributing relevant data, 29 described themselves as men, 36 as women, and one as “other”. Their average age was 26.6 years (SD = 4.3 years). Prior to our behavioural procedures, participants completed a “check-in” questionnaire in which we gathered information related to their current state. No participants reported feeling drunk, high, or indicated they would have trouble remaining awake. Accordingly, none were excluded from analysis. 2.3. Behavioral measures Over two separate sessions, we administered the CBI, CAQ, and the ATTA. We scored each participant’s responses to each instrument using standard procedures. First, the CBI consists of a list of creative behaviors (Dollinger, 2003). We assigned values of 0 to 3 to participants’ self-estimates of how often they performed each one (0 = ”Never did this”; 1 = ”Did this once or twice”; 2 = ”3-5 times”; 3 = ”More than 5 times”). To obtain a single score for each participant, we took the mean resulting value across all items. The CAQ consists of questions regarding participants’ accomplishments across a variety of creative domains (e.g., writing, visual art, music, cooking, etc.; Carson et. al, 2005). Each question involves a true/false assessment of whether the participant has attained a certain creative achievement. If participants attest to at least a low level of achievement in a given domain, further questions are administered for that domain. We administered standard scoring of this instrument, assigning greater points to more substantial accomplishments, and then computing the sum of these values across all domains to obtain a single achievement score for each participant. As CAQ scores are skewed in the general population (Carson et. al, 2005) as well as in our sample (Table 1), we transformed participants’ CAQ scores prior to subsequent analysis. We found

that a square-root transformation best restored normality in the data, yielding a final skewness of 0.23. We tested divergent thinking via the ATTA, a validated abbreviated measure of the TTCT (Althuizen et al., 2010), which yields both verbal and visual estimates of creativity. Two raters implemented standard scoring procedures by evaluating each response in terms of various creative elements (see Goff, 2002, for details), using a software program (Mountjoy & Poppenk, 2015) to blind them as to which participant made which response. Briefly, visual and verbal creativity scores were obtained by first summing within-criterion scores within norm-referenced indicators, doing so separately within each visual and verbal domain. The raw score for the criterion indicators was then added to the total scaled score for the norm-referenced measures, giving a total creativity score within each of the verbal and visual domains. Agreement was high across the two raters, with an Intraclass Correlation Coefficient (ICC) of 0.903 for verbal creativity scores, and 0.943 for visual creativity scores. Finally, in a separate session in a private room, participants completed core subtests of the WAIS-IV (Canadian Edition; Sattler & Ryan, 2009), a widely-used measure of full-scale intelligence. Testing was administered by students trained in its administration. After applying standard scoring procedures, we applied Canadian norms to obtain a full-scale IQ and index scores for each participant. 2.4. Neuroimaging procedure Later, the day prior to each participants’ MRI scan, participants completed a biofeedback session in the simulated MRI scanner, both to become better habituated to an MRI-like environment and to learn to reduce their head movement (and thereby improve the signal quality of their brain images sampled the next day). During the biofeedback session, participants viewed a 45-minute documentary with a live readout of their head motion overlaid. When their head motion exceeded an adaptive threshold, the documentary was paused for several seconds while static was played on the screen along with loud and unpleasant white noise. Following the documentary, a brief memory test was administered outside of the mock scanner to ensure participants were paying attention to the film content rather than just their motion (not analyzed here). The next day, we used a whole-body MRI scanner (Magnetom Tim Trio; Siemens Healthcare) to gather a variety of image sequences over the course of a 1.5-hour scan. For purposes of measuring whole-brain anatomy and registration of functionals to standard fMRI space, we gathered high-resolution whole-brain T1-weighted (T1w) and T2-weighted (T2w) anatomical images (in-plane resolution 0.7 x 0.7 mm2; 320 x 320 matrix; slice thickness 0.7 mm; 256 AC-PC transverse slices; anterior-to-posterior encoding; 2 x acceleration factor; T1w TR 2400 ms; TE 2.13 ms; flip angle 8°; echo spacing 6.5 ms; T2w TR 3200 ms; TE 567 ms; variable flip angle; echo spacing 3.74 ms). To measure the hippocampus in greater detail, we gathered an ultra-high resolution T2-weighted volume centered on the medial temporal lobes (resolution 0.5 x 0.5 mm2; 384 x 384 matrix; slice thickness 0.5 mm; 104 transverse slices acquired parallel to the hippocampus long axis; anterior-to-posterior encoding; 2 x acceleration factor; TR 3200 ms; TE 351 ms; variable flip angle; echo spacing 5.12 ms). For purposes of measuring resting-state functional connectivity, we gathered two T2*weighted scans (158 volumes; in-plane resolution 1.5 x 1.5 mm2; 128 x 128 matrix; slice

thickness 1.5 mm; 90 AC-PC transverse slices; anterior-to-posterior encoding; 6 x multiband acceleration factor; TR 1900 ms; TE 36.8 ms; flip angle 75°; echo spacing 0.88 ms). For investigation of structural integrity of white matter, we gathered diffusion tensor imaging (DTI) scans (64 diffusion directions; b-value 1200; in-plane resolution 1.5 x 1.5 mm2, 128 x 128 matrix; slice thickness 1.5 mm; 93 AC-PC transverse slices; 3 x multiband acceleration factor; TR 5180 ms; TE 103.4 ms; flip angle 78°; echo spacing 0.77 ms). Complementary DTI acquisitions were gathered using right-to-left and left-toright encoding directions. Three b0 scans were gathered for each encoding direction. Segmentation of brain regions was conducted using Freesurfer (v6, Fischl et al., 2004) applied our 0.7 mm whole-brain T1w and T2w brain scans to create segmentation labels for most of our regions of interest (ROIs), as well as to obtain Estimated Intracranial Volumes. We obtained high quality segmentations of aHPC and pHPC by drawing upon the ultra-high-resolution T2w 0.5mm isotropic medial temporal lobe scans. We submitted these to automated segmentation using HIPS, an algorithm previously validated to human raters specialized in segmenting detailed neuroanatomical scans of the hippocampus (Romero et al., 2017). Three independent raters were trained on segmenting the hippocampus at the uncal apex into aHPC and pHPC segments; achieving a Dice coefficient of absolute agreement of 80%. Two of these raters independently segmented all participants in this study using the 0.5mm T2w scans. The 0.5 mm T2w medial temporal lobe scans were registered to the 0.7 mm T2w whole-brain scans, which were in turn registered to the T1w whole-brain scans, and the combined transform was used to place the rater landmarks on the detailed medial temporal lobe scans. All voxels belonging to the hippocampus posterior to this coronal plane containing this point were classified as pHPC, whereas all points anterior to and including this plane were classified as aHPC. These classifications were projected back on the T1w images to obtain aHPC and pHPC segmentations. For each ROI, the total number of voxels was multiplied by the volume of each voxel to obtain a total volume estimate. Intracranial volume was then residualized from the resulting group vector to control for head size. These steps were followed separately in each hemisphere. To support voxel-based morphometry (VBM) analysis, we also transformed each participants’ grey matter mask, which we constructed by assembling all grey-matter regions obtained from their Freesurfer segmentation, to native space using nearest-neighbour interpolation, and smoothed the mask using a 6 mm full-width half maximum (FWHM) gaussian filter. These procedures amounted to “unmodulated” VBM (in which no correction is applied for global rescaling to native space), which is understood to reflect concentration, rather than absolute volume of structures (Good et al., 2001). We took this approach because, as in our ROI-based analyses, overall head size was a variable of no interest. Next, to prepare the T2*-weighted fMRI images for analysis, we first used reference scans to identify encoding direction-based spatial distortions and used this distortion map to unwarp the images. This step was performed simultaneously with a motion correction step and warp to MNI space using scripts adapted from the Human Connectome Project (Glasser et al., 2013) and tools the FSL software suite (v.5, (Jenkinson et al., 2012). We then dropped the first five “stabilization” volumes gathered in each run, conducted linear trend removal, and ran ICA-FIX (Salimi-Khorshidi et al., 2014) from the FSL package to denoise our data (the model was trained using

consensus component ratings from two raters). We used the full group’s Freesurfer segmentations to ascertain which voxels corresponded to grey matter in at least 50% of participants. We zeroed out voxels falling outside of this mask and removed those coordinates from further analysis. After cleaning the data, we took several steps to generate functional connectivity estimates for each participant and seed region/target region pairing. First, we needed to establish a timecourse for each ROI. We transformed our segmentation masks into MNI space, then z-scored the timecourse of the averaged signal each mask, doing so separately for each run and participant. We then took the median value within the normalized timeseries within run and participant. This yielded a vector corresponding to the average activity of that ROI across that run. Then, we correlated each seed ROI vector against each target ROI vector and retained the correlation coefficient. We averaged the coefficients obtained across the two resting state runs, such that each participant had one coefficient for each pairing of ROIs. To prepare the DTI images, we again applied a de-warping step, similar to the one described for fMRI images. We then applied the FDT preprocessing stream of the FSL toolbox, consisting of correction for susceptibility-induced distortions, correction for eddy currents and participant movements, and tensor fitting to obtain a native-space FA map for each individual (Jenkinson et al., 2012). Using the segmentation masks obtained in steps above, we sampled the mean intensity of all voxels falling within each ROI and took their mean to obtain an FA value for that ROI and participant. 2.5. General statistical approach As reported in our pre-registration, based on our sample size of 66, we anticipated 80% power for detecting medium-sized effects of r = 0.34, and 80% power for replicating slightly smaller (r = .30) correlations with one-way tests; we targeted this sample size as the largest neuroimaging sample possible based on available funds. To diminish the influence of outliers, for each variable, we removed values falling more than three median absolute deviations (MAD) from the mean. The MAD statistic is similar to standard deviation, but is more resistant to influence by outliers; this quality was ideal for our use of the statistic to screen possible outliers (Leys et al., 2013). As part of an effort to exclude as few cases as possible, we performed these outlier-based exclusions on each variable, rather than in a listwise fashion. For this reason, as well as the fact that not all participants completed all tests, our degrees of freedom varied slightly from one test to another and are reported alongside all analyses. To improve our sensitivity to small but reliable effects (and decrease sensitivity to larger, unreliable effects), we employed non-parametric resampling procedures (Efron & Tibshirani, 1986) for our statistical testing. Briefly, our approach relied on a bootstrap resampling procedure to obtain the standard error around measured statistical parameters. After obtaining a correlation coefficient, we sampled with replacement from the original pool of participants 1000 times, recomputing the parameter each time and assessing the standard deviation of this distribution. We then took the ratio of the parameter to this distribution to obtain a bootstrap ratio (BSR). Because the distribution of BSR values corresponds approximately to a z-distribution, we used it to derive pvalues.

To conduct whole-brain analyses, we used this same approach, regressing each variable and voxel’s values across participants and obtaining a BSR. To correct for the large number of comparisons necessary, we employed the FSL cluster tool to obtain Gaussian Random Field (GRF) statistics, applying a voxel-wise threshold of p < .005 to obtain a corrected cluster-wise threshold of .05. We addressed recent concerns of inflated false-positive rates arising from problems with estimation of smoothness in such fMRI clustering procedures by incorporating ICA-FIX, two-way tests and groupresampling into our procedures (see Eklund et al., 2019, for discussion).

3.0 Results 3.1. Cognition We began our analysis by investigating the relationships among our three measures of creativity: divergent thinking (ATTA), creative behaviour (CBI), and creative achievement (CAQ; see Table 1 for descriptive statistics describing our sample), anticipating significant inter-relations. In a cognitive study, Jauk and colleagues (2014) showed that creative achievement and creative behaviour were related, r(296) = .52, p < .001; our results reflected this pattern with a positive association between CBI and CAQ scores, r(64) = .31, p = .003. Jauk and colleagues differed in employing the activities scale of the Inventory of Creative Activities and Achievements (Diedrich et al. 2018) to measure creative behaviour, but this measure is analogous to the CBI. Jauk and colleagues also found creative behaviour to be linked to fluency, r(296) = .32, p < .001, and originality, r(296) = .28, p < .001, both of which are measures of divergent thinking. We, too, found divergent thinking and CBI to be correlated, r(63) = .26, p = .009. Completing this triangle of relations, Zhu and colleagues (2016) observed a correlation between the visual TTCT and the CBI, r(163) = .21, p < .05, a pattern that we too observed between the visual ATTA and CBI, r(60) = .35, p < .001. We next attempted to replicate findings linking these creativity scores to measures of cognitive ability. A meta-analysis by Kim (2005) revealed a modest correlation across studies between divergent thinking and intelligence, r(45, 880) = .174, 95% CI = [.165 .183]: similarly, we observed a relationship between ATTA verbal creativity and WAIS FSIQ, r(60) = .29, p = .005, but not ATTA visual creativity and FSIQ, r(60) = .16, p = .185. This finding is consistent with several observations that the verbal, but not visual version of other divergent thinking tasks is related to intelligence (Kershner & Ledger, 1985; Li et al., 2016; Moore et al., 2009). Other measures of creativity have also previously been linked to intelligence: Jauk et al. (2014) found that creative activities were reliable predictors, r(296) = .13, p < .05, as was creative achievement, r(296) = .17, p = < .01. We found this pattern as well, observing that FSIQ predicted CAQ scores, r(58) = .31, p = .006, but not CBI scores, r(59) = .13, p = .178. Because intelligence predicted several of our creativity measures, and because we were interested in neural contributions to creativity that took place above and beyond intelligence, we residualized intelligence from our behavioural creativity measures in all of our subsequent analyses.

3.2. Volumetry To evaluate volumetric predictors, we pre-registered an ROI-based approach, in which volumes of regions containing effects described in prior reports were evaluated as a whole and regressed against individual differences in creativity. While appealing in testing past effects using a limited number of comparisons, we found this approach to be broadly insensitive. We speculated that whole-ROI volume could have been too broad of a measure to isolate sub-ROI effects reported previously: definitions of, and delineations within ROIs are highly variable and can dramatically alter structural and functional connectivity profiles (Li et al., 2010). To exhaust every reasonable opportunity for replication, we implemented an alternate hybrid approach, counting the number of voxels exceeding a modest statistical threshold within each ROI and, using a chisquared test, then comparing this tally against the number that would be expected by chance. We performed these tests in a one-way fashion, evaluating only the direction that would align with past results. To support possible generalization, we then repeated this step in a post-hoc fashion for each creativity measure (see Table 2). Using this approach, we found two of ten volumetric predictors of verbal divergent thinking – left and right IFG – to be reliable predictors, finding both to generalize to all creativity measures other than the visual ATTA. We also replicated two out of three CAQ predictors – left and right vmPFC – with one or the other generalizing across all other creativity instruments. For ROIs where we failed to obtain creativity effects using the original creativity construct, seven of nine were associated with at least one creativity measure, even though the creativity measure did not necessarily map onto the analogous measure used in the original study. For example, although we did not find the predicted link to creativity within left vmPFC and verbal ATTA, we did see such a link between the left vmPFC and both the visual ATTA and CAQ). In addition to the above tests, we pre-registered evaluation of ACC and PCC associations with creativity. We present these results here separately, as we included them in our pre-registration on the basis of their membership in the DMN, rather than appearance in any past studies. We found volumetric features of both ROIs predicted our creativity measures. For ACC, these included visual ATTA (χ2 = 5.71, p < .001), CBI (χ2 = 7.28, p < .001), and CAQ (χ2 = 4.92, p < .001), but not verbal ATTA (χ2 = 12.70, p > .999). For PCC, these also included visual ATTA (χ2 = 5.75, p < .001), CBI (χ2 = 5.21, p < .001), and CAQ (χ2 = 6.42, p < .001), but not verbal ATTA (χ2 = .73, p = .173). 3.3. Connectivity Of thirteen white-matter ROIs (all derived from verbal divergent thinking tasks), we found that five encompassed FA predictors of verbal ATTA. These included right inferior frontal gyrus, the left superior longitudinal fasciculus, the left anterior internal capsule, left basal ganglia, and the body of the corpus callosum (see Table 3). Each of these ROIs featured generalization to at least one other creativity measure. The left superior longitudinal fasciculus (SLF) was particularly noteworthy in this respect, as it encompassed local predictors of every one of our four creativity instruments, and was the only ROI to do so. As with our volumetric ROIs, a majority of the remaining eight white matter ROI’s included predictors of at least one other measure of creativity in the expected direction.

With respect to functional connectivity, out of eight ROI pairs previously found to predict verbal ATTA in rs-fc studies, we found two were significant predictors in the current study (see Table 4). In particular, we found ipsilateral connections between the IFG and IPL. The left-hemispheric instance of this pairing generalized to verbal ATTA. The only other functional connectivity pair that significantly predicted any form of creativity was ACC-left MTG, which predicted both visual ATTA and the CAQ, although functional connectivity between left ACC and PCC showed a trend towards being a significant predictor of both verbal and visual creativity. 3.4. Whole-brain analyses Our pre-registered analysis plan included inspection of whole-brain VBM and FA maps to identify sources of variance in our dataset not captured by our a priori predictions. We deemed this step necessary in part because we were aware of no past reports evaluating the white matter correlates of both the CBI and the CAQ; and in part to pursue identification of multimodal creativity predictors. For a comprehensive list of identified predictive VBM and FA clusters, see Tables 5 and 6, respectively. We found the cerebellum to be a prominent predictor across all creativity tests, as expressed in terms of positive predictions of either greater grey matter density, white matter integrity, or both (see Fig. 1A). Other cerebellar regions were almost always also implicated, but their effects varied in location (most frequently within one of lobules VIVIII) as well as direction. For example, mainly positive other associations were observed between cerebellar volume and verbal creativity, creative behaviour, and creative achievement, but mainly negative other associations arose between cerebellar volume and visual creativity. Similarly present as a predictor of all creativity tests, albeit with lower consistency in localization, was the parahippocampal gyrus (see Fig. 1B), where grey matter density positively predicted ATTA verbal, ATTA visual and CAQ scores and white matter integrity predicted CBI scores. Generalizing across smaller subsets of creativity instruments was the grey matter density of the anterior insula, which was positively associated with verbal ATTA and creative behaviour. Different portions of the thalamus had both positive and negative grey matter correlates within verbal creativity; thalamus grey matter density and FA values also negatively predicted creative achievement. Finally, grey matter density of the cuneus was a negative predictor of verbal ATTA and CBI, but a positive predictor of visual ATTA (proximal to the cuneus, positive FA predictors were observed in a posterior thalamic radiation for visual ATTA, and in the forceps major for the CAQ). Among regions predicted only one creativity instrument, most notable were predictions of visual ATTA by a set of dopaminergic structures, which included grey matter density in the SN / VTA (positive) and left anterior hippocampus (negative), as well as FA of the internal capsule at the caudate/anterior commissure, although a VBM predictor of the caudate was negative for CAQ. Cross-hemispheric white matter pathways were also noteworthy, with FA in a rostral corpus callosum region extending into the anterior corona radiata positively predicting verbal ATTA, and FA in the aforementioned anterior commissure region positively predicting visual ATTA.

4.0. Discussion

We conducted a pre-registered conceptual replication of cognitive and neuroimaging predictors of individual differences in creativity. Our results affirm the reliability of relationships among creativity measures and their cognitive predictors. They also also affirm the reliability of many neural predictors, including elements of executive, DMN and dopaminergic networks. In line with recent discoveries implicating the cerebellum in creative processes, exploratory analysis also revealed multimodal contributions of the cerebellar anterior lobe, as well as less spatially and directionally consistent contributions from lobules VI-VIII, to multiple measures of creativity. Consistent with a role for long-term memory, we also found reliable predictions by features of the parahippocampal gyrus of each of our measures of creativity. Overall, our findings underscore the relevance of executive, memory, motor and reward systems to creative processes; support proposals that ECN-DMN interactions may facilitate creative processes; and confirm that neuroimaging biomarkers can be used to predict individual differences in creativity. 4.1. Relationship among creativity and intelligence measures Behaviourally, our measures followed patterns that would be expected based on earlier publications. In particular, ATTA scores were moderately correlated with both the CBI and CAQ, consistent with a link between divergent thinking, creative behaviour, and creative achievement described by Jauk and colleagues (2014). This is, however, the first time we are aware of that this link has been established using the ATTA. Consistent with the idea that these dimenions of creativity are influenced by cognitive factors (Amabile, 2012), as well as a meta-analysis by Kim (2005), we found that ATTA scores were correlated with FSIQ. Intelligence also correlated with creative achievement scores, but not creative behaviour, also consistent with similar findings by Jauk and colleagues (2014). This pattern reinforces the view that cognitive ability is relevant to the quality of creative output, but not to its spontaneous production. Although our measures of creativity were designed to access different dimensions of the construct, all were contrived with a common underlying construct of “creativity” in mind. The moderate size of cognitive correlations supported the convergent and divergent validity of these measures, apparently reflecting both core creativity variance as well as multi-dimensional variance (Croply, 2000). We therefore next sought to ascertain whether we would find similarly overlapping and distinctive features of these dimensions in terms of their neural substrates. 4.2. Network predictors of creativity Turning to our pre-registered neuroimaging hypotheses, we observed a robust confirmation of predictions concerning the involvement of IFG and its connectivity with the IPL in creativity. We observed multi-modal involvement of ipsilateral fronto-parietal links between them that included a positive correlation between creativity and IFG grey matter density; FA near IFG; IFG – IPL functional connectivity; and FA of the SLF (a white matter tract that has been proposed to link IFG and IPL; Yagmurlu et al., 2016). Post-hoc analysis using tract-based statistics further confirmed the association between creativity and the SLF.

This set of findings replicates and extend earlier discoveries, including a finding by Zhu and colleagues (2013), who first observed IFG predictors involving volumetric and white matter integrity; by Beaty and colleagues (2014), who noted the importance of functional connectivity between IFG and IPL in promoting creativity; and by Takeuchi and colleagues, who observed a white matter predictor of creativity in the SLF (2010a) and right IPL (2017). In so doing, our findings underscore not only the relevance of the IFG, which has a variety of important roles, including inhibition, memory retrieval, and executive processes (Badre & Wagner, 2007; Aron, 2007), but also communication of the IFG with IPL, a core DMN region associated with internally-directed attention (Fink & Benedek, 2014). Interactions between the ECN and DMN may act to limit off-task thoughts during creative idea generation, thereby facilitating novel idea generation (Beaty et al., 2015; 2016); under this model, the DMN contributes to idea generation, but the ECN evaluates these ideas and constrains them to meet task-specific goals. The IFG-IPL communication confirmed in our dataset appears to reflect a possible conduit through which this ECN suppression of the DMN could take place. We also evaluated ROIs relevant to dopaminergic and salience networks, which mediate reward-seeking behaviour and to detect and filter focal stimuli, respectively. Parts of the dopaminergic system overlap with the salience network (the ventral striatum, insula, and the SN/VTA), and both have been implicated in creativity to some extent (e.g., Takeuchi et al., 2010a, Beaty et al., 2015). In our study, dopaminergic structures including the insula, caudate nucleus, and SN/VTA were each predicted at least one measure of creativity, as did salience structures such as the ACC. Notably, these networks have links to the DMN and ECN: there is diffuse dopaminergic projection from the midbrain to the frontal lobe where nerve terminals supply dopamine (Middleton & Strick, 2002), and this neurotransmission improves performance in tasks associated with creativity such as cognitive flexibility tasks (Nagano-Saito et al., 2008). Sridharan and colleagues (2008) propose an integration of the salience, executive control, and default mode network such that the salience network plays a causal role in switching between the ECN and DMN. Thus, while the DMN may be involved in generating novel and useful ideas and the ECN may constrain creative cognition to meet task demands, the salience network may modulate between the two networks. Although we found various elements of these networks to be associated with creativity, this evidence was less consistent across measures and neuroimaging modalities than in the IFG-IPL case, possibly resulting from the small volume of some of the constituent regions. 4.3. Regional predictors of creativity The anterior lobe of the cerebellum was distinctive for predicting multiple measures of creativity across multiple neuroimaging modalities. Other aspects of the cerebellum were also regular predictors – typically falling in lobules VI-VIII – but were less consistent in their spatial position and direction. Anterior cerebellum has been linked with somatomotor cortex (Buckner, 2013), and is itself regarded as a primary motor coordination area of the cerebellum (Koziol et al., 2009; Stoodley & Shoeman, 2016). Along these lines, lobule X connects primarily to the vestibular system and is associated with proprioception. By contrast, lobules VI-VIII feature strong functional

connectivity with associative cortex, and are believed to be more cognitive in nature, with hypothesized roles in executive function and working memory. Our results complement a number of other recent discoveries implicating the cerebellum in creative processes (e.g., Sun et al. 2018, Neumann et al., 2018; Ogawa et al. 2018). Although the cerebellum has not been historically regarded as a core contributor to creativity, it is clear how sensorimotor systems could play some role, considering that the CAQ contains overtly sensorimotor subscales, including dance, music, and culinary arts. Motor coordination for drawing or writing is also important for interpretable and fluent output (relevant e.g., for divergent thinking and creative achievement), factors that could make creative activity less costly and more rewarding (relevant to creative behaviour). Cognitive cerebellar contributions have also been proposed: for example, Vandervert, Schimpf, and Liu (2007) proposed that unconscious cerebellar inputs contribute to generating novel solutions; also, like the repetitive movements that the cerebellum is thought to control, the cerebellum-driven repetitive manipulation and rehearsal of working memory could make formation of novel ideas more efficient (Welling, 2007). We also found evidence for contributions to creativity of ROIs known for their mnemonic role. In particular, every one of our creativity metrics had a grey or whitematter predictor in the parahippocampal gyrus. Madore et al. (2015, 2017) found that following episodic memory induction, divergent thinking was enhanced and associated with greater neural activity in the hippocampus and bilateral parahippocampal gyri, structures known to be important for long-term memory (e.g., Bauer et al., 2017; Madore et al., 2017). Structurally, Jung et al. (2015) linked greater thickness in the parahippocampal gyrus to higher creativity scores. This pattern is consistent with the idea that long-term memory can be a contributor to creative ideation (Schacter, Addis & Szpunar, 2017; Shen et al. 2017). In contrast to robust parahippocampal prediction of creativity, the only creativity link we observed with the hippocampus was a negative association between local gray matter density in the anterior hippocampus and visual ATTA. We note that more is not always better when it comes to the hippocampus, however; similar to the current study, we (Poppenk et al., 2011) found that a smaller anterior hippocampus predicted better source and recollection memory. Thus, it is possible that, as in that study, a smaller anterior hippocampus was associated with superior episodic memory detail. However, in the 2011 study, we also observed a positive correlation between our memory measures and the posterior hippocampus; in the current case, no such positive correlate was observed. Moreover, the negative effect that was found was limited to visual ATTA. The implication of this finding, if any, is unclear. 4.4. Limitations and future directions Of studies published in the past ten years on prediction of individual differences in creativity from neural variables, our sample size of n = 66 fell in the 48th percentile (see Table S1) and was sufficient for detection of moderate-sized effects, but was nonetheless lower than would be required for a definitive confirmatory analysis. Accordingly, we do not make the claim that those reports we failed to substantiate were originally results of type I error, as it is just as possible that our null findings arose due to type II error. Also, because our aim was to identify predictors that generalized across

the idiosyncrasies of different neuroimaging and behavioural tests, and because past studies vary widely in these and other respects, we did not attempt to precisely align our procedures to any one past studies in terms of participant populations, neuroimaging methods, and data analysis choices. As such, one should not infer that regions and connections we did not identify are irrelevant to the neuroscience of creativity, only that the factors we did replicate are among the most robust and generalizable, and perhaps therefore among the most useful for predictive purposes. It should also be noted that the ATTA, our measure for divergent thinking, is an abbreviated measure of divergent thinking. Although this instrument has been validated (Althuizen et al., 2010), it is, due to its brevity, a less precise measure of divergent thinking than its more comprehensive counterpart, the TTCT. However, there are also advantages to using abbreviated measures – for example, the use of abbreviated measures increase the accessibility of this research in samples used by large consortia gathering thousands of brain scans (but with little time available for each participant). We note that many relationships, including notoriously weak ones such as that between divergent thinking and intelligence, were borne out with this abbreviated measure, suggesting the instrument has sensitivity to the construct of divergent thinking that was adequate for sensitivity to the same neural substrates. Also, although we hoped to learn which predictors would generalize across measures of creativity, our rationale for using multiple predictors is that creativity is believed to be a multidimensional construct, and that these measures should have divergent, as well as convergent validity (Aron, 2000). Just as various authors are attempting to identify neural predictors that distinguish facets of divergent thinking (e.g., Takeuchi et al., 2017), a complementary future goal will be to understand what features best distinguish divergent thinking, creative behaviour, and creative achievement predictors in the brain and how these patterns could be mapped back to traits that further refine a cognitive understanding of these constructs. Finally, as the reliability of particular neural predictors becomes clearer, one question becomes: is the predictive power over creativity cumulative, or overlapping? We considered, but did not conduct, a multiple-regression analysis, as we wished to avoid the circumstance where we evaluated the predictive power of the same data we used to ascertain which predictors were the most reliable. However, with numerous reliable predictors as conceptually and spatially distinct as fronto-parietal network, the cerebellum, parahippocampal gyrus, parietal cortex, and still others, it is possible that neural biomarkers of creativity could cumulatively explain a large amount of variance of individual differences in creativity. 5.0 Conclusion We analyzed a set of pre-registered hypotheses aimed at replicating and generalizing previously observed cognitive and neural correlates of creativity. Cognitive predictors were largely confirmed, and many neural predictors were also effective, most notably the IFG and its connections to the IPL, but also including the basal ganglia, and several discovered predictors including the cerebellum and parahippocampal cortex. To varying degrees, we further replicated volumetic, white-matter, and rs-fc effects. These results support the perspective that creative ideation is supported by cerebellar operations of working memory, parahippocampal contributions to long-term memory

recall, and IFG-IPL connectivity as a means of the ECN dynamically exerting control over the DMN to fine-tune its contributions. Overall, our findings help to consolidate previously hypothesized substrates of individual differences in creativity, affirming cognitive and neural predictors that will be suitable targets for analysis in future investigations.

Acknowledgements We gratefully acknowledge Nelly Matorina, Julie Tseng, Natalie Doan, Lauren DeMone, Nigel Barnim, Sarah Berger, Megan Fleming, Maddie Gillis, Sophie Kinley, Lindsay Lo, Lydia Luchich, and Gillian Marvel for assistance with behavioural data collection. We also thank Julie Tseng, Lauren DeMone, and Natalie Doan with scheduling; Julie Tseng and Don Brien with MRI data acquisition; and Justin Siu, Roland Dupras and Mike Lewis with technical support. This research was funded by Natural Sciences & Engineering Research Council Discovery Grant 03637 (J.P.). Infrastructure funding was provided by Canada Foundation for Innovation – John R. Evans Leaders Fund (J.P.), and a Queen’s University Research Initiation Grant to J.P., who was supported by the Canada Research Chairs program. Declarations of interest: none

References Alexander, G. E., DeLong, M. R., & Strick, P. L. (1986). Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Annual review of neuroscience, 9(1), 357-381. https://doi.org/10.1146/annurev.ne.09.030186.002041 Amabile, T. M. (2012). Componential theory of creativity. Harvard Business School, 12(96), 1-10. Althuizen, N., Wierenga, B., & Rossiter, J. (2010). The validity of two brief measures of creative ability. Creativity Research Journal, 22(1), 53-61. https://doi.org/10.1080/10400410903579577 Andrews-Hanna, J. R., Reidler, J. S., Sepulcre, J., Poulin, R., & Buckner, R. L. (2010). Functional-anatomic fractionation of the brain's default network. Neuron, 65(4), 550562. https://doi.org/10.1016/j.neuron.2010.02.005 Aron, A. R. (2007). The neural basis of inhibition in cognitive control. The Neuroscientist, 13(3), 214–228. Badre, D., & Wagner, A. D. (2007). Left ventrolateral prefrontal cortex and the cognitive control of memory. Neuropsychologia, 45(13), 2883–2901. Batey, M., & Furnham, A. (2006). Creativity, intelligence, and personality: A critical review of the scattered literature. Genetic, social, and general psychology monographs, 132(4), 355-429. https://doi.org/10.3200/MONO.132.4.355-430 Beaty, R. E., Benedek, M., Kaufman, S. B., & Silvia, P. J. (2015). Default and executive network coupling supports creative idea production. Scientific reports, 5, 10964. https://doi.org/10.1038/srep10964 Beaty, R. E., Benedek, M., Silvia, P. J., & Schacter, D. L. (2016). Creative cognition and brain network dynamics. Trends in cognitive sciences, 20(2), 87-95. https://doi.org/10.1016/j.tics.2015.10.004 Beaty, R. E., Benedek, M., Wilkins, R. W., Jauk, E., Fink, A., Silvia, P. J., ... & Neubauer, A. C. (2014). Creativity and the default network: A functional connectivity analysis of the creative brain at rest. Neuropsychologia, 64, 92-98. https://doi.org/10.1016/j.neuropsychologia.2014.09.019 Beaty, R. E., Kenett, Y. N., Christensen, A. P., Rosenberg, M. D., Benedek, M., Chen, Q., ... & Silvia, P. J. (2018). Robust prediction of individual creative ability from brain functional connectivity. Proceedings of the National Academy of Sciences, 115(5), 1087-1092. https://doi.org/10.1073/pnas.1713532115 Beaty, R. E., & Schacter, D. L. (2018). 14 Episodic Memory and Cognitive Control: Contributions to Creative Idea Production. The Cambridge Handbook of the Neuroscience of Creativity, 249. https://doi.org/10.1017/9781316556238.015 Bauer, P. J., Pathman, T., Inman, C., Campanella, C., & Hamann, S. (2017). Neural correlates of autobiographical memory retrieval in children and adults. Memory, 25(4), 450-466. https://doi.org/10.1080/09658211.2016.1186699 Buckner, R. L., Andrews‐Hanna, J. R., & Schacter, D. L. (2008). The brain's default network: anatomy, function, and relevance to disease. Annals of the New York Academy of Sciences, 1124(1), 1-38. https://doi.org/10.1196/annals.1440.011 Buckner, R. L. (2013). The cerebellum and cognitive function: 25 years of insight from anatomy and neuroimaging. Neuron, 80(3), 807-815. https://doi.org/10.1016/j.neuron.2013.10.044

Carson, S. H., Peterson, J. B., & Higgins, D. M. (2005). Reliability, validity, and factor structure of the creative achievement questionnaire. Creativity Research Journal, 17(1), 37-50. http://dx.doi.org/10.1207/s15326934crj1701_4 Chen, Q., Yang, W., Li, W., Wei, D., Li, H., Lei, Q., ... & Qiu, J. (2014). Association of creative achievement with cognitive flexibility by a combined voxel-based morphometry and resting-state functional connectivity study. Neuroimage, 102, 474483. https://doi.org/10.1016/j.neuroimage.2014.08.008 Chen, Q. L., Xu, T., Yang, W. J., Li, Y. D., Sun, J. Z., Wang, K. C., ... & Qiu, J. (2015). Individual differences in verbal creative thinking are reflected in the precuneus. Neuropsychologia, 75, 441-449. https://doi.org/10.1016/j.neuropsychologia.2015.07.001 Christoff, K., Irving, Z. C., Fox, K. C., Spreng, R. N., & Andrews-Hanna, J. R. (2016). Mind-wandering as spontaneous thought: a dynamic framework. Nature Reviews Neuroscience, 17(11), 718. https://doi.org/10.1038/nrn.2016.113 Christoff, K., Gordon, A. M., Smallwood, J., Smith, R., & Schooler, J. W. (2009). Experience sampling during fMRI reveals default network and executive system contributions to mind wandering. Proceedings of the National Academy of Sciences, 106(21), 8719-8724. https://doi.org/10.1073/pnas.0900234106 Cropley, A. J. (2000). Defining and measuring creativity: Are creativity tests worth using? Roeper Review, 23(2), 72–79. https://doi.org/10.1080/02783190009554069 Cools, R., Sheridan, M., Jacobs, E., & D'Esposito, M. (2007). Impulsive personality predicts dopamine-dependent changes in frontostriatal activity during component processes of working memory. Journal of Neuroscience, 27(20), 5506-5514. https://doi.org/10.1523/JNEUROSCI.0601-07.2007 Destrieux, C., Fischl, B., Dale, A., & Halgren, E. (2010). Automatic parcellation of human cortical gyri and sulci using standard anatomical nomenclature. Neuroimage, 53(1), 1-15. https://doi.org/10.1016/j.neuroimage.2010.06.010 Diedrich, J., Jauk, E., Silvia, P. J., Gredlein, J. M., Neubauer, A. C., & Benedek, M. (2018). Assessment of real-life creativity: The Inventory of Creative Activities and Achievements (ICAA). Psychology of Aesthetics, Creativity, and the Arts, 12(3), 304. http://dx.doi.org/10.1037/aca0000137304 Dietrich, A., & Kanso, R. (2010). A review of EEG, ERP, and neuroimaging studies of creativity and insight. Psychological bulletin, 136(5), 822. https://doi.org/10.1037/a0019749 Dollinger, S. J. (2003). Need for uniqueness, need for cognition, and creativity. Journal of Creative Behavior, 37, 99–116. https://doi.org/10.1002/j.21626057.2003.tb00828.x Efron, B., & Tibshirani, R. (1986). Bootstrap methods for standard errors, confidence intervals, and other measures of statistical accuracy. Statistical science, 54-75. https://doi.org/10.1214/ss/1177013817 Eklund, A., Knutsson, H., & Nichols, T. E. (2019). Cluster failure revisited: Impact of first level design and physiological noise on cluster false positive rates. Human brain mapping, 40(7), 2017-2032. https://doi.org/10.1002/hbm.24350 Fink, A., & Benedek, M. (2014). EEG alpha power and creative ideation. Neuroscience & Biobehavioral Reviews, 44, 111–123. https://doi.org/10.1016/j.neubiorev.2012.12.002

Fischl, B., van der Kouwe, A., Destrieux, C., … & Dale, A. (2004). Automatically parcellating the human cerebral cortex. Cerebral cortex, 14(1), 11-22. https://doi.org/10.1093/cercor/bhg087 Flaherty, A. W. (2005). Frontotemporal and dopaminergic control of idea generation and creative drive. Journal of Comparative Neurology, 493(1), 147-153. https://doi.org/10.1002/cne.20768 Fox, M. D., Snyder, A. Z., Vincent, J. L., Corbetta, M., Van Essen, D. C., & Raichle, M. E. (2005). The human brain is intrinsically organized into dynamic, anticorrelated functional networks. Proceedings of the National Academy of Sciences, 102(27), 9673-9678. https://doi.org/10.1073/pnas.0504136102 Furnham, A., & Bachtiar, V. (2008). Personality and intelligence as predictors of creativity. Personality and individual differences, 45(7), 613-617. https://doi.org/10.1016/j.paid.2008.06.023 Glasser, M. F., Sotiropoulos, S. N., Wilson, J. A., Coalson, T. S., Fischl, B., Andersson, J. L., ... & Van Essen, D. C. (2013). The minimal preprocessing pipelines for the Human Connectome Project. Neuroimage, 80, 105-124. https://doi.org/10.1016/j.neuroimage.2013.04.127 Goff, K. (2002). Abbreviated torrance test for adults. Bensenville, IL: Scholastic Testing Service. Gonen-Yaacovi, G., De Souza, L. C., Levy, R., Urbanski, M., Josse, G., & Volle, E. (2013). Rostral and caudal prefrontal contribution to creativity: a meta-analysis of functional imaging data. Frontiers in human neuroscience, 7, 465. https://doi.org/10.3389/fnhum.2013.00465 Good, C. D., Johnsrude, I. S., Ashburner, J., Henson, R. N., Friston, K. J., & Frackowiak, R. S. (2001). A voxel-based morphometric study of ageing in 465 normal adult human brains. Neuroimage, 14(1), 21-36. https://doi.org/10.1006/nimg.2001.0786 Guilford, J. P. (1967). The nature of human intelligence. New York: McGraw-Hill. Jauk, E., Benedek, M., & Neubauer, A. C. (2014). The road to creative achievement: A latent variable model of ability and personality predictors. European journal of personality, 28(1), 95-105. https://doi.org/10.1002/per.1941 Jenkinson, M., Beckmann, C. F., Behrens, T. E., Woolrich, M. W., & Smith, S. M. (2012). Fsl. Neuroimage, 62(2), 782-790. https://doi.org/10.1016/j.neuroimage.2011.09.015 Jung, R. E., Segall, J. M., Jeremy Bockholt, H., Flores, R. A., Smith, S. M., Chavez, R. S., & Haier, R. J. (2010). Neuroanatomy of creativity. Human brain mapping, 31(3), 398-409. https://doi.org/10.1002/hbm.20874 Jung, R. E., Wertz, C. J., Meadows, C. A., Ryman, S. G., Vakhtin, A. A., & Flores, R. A. (2015). Quantity yields quality when it comes to creativity: a brain and behavioral test of the equal-odds rule. Frontiers in psychology, 6, 864. https://doi.org/10.3389/fpsyg.2015.00864 Kim, K. H., Cramond, B., & VanTassel-Baska, J. (2010). The relationship between creativity and intelligence. The Cambridge handbook of creativity, 395-412. Kim, K. H. (2005). Can only intelligent people be creative? A meta-analysis. Journal of Secondary Gifted Education, 16(2-3), 57-66. https://doi.org/10.4219%2Fjsge-2005473

Kershner, J. R., & Ledger, G. (1985). Effect of sex, intelligence, and style of thinking on creativity: A comparison of gifted and average IQ children. Journal of personality and social psychology, 48(4), 1033. http://dx.doi.org/10.1037/0022-3514.48.4.1033 Koziol, L. F., & Budding, D. E. (2009). The Cerebellum: Quality control, creativity, intuition, and unconscious working memory. In Subcortical Structures and Cognition (pp. 125-165). Springer, New York, NY. https://doi.org/10.1007/978-0-387-84868-6_5 Kühn, S., Ritter, S. M., Müller, B. C., Van Baaren, R. B., Brass, M., & Dijksterhuis, A. (2014). The importance of the default mode network in creativity—A structural MRI study. The Journal of Creative Behavior, 48(2), 152-163. https://doi.org/10.1002/jocb.45 Lee, S. Y., Florida, R., & Acs, Z. (2004). Creativity and entrepreneurship: a regional analysis of new firm formation. Regional studies, 38(8), 879-891. https://doi.org/10.1080/0034340042000280910 Leys, C., Ley, C., Klein, O., Bernard, P., & Licata, L. (2013). Detecting outliers: Do not use standard deviation around the mean, use absolute deviation around the median. Journal of Experimental Social Psychology, 49(4), 764-766. https://doi.org/10.1016/j.jesp.2013.03.013 Li, K., Guo, L., Faraco, C., Zhu, D., Deng, F., Zhang, T., ... & Miller, S. (2010). Individualized ROI optimization via maximization of group-wise consistency of structural and functional profiles. In Advances in Neural Information Processing Systems (pp. 1369-1377). https://doi.org/10.1007/s12021-012-9142-5 Li, W., Yang, J., Zhang, Q., Li, G., & Qiu, J. (2016). The Association between resting functional connectivity and visual creativity. Scientific reports, 6, 25395. https://doi.org/10.1038/srep25395 Lieberman, M. D., & Cunningham, W. A. (2009). Type I and Type II error concerns in fMRI research: re-balancing the scale. Social cognitive and affective neuroscience, 4(4), 423-428. https://doi.org/10.1093/scan/nsp052 Madore, K. P., Addis, D. R., & Schacter, D. L. (2015). Creativity and memory: Effects of an episodic-specificity induction on divergent thinking. Psychological science, 26(9), 1461-1468. https://doi.org/10.1177/0956797615591863 Madore, K. P., Thakral, P. P., Beaty, R. E., Addis, D. R., & Schacter, D. L. (2017). Neural mechanisms of episodic retrieval support divergent creative thinking. Cerebral Cortex, 29(1), 150-166. https://doi.org/10.1093/cercor/bhx312 Martindale, C. (1999). Biological bases of creativity. Handbook of creativity, 2, p 137152. Mason, M. F., Norton, M. I., Van Horn, J. D., Wegner, D. M., Grafton, S. T., & Macrae, C. N. (2007). Wandering minds: the default network and stimulus-independent thought. Science, 315(5810), 393-395. https://doi.org/10.1126/science.1131295 Matorina, N., & Poppenk, J. (2019). Sleep promotes relational overlapping memories for long-term generalization. bioRxiv, 578492. Menon V. (2015) Salience Network. In: Arthur W. Toga, editor. Brain Mapping: An Encyclopedic Reference, vol. 2, pp. 597-611. Academic Press: Elsevier. Middleton, F. A., & Strick, P. L. (2002). Basal-ganglia ‘projections’ to the prefrontal cortex of the primate. Cerebral Cortex, 12(9), 926-935. https://doi.org/10.1093/cercor/12.9.926

Moore, D. W., Bhadelia, R. A., Billings, R. L., Fulwiler, C., Heilman, K. M., Rood, K. M., & Gansler, D. A. (2009). Hemispheric connectivity and the visual–spatial divergentthinking component of creativity. Brain and Cognition, 70(3), 267-272. https://doi.org/10.1016/j.bandc.2009.02.011 Mountjoy, J., & Poppenk, J. (2015). Introducing SuperPsychToolbox: An Open-Source Tool to Facilitate Coding and Analysis of Psychology Experiments. In Canadian Journal of Experimental Psychology, 69(4), 332–332. Nagano-Saito, A., Leyton, M., Monchi, O., Goldberg, Y. K., He, Y., & Dagher, A. (2008). Dopamine depletion impairs frontostriatal functional connectivity during a set-shifting task. Journal of Neuroscience, 28(14), 3697-3706. 10.1523/JNEUROSCI.392107.2008 Neumann, N., Domin, M., Erhard, K., & Lotze, M. (2018). Voxel‐based morphometry in creative writers: Grey matter increase in a prefronto‐thalamic‐cerebellar network. European Journal of Neuroscience, 48(1), 1647-1653. https://doi.org/10.1111/ejn.13952 Nijstad, B. A., De Dreu, C. K., Rietzschel, E. F., & Baas, M. (2010). The dual pathway to creativity model: Creative ideation as a function of flexibility and persistence. European Review of Social Psychology, 21(1), 34-77. https://doi.org/10.1080/10463281003765323 Ogawa, T., Aihara, T., Shimokawa, T., & Yamashita, O. (2018). Large-scale brain network associated with creative insight: combined voxel-based morphometry and resting-state functional connectivity analyses. Scientific reports, 8. https://dx.doi.org/10.1038%2Fs41598-018-24981-0 Piffer, D. (2012). Can creativity be measured? An attempt to clarify the notion of creativity and general directions for future research. Thinking Skills and Creativity, 7(3), 258-264. https://doi.org/10.1016/j.tsc.2012.04.009 Poppenk, J., Evensmoen, H. R., Moscovitch, M., & Nadel, L. (2013). Long-axis specialization of the human hippocampus. Trends in cognitive sciences, 17(5), 230240. https://doi.org/10.1016/j.tics.2013.03.005 Poppenk, J., & Moscovitch, M. (2011). A hippocampal marker of recollection memory ability among healthy young adults: contributions of posterior and anterior segments. Neuron, 72(6), 931-937. https://doi.org/10.1016/j.neuron.2011.10.014 Richards, R. (Ed.). (2007). Everyday creativity and new views of human nature: Psychological, social, and spiritual perspectives. Washington, DC, US: American Psychological Association. http://dx.doi.org/10.1037/11595-000 Romero, J. E., Coupé, P., & Manjón, J. V. (2017). HIPS: A new hippocampus subfield segmentation method. NeuroImage, 163, 286-295. https://doi.org/10.1016/j.neuroimage.2017.09.049 Salimi-Khorshidi, G., Douaud, G., Beckmann, C. F., Glasser, M. F., Griffanti, L., & Smith, S. M. (2014). Automatic denoising of functional MRI data: combining independent component analysis and hierarchical fusion of classifiers. Neuroimage, 90, 449-468. https://doi.org/10.1016/j.neuroimage.2013.11.046 Sattler, J. M., & Ryan, J. J. (2009). Assessment with the WAIS-IV. Jerome M Sattler. Schacter, D. L., Addis, D. R., & Szpunar, K. K. (2017). Escaping the Past: Contributions of the Hippocampus to Future Thinking and Imagination. In D. E. Hannula & M. C. Duff (Eds.), The Hippocampus from Cells to Systems: Structure, Connectivity, and

Functional Contributions to Memory and Flexible Cognition (pp. 439–465). https://doi.org/10.1007/978-3-319-50406-3_14 Schouten, K. A., de Niet, G. J., Knipscheer, J. W., Kleber, R. J., & Hutschemaekers, G. J. (2015). The effectiveness of art therapy in the treatment of traumatized adults: a systematic review on art therapy and trauma. Trauma, violence, & abuse, 16(2), 220228. https://doi.org/10.1177/1524838014555032 Shen, W., Yuan, Y., Liu, C., & Luo, J. (2017). The roles of the temporal lobe in creative insight: an integrated review. Thinking & Reasoning, 23(4), 321-375. http://dx.doi.org/10.1080/13546783.2017.1308885 Silvia, P. J. (2008). Another look at creativity and intelligence: Exploring higher-order models and probable confounds. Personality and Individual differences, 44(4), 10121021. https://dx.doi.org/10.1016/j.paid.2007.10.027 Spreng, R. N., Mar, R. A., & Kim, A. S. (2009). The common neural basis of autobiographical memory, prospection, navigation, theory of mind, and the default mode: a quantitative meta-analysis. Journal of cognitive neuroscience, 21(3), 489510. https://doi.org/10.1162/jocn.2008.21029 Spreng, R. N., Stevens, W. D., Chamberlain, J. P., Gilmore, A. W., & Schacter, D. L. (2010). Default network activity, coupled with the frontoparietal control network, supports goal-directed cognition. Neuroimage, 53(1), 303-317. https://doi.org/10.1016/j.neuroimage.2010.06.016 Sridharan, D., Levitin, D. J., & Menon, V. (2008). A critical role for the right fronto-insular cortex in switching between central-executive and default-mode networks. Proceedings of the National Academy of Sciences, 105(34), 12569-12574. https://doi.org/10.1073/pnas.0800005105 Stoodley, C. J., & Schmahmann, J. D. (2016). Functional topography of the human cerebellum. In Essentials of Cerebellum and Cerebellar Disorders (pp. 373-381). Springer, Cham. https://doi.10.1016/B978-0-444-63956-1.00004-7 Sun, J., Liu, Z., Rolls, E. T., Chen, Q., Yao, Y., Yang, W., ... & Qiu, J. (2018). Verbal creativity correlates with the temporal variability of brain networks during the resting state. Cerebral Cortex, 29(3), 1047-1058. https://doi.org/10.1093/cercor/bhy010 Takeuchi, H., Taki, Y., Sassa, Y., Hashizume, H., Sekiguchi, A., Fukushima, A., & Kawashima, R. (2010a). Regional gray matter volume of dopaminergic system associate with creativity: evidence from voxel-based morphometry. Neuroimage, 51(2), 578-585. https://doi.org/10.1016/j.neuroimage.2010.02.078 Takeuchi, H., Taki, Y., Sassa, Y., Hashizume, H., Sekiguchi, A., Fukushima, A., & Kawashima, R. (2010b). White matter structures associated with creativity: evidence from diffusion tensor imaging. Neuroimage, 51(1), 11-18. https://doi.org/10.1016/j.neuroimage.2010.02.035 Takeuchi, H., Taki, Y., Hashizume, H., Sassa, Y., Nagase, T., Nouchi, R., & Kawashima, R. (2012). The association between resting functional connectivity and creativity. Cerebral Cortex, 22(12), 2921-2929. https://doi.org/10.1093/cercor/bhr371 Takeuchi, H., Taki, Y., Nouchi, R., Yokoyama, R., Kotozaki, Y., Nakagawa, S., … Hanawa, S. (2017). Creative females have larger white matter structures: Evidence from a large sample study. Human Brain Mapping, 38(1), 414–430.

Torrance, E. P. (1966). The Torrance Tests of Creative Thinking-Norms-Technical Manual Research Edition-Verbal Tests, Forms A and B-Figural Tests, Forms A and B. Princeton, NJ: Personnel Press. Vandervert, L. R., Schimpf, P. H., & Liu, H. (2007). How working memory and the cerebellum collaborate to produce creativity and innovation. Creativity Research Journal, 19(1), 1-18. https://doi.org/10.1080/10400410709336877 Wallach, M. A., & Kogan, N. (1965). Modes of thinking in young children: a study of the creativity–intelligence distinction. Oxford, England: Holt, Rinehart & Winston. Wei, D., Yang, J., Li, W., Wang, K., Zhang, Q., & Qiu, J. (2014). Increased resting functional connectivity of the medial prefrontal cortex in creativity by means of cognitive stimulation. Cortex, 51, 92-102. https://doi.org/10.1016/j.cortex.2013.09.004 Welling, H. (2007). Four mental operations in creative cognition: The importance of abstraction. Creativity Research Journal, 19(2-3), 163-177. https://doi.org/10.1080/10400410701397214 Wu, X., Yang, W., Tong, D., Sun, J., Chen, Q., Wei, D., ... & Qiu, J. (2015). A meta‐ analysis of neuroimaging studies on divergent thinking using activation likelihood estimation. Human brain mapping, 36(7), 2703-2718. https://doi.org/10.1002/hbm.22801 Yagmurlu, K., Middlebrooks, E. H., Tanriover, N., & Rhoton, A. L. (2016). Fiber tracts of the dorsal language stream in the human brain. Journal of neurosurgery, 124(5), 1396-1405. https://doi.org/10.3171/2015.5.JNS15455 Zhu, F., Zhang, Q., & Qiu, J. (2013). Relating inter-individual differences in verbal creative thinking to cerebral structures: an optimal voxel-based morphometry study. PLoS One, 8(11), e79272. Zhu, W., Chen, Q., Tang, C., Cao, G., Hou, Y., & Qiu, J. (2016). Brain structure links everyday creativity to creative achievement. Brain and Cognition, 103, 70-76. https://doi.org/10.1016/j.bandc.2015.09.008

Figure Captions Fig. 1. Multimodal predictors of creativity. (A) The anterior lobe of the cerebellum was implicated as a volumetric predictor of ATTA scores (verbal and visual) as well as CBI scores. Greater white matter integrity within the anterior lobe of the cerebellum similarly predicted greater creativity scores of visual ATTA and CBI scores. (B) We observed a parahippocampal predictor of creativity scores obtained using each of our four creativity inventories, although the localization of predictive voxels within the parahippocampal cortex or adjacent white matter varied. Positive correlations between creativity scores and grey matter density or structural integrity are shown in orange, and negative correlations in blue. Clusters were thresholded at a corrected cluster-wise threshold of p < 0.05. Right = anterior in all images. See also Tables 5 and 6 for detailed cluster statistics.

Supplemental results We pre-registered additional neural predictors of creativity that do not conform well to the neuroimaging focus of our main text. For purposes of completeness and adherence to our pre-registration plan, those analyses are reported below. S1.0. Sleep EEG predictors Sleep has often been implicated in creative processes, with many people noticing creative insights directly after falling asleep or upon waking (Heilman, 2016). Additionally, sleep-deprived children obtained lower verbal TTCT scores than children in the control condition (Randazzo et al., 1998). REM sleep has received special focus in proposals linking sleep to creativity, as it is characterized by excitability, plasticity, and connectivity; according to one proposal, memory replay in REM promotes the recombination of novel associations, whereas memory replay in non-REM sleep facilitates the abstraction of gist information (Lewis et al., 2018). Consistent with this idea, Cai and colleagues (2009) found that REM sleep enhanced creative problem solving for primed items on the Remote Associations Task (Mednick, 1962) relative to quiet rest and non-REM sleep. Accordingly, we pre-registered a series of one-tailed tests aimed at replicating previously observed effects of REM sleep time on scores from creativity tests. S1.1. Results We observed a trend towards prediction of ATTA visual scores by total REM sleep time, r(51) = .22, p = .061, and significant prediction by the proportion of REM to non-REM sleep, r(52) = .30, p = .020, but not total sleep time, r(51) = -.03, p = .596. However, ATTA verbal scores were not reliably predicted by any sleep predictors, including REM sleep time, r(53) = .08, p = .49, the proportion of REM to non-REM sleep, r(54) = .06, p = .276, and total sleep time, r(52) = .04, p = .381. As a post-hoc analysis, we also correlated the CAQ and CBI scores with our sleep measures, testing in the same direction. CAQ scores were not related to any sleep measure, including REM sleep time, r(55) = .09, p = .280, the proportion of REM to non-REM sleep, r(56) = .04, p = .381, and total sleep time, r(54) = .00, p = .499. CBI scores also were not related to REM sleep time, r(56) = .16, p = .11, or to total sleep time, r(55) = -.09, p = .744, but were significantly related to the proportion of REM to non-REM sleep, r(57) = .23, p = .020. Overall, our sleep results provide some limited support for the idea that REM sleep facilitates creative processes. In our study, the proportion of REM was a better predictor than total REM sleep or total sleep time. It is worth noting that our sleep recordings originated from a portable home sleep EEG device, the Sleep Profiler: While providing for more naturalistic sleep at home than participants sleeping in a laboratory, this device is also likely less sensitive than an in-laboratory polysomnography setup (Matorina & Poppenk, 2019; Lucey et al., 2016). S2.0. Hippocampal subfield predictors Beaty et al. (2016) described evidence linking creative cognition and episodic memory; Beaty et al. (2018) later found an emerging core of brain regions that included

the bilateral hippocampus that were activated when engaging in recalling past experiences, stimulating future ones, and engaging in divergent thinking. We therefore further hypothesized that the hippocampus might contribute to the prediction of creativity. We tailored our hypotheses to the detailed properties of the hippocampus. Openness refers to being able to delay closure; completing incomplete figures with straight or somewhat curved lines is evidence of premature closure and the 'completion' of a pattern. The CA1 and CA3 subfields are hippocampal regions associated with temporal and spatial pattern completion, respectively (Klausberger & Somogyi, 2008). Because of this role, we posited that participants with a larger CA1 and CA3 might be predisposed to close figures sooner (lower openness) . Flexibility in the ATTA is the ability to process information in distinct ways, given the same stimulus. In activity three, participants draw distinct figures from nine triangles. Participants with a smaller DG may be more likely to become “fixated” in a particular interpretation, and thus have lower flexibility. Accordingly, we posited a positive correlation between DG volume and flexibility. Finally, we (Poppenk et al., 2013) argued that the anterior hippocampus is implicated in gist representations of memory, which although less exact, are more accessible for retrieval due to a lower requirement for exact pattern-matching. We hypothesized that this easier pattern-matching would be associated with higher fluency, based on the idea that ideation is supported by memory. By contrast, the posterior hippocampus features detailed representations of memory. We proposed a positive correlation between elaboration and the posterior hippocampus, because embellishment of details would be supported by superior detail memory. S2.1. Results We pre-registered exploratory predictions regarding the anterior and posterior hippocampus and the fluency and originality facets of the ATTA. We observed no significant relations between any left or right hippocampal segment and any of these facets, all p’s > .365. We also observed no relationship between pre-registered associations between CA1 or CA3 and the openness facet, nor DG and the flexibility facet, all p’s > 0.166.

Supplemental References Beaty, R. E., Benedek, M., Silvia, P. J., & Schacter, D. L. (2016). Creative cognition and brain network dynamics. Trends in cognitive sciences, 20(2), 87-95. https://doi.org/10.1016/j.tics.2015.10.004 Beaty, R. E., Thakral, P. P., Madore, K. P., Benedek, M., & Schacter, D. L. (2018). Core network contributions to remembering the past, imagining the future, and thinking creatively. Journal of cognitive neuroscience, 30(12), 1939-1951. https://doi.org/10.1162/jocn_a_01327 Cai, D. J., Mednick, S. A., Harrison, E. M., Kanady, J. C., & Mednick, S. C. (2009). REM, not incubation, improves creativity by priming associative

networks. Proceedings of the National Academy of Sciences, 106(25), 10130-10134. https://doi.org/10.1073/pnas.0900271106 Heilman, K. M. (2016). Possible brain mechanisms of creativity. Archives of Clinical Neuropsychology, 31(4), 285-296. https://doi.org/10.1093/arclin/acw009 Klausberger, T., & Somogyi, P. (2008). Neuronal diversity and temporal dynamics: The unity of hippocampal circuit operations. Science, 321(5885), 53–57. Lewis, P. A., Knoblich, G., & Poe, G. (2018). How Memory Replay in Sleep Boosts Creative Problem-Solving. Trends in cognitive sciences, 22(6), 491-503. https://doi.org/10.1016/j.tics.2018.03.009 Lucey, B. P., Mcleland, J. S., Toedebusch, C. D., Boyd, J., Morris, J. C., Landsness, E. C., ... & Holtzman, D. M. (2016). Comparison of a single‐channel EEG sleep study to polysomnography. Journal of sleep research, 25(6), 625-635. https://doi.org/10.1111/jsr.12417 Mednick, S. (1962). The associative basis of the creative process. Psychological review, 69(3), 220. http://dx.doi.org/10.1037/h0048850 Randazzo, A. C., Muehlbach, M. J., Schweitzer, P. K., & Waish1, J. K. (1998). Cognitive function following acute sleep restriction in children ages 10–14. Sleep, 21(8), 861868. https://doi.org/10.1093/sleep/21.8.861 Sheldon, S., & Moscovitch, M. (2012). The nature and time‐course of medial temporal lobe contributions to semantic retrieval: An fMRI study on verbal fluency. Hippocampus, 22(6), 1451-1466. https://doi.org/10.1002/hipo.20985

Table 1. Descriptive statistics of behavioural test performance. Construct Divergent thinking (verbal) Divergent thinking (visual) Creative achievement Creative behaviour Intelligence

Test Abbreviated Torrance Test for Adults (ATTA) Abbreviated Torrance Test for Adults (ATTA) Creative Achievement Questionnaire (CAQ) Creative Behaviour Inventory (CBI) Wechler Adult Intelligence Scale (WAIS-IV)

N 65 65 66 66 60

M

SD

2.9 1.6 7.6 3.1 18.4 14.1 0.9 0.5 113.1 12.4

*Because of substantial skew in this instrument, CAQ scores were transformed prior to further analysis (see also Carson, Peterson, & Higgins, 2005).

Skew 0.58 -0.31 1.59* 0.48 -0.35

Table 2 . ROI-based volume predictors of creativity, controlling for intelligence. Reported result Brain structure

Inferior frontal g.

Hemi.

Creativity test

L R

TTCT-Verbal

L Ventromedial prefrontal ctx.

Alternate Uses Task R

Temporoparietal junction

Alternate Uses Task

Kühn et al. (2014)

L

S-A Creativity Test

Takeuchi et al. (2010a) Chen et al. (2015)*

R

Caudate n. Brainstem Inferior parietal l.

CAQ

Zhu et al. (2013)* Jung et al. (2010) Kühn et al. (2014) Chen et al. (2014)*

R

Precuneus

Insula

CAQ

Source Paper

L L R L/R R

TTCT-Verbal

S-A Creativity Test

Takeuchi et al. (2010a)

CAQ

Jung et al. (2010)

Verbal ATTA

dv

n

Stat.

% beyond thresh.

vol

283

r = 0.25 r = 0.24

4.42 3.78

thick

Visual ATTA

CBI

CAQ

% beyond thresh.

X

p

% beyond thresh.

X

30.1 < 0.001 13.5 < 0.001

2.26 1.69

0.4 5.3

0.802 0.983

3.54 4.55

8.8 34.8

0.003 4.82 < 0.001 6.29

43.7 < 0.001 119.1 < 0.001

61 z = -3.21 0.67

1.0

0.845

4.70

13.4

< 0.001

2.69

20.3

< 0.001 6.10

52.7 < 0.001

z = 4.26 2.66

0.3

0.212

3.54

13.7

< 0.001

0.00

8.1

0.996

7.23

285.9 < 0.001

t = 4.23

80.7 < 0.001

0.90

10.3

0.999

0.00

1.4

0.874

6.59

66.9 < 0.001

20.7

0.999

2.05

4.7

0.976

2.77

1.7

0.108

6.18

318.9 < 0.001

0.00

52.4

0.999

4.15

22.8

< 0.001

0.00

2.1

0.911

2.53

0.0

0.249

2.37

0.1

0.765

3.27

4.2

0.031

0.00

1.2

0.865

3.38

5.5

0.016

2.71 3.21 0.00 0.98

0.2 1.8 23.9 62.6

0.223 0.100 0.999 0.999

1.39 0.00 0.00 6.20

6.3 22.6 23.9 372.5

0.989 0.999 0.999 < 0.001

0.36 10.21 0.00 2.81

61 z = 3.99 0.52

50.7

0.999

2.65

0.3

0.217

0.00

21 366 21

6.99

z = 4.76 1.56

2

X

p

2

2

p

% beyond thresh.

2

X

p

z = 3.99 vol

55

t = 4.15

268 r = .192 t = 4.03 z = 3.84 55 t = 4.57 t = 3.93 t = 5.10 thick

5.1 0.981 274.7 < 0.001 27.8 0.999 2.8 0.062 2.2

0.917

8.29 0.11 2.10 2.01 0.00

173.0 < 0.001 20.6 0.999 0.6 0.817 6.5 0.990 2.2

0.918

Note: % beyond-threshold voxels was calculated as the proportion of voxels exceeding a BSR of 1.96 (p < 0.05) in the direction of the original result. Highlighting designates ROI / test combinations where more voxels fell beyond threshold than expected by chance. Outlined areas designate tests most relevant to prior investigations. *Represents studies related to our replication efforts but those that we did not pre-register. Abbreviations: ctx, cortex; dv, dependent variable; g, gyrus; hemi, hemisphere; L, left; n, nucleus; R, right; vol, volume.

Table 3 . ROI-based FA predictors of creativity, controlling for intelligence. Reported result

Brain structure

Inferior frontal g. Superior frontal g. Superior logitudinal f. Internal capsule (anterior) Caudate n. Inferior parietal l. Basal ganglia Temporoparietal junction Corpus callosum (body) Uncinate fasciculus

Creativity Hemi. test

L R L R L R

TTCTVerbal

Source Paper

Zhu et al. (2013)*

dv

n

wmv 283

L

Verbal ATTA

Stat.

p

r = 0.23 r = 0.23 t = 4.07 t = 2.71 t = 2.71 t = 3.14

< .001 < .001 0.020 0.045 0.045 0.033

% beyon d thresh.

2

X

Visual ATTA

p

0.65 2.1 0.912 5.67 7.8 0.005 1.02 32.4 1.000 0.16 85.1 1.000 19.74 2251.9 < 0.001 2.42 0.1 0.757

% beyon d thresh. 0 0 2.29 2.945 7.177 4.414

2

X

CBI

p

3.8 0.963 4.9 0.978 0.7 0.820 3.1 0.054 165.8 < 0.001 27.9 < 0.001

19.2

1.000

0.975 0 < 0.001 0.078 < 0.001 8.989 < 0.001 1.328

27.8 1.000 0 30.1 1.000 0.781 676.4 < 0.001 0.822 21.9 1.000 0.05

27.8 15.1 45.2 95.8

1.000 1.000 1.000 1.000

27.8 1.000 28.2 1.000 317.7 < 0.001 31.2 1.000

L

< 0.05^

1.79

4.0

L/R

< 0.05^

6.39

57.7 < 0.001

0

23.9

1.000

8.377

0.00

2.2

0

2.2

0.999

0

L

Composite DT index

Jung et. al (2010)

fa

72 t = -5.36 0.010

0.966

0.999

0.035 4.6 5.543 47.4 9.711 835.3 17.59 3633.7 7.478

p

951.0 < 0.001 0.217

0.00 0.16 6.95 1.10

55

X2

18.57

t = 3.27 0.031 t = 2.8 0.042 < 0.05^ < 0.05^

fa

p

3.8 0.963 0.7 0.825 36.0 1.000 70.8 1.000 120.7 < 0.001 60.0 1.000

104.2 < 0.001 1.629

Takeuchi et al. (2010b)

X

% beyon d thresh.

0.9 0.838 0 13.7 <0.001 1.55 0.7 0.820 0.943 156.9 < 0.001 0.362 9.9 0.002 6.491 47.3 < 0.001 0.053

7.82

R R L R

2

1.31 6.7 2.29 5.683 3.641 4.992

t = 3.26 0.031 S-A Creativity Test

2.8

% beyon d thresh.

CAQ

0.939

198.8 < 0.001 2.642

0.2

0.616

131.9 < 0.001 2.2

0.999

4.586

34.9 < 0.001

0

23.9

1.000

0

2.2

0.999

Note: % beyond-threshold voxels was calculated as the proportion of voxels exceeding a BSR of 1.96 (p < 0.05) in the direction of the original result. Highlighting designates ROI / test combinations where more voxels fell beyond threshold than expected by chance. Outlined areas indicate tests most relevant to prior investigations. *Represents studies related to our replication efforts but those that we did not pre-register. ^Not reported in the source paper. Abbreviations: ctx, cortex; dv, dependent variable; g, gyrus; hemi, hemisphere; fa, functional anisotropy; L, left; l, lobe; n, nucleus; R, right; wmv, white matter volume.

Table 4 . ROI-based functional connectivity predictors of creativity, controlling for intelligence. Reported result Seed

Seed hemi. L

Inferior frontal g.

R

L

Inferior parietal l. Anterior cingulate ctx. Posterior cingulate ctx.

L

Middle temporal g.

L L

Anterior cingulate ctx.

Dorsolateral prefrontal ctx.

Target

L

Targe t Source paper Creativity test n hemi. L R L R L/R

Beaty et al. (2014)

Alternate Uses & Instances Task

L/R

L

Wei et al. (2014)*

TTCTVerbal

Posterior cingulate ctx.

L/R

Takeuchi et al. (2012)*

S-A creativity test

Precuneus

L/R

Dorsolateral prefrontal ctx.

24

R

Li et al. (2016)*

ATTA (verbal) ATTA (visual)

CBI

CAQ

Stat.

p

r

p

r

p

r

p

r

p

t = 3.00 t = 3.75 t = 2.76 t = 3.36

0.008 0.004 0.022 0.011

0.24 0.15 0.11 0.35

0.038 0.139 0.205 0.004

0.19 0.10 -0.14 0.07

0.049 0.227 0.167 0.308

0.13 0.06 -0.13 0.03

0.176 0.315 0.812 0.400

0.11 -0.02 -0.15 0.00

0.213 0.544 0.818 0.508

t = 2.38 0.026

0.04 0.394 0.11 0.255 0.00 0.013 0.03 0.415

t = 3.34 0.005

0.09 0.273 0.04 0.378 0.13 0.177 0.11 0.233

269 r = .33 < 0.001 0.00 0.501 0.23 0.045 0.15 0.096 0.24 0.043

159 t = 5.16 0.038

0.22 0.071 0.24 0.068 0.16 0.144 0.15 0.120

t = -4.78 0.039

0.08 0.726 0.03 0.584 0.21 0.973 0.06 0.665

t = 5.18 0.006

0.17 0.157 0.14 0.174 0.11 0.261 0.07 0.303

TTCT-visual 304

Note: Highlighting designates ROI / test combinations where more voxels fell beyond threshold than expected by chance. Outlined areas designate tests most relevant to prior investigations. *Represents studies related to our replication efforts but those that we did not pre-register. Abbreviations: ctx, cortex; dv, dependent variable; hemi, hemisphere; L, left; R, right.

Table 5 VBM voxel clusters correlated with creativity measures, after controlling for intelligence. Region

Hemi.

Peak MNI coordinates Spatial Cluster Peak z extent corrected estimate X Y Z (mm³) p

ATTA (Verbal) Temporal pole Insula (anterior) Parahippocampal g. Parahippocampal g. Thalamus Cerebellum (lobules V-VI) Cerebellum (anterior l.) Cerebellum (lobule VI) Cerebellum (lobule VIII) Cerebellum (lobule VII) Dorsolateral prefrontal ctx. Fusiform g. (anterior) Thalamus Posterior cingulate g. (dorsal) Temporoparietal junct. Cuneus Cuneus

R R L R L R R R R L R R R L/R L L/R R

45 30 -20 33 -12 21 8 30 27 -20 47 39 11 3 -53 -3 5

14 8 -18 -20 -20 -42 -44 -60 -74 -84 5 -11 -26 -35 -59 -68 -74

-20 -15 -23 -29 15 -23 -12 -29 -44 -30 33 -26 -3 48 18 8 21

624 520 385 273 253 253 250 250 240 223 402 361 334 257 253 233 226

0.026 0.019 0.021 0.001 0.019 0.039 0.000 0.021 0.000 0.012 0.002 0.003 0.031 0.018 0.001 0.019 0.036

4.02 4.95 3.93 5.06 8.49 4.07 4.39 4.18 4.10 5.62 -6.56 -4.54 -5.04 -5.73 -5.55 -4.19 -4.91

L/R L/R L L/R R L

3 -6 -6 0 26 -30

-21 -45 -51 -53 -60 -9

-15 -6 2 -36 -47 -20

969 628 452 375 250 851

0.001 < 0.001 0.019 < 0.001 < 0.001 0.028

3.86 6.66 5.27 5.36 3.71 -5.18

ATTA (Visual) Substantia nigra / Ventral tegmental area parahippocampal g. Cuneus Cerebellum (flocculonodular l.) Cerebellum (lobule VIII) Hippocampus (anterior)

Cerebellum (lobule X) Cerebellum (lobule VIII) Cerebellum (lobule VI) Cerebellum (lobule VI)

L/R R R L

-6 33 26 -29

-44 -65 -69 -71

-24 -59 -30 -26

496 253 233 230

< 0.001 0.017 < 0.001 0.030

-4.22 -3.47 -4.89 -4.35

L/R R R L/R L L L L/R

-2 44 54 0 -35 -53 -33 -3

33 20 -14 -44 -47 -56 -45 -68

-9 3 -29 -14 -23 -29 -33 8

614 621 311 392 273 223 790 635

0.007 0.024 < 0.001 0.003 < 0.001 < 0.001 < 0.001 < 0.001

4.63 4.60 5.82 5.52 4.54 4.07 -5.63 -5.58

R L R R R L R R

44 -21 45 50 11 -14 12 6

14 -38 -42 -59 -59 -71 6 -20

30 -12 47 -42 -57 -20 15 8

726 523 500 287 257 196 334 213

< 0.001 < 0.001 0.004 < 0.001 0.009 0.046 0.030 0.001

6.79 6.09 4.21 5.42 5.53 5.23 -6.36 -4.41

CBI Anterior cingulate g. Insula (anterior) Inferior temporal g. Cerebellum (anterior l.) Cerebellum (lobule VI) Cerebellum (lobule VII) Cerebellum (lobule VI) Cuneus CAQ Medial frontal g. Parahippocampal g. Supramarginal g. Cerebellum (lobule VII) Cerebellum (lobule X) Cerebellum (lobule V-VI) Caudate Thalamus

Note: coordinates are displayed in MNI space. Names of negative predictors are in italics.

Table 6 . FA voxel clusters correlated with creativity measures, after controlling for intelligence. Peak MNI coordinates Spatial Cluster Peak z Region Hemi. extent corrected estimate X Y Z p (mm³) ATTA (Verbal) Anterior corpus callosum / corona radiata Anterior corona radiata Anterior corona radiata Pons

L/R

-27

36

-3

2417

< 0.001

6.04

R L L/R

27 -24 -3

24 18 -20

-12 17 -38

1178 783 510

0.002 0.025 < 0.001

4.36 4.79 4.75

R R L/R L//R L R L L

12 23 -3 9 -24 23 -8 -20

-6 -35 -51 -53 -66 -26 -29 -84

-5 29 -8 -42 17 -6 -20 -48

847 18461 1036 975 678 1016 736 1431

0.045 0.008 < 0.001 0.006 0.016 0.001 0.032 0.006

4.57 4.61 7.83 5.95 5.24 -5.36 -4.72 -5.68

L R L/R L R

-15 36 8 -24 45

-51 -53 -53 -62 -32

-57 -3 -5 -3 -15

1944 503 537 584 655

0.014 0.019 < 0.001 < 0.001 0.004

4.41 5.48 5.23 6.88 -5.12

ATTA (Visual) capsule Superior longitudinal f. Cerebellum (anterior l.) Cerebellum (lobule X) Posterior thalamic radiation Cerebral peduncle Cerebral peduncle Cerebellum (lobule VII/VIII) CBI Cerebellum (lobule X) Inferior longitudinal f. Cerebellum (anterior l.) Inferior longitudinal f. Saggital stratum

Cerebellum (lobule VII)

R

41

-63

-39

584

0.008

-5.50

L R L/R

-32 6 6

-78 -21 -35

14 14 -39

537 2383 381

0.002 0.014 < 0.001

5.85 -4.24 -5.74

CAQ Forceps major Thalamus Cerebellar peduncles / pons

Note: coordinates are displayed in MNI space. Names of negative predictors are in italics.