Mapping Neuroimaging Findings of Creativity and Brain Disease Onto a Common Brain Circuit

Abstract

Importance: Creativity is important for problem solving, adaptation to a changing environment, and innovation. Neuroimaging studies seeking to map creativity have yielded conflicting results, and studies of patients with brain disease have reported both decreases and paradoxical increases in creativity, leaving the neural basis of creativity unclear.

Objective: To investigate the brain circuit underlying creativity and assess its association with brain injury and neurodegenerative disease.

Design, setting, and participants: This study examined neuroimaging coordinates from a meta-analysis of 36 studies published between 2004 and 2019 associated with increased activity during creative tasks in healthy participants. A validated method termed coordinate network mapping and a database of resting-state functional connectivity from 1000 healthy individuals were used to test whether these coordinates mapped to a common brain circuit. Specificity was assessed through comparison to random coordinates and coordinates from working memory tasks in healthy participants. Reproducibility was assessed using an independent dataset of coordinates from additional studies of creativity in healthy participants. Finally, alignment with effects of focal brain damage on creativity was tested using data from patients with brain lesions and coordinates of brain atrophy from 7 different neurodegenerative disorders.

Main outcomes and measures: The primary outcomes were creativity or no creativity and alignment with a creativity circuit or no alignment.

Results: Creativity tasks activated heterogenous locations, with coordinates scattered across many different brain regions (415 coordinates derived from 857 healthy participants; pooled mean [SD] age, 24.1 [6.91] years; 461 [54%] female). However, these activation coordinates were part of a common brain circuit, defined by negative connectivity to the right frontal pole. This result was consistent across creative domains, reproducible in an independent dataset (383 coordinates derived from 691 participants) and specific to creativity when compared with random gray matter coordinates (n = 415) or coordinates activated by working memory tasks (3072 coordinates derived from 2900 healthy participants). Damage to this creativity circuit by lesions (n = 56 patients) or neurodegenerative disease (2262 coordinates derived from 4804 patients) aligned with both decreases and increases in creativity observed in these disorders.

Conclusions and relevance: Findings from this study suggest that brain regions activated by creativity tasks map to a brain circuit defined by negative functional connectivity to the right frontal pole. Damage to this circuit aligned with changes in creativity observed in individuals with certain brain diseases, including paradoxical creativity increases.

Figure 1.. Neuroimaging Studies on Creativity Displaying High Heterogeneity of Results

A, Brain maps showing peak coordinates of task activation (red) from individual studies of creativity (showing results from 3 of 36 studies). B, Brain maps showing all included coordinates.

Figure 2.. Methodological Overview

A, Brain maps showing peak coordinates of task activation (red) from individual studies of creativity were used as seed regions. For each study/seed, functional connectivity maps were obtained using a 1000-participants resting-state functional connectome. B, Individual connectivity maps were thresholded (t ≥ 5) and binarized to obtain a network overlap map in which peak overlaps indicate the regions most consistently connected to the activation seeds. C, Significant results from a 1-sample t test at a family-wise error rate–corrected P value threshold of <.05. D and E, Specificity was assessed via statistical comparison with permutation analysis of linear models at 10 000 permutations against random gray matter and working memory coordinates. Reported clusters are significant at false discovery rate–corrected P < .05 with threshold-free cluster enhancement (TFCE). The right frontal pole (red arrows) was a consistent finding across all sensitivity and specificity analyses.

Figure 3.. Connectivity of Creativity Activation Foci to the Right Frontal Pole (rFP)

A, Functional connectivity with the rFP (green, left) defines a distributed brain circuit (right) that will encompass coordinates activated by creativity tasks. B, Brain maps showing coordinates activated in 36 studies of creativity (white dots) fall within the circuit (warm colors) defined by negative functional connectivity to the rFP. C, Brain maps showing coordinates activated in 30 independent studies of creativity (white dots) fall within the same circuit.

Figure 4.. Lesions Affecting Creativity Task Performance and the Creativity Circuit

A, Lesion overlap maps from Abraham et al showing groups of patients with lesions in different neuroanatomical locations. Patients with lesions to the lateral frontal lobe (right) had impaired performance on creativity tasks, while patients with lesions to the frontal pole (left) had higher performance. B, Topographical relation of overlap maps in (A, white outlines) to the creativity circuit. Intersection of the lesion locations with our creativity circuit was significantly associated with lesion-induced effects on creativity task performance (eFigure 4 in Supplement 1). rFP indicates right frontal pole.

Figure 5.. Neurodegenerative Atrophy Patterns and the Creativity Circuit

A, Brain maps showing atrophy coordinates (red) in different neuroanatomical locations in studies on semantic (svPPA), logopenic (lvPPA), and nonfluent (nfvPPA) variants of primary progressive aphasia, behavioral variant of frontotemporal dementia (bvFTD), typical Alzheimer disease (AD), Parkinson disease (PD), and amyotrophic lateral sclerosis (ALS). Patients with svPPA and bvFTD (left) have been reported to experience increases in creativity, while patients with nfvPPA (right) have been reported to experience decreased creativity. B, Topographical relation of atrophy coordinates in panel A to the creativity circuit (eFigure 5B in Supplement 1). For svPPA, 75% of the atrophy coordinates hit regions positively connected to the right frontal pole (rFP) (cool colors). By contrast, for nfvPPA, 68% of the atrophy coordinates hit regions negatively connected to the rFP (warm colors).