Evaluative and generative modes of thought during the creative process

Abstract

Psychological theories have suggested that creativity involves a twofold process characterized by a generative component facilitating the production of novel ideas and an evaluative component enabling the assessment of their usefulness. The present study employed a novel fMRI paradigm designed to distinguish between these two components at the neural level. Participants designed book cover illustrations while alternating between the generation and evaluation of ideas. The use of an fMRI-compatible drawing tablet allowed for a more natural drawing and creative environment. Creative generation was associated with preferential recruitment of medial temporal lobe regions, while creative evaluation was associated with joint recruitment of executive and default network regions and activation of the rostrolateral prefrontal cortex, insula, and temporopolar cortex. Executive and default regions showed positive functional connectivity throughout task performance. These findings suggest that the medial temporal lobe may be central to the generation of novel ideas and creative evaluation may extend beyond deliberate analytical processes supported by executive brain regions to include more spontaneous affective and visceroceptive evaluative processes supported by default and limbic regions. Thus, creative thinking appears to recruit a unique configuration of neural processes not typically used together during traditional problem solving tasks.

Introduction

What did Leonardo da Vinci, Albert Einstein, and Thomas Edison possess that allowed them to produce works and ideas that changed how we live our lives and understand the world? Creativity is a quintessential and uniquely human characteristic manifested in art galleries, concert halls, and science laboratories, as well as in everyday activities (Runco, 2004). Often defined in terms of its product, creativity is the ability to produce ideas that are both novel (original and unique) and useful (appropriate and meaningful) (Amabile and Tighe, 1993, Besemer and Treffinger, 1981, Bruner, 1962, Gardner, 1989, Sternberg, 1985). Paralleling this twofold definition of creativity, psychological findings have suggested a twofold creative process that includes generative and evaluative components (Basadur et al., 1982, Campbell, 1960, Finke et al., 1992, Israeli, 1962, Wallas, 1926). For example, one model of the creative process proposes that it begins with the generation of crudely formed ideas and associations, followed by their exploration through evaluation and testing (Basadur et al., 1982, Finke et al., 1992). Similarly, the creative process has been described as the “mutation” of a thought into many different variants to generate ideas and the evaluation of the ideas to select the “fittest” or best variant (Campbell, 1960). Thus, the dichotomy between generation and evaluation appears to be ubiquitous in psychological theories of the creative process, with novel ideas produced during generative phases and their utility assessed during subsequent evaluative phases. This dichotomy is also present in artists' accounts of their own creative process, which they often describe as alternating between rough sketching of ideas and critiquing those ideas to guide the next cycle of sketching and critiquing (Fox, 1997, McMullan, 2010 Dec 2, Victore, 1997). The neural correlates of this distinction, however, have remained largely unknown, and revealing them could further our understanding of creativity and its component processes.

Do different neural networks and brain regions contribute differentially to creative generation and evaluation? While earlier neuroscientific studies of creativity emphasized large-scale brain distinctions such as hemispheric differences (Bekhtereva et al., 2000, Bekhtereva et al., 2001, Carlsson et al., 2000, Finkelstein et al., 1991, Martindale et al., 1984, Sperry, 1964) and frontal versus parietal lobe engagement in creative thinking (Bekhtereva et al., 2004, Chavez-Eakle et al., 2007, Fink et al., 2009, Fink and Neubauer, 2006, Geake and Hansen, 2005, Jung et al., 2010, Miller et al., 1998, Miller et al., 2000, Molle et al., 1996, Molle et al., 1999, Razoumnikova, 2000, Razumnikova, 2007, Starchenko et al., 2003), recent findings point to the possible involvement of specific networks and brain regions in the different components of the creative process.

One network that appears to contribute to creative thinking through its robust association with cognitive control is the executive network. Including, most prominently, the dorsolateral PFC (DLPFC) and dorsal anterior cingulate cortex (dACC), the executive network is specifically recruited during conditions of high cognitive control (Desimone and Duncan, 1995, Miller and Cohen, 2001). The DLPFC and dACC are known to be activated during a variety of creative tasks, including piano improvisation (Bengtsson et al., 2007, Berkowitz and Ansari, 2008), creative story generation (Bekhtereva et al., 2000, Bekhtereva et al., 2001, Howard-Jones et al., 2005), word association (Bekhtereva et al., 2004), divergent thinking (Carlsson et al., 2000, Seger et al., 2000), fluid analogy formation (Geake and Hansen, 2005), insight problem solving (Geake and Hansen, 2005, Kounios et al., 2008, Subramaniam et al., 2009), and visual art design (Kowatari et al., 2009). During these creative tasks, high cognitive control enables a deliberate, analytic mode of information processing that facilitates the evaluation of the utility of novel ideas (Howard-Jones and Murray, 2003) and allows individuals to focus on the pertinent task details and to select the relevant generated ideas (Dorfman et al., 2008, Gabora, 2010, Heilman et al., 2003, Lepine et al., 2005, Vartanian et al., 2007). Therefore, the executive network may contribute specifically to the evaluative mode of creative thought.

Another network that plays an important role during creative thought is the default network. It includes, most prominently, the medial prefrontal cortex (MPFC), posterior cingulate cortex (PCC)/precuneus, and temporoparietal junction (TPJ) (Raichle et al., 2001), which are frequently activated in creativity experiments. For example, enhanced TPJ activity has been found during divergent thinking tasks (Fink and Neubauer, 2006, Grabner et al., 2007), creative story generation (Bekhtereva et al., 2004), hypothesis generation (Jin et al., 2006), fluid analogy formation (Geake and Hansen, 2005), and remote associate insight problems (Jung-Beeman et al., 2004, Subramaniam et al., 2009). Creative story generation also recruits the MPFC (Howard-Jones et al., 2005), while insight problems activate both the MPFC and PCC/precuneus (Geake and Hansen, 2005, Jung-Beeman et al., 2004, Kounios et al., 2008, Subramaniam et al., 2009). Moreover, Limb and Braun (2008) found activation of the default network and deactivation of the executive network during improvisation by professional jazz pianists. However, with the exception of Limb and Braun's (2008) study, only parts of the default network, rather than the whole network, have been associated with creativity tasks.

There are two possibilities regarding the default network's role in creative generation and evaluation. On the one hand, because it is specifically activated during conditions of low cognitive control (Raichle et al., 2001, Shulman and Fiez, 1997), it may facilitate an associative mode of processing that supports the generation of novel ideas (Dorfman et al., 2008, Howard-Jones and Murray, 2003, Vartanian et al., 2007), thereby contributing to creative generation more than to creative evaluation. On the other hand, the default network may contribute more to creative evaluation than to generation, through its role in affective and visceroceptive evaluative processes demonstrated during emotional paradigms. It is activated during the evaluation of emotional reactions (Fossati et al., 2003, Ochsner et al., 2004, Ruby and Decety, 2004) and internally generated affective information (Damasio et al., 2000, Gusnard et al., 2001, Lane et al., 1997, Zysset et al., 2002), which may facilitate the formation and awareness of “gut reactions” that individuals monitor during creative work (de Bono, 2000).

In addition to default network regions, medial temporal lobe (MTL) memory regions such as the hippocampus and the parahippocampus have also frequently been reported during creativity experiments, although they have received relatively little attention to date. For example, the hippocampus exhibits greater recruitment during visual art design (Kowatari et al., 2009) and divergent thinking (Fink et al., 2009). In general, studies have linked the MTL to memory retrieval (Squire et al., 2004) and associative processing (Eichenbaum, 2000). The MTL is activated during the formation and retrieval of semantic and episodic associations (Aminoff et al., 2007, Bar et al., 2008, Henke et al., 1997, Henke et al., 1999, Rombouts et al., 1997), as well as during mental simulations of past, future, and novel events that require the recombination of stored associations (Addis et al., 2007, Botzung et al., 2008, Hassabis et al., 2007, Okuda et al., 2003, Szpunar et al., 2009). The associative function of the MTL implies that it may be particularly important for creative thought by facilitating the generation of novel ideas and associations and the recombination of old ones.

While all three aforementioned networks appear to play important roles in creative thought, their distinct contributions to the different components of the creative process remain unclear. On the basis of the previously reviewed neuroscientific findings, we could hypothesize that: (i) the MTL memory network may contribute to associative processes that would enable creative generation; (ii) the default network may contribute either to creative generation through its role in low cognitive control or to creative evaluation through its role in affective and visceral evaluative processing; and (iii) the executive network may contribute to the analytical evaluative processes required during creative evaluation.

To identify the specific contributions of various brain areas to the creative process, we developed a novel paradigm that separated and alternated between generative and evaluative modes during the performance of a visual book cover design task, used as a creativity exercise in visual arts and design programs. To complete the task, participants used a functional magnetic resonance imaging (fMRI)-compatible drawing tablet (Tam et al., 2010) that allowed them to actively draw and write their ideas and evaluations while in an fMRI scanner. By doing so, the study used an approach closer to real-life creative activities compared to previous studies, most of which required participants to only imagine their solutions and designs during creativity tasks (Fink et al., 2007).