Cerebral Cortex, July 2017;27: 3542–3552

doi: 10.1093/cercor/bhw171
Advance Access Publication Date: 15 June 2016
Original Article

ORIGINAL ARTICLE

Changes in Brain Activation Associated with

Spontaneous Improvization and Figural Creativity

Downloaded from https://academic.oup.com/cercor/article/27/7/3542/3056437 by guest on 04 May 2024

After Design-Thinking-Based Training: A Longitudinal
fMRI Study
Manish Saggar1, Eve-Marie Quintin1, Nicholas T. Bott1,2, Eliza Kienitz1,2,
Yin-hsuan Chien1,3, Daniel W-C. Hong4, Ning Liu1, Adam Royalty5,
Grace Hawthorne5 and Allan L. Reiss1,6
1
Center for Interdisciplinary Brain Sciences Research, Department of Psychiatry and Behavioral Sciences,
Stanford University School of Medicine, 401 Quarry Road, Stanford, CA 94305, USA, 2Pacific Graduate School of
Psychology-Stanford University Psy.D. Consortium, 1791 Arastradero Road, Palo Alto, CA 94304, USA,
3
Department of Pediatrics, Taipei City Hospital Zhong-Xing Branch, No. 145 Zhengzhou Rd, Datong area, Taipei,
Taiwan, 4Institute of Biomedical Engineering, National Taiwan University, No. 1, Sec. 4, Roosevelt Road, Taipei
10617, Taiwan, 5Hasso Plattner Institute of Design, Stanford University, Building 550, 416 Escondido Mall,
Stanford, CA 94305, USA and 6Department of Radiology, Stanford University School of Medicine, 300 Pasteur Road,
Stanford, CA 94305, USA
Address correspondence to Manish Saggar, PhD, Center for Interdisciplinary Brain Sciences Research, Department of Psychiatry and Behavioral Sciences,
Stanford University School of Medicine, 401 Quarry Road, Stanford, CA 94305, USA. Email: [email protected]

Abstract
Creativity is widely recognized as an essential skill for entrepreneurial success and adaptation to daily-life demands. However,
we know little about the neural changes associated with creative capacity enhancement. For the first time, using a prospective,
randomized control design, we examined longitudinal changes in brain activity associated with participating in a five-week
design-thinking-based Creative Capacity Building Program (CCBP), when compared with Language Capacity Building Program
(LCBP). Creativity, an elusive and multifaceted construct, is loosely defined as an ability to produce useful/appropriate and novel
outcomes. Here, we focus on one of the facets of creative thinking—spontaneous improvization. Participants were assessed pre-
and post-intervention for spontaneous improvization skills using a game-like figural Pictionary-based fMRI task. Whole-brain
group-by-time interaction revealed reduced task-related activity in CCBP participants (compared with LCBP participants) after
training in the right dorsolateral prefrontal cortex, anterior/paracingulate gyrus, supplementary motor area, and parietal
regions. Further, greater cerebellar–cerebral connectivity was observed in CCBP participants at post-intervention when
compared with LCBP participants. In sum, our results suggest that improvization-based creative capacity enhancement is
associated with reduced engagement of executive functioning regions and increased involvement of spontaneous implicit
processing.

Key words: creativity, fMRI, functional connectivity, intervention, randomized control trial

© The Author 2016. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: [email protected]
 Neural Correlates of Enhanced Creativity Saggar et al. | 3543

Introduction inferior parietal cortex and the left middle temporal gyrus. In an-
other study, Cousijn et al. examined changes in resting-state
Creativity is considered the driving force behind all human
functional connectivity (RSFC) associated with a short 2-week
progress—in science, business, politics, arts and even interper-
at-home computerized divergent-thinking-based training pro-
sonal relationships. It is no surprise that creativity is now widely
gram (comprised of 8 20-min sessions) in adolescents (Cousijn
recognized as an essential skill for personal and entrepreneurial
et al. 2014). Although no training-related changes in RSFC were
success (Stavridou and Furnham 1996; Amabile 1997; Kern 2010),
observed, Cousijn et al. reported that performance over time
successful adaptation to daily-life demands (Csikszentmihalyi
was predicted by connectivity between left supramarginal gyrus
1997; Reiter-Palmon et al. 1998; Sternberg and O’Hara 1999;
and right occipital cortex at baseline.
Carson et al. 2003), psychological well-being (Cropley 1990; Solso
Taken together, these studies provide first steps towards un-
2001; Runco 2004; Gobel et al. 2011) and resilience (Runco 1991; derstanding changes in brain activity/connectivity associated
Metzl 2009). Given the importance of this cognitive faculty, it with creative capacity enhancement. However, several knowledge
is vital to examine the neural correlates of creative capacity en- gaps need to be addressed to further advance our understanding
hancement so that effective augmentation strategies, informed of this area. For example, both the Fink et al. and Cousijn et al.
by cognitive neuroscience, can be developed.

Downloaded from https://academic.oup.com/cercor/article/27/7/3542/3056437 by guest on 04 May 2024

studies used short, at-home computerized training programs to
Although the lay definition of creativity can be simply stated enhance creativity in individuals. As such, the results from these
as the ability to create novel and useful outcomes, the construct studies may not be germane to understanding the neural basis
of creativity is elusive and complex (Abraham 2014). Creativity is of creativity occurring in group-settings or involving “hands-on”
not a unitary construct and researchers agree that a multifaceted activities. Thus, it can be argued that brain activity changes asso-
approach is required to fully understand creativity. Thus, creativ- ciated with training programs such as those employed in the study
ity has been a subject of intense research across disciplines. To described here could be more relevant to naturally occurring edu-
evaluate creative potential in individuals, researchers have stud- cational and vocational settings. Further, the training programs as
ied novelty and appropriateness of products or outcomes gener- well as assessments of creative capacity enhancement in both
ated by established artists and writers (Colangelo et al. 1992; these studies were specifically limited to the divergent thinking
Carson et al. 2005). Prior work has also focused on understanding component of creativity. Thus, it is not clear whether the knowl-
how individual differences in personality and outlook affect cre- edge or expertise gained in such focused training programs
ativity (Csikszentmihalyi 1997; Wolfradt and Pretz 2001). Similar- would apply broadly to other domains of creativity such as impro-
ly, different environmental factors (e.g., upbringing or work vization, imagination, etc. Lastly, to tease apart the longitudinal
culture) can affect creative capacity in individuals (Runco 2004). changes in brain activity associated with creative capacity en-
From the perspective of brain mechanisms underlying creative hancement from those associated with training regimen itself
thinking, neuroscientific research has largely focused on studying (i.e., cognitive engagement and motivation), a parallel active con-
individual components of creative thinking (Dietrich and Kanso trol-training program is required.
2010; Kozbelt et al. 2010; Abraham 2014). For example, investiga- To address some of these gaps, we conducted a 5-week de-
tors have used innovative experimental designs to study insight sign-thinking-based Creative Capacity Building Program (CCBP)
(or Aha! Moments) (Jung-Beeman et al. 2004; Aziz-Zadeh et al. and a parallel, control training program (Language Capacity
2009), conceptual expansion (Abraham et al. 2012), mental im- Building Program [LCBP]) with healthy adults. Participants were
agery (LeBoutillier and Marks 2003), and spontaneous improviza- randomly assigned to either training program after initial behav-
tion (Limb and Braun 2008; Saggar et al. 2015). ioral and neuroimaging assessments (Fig. 1). The CCBP was de-
Given the fundamental importance of creativity in everyday signed as an interactive group studio where students learned to
life, as well as the fact that creativity can decline during early improve their improvizational skills by increasing their bias to-
childhood (Torrance 1968; Kim 2011), a wide array of experimen- wards taking action through hands-on experiences, rapid proto-
tal studies have quantitatively examined and shown that creativ- typing, and fail-faster exercises (see Materials and Methods;
ity can be enhanced at the individual and team level (see Scott Hawthorne et al. 2013). The LCBP was designed as active control
et al. 2004 for review). However, it can be argued that describing training where participants learned basic Chinese (Mandarin) vo-
the neural correlates of creativity and its enhancement are es- cabulary, character writing, and commonly used phrases via
sential for developing more neuroscience-informed approaches hands-on exercises. The LCBP was also performed in a group set-
for fostering and sustaining creativity across the lifespan. For ex- ting to simulate the shared environment atmosphere of the CCBP
ample, in future, neuroscience-informed training approaches studio. We endeavored to keep CCBP and LCBP trainings identical
could help us understand how brain development affects creative in terms of hands-on exercises, stimulating group environment
thinking, so that efficient interventions can be developed to re- and instructor’s motivation, while minimizing creative idea gen-
duce the impact of the so-called creative “slump” during middle eration and consolidation during the LCBP training (Hawthorne
childhood (Torrance 1968; Claxton et al. 2005). Additionally, re- et al. 2013). The overall study design incorporated a crossover
searchers have also argued that better understanding of the component such that participants enrolled in the LCBP later
neuroscience of creativity and including those insights into train- received the CCBP training and vice versa (see Materials and
ing is crucial for enhancing creative confidence in individuals Methods). However, the focus of current paper is limited to com-
(Onarheim and Friis-Olivarius 2013). However, very little is known parison of the initial parallel-group phase (i.e., before crossover).
about the neural correlates of creative capacity enhancement. Using a standardized figural creative thinking assessment
Recently, Fink et al. employed functional Magnetic Resonance (i.e., the Torrance Test of Creative Thinking [TTCT-Figural;
Imaging (fMRI) to examine longitudinal task-related changes in Torrance 1998]), we recently showed that participation in the
brain activity associated with a 3-week (20 min a day) at-home CCBP, when compared with LCBP, led to increased creative cap-
computerized verbal divergent thinking training program (Fink acity at post-intervention (Kienitz et al. 2014). For the TTCT-F,
et al. 2015). Whole-brain voxel-wise analysis revealed longitudin- participants are given a set of incomplete figures and are asked
al changes in brain activity during a divergent thinking task to complete them in a set amount of time such that the final fig-
(when compared with a control task) in regions within the left ure should portray a unique and interesting story (Torrance 1998).
 3544 | Cerebral Cortex, 2017, Vol. 27, No. 7

participation in the CCBP led to increased productivity of auto-

matic or implicit processes.
As creativity is not a unitary construct, examining the neural
correlates of creative thinking or its enhancement is a daunting
task. Thus, researchers are increasingly adopting a “process-
driven” or component-based approach for examining neural cor-
relates of creative thinking (Dietrich and Kanso 2010; Kozbelt
et al. 2010; Abraham 2014). In this paper, to examine the longitu-
dinal changes in brain activity associated with CCBP when com-
pared with LCBP, we focused specifically on the spontaneous
improvization component of creative thinking. Though just a com-
ponent of creative thinking, improvization can also encapsulate
prototypical creative behavior as a whole (Bengtsson et al.
2007). For example, during a musical improvization, a musician

Downloaded from https://academic.oup.com/cercor/article/27/7/3542/3056437 by guest on 04 May 2024

can freely generate behavioral choices, adapt these choices to
the ongoing performance, monitor outcomes through auditory
and somatosensory feedback, and preserve the overall esthetic
goal (Pressing 1988). Consistent with this view, several studies
have employed free-form spontaneous improvization in musi-
cians to examine the neural correlates of creative thinking
(Bengtsson et al. 2007; Limb and Braun 2008; Pinho et al. 2014).
In previous work, we investigated the neural correlates of
drawing-based spontaneous improvization in the same nonartist/
musician cohort described here before intervention. To engage par-
ticipants in drawing-based improvization, we developed a novel
fMRI task based on the social game of Pictionary™ (Saggar et al.
2015). In this task, using an MR-safe drawing tablet, participants
were asked to improvize and draw representations of a given verb
in 30 s with the caveat that others would later guess the word by
their drawing alone. In line with musical improvization, during
Figure 1. Randomized control design to examine the neural correlates of creative this Pictionary task, participants had to generate free choices to re-
capacity enhancement. present the given verb, adapt and monitor their sketching, while
preserving the overall goal that the given verb should later be easily
guessed by their drawing alone. To reduce anxiety and perform-
Out of the 5 subscales on which the final figures are (blindly) ance pressure, no explicit instructions were given to the partici-
scored, 2 subscales (i.e., elaboration and resistance to closure) pants to be “creative” during the task. However, the drawings
showed significant enhancement in the CCBP when compared were later scored for creative content and subjective ease of gues-
with the LCBP participants after training. As the elaboration sing by two creativity experts from Stanford’s Design School (A.R.
score on the TTCT-F captures details and imagination of re- and G.H.). The neuroimaging results showed that higher expert-
sponses to stimuli (Torrance 1998), enhanced elaboration scores rated creative content in the drawings was associated with
suggest that CCPB participants generated detailed and more im- increased engagement of bilateral cerebellum, while task-related
aginative responses than participants in the LCBP. Furthermore, activation in the cingulate and dorsolateral prefrontal cortices
the resistance to premature closure on the TTCT-F is associated negatively influenced task performance (i.e., ease of guessing the
with the tendency to persist with a creative undertaking in the word) (Saggar et al. 2015). These results suggested a putative nega-
service of finding novel outcomes as opposed to stopping at tive role of conscious monitoring and volitional control and a po-
more routine or predictable solutions (Torrance 1998). Thus, an tentially positive role of implicit processing via cerebral-cerebellar
increase in resistance to closure in CCBP participants after train- interaction during spontaneous improvization (Saggar et al. 2015).
ing suggests that they demonstrated greater persistence in arriv- Extending this previous work, for the first time, here we report
ing at novel outcomes. longitudinal changes in brain activity associated with a “live”
As part of the overall study, we also examined changes in ex- group-based and domain-general creativity-training program
ecutive functioning associated with participation in the CCBP (i.e., CCBP) when compared with an active control-training
when compared with LCBP training (Bott et al. 2014). The Color- program (i.e., LCBP). The Pictionary-based fMRI task described
Word Interference Test (CWIT) was used as our primary out- above was used to assess the behavioral correlates of spontan-
come measure (Delis 2001). The CWIT is based on the Stroop eous improvization at both pre- and post-intervention. Based on
measure (Stroop 1935) and it consists of four conditions. The the nature of the CCBP training (see Materials and Methods) as
first two conditions (color-naming and word-reading) assess pro- well as previous findings pertaining to the neural correlates of
cessing speed, whereas the last two assess “higher-level” inhib- improvization (Limb and Braun 2008; Pinho et al. 2014; Saggar
ition and cognitive flexibility. Interestingly, while no significant et al. 2015), we predicted that enhanced creativity in CCBP relative
group by time interaction was observed for inhibition and cogni- to LCBP participants would be associated with increased engage-
tive flexibility scores, significant improvements in processing ment of implicit cognitive processes and reduced engagement of
speed was observed in CCBP participants when compared with conscious monitoring and volitional control processes. Specific-
LCBP participants after training (Bott et al. 2014). As the color- ally, at post-intervention, we predicted increased cerebellar–
naming and word-reading conditions of CWIT are highly auto- cerebral interaction and reduced prefrontal activation in CCBP
matic and prepotent processes in adults, we argued that participants relative to LCBP participants.
 Neural Correlates of Enhanced Creativity Saggar et al. | 3545

Materials and Methods the given 30 s in each block and continue to add elements to
the illustration in case they wanted to finish early. Each block
Participants and Study Design was separated by a fixation period with a random duration within
Thirty-six healthy adults (18M, 18F) were enrolled in the study. the range of 10–15 s. There were a total of 10 blocks per condition
Half of the participants were randomly assigned to either a and the total duration of task was approximately 14.5 min. For
5-week CCBP or a parallel 5-week LCBP. Participants were as- more details, see Supplementary Methods.
sessed before and after intervention. See Figure 1 for a visual re-
presentation of the study design. Additional details regarding
Behavioral Assessments
assessments are provided in Supplementary Methods.
Two task-related behavioral assessments were employed in this
study: 1) expert-rated representation and creativity ratings for
Interventions each drawing (from the word-drawing condition) and 2) self-
Details of the CCBP and LCBP interventions have been described rated subjective rating of difficulty (as difficult, medium, or
elsewhere (Hawthorne et al. 2013; Bott et al. 2014; Kienitz et al. easy) in drawing each word by the participants during postscan

Downloaded from https://academic.oup.com/cercor/article/27/7/3542/3056437 by guest on 04 May 2024

2014). To summarize, both interventions included 5 2-h group questionnaire. Two raters from the Stanford Design School
classes conducted weekly. The CCBP was an abbreviated version (authors A.R. and G.H.) were chosen to blindly rate each drawing
of a highly popular class offered at the Stanford’s Hasso Plattner (from the word-drawing condition) on the scales of representa-
Institute of Design called “Creative Gym” (http://dschool.stanford. tion and creativity. The instructions for the “representation”
edu/classes/#creative-gym-a-design-thinking-skills-studio), taught scale were as follows: “how easily do you think another person
by the author G.H. Creative Gym course immerses students in a can guess the word represented by the drawing”. The ratings
learning environment based in experimentation with an overall were obtained on a 5-point scale (1–5).
goal to enhance improvization and creative skills. To improvize The creativity ratings of each drawing were assessed based on
novel solutions, students rapidly cycle through a series of phases: the 3 subscales of—fluency, elaboration, and originality. These
observe, brainstorm, synthesize, prototype, and implement; subscales were chosen based on the assessment scheme used
repeating as necessary. The class includes many hands-on in the standardized test of creativity (TTCT-F; Torrance 1998).
exercises based on the following elements: 1) “see” (reducing per- Scores from these three subscales were averaged to get the
ceptual bias to identify multiple perspectives to a give problem/ final score of creativity. Each subscale was defined as follows:
issue); 2) “start/build” (inducing bias toward action by rapidly 1) Fluency—total number of elements/ideas in the drawing;
prototyping using limited everyday materials); 3) “feel” (enhan- 2) Elaboration—imagination and exposition of detail; and 3) Ori-
cing perspective taking and empathy); 4) “communicate/inspire” ginality—the statistical infrequency and unusualness/uniqueness
(seeking active inspiration from everyday incidents and of the response. The rating for each subscale was also done on a
situations) and 5) “synthesize/navigate” (combining disparate 5-point scale (1–5). The 2 raters were trained on a small sample
constructs to transform independent sets of ideas into new direc- of drawings (36 drawings) and their inter-rater reliability index
tions). Taken together, the underlying goal of CCBP training was for all the drawings (as measured by Intra-class Correlation Coef-
to make participants repeatedly exercise their improvizational ficient [ICC]) were 0.80 for representation and 0.884 for the creativ-
skills in a group setting, thereby allowing them to improvize cre- ity scale.
ative outcomes, even in situations where resources were limited Although participants were highly encouraged to attend all 5
or the results were uncertain. sessions (i.e., 10 h), a few participants still missed sessions partly
We endeavored to keep CCBP and LCBP conditions comparable or fully due to unanticipated family or work events. Thus, we re-
in terms of hands-on exercises, stimulating group environment corded attendance based on the number of hours participants
and instructor’s motivation, while minimizing improvization were present in these sessions. Further, for each 2-h session
and creative thinking during LCBP training. As an active control, missed, participants were given 1-h worth of make-up home-
the LCBP was of same duration as CCBP and was conducted in work to be completed before the next assessment. Overall, within
parallel (at the same time) in a group setting and in a similar inter- the CCBP, 1 participant received 8 h of training, 6 received 9 h of
active studio as CCBP. The LCBP classes consisted of many hands- training, and the rest received full 10 h of training.
on exercises to learn basic Chinese vocabulary, character writing
(Calligraphy), commonly used phrases (e.g., How are you?) and
MRI Image Acquisition and Data Analysis
language learning group games. The author N.L., who is a bilingual
native Mandarin speaker, was the LCBP instructor. Participants were imaged on a 3Tesla scanner (GE MR750, Mil-
waukee, WI, USA) at Stanford University’s Center for Cognitive
and Neurobiological Imaging (CNI) using a 32-channel radiofre-
fMRI Task Design
quency receive head coil (Nova Medical, Inc., Wilmington, MA,
Details of this task are presented elsewhere (Saggar et al. 2015). USA). A total of 435 whole-brain volumes were collected on 42
Briefly, the word-guessing fMRI task is based on a block-design axial-oblique slices (2.9 mm thick) prescribed parallel to the in-
paradigm, with 30 s block duration for each of the two conditions tercommissural (AC-PC) line, using a T2*-weighted gradient
(word-drawing and zigzag-drawing). In the first condition, word- echo pulse sequence sensitive to blood oxygen level-dependence
drawing, participants were asked to draw a given word (mainly (BOLD) contrast with the following acquisition parameters: echo
actions/verbs) to the best of their ability using the MR-safe draw- time (TE) = 30 ms, repetition time (TR) = 2000 ms, flip angle = 77°,
ing tablet, with the caveat that others would later try to guess the FOV = 23.2 cm, acquisition matrix = 80 × 80, approximate voxel
word by their drawing alone (see Supplementary Fig. 2 for a set of size = 2.9 × 2.9 × 2.9 mm. A high-resolution T1-weighted 3D BRAVO
sample drawings). To control for the basic motor and visuospatial pulse sequence acquisition was acquired for co-registration with
aspect during the word-drawing condition, participants were the following parameters: echo time (TE) = 2.8 ms, repetition time
also asked to make a drawing representing a control word (zigzag) (TR) = 7.2 ms, flip angle = 12°, FOV = 23 cm, slice thickness = 0.9
in the second condition. Participants were asked to fully utilize mm, 190 slices in the sagittal plane; matrix = 256 × 256; acquired
 3546 | Cerebral Cortex, 2017, Vol. 27, No. 7

resolution = 0.9 × 0.9 × 0.9 mm. The images were reconstructed as size for between-group longitudinal differences in creativity rat-
a 256 × 256 × 190 matrix. ings. We observed a medium effect size (Cohen’s d = 0.44) for be-
Functional MRI data processing was carried out using FEAT tween-group longitudinal differences in creativity ratings,
(FMRI Expert Analysis Tool) Version 6.00, part of FSL (FMRIB’s suggesting a tendency towards larger increase in creativity rat-
Software Library, www.fmrib.ox.ac.uk/fsl). Details regarding the ings in CCBP when compared with LCBP. Table 1 reports descrip-
data preprocessing and functional analysis are provided in tive statistics for pre- and post-intervention scores for each rating.
Supplementary Methods.
Task-Related Correlates of Changes in Brain Activity
Results Associated with CCBP
To reveal task-related changes in brain activity associated with
Task-Related Behavioral Performance
CCBP when compared with LCBP, group (CCBP vs. LCBP) × time
To examine the intervention-related changes in behavioral per- ( pre- vs. post-intervention) interaction was performed for the
formance, we ran repeated measures MANOVA on the factors of primary task contrast (i.e., word- vs. zigzag-drawing). Five signifi-
time ( pre- and post-intervention) and rating scores (expert cre- cant clusters were observed encompassing the supplementary

Downloaded from https://academic.oup.com/cercor/article/27/7/3542/3056437 by guest on 04 May 2024

ativity score, expert representation score and subjective difficulty motor area (SMA), anterior cingulate cortex (ACC), paracingulate
ratings) between the two groups (CCBP and LCBP). Only a trend to- gyrus, right dorsolateral prefrontal cortex (DLPFC), bilateral lat-
wards significance was observed for the main interaction of time eral occipital complexes (LOC) and parietal lobules, left tem-
x rating scores x group (F2,27 = 2.945, P = 0.07, η 2 = 0.18). As an ex- poral-occipital fusiform, lingual gyrus, and cerebellum (Fig. 2A
ploratory post hoc analysis, we separately examined the effect and Supplementary Table 1). The same pattern of direction was

Table 1 Task-related behavioral performance

CCBP LCBP

Pre-intervention Post-intervention Pre-intervention Post-intervention

Mean expert creativity ratings (scale of 1–5) 2.66 (SD = 0.30) 2.78 (SD = 0.28) 2.70 (SD = 0.21) 2.74 (SD = 0.26)
Mean expert representation ratings (scale of 1–5) 3.43 (SD = 0.72) 3.25 (SD = 0.78) 3.69 (SD = 0.77) 3.24 (SD = 0.91)
Mean subjective difficulty in drawing each word (scale of 1−3) 1.83 (SD = 0.23) 1.78 (SD = 0.24) 1.84 (SD = 0.26) 1.95 (SD = 0.18)

Figure 2. Neural correlates of creative capacity enhancement. (A) Shows the clusters found for the group by time interaction. (B) Average time course for a representative
cluster (R. dorsolateral prefrontal) before and after training. Similar pattern of longitudinal change in activation, across groups, was observed in other 4 clusters. Band
around the mean time course represents SEM across participants. (C) Mean change in activation for each of the 5 clusters, bars represent SEM across participants. (D)
Observed relation between reduction in activity and hours of CCBP training (r(15) = −0.563, P = 0.029).
 Neural Correlates of Enhanced Creativity Saggar et al. | 3547

evident across all clusters, i.e., the CCBP participants had reduced we also estimated the effect size for between-group longitudinal
activity after training relative to baseline, whereas LCBP partici- differences in cerebellar–cerebral connectivity. A medium-size
pants had increased activity after training (Fig. 2B,C). effect was observed (Cohen’s d = 0.52), suggesting a tendency
Although we asked participants to attend all sessions, due to toward larger increase in connectivity after participating in the
unavoidable circumstances, a few participants received less than CCBP when compared with LCBP.
10 h of training (see Materials and Methods). This discrepancy,
however, provided an opportunity to assess whether longitudin-
al changes in neural activity were related to the amount of CCBP Discussion
training received. Significant negative association was observed Using a randomized control study design, we previously showed
between training hours and reduced activity in the left occipi- that participation in CCBP relative to LCBP was associated with: 1)
tal/parietal regions [r(15) = −0.563, P = 0.029], suggesting that the increased creative capacity as measured by the standardized test
“dose” of CCBP training affected task-related reductions in of figural creativity (TTCT-F; Kienitz et al. 2014) and 2) increased
brain activity at post-training (Fig. 2D). Negative associations automatic or implicit processing as measured by standardized
were also observed between training hours and reduced activity neuropsychological testing (Bott et al. 2014). In this paper, using

Downloaded from https://academic.oup.com/cercor/article/27/7/3542/3056437 by guest on 04 May 2024

in the other four clusters but did not reach significance. the same cohort from whom these behavioral data were obtained, we
examined longitudinal changes in task-related brain activity/
connectivity associated with improvization-based design-think-
Changes in Cerebellar-Cerebral Connectivity with
ing training to enhance creativity when compared with an active
Training
control language-based training. Participants were assessed pre-
To examine longitudinal changes in cerebellar–cerebral connect- and post-intervention for spontaneous improvization while per-
ivity associated with participation in the CCBP when compared forming the word-drawing social game of Pictionary during an
with LCBP, longitudinal between-group PPI analyses were per- fMRI scan. As predicted, at post-intervention, widespread task-
formed. Similar to examining neural correlates of creative related reduction in prefrontal and parietal brain activity was
capacity enhancement, we first calculated within-participant dif- observed in the CCBP participants when compared with LCBP
ferences in cerebellar–cerebral connectivity across time (i.e., pre- participants. Furthermore, the amount of decrease in parietal ac-
vs. post-intervention), followed by comparing these longitudinal tivation was associated with the quantity of training received
differences between groups (CCBP vs. LCBP). No significant group (in hours), suggesting a “dose–response” relationship from
x time interaction was observed for the PPI analysis. How- CCBP training. Greater cerebellar–cerebral connectivity was also
ever, at post-intervention, differences in the cerebellar–cerebral observed in the CCBP group post-intervention relative to LCBP.
connectivity between groups were observed to be significant However, despite the fact that the two groups did not show sig-
(Fig. 3 and Supplementary Table 2). Specifically, increased nificant differences at baseline, the group by time interaction
post-intervention cerebellar–cerebral connectivity was observed for cerebellar–cerebral connectivity did not reach significance.
in the CCBP participants when compared with LCBP participants The lack of a significant group by time interaction, despite a me-
for the right ACC, bilateral paracingulate gyrus, right inferior dium estimated effect size, may be related to the sample sizes
frontal gyrus, bilateral DLPFC, bilateral frontal pole, and bilateral used in this first of its kind study to examine the neural correlates
superior frontal gyrus. A similar between-group analysis at of creative capacity enhancement. In sum, our data suggest
pre-intervention revealed no significant difference between that reduced engagement of executive functioning and volitional
the groups. Lastly, to visualize change in cerebellar–cerebral control, and increased involvement of implicit processing
connectivity, beta estimates were extracted from pre- and (via cerebellar–cerebral connectivity) might be associated with
post-intervention data (separately for each group) using the sig- improvization-based creativity training.
nificant group-difference cluster observed at post-intervention. A significant group by time interaction revealed widespread
Qualitatively, an increase in cerebellar–cerebral connectivity reduction of task-related activity in the CCBP participants when
was observed in CCBP participants, while a reduction was compared with LCBP participants at post-intervention. This
observed for LCBP participants (Fig. 3). Using the significant training-related drop in activity was observed in several brain re-
group-difference cluster observed at post-intervention as a mask, gions including the ACC/DLPFC, SMA, paracingulate gyrus, LOC

Figure 3. Differences in cerebellar–cerebral (PPI) connectivity between groups (CCBP > LCBP) at post-intervention. No significant group differences in PPI connectivity were
observed at pre-intervention. (A) Shows the cluster with significant PPI differences between-group (CCBP > LCBP). (B) Visualization of the change in cerebellar–cerebral
connectivity over time (i.e., before and after training). The cluster found at post-intervention (see A) was used to extract values at pre-intervention.
 3548 | Cerebral Cortex, 2017, Vol. 27, No. 7

and parietal lobules, left temporal-occipital fusiform, lingual overall intelligence might serve as a proxy (Sternberg and O’Hara
gyrus, and cerebellum. The ACC and DLPFC regions are known 1999; Carson et al. 2003). Building upon these previous results and
to preferentially activate during effortful monitoring, evaluation the fact that our participant sample had high intelligence scores
and selection of relevant responses (MacDonald et al. 2000; (∼120; Table 2), reduced activity in the top-down cognitive control
Koechlin et al. 2003). Reduced activity in these regions after train- regions after CCBP training (when compared with LCBP training)
ing could be explained by two plausible mechanisms. First, it is might have enhanced improvization and creative capacity by
possible that CCBP training led to efficient utilization of cognitive facilitating diverse associations between ideas.
control regions and thus, after design-thinking-based training, It is important to question, however, whether creativity re-
participants could exert such control effortlessly, that is, with quires evaluation or selection of ideas at all. Some might argue
less neural resources. Previous research using various cognitive that to choose an original or unusual response, one has to evalu-
tasks indicated that “efficient” utilization of neural resources ate and reject other less novel responses (Runco 1991). Thus,
occur with other kinds of training (a.k.a. expertise or practice ef- evaluation seems to be a necessary component in creative think-
fect) (Gobel et al. 2011). Alternatively, it is also possible that after ing. In a recent study, Beaty et al. (2015) showed engagement of
CCBP training, participants focused less on monitoring, evaluat- large-scale brain networks of executive functioning, in addition

Downloaded from https://academic.oup.com/cercor/article/27/7/3542/3056437 by guest on 04 May 2024

ing or selecting ideas and more on generating and synthesizing to spontaneous thought processes, during a divergent thinking
ideas. We argue that the second interpretation is more likely, be- task (i.e., to find alternate uses of a given object). Several other
cause of the fact that CCBP was specifically designed to enhance studies have also reported engagement of executive function
improvization skills in participants by increasing their bias and/or cognitive control regions during creative thinking for
toward action. Thus, by regularly practicing improvization and similar divergent thinking tasks (Kowatari et al. 2009; Aziz-
rapid prototyping, participants in the CCBP training learned to Zadeh et al. 2013). Our finding of reduced engagement of execu-
increase engagement of implicit processing while reducing the tive functioning regions with improvization-based training and
involvement of volitional cognitive processing during improviza- associated creative capacity enhancement may sound contradic-
tion and creative thinking. Further, in our previous work in the tory at first, but the observed differences could be mainly due to
same cohort, using only baseline ( pre-intervention) data, we the task design. Instead of using divergent thinking tasks where
showed that activation in the ACC/DLPFC regions was negatively participants are explicitly asked to produce an “alternate” or “un-
associated with representation ratings of the drawings (Fig. 2 in usual” answer to a given problem or question, we used spontan-
Saggar et al. 2015). Thus, a drop of task-related brain activity in eous improvization as a means to examine the neural correlates
these regions after CCBP training also suggests that participants of implicit creative thinking. We argued that implicit paradigms
may have focused less on selecting or inhibiting ideas and more could play an essential role in reducing variation in creativity
on generating and synthesizing ideas during the word-drawing neuroimaging results by minimizing confounding influences of
condition. cognitive processes that might not be directly related to creative
Other studies have also indicated that greater creativity is thinking, but are employed due to task design (Saggar et al. 2015).
associated with reduced cognitive inhibition (Stavridou and Thus, it is possible that different components of creativity (e.g.,
Furnham 1996; Carson et al. 2003; Takeuchi et al. 2011). For ex- improvization vs. divergent thinking) may exploit different cog-
ample, using a latent-inhibition paradigm, Carson et al. (2003) nitive resources. Future research is required to systematically
showed that highly creative individuals have reduced early select- compare longitudinal changes in brain activity associated with
ive attention processing, thereby providing access to a larger creativity trainings that are focused on different components of
inventory of unfiltered stimuli and increasing opportunities for creativity.
synthesis of original (or unusual) ideas. Similarly, Stavridou and Our finding of reduced engagement of executive functioning
Furnham (1996) used a negative-priming paradigm to suggest areas with creative capacity enhancement in nonartists/musi-
that reduced cognitive inhibition during selective attention is re- cians is in line with the previous literature in expert musicians
sponsible for “widening” of associative connections, thus allowing where improvization tasks were used to assess creativity. For
creative individuals to attend to many aspects of a given stimulus example, using a musical improvization task, Limb and Braun
and produce more diverse associations. To deal with the large (2008) have found that enhanced creativity in expert Jazz musi-
inventory of unfiltered stimuli, however, creative individuals cians was associated with reduced recruitment of executive func-
are thought to engage additional intellectual qualities, for which tioning regions. Similarly, more recent studies have shown

Table 2 Participant demographics

Number of participants Total CCBP LCBP t P

30 (16F; 8 students) 15 (8F; 4 students) 15 (8F; 4 students)

Age (years) Mean 28.77 27.67 29.87 −1.09 0.28

SD 5.54 5.15 5.87
I.Q. Mean 120.67 119 122.33 −0.86 0.40
SD 10.52 11.81 9.16
Gross income Mean 5.23 5.2 5.27 −0.08 0.94
SD 2.22 1.61 2.76
Creative achievement (CAQ) N 20.93 23.2 18.67 0.87 0.39
SD 14.14 13.27 15.08
TTCT-F average Mean 110.31 111.267 109.286 0.49 0.63
SD 10.7 9.25 12.33

The two groups had equal number of participants, females, and students. Groups did not differ at baseline in terms of age, intelligence, gross household income, previous
creative achievements and performance on the standardized test of creativity (TTCT-F).
 Neural Correlates of Enhanced Creativity Saggar et al. | 3549

deactivations in the executive functioning regions (especially the capacity enhancement. Using Ito’s model for the cerebellum’s
DLPFC) during musical improvization (Liu et al. 2012; Pinho et al. role in mental activities (Ito 2008), we can speculate that an in-
2014) and poetry generation (Liu et al. 2015) as a sign of reduced crease in functional connectivity between dorsolateral prefrontal
monitoring and volitional control. cortex and cerebellum would allow the cerebellum to emulate
Interestingly, although we observed a reduction in task- the “controller” function of prefrontal cortex using inverse in-
related activity in some cerebellar areas associated with CCBP ternal models and, in turn, facilitate efficient improvization of
training, we also observed greater functional connectivity be- word-drawings.
tween the dorsolateral prefrontal cortex and cerebellum in CCBP Previous neuroimaging work has shown activations in specif-
(when compared with LCBP) participants at post-intervention. ic cerebellar regions while participants learned to acquire input–
Cerebellar–cerebral functional connectivity has been previously output properties of controlled objects, including tools (see
hypothesized to facilitate implicit processing during creative Imamizu and Kawato 2012 for review). In the current work, parti-
thinking (Vandervert et al. 2007; Ito 2008). Especially when con- cipants used an MR-safe drawing tablet for sketching word
fronted with novel situations (as in the case of improvization), representation. Thus, it is possible that some of the longitudinal
prefrontal-cerebellar connectivity has been postulated to provide changes in cerebellar activation/connectivity could be due to
“rapid” manipulation of conceptual ideas for making a quick

Downloaded from https://academic.oup.com/cercor/article/27/7/3542/3056437 by guest on 04 May 2024

learning how to dexterously use the tablet. However, we argue
decision (Leiner et al. 1986). These hypotheses are based on that it is unlikely that the primary driver for observed longitudin-
the premise that the cerebellum might facilitate efficient mani- al changes in cerebellar activity/connectivity is tablet-related
pulation of mental representations as well as movements; the dexterity, due to following reasons. First, we endeavored to re-
widespread anatomical connectivity between cerebellum and pre- duce tablet-related effects on results by training the participants
frontal regions in humans supports this premise (Schmahmann on how to use the drawing tablet before they entered the MR
1991; Ito 1993, 2006; Vandervert et al. 2007; Buckner 2013). scanner, both at pre- and post-intervention assessments. Se-
To better understand the role of cerebellar–cerebral connect- cond, due to the study design, any longitudinal effects associated
ivity and implicit processing during creative thinking, Ito pro- with tablet-use should get canceled as both groups (CCBP and
posed a theoretical analog of how the cerebellum facilitates LCBP) employed the tablet during the fMRI task at pre- and
motor learning from the viewpoint of control (Ito 2008). To post-intervention. Third, and specific to cerebellar–cerebral
achieve speed, accuracy, and automaticity in executing motor connectivity results, if dexterity of tablet-use was driving longitu-
commands, researchers have proposed that the commands di- dinal changes in cerebellum, we should have observed changes
rected towards movement control regions in the cerebral cortex in connectivity between sensory/motor areas and cerebellum. In-
also get copied as “internal models” (forward or inverse) in the stead, we observed greater connectivity between cerebellum and
cerebellum for simulating natural body movements (Ito 2005). prefrontal regions, thereby suggesting that cerebellum was facili-
Through repeated and parallel simulations, the cerebellum facil- tating idea manipulation rather than movement manipulation.
itates acquisition of advanced motor skills via forward models Cerebro-cerebellar connectivity has also been previously hy-
and eventually provides automaticity via inverse models. In hu- pothesized to facilitate creative thinking and innovation by en-
mans, based on the extensive cerebellar–cerebral anatomical hancing working memory functioning (Vandervert 2003, 2015;
connectivity (Ito 1997), it can be speculated that this theoretical Vandervert et al. 2007). As per Baddeley’s well-known working
model of motor learning can be extended to higher order memory model, three components—central executive, phono-
cognitive functioning and thought processing (Ramnani 2006; logical loop, and visuospatial sketchpad—work together for suc-
Vandervert et al. 2007; Ito 2008; Koziol et al. 2014). Thus, analo- cessful implementation of working memory (Baddeley 2003).
gous to voluntary movement, one could speculate the forward Using the internal models of cerebellum (i.e., forward and inverse
models in cerebellum could facilitate faster acquisition and pro- modeling), the cerebro-cerebellar connectivity has been hy-
cessing of information (e.g., quick arithmetic calculation), where- pothesized to enhance information processing in each of the
as the inverse models in cerebellum could enable automaticity in three components of working memory; such increase in efficient
thinking and idea generation (e.g., “aha” moments) (Ito 1997). processing of working memory operations is in turn thought to
Although the postulated role of cerebellum in thought control increase the likelihood of creative solutions (Vandervert 2003;
mechanism by Ito et al. is appealing, admittedly we know very lit- Koziol et al. 2014). Recent neuroimaging findings support the
tle about how our brain stores and manipulates a thought (Ito premise that cerebro-cerebellar connectivity plays a role in effi-
2011, pp. 19–20). Empirical evidence, however, is beginning to cient processing of working memory operations (Marvel and
confirm the role of cerebellar–cerebral connectivity in thinking Desmond 2010, 2012; Sobczak-Edmans et al. 2016). Thus, from
in general and creative improvization in particular. In previous the perspective of working memory, it can be argued that our fig-
work, also using PPI analysis, de Manzano and Ullén (2012) de- ural task could have enlisted the visuospatial sketchpad for tem-
monstrated cerebellar–cerebral connectivity was higher during porary storage and manipulation of visuospatial information as
improvization when compared with other control conditions. well as for bottom-up planning of hand movements to draw the
Similarly, in a recent study, Pinho et al. (2014) also showed representation. Additionally, our task could have also employed
increased cerebral-cerebellar functional connectivity in expert the central executive component to facilitate coordination be-
musicians during improvization. In our previous work, using tween the low-level components and in selectively attending
the same participant-set and Pictionary-based fMRI task as and inhibiting ideas for drawing the word. The neuroimaging re-
presented in this paper, we showed that activation in the cerebel- sults for task-related activation, using just the pre-intervention
lum uniquely and linearly increased with increasing expert cre- data, supports the enlisting of visuospatial and central executive
ativity ratings at pre-intervention (Saggar et al. 2015). When components of working memory during the word-drawing task
considered in the light of previous theoretical and empirical find- (Saggar et al. 2015). The longitudinal analysis presented here
ings, our current result of higher cerebellar–cerebral connectivity shows greater cerebro-cerebellar connectivity between the dor-
associated with CCBP training (when compared with LCBP) solateral prefrontal regions and cerebellum associated with cre-
provides preliminary evidence for a potentially direct role of ativity training. As the dorsolateral prefrontal region has been
cerebellar–cerebral connectivity in improvization-based creative previously implicated to host the central executive component
 3550 | Cerebral Cortex, 2017, Vol. 27, No. 7

of working memory (Baddeley 2003), our results again suggest 98 participants may be required to observe significant differences
that creative capacity enhancement could be associated with for the group by time interaction of interest. Nonetheless, sig-
emulation of the prefrontal “controller” function by internal cere- nificant group differences in cerebellar–cerebral connectivity
bellar models. at post-intervention provide an interesting avenue for future
Although we mainly focused on examining the neural corre- research.
lates of creative capacity enhancement, a brief discussion is war- Lastly, although we endeavored to keep the CCBP and LCBP
ranted regarding the longitudinal changes associated with LCBP. trainings comparable in all aspects, except that of creativity, it
Opposite to the pattern observed in CCBP participants, the group is possible that some of the results presented in this study
× time interaction revealed an increase in task-related activity in could be associated with differences in instructor motivation
LCBP participants after training. Furthermore, the PPI analysis re- and teaching style.
vealed reduced cerebro-cerebellar task-related connectivity at Altogether, for the first time, we reveal the neural correlates of
post-intervention in LCBP (when compared with CCBP) partici- improvization-based creative capacity enhancement in adults
pants. Given that the LCBP included exercises on learning using a randomized control design. Our data suggest reduced
Mandarin language, which presumably requires higher mental engagement of cognitive monitoring and volitional control and

Downloaded from https://academic.oup.com/cercor/article/27/7/3542/3056437 by guest on 04 May 2024

control (Tu et al. 2015), it can be argued that the task-related in- putatively higher cerebellar–cerebral connectivity associated
crease in activity in the cognitive control regions and SMA might with improvization-based creativity training. We anticipate that
be due to increased evaluation and motor planning. Furthermore, these results can guide future efforts to develop and measure the
as the LCBP did not include improvization-based exercises, we efficacy of interventions to enhance improvization-based cre-
did not expect cerebro-cerebellar connectivity to be stronger at ativity in adults and children across lifespan.
post-intervention.
It is important to note that due to the nature of our fMRI task
(i.e., improvization and drawing) and our design-thinking-based
Supplementary Material
training paradigm, our results might not generalize to other com- Supplementary material can be found at: http://www.cercor.
ponents of the broadly defined construct of creativity. Future oxfordjournals.org/.
work is required to extend our creativity-conducive task and
training paradigms to other components of creative thinking. In
designing our fMRI task, we used zigzag-drawing as a control for
Funding
basic motor and visuospatial processing that is also required dur- This work was supported by a Hasso Plattner Design Thinking Re-
ing the word-drawing condition. However, it is possible that zig- search Program grant to A.L.R. Additional support was provided
zag-drawing might not provide an optimal control for the degree by the NIH K99-MH104605 grant to M.S.
of language processing required for the word-drawing condition.
Furthermore, it might be argued that the overall cognitive load re-
quired during the word-drawing condition was not matched in
Notes
the zigzag-drawing condition. The previous work has suggested We thank the staff at Stanford’s Design School, Center for Inter-
that such imbalance between creative and control conditions disciplinary Brain Sciences Research, and Center for Cognitive
can produce activation maps that are not necessarily specific to and Neurobiological Imaging for their support. Conflict of Interest:
creative thinking (Abraham et al. 2012). However, this potential None declared.
limitation of the control condition should not affect the be-
tween-group task-related results presented in this paper as
both groups participated in the study across both time points.
References
Furthermore, as a post hoc analysis, we also compared be- Abraham A. 2014. Creative thinking as orchestrated by semantic
tween-group longitudinal changes in brain activity associated processing vs. cognitive control brain networks. Front Hum
with just the word-drawing blocks (i.e., without the zigzag- Neurosci. 8:95.
drawing condition). Comparing just the word-drawing blocks Abraham A, Pieritz K, Thybusch K, Rutter B, Kröger S,
across time and groups resulted in similar reduction in task- Schweckendiek J, Stark R, Windmann S, Hermann C. 2012.
related activity in the cingulate and prefrontal regions in CCBP Creativity and the brain: uncovering the neural signature of
(when compared with LCBP) participants as observed with includ- conceptual expansion. Neuropsychologia. 50:1906–1917.
ing zigzag-drawing as a control condition (Supplementary Fig. 3). Amabile TM. 1997. Entrepreneurial creativity through motiv-
We did not observe a significant group by time interaction for ational synergy. J Creative Behavior. 31:18–26.
expert creativity ratings for the drawings made during the fMRI Aziz-Zadeh L, Kaplan JT, Iacoboni M. 2009. “Aha!”: The neural corre-
task. The lack of significance could be attributed to low statistical lates of verbal insight solutions. Hum Brain Mapp. 30:908–916.
power or the fact that our task was not standardized. However, Aziz-Zadeh L, Liew S-L, Dandekar F. 2013. Exploring the neural
using the same set of individuals, we previously showed that par- correlates of visual creativity. Soc Cogn Affect Neurosci.
ticipation in the CCBP led to increased standardized scores of cre- 8:475–480.
ativity (TTCT-F) after training compared with LCBP participation Baddeley A. 2003. Working memory: looking back and looking
(Kienitz et al. 2014). Additionally, we did not observe a significant forward. Nat Rev Neurosci. 4:829–839.
group by time interaction for the whole-brain PPI analysis, which Beaty RE, Benedek M, Barry Kaufman S, Silvia PJ. 2015. Default
was conducted to examine between-group longitudinal changes and executive network coupling supports creative idea pro-
in cerebellar–cerebral connectivity. PPI analyses tend to lack duction. Sci Rep. 5:10964.
power and hence are prone to higher false negatives (O’Reilly Bengtsson SL, Csíkszentmihályi M, Ullén F. 2007. Cortical regions
et al. 2012), as the psychophysiological interaction time series is involved in the generation of musical structures during im-
likely correlated with both the psychological and seed ROI time provisation in pianists. J Cogn Neurosci. 19:830–842.
series. Post hoc power analysis revealed that given α = 0.05 and Bott N, Quintin E-M, Saggar M, Kienitz E, Royalty A, Hong DW-C,
observed effect size of 0.52, future studies with a group size of liu N, Chien Y-H, Hawthorne G, Reiss AL. 2014. Creativity
 Neural Correlates of Enhanced Creativity Saggar et al. | 3551

training enhances goal-directed attention and information Kern F. 2010. What Chief Executives Really Want. Bloomberg
processing. Thinking Skills Creat. 13:120–128. Businessweek.
Buckner RL. 2013. The cerebellum and cognitive function: 25 Kienitz E, Quintin E-M, Saggar M, Bott NT, Royalty A, Hong DW-C,
years of insight from anatomy and neuroimaging. Neuron. liu N, Chien Y-H, Hawthorne G, Reiss AL. 2014. Targeted inter-
80:807–815. vention to increase creative capacity and performance: a rando-
Carson SH, Peterson JB, Higgins DM. 2003. Decreased latent inhib- mized controlled pilot study. Thinking Skills Creat. 13:57–66.
ition is associated with increased creative achievement in Kim KH. 2011. The creativity crisis: the decrease in creative think-
high-functioning individuals. J Pers Soc Psychol. 85:499–506. ing scores on the torrance tests of creative thinking. Creat Res
Carson SH, Peterson JB, Higgins DM. 2005. Reliability, validity, and J. 23:285–295.
factor structure of the creative achievement questionnaire. Koechlin E, Ody C, Kouneiher F. 2003. The architecture of cogni-
Creat Res J. 17:37–50. tive control in the human prefrontal cortex. Science.
Claxton AF, Pannells TC, Rhoads PA. 2005. Developmental trends 302:1181–1185.
in the creativity of school-age children. Creat Res J. 17:327–335. Kowatari Y, Lee SH, Yamamura H, Nagamori Y, Levy P, Yamane S,
Colangelo N, Kerr B, Hallowell K. 1992. The Iowa inventiveness in- Yamamoto M. 2009. Neural networks involved in artistic cre-

Downloaded from https://academic.oup.com/cercor/article/27/7/3542/3056437 by guest on 04 May 2024

ventory: toward a measure of mechanical inventiveness. ativity. Hum Brain Mapp. 30:1678–1690.
Creat Res. 5:157–163. Kozbelt A, Beghetto RA, Sternberg RJ. 2010. Theories of creativity.
Cousijn J, Zanolie K, Munsters RJM, Kleibeuker SW, Crone EA. In: Kaufman JC, Sternberg RJ, editors. The Cambridge
2014. The relation between resting state connectivity and cre- Handbook of Creativity. Cambridge. p. 20–47.
ativity in adolescents before and after training. PLoS ONE. 9: Koziol LF, Budding D, Andreasen N, D’Arrigo S, Bulgheroni S,
e105780. Imamizu H, Ito M, Manto M, Marvel C, Parker K, et al. 2014.
Cropley AJ. 1990. Creativity and mental health in everyday life. Consensus paper: the cerebellum’s role in movement and
Creat. Res. J. 3:167–178. cognition. In: Presented at the Cerebellum. London, England
Csikszentmihalyi M. 1997. Creativity: Flow and the Psychology of p. 151–177.
Discovery and invention. New York: HarperCollins Publications. LeBoutillier N, Marks DF. 2003. Mental imagery and creativity: a
Delis DC. 2001. Delis-Kaplan Executive Functioning System meta-analytic review study. Br J Psychol. 94:29–44.
(D-KEFS). San Antonio: The Psychological Corporation. Leiner HC, Leiner AL, Dow RS. 1986. Does the cerebellum contrib-
de Manzano Ö, Ullén F. 2012. Activation and connectivity ute to mental skills? Behav Neurosci. 100:443.
patterns of the presupplementary and dorsal premotor Limb CJ, Braun AR. 2008. Neural substrates of spontaneous mu-
areas during free improvisation of melodies and rhythms. sical performance: an FMRI study of jazz improvisation.
NeuroImage. 63:272–280. PLoS ONE. 3:e1679.
Dietrich A, Kanso R. 2010. A review of EEG, ERP, and neuroimaging Liu S, Chow HM, Xu Y, Erkkinen MG, Swett KE, Eagle MW, Rizik-
studies of creativity and insight. Psychol Bull. 136:822–848. Baer DA, Braun AR. 2012. Neural correlates of lyrical impro-
Fink A, Benedek M, Koschutnig K, Pirker E, Berger E, Meister S, visation: an FMRI study of freestyle rap. Sci Rep. 2:834.
Neubauer AC, Papousek I, Weiss EM. 2015. Training of verbal Liu S, Erkkinen MG, Healey ML, Xu Y, Swett KE, Chow HM,
creativity modulates brain activity in regions associated Braun AR. 2015. Brain activity and connectivity during poetry
with language- and memory-related demands. Hum Brain composition: Toward a multidimensional model of the cre-
Mapp. 36:4104–4115. ative process. Hum Brain Mapp. 36:3351–3372.
Gobel EW, Parrish TB, Reber PJ. 2011. Neural correlates of skill ac- MacDonald AW, Cohen JD, Stenger VA, Carter CS. 2000. Dissociat-
quisition: decreased cortical activity during a serial intercep- ing the role of the dorsolateral prefrontal and anterior cingu-
tion sequence learning task. NeuroImage. 58:1150–1157. late cortex in cognitive control. Science. 288:1835–1838.
Hawthorne G, Quintin E-M, Saggar M, Bott N, Keinitz E, liu N, Marvel CL, Desmond JE. 2010. The contributions of cerebro-cere-
Chien Y-H, Hong D, Royalty A, Reiss AL. 2013. Impact and bellar circuitry to executive verbal working memory. Cortex.
Sustainability of Creative Capacity Building: The Cognitive, 46:880–895.
Behavioral, and Neural Correlates of Increasing Creative Marvel CL, Desmond JE. 2012. From storage to manipulation:
Capacity. Cham: Springer International Publishing. p. 65–77. How the neural correlates of verbal working memory reflect
Imamizu H, Kawato M. 2012. Cerebellar internal models: implica- varying demands on inner speech. Brain and Language.
tions for the dexterous use of tools. Cerebellum. 11:325–335. 120:42–51.
Ito M. 2005. Bases and implications of learning in the cerebellum– Metzl ES. 2009. The role of creative thinking in resilience after
adaptive control and internal model mechanism. Prog Brain hurricane Katrina. Psychol. Aesthet Creat. Arts. 3:112–123.
Res. 148:95–109. Onarheim B, Friis-Olivarius M. 2013. Applying the neuroscience
Ito M. 2006. Cerebellar circuitry as a neuronal machine. Prog of creativity to creativity training. Front Hum Neurosci. 7:656.
Neurobiol. 78:272–303. O’Reilly JX, Woolrich MW, Behrens TEJ, Smith SM, Johansen-
Ito M. 1997. Cerebellar microcomplexes. In: Schmahmann JD, Berg H. 2012. Tools of the trade: psychophysiological interac-
editor. The cerebellum and cognition. New York: Academic. tions and functional connectivity. Soc Cogn Affect Neurosci.
p. 475–487. 7:604–609.
Ito M. 2008. Control of mental activities by internal models in the Pinho AL, de Manzano Ö, Fransson P, Eriksson H, Ullén F. 2014.
cerebellum. Nat Rev Neurosci. 9:304–313. Connecting to create: expertise in musical improvisation is
Ito M. 1993. Movement and thought: identical control mechan- associated with increased functional connectivity between
isms by the cerebellum. Trends Neurosci. 16:448–450. premotor and prefrontal areas. J Neurosci. 34:6156–6163.
Ito M. 2011. The cerebellum: brain for an implicit self. Upper Ramnani N. 2006. The primate cortico-cerebellar system: anat-
Saddle River: FT Press. omy and function. Nat Rev Neurosci. 7:511–522.
Jung-Beeman M, Bowden EM, Haberman J, Frymiare JL, Arambel- Reiter-Palmon R, Mumford MD, Threlfall KV. 1998. Solving every-
Liu S, Greenblatt R, Reber PJ, Kounios J. 2004. Neural activity day problems creatively: the role of problem construction and
when people solve verbal problems with insight. PLoS Biol. 2:E97. personality type. Creat Res J. 11:187–197.
 3552 | Cerebral Cortex, 2017, Vol. 27, No. 7

Runco MA. 2004. Creativity. Annu Rev Psychol. 55:657–687. Takeuchi H, Taki Y, Hashizume H, Sassa Y, Nagase T, Nouchi R,
Runco MA. 1991. The evaluative, valuative, and divergent think- Kawashima R. 2011. Failing to deactivate: the association be-
ing of children. J Creat Behav. 25:311–319. tween brain activity during a working memory task and cre-
Saggar M, Quintin E-M, Kienitz E, Bott NT, Sun Z, Hong W-C, ativity. NeuroImage. 55:681–687.
Chien Y-H, liu N, dougherty RF, Royalty A, et al. 2015. Piction- Torrance EP. 1968. A longitudinal examination of the fourth grade
ary-based fMRI paradigm to study the neural correlates of slump in creativity. Gifted Child Q. 12:195–199.
spontaneous improvisation and figural creativity. Sci Rep. Torrance EP. 1998. Torrance tests of creative thinking: Norms-
5:10894. technical manual: Figural (streamlined) forms A & B. Earth
Schmahmann JD. 1991. An emerging concept. The cerebellar con- City: Scholastic Testing Services.
tribution to higher function. Arch Neurol. 48:1178–1187. Tu L, Wang J, Abutalebi J, Jiang B, Pan X, Li M, Gao W, Yang Y, Liang B,
Scott G, Leritz LE, Mumford MD. 2004. The effectiveness of cre- Lu Z, et al. 2015. Language exposure induced neuroplasticity in
ativity training: A quantitative review. Creat Res J. 16:361–388. the bilingual brain: a follow-up fMRI study. Cortex. 64:8–19.
Sobczak-Edmans M, Ng THB, Chan YC, Chew E, Chuang KH, Vandervert L. 2015. How music training enhances working mem-
Chen SHA. 2016. Temporal dynamics of visual working mem- ory: a cerebrocerebellar blending mechanism that can lead

Downloaded from https://academic.oup.com/cercor/article/27/7/3542/3056437 by guest on 04 May 2024

ory. NeuroImage. 124:1021–1030. equally to scientific discovery and therapeutic efficacy in
Solso RL. 2001. Brain activities in a skilled versus a novice artist: neurological disorders. Cerebellum Ataxias. 2:11.
an fMRI study. Leonardo. 34:31–34. Vandervert LR. 2003. How working memory and cognitive model-
Stavridou A, Furnham A. 1996. The relationship between psycho- ing functions of the cerebellum contribute to discoveries in
ticism, trait-creativity and the attentional mechanism of cog- mathematics. New Ideas Psychol. 21:159–175.
nitive inhibition. Pers Individual Differ. 21:143–153. Vandervert LR, Schimpf PH, Liu H. 2007. How working memory
Sternberg R, Ohara PT. 1999. Handbook of Creativity. Cambridge and the cerebellum collaborate to produce creativity and in-
University Press. novation. Creat Res J. 19:1–18.
Stroop JR. 1935. Studies of interference in serial verbal reactions. Wolfradt U, Pretz JE. 2001. Individual differences in creativity: per-
J Exp Psychol. 18:643–662. sonality, story writing, and hobbies. Eur J Pers. 15:297–310.