OPEN Neurocognitive differences in sketching between design tasks and creativity tests

OPEN Neurocognitive differences in sketching between design tasks and creativity tests

This study investigates differences in brain activity among thirty-seven engineering students during sketching in a creativity test and a design task. Using EEG data, we first examined the impact of baseline selection (eyes-open vs. eyes-closed) on the representation of results in terms of de/synchronization across frequency bands from theta (four to seven Hertz) to lower gamma (thirty to forty-five Hertz). Specifically, an eyes-closed baseline allowed for the observation of alpha desynchronization during sketching, whereas an eyes-open baseline, which better aligns with real-world design conditions, shifted the pattern to synchronization. Under the eyes-open baseline, the comparisons between the two sketching tasks revealed the following results: One. Regions that showed significant bilateral TRP asymmetries in creativity test sketching also exhibited differences in design sketching across frequency bands from theta to beta band. Areas that exhibited significant bilateral TRP asymmetry only in design sketching were located in the temporal sites (theta and beta), suggesting these areas within the corresponding frequency bands could serve as indicators for distinguishing between the two cognitive tasks. Two. Statistically significant differences between the two sketching tasks in channel comparisons were primarily observed in the theta and sub-alpha bands, especially in the left frontal and right hemisphere areas. Three. The beta band exhibited similar behavior across both sketching tasks, indicating shared cognitive processes in different sketching contexts. Four. While the lower gamma band did not show significant differences in channel comparisons between tasks, it exhibited distinct bilateral TRP asymmetry in the frontocentral area only in the creativity test, highlighting the potential of bilateral asymmetry as a more sensitive measure for distinguishing between design and creativity tasks.

In the design domain, sketching has long been acknowledged as a fundamental and powerful tool in the ideation and conceptualization processes, as it externalizes mental representations into visual forms and supports idea exploration and refinement.

Much of the design literature has examined how creativity is manifested during sketching, by means of sketching, and through the outcomes of sketching, often through observational approaches that inherently introduce subjectivity and pose interpreting challenges in unfolding the underlying cognitive behavior during sketching. To overcome these limitations, neurophysiological methods (e.g., EEG) have increasingly been adopted to study creativity and design cognition. However, these investigations are mainly based on standard creativity tests rather than actual design tasks. In particular, the Torrance Test of Creative Thinking-Incomplete Figures has been widely employed to study visual creativity. While informative, it emphasizes figural imagination and sketching activities under simplified and controlled conditions. This raises questions about whether the neural patterns observed in it can be generalized to design sketching, which is often constrained by functional requirements, requires integrative composition of multiple features, and demands more complex reasoning and visuospatial integration. The present study aims to address this gap by investigating whether cognitive activities during different sketching tasks can be distinguished through electroencephalogram data, specifically comparing sketching in a design task and sketching in a standard creativity test (i.e., Torrance Test of Creative Thinking-Incomplete Figures).

Following the introduction section, this paper first presents the background of the study. Then, it describes the employed methodology, detailing the experimental protocol, signal processing algorithms, and how the statistical hypothesis testing provides elements to answer the research question. The results section then showcases the EEG behavior observed from the experiments. The paper then discusses these results and concludes with a final section summarizing the study's key findings.

State of the art

Sketching and sketches are crucial in various stages of the design process. By promoting information processing and externalizing the designer's ideas, sketching can serve as a cognitive tool to support design reasoning. It facilitates rapid processing of spatial and visual information and enhances memory and thought. It can store the created idea while simultaneously offering the designers immediate feedback, providing the designer with new insights that stimulate further ideas, and boosting exploration's fluidity. Sketches may enhance fixation effects by reducing the variety and originality of ideas generated during creativity sessions. Research on design cognition demonstrated that novice sketchers are not more creative when they sketch and that the effectiveness of sketches to increase creativity greatly depends on the design strategy followed by the designer.

Sketches are used in both divergent thinking and convergent thinking, which are commonly recognized design reasoning processes involved in idea exploration. Divergent thinking fosters diversity and exploration in idea generation, resulting in a wider variety of screened possibilities, while convergent thinking helps focus and refine ideas. During sketching, there might be a dynamic interplay of convergent and divergent thinking, which enables designers to navigate the complexity of the design space effectively. A good design often results from effectively balancing both processes. However, novice designers may find it challenging to apply this balance in their early practice. Recent studies focus on integrating neurophysiological data with traditional observational approaches to distinguish different cognitive processes across various types of creativity tests. This could potentially allow the design process to be monitored and guided in real-time using biofeedback, which may reduce the need for additional guidance during the ideation process for novice designers. However, unlike creativity studies, there has been limited exploration of the design activities incorporating sketching through neurophysiological measures. Furthermore, it has been discovered that sketching differs cognitively from other design tasks (e.g. verbalization and typing); hence, the findings from other studies cannot be applied directly to sketching-based design activities.

Recent reviews indicate that the neurophysiological signal measurements mainly employed in design cognition studies include: functional magnetic resonance imaging, which necessitates that individuals remain motionless in the scanning tunnel until data collection is complete; near-infrared spectroscopy, which is restricted to the frontal areas of the scalp, with interpretation challenges related to multiple sources of the vascular signal; and electroencephalogram, which primarily reflects scalp-level brain activity, while sensitive to various sources of noise, it offers excellent time resolution and greater flexibility for body movements compared to functional magnetic resonance imaging. Additionally, electroencephalogram typically covers more areas of the scalp than near-infrared spectroscopy, providing a more comprehensive dataset for analysis. Of all the available techniques, electroencephalogram measurement was considered the most appropriate for the current investigation given its strengths and the presence of existing research evidence to evaluate the data quality.

Activities involving sketching as part of creativity tests (mostly in psychology studies) or freehand design sketching have previously drawn attention and have been investigated by several EEG data-based research studies. The review provided insights into visual creativity, specifically how EEG was used to examine brain activity during creativity tests involving sketching. Visual creativity, in this context, refers to generating new and practical visual ideas by transforming mental images into visual forms through sketching or painting. The review found that the task most commonly employed in EEG-based studies was the Torrance Test of Creative Thinking-Incomplete Figures. The task requires the participant to generate a complete image from a provided fragment of a drawing (i.e., basic geometric shapes like lines, circles, and rectangles). Most of the reviewed contributions considered EEG power, or event/task-related power and EEG coherence of lower alpha (alpha one: seven to ten Hertz) and upper alpha (alpha two: ten to thirteen Hertz) sub-frequency bands. Although the observations were not always consistent within each measure (power or coherence), the studies utilizing power analysis revealed a more pronounced trend in brain signal behavior compared to those focusing on coherence. Table one presents the studies selected from the review, along with additional studies that focus on EEG-based analysis of the Torrance Test of Creative Thinking-Incomplete Figures test, which considered the use of TRP or POW as (part of) the data treatment method.

All the studies presented in Table one, except the recent work by Jia and Zeng, instructed participants to generate sketching ideas mentally, focusing their EEG analysis on the imaginative phase. As a result, these studies primarily emphasized figural imagination abilities rather than the actual progression of idea development through sketching. A common observation in this body of research was task-related alpha desynchronization when compared to the baseline period. This desynchronization was also observed in Jia and Zeng's study. However, it is important to note that Jia and Zeng used an eyes-closed period as the baseline, which differs significantly in power from the eyes-open condition. Notably, the eyes-open condition could offer greater continuity and comparability with task activities like sketching, as it more closely reflects the neural state associated with visual engagement and active cognitive processing during such task activities. This highlights a gap in the literature, as studies have yet to use an eyes-open baseline while incorporating actual sketching as the test activity. Therefore, the first research question this paper would like to address is what the de/synchronization brain behavior of engineering students when sketching in TTCT-IF with reference to the different eye conditions used as a baseline for TRP estimation is, and how these observations span a wider frequency range.

Most of these studies have primarily focused on the alpha band due to its well-established correlation with creativity, while notable results were also observed in the beta band. However, higher frequencies, such as the gamma band, have not been reported in any studies using the TTCT-IF task. On the one hand, the existing findings about the alpha band in sketching activities require extending the analysis to EEG behavior using the eyes-open baseline condition. On the other hand, broadening the analysis to include more frequency bands can offer a more comprehensive understanding of brain behavior in the accomplishment of this task and more comparison opportunities with data collected in other relevant tasks. Additionally, the participants in these studies were from different backgrounds (mostly in psychology), raising the question of whether the observations of sketching activities in the TTCT-IF test can be generalized to design sketching that requires specialized technical knowledge (e.g. engineering).

Regarding the EEG-based research that focuses on design sketching rather than creativity tests, through a series of studies, Vieira and others compared freehand design sketching within an open design task with a constrained problem-solving design task from diverse viewpoints. Their research demonstrated that EEG data exhibit a main effect of gender in the theta, higher alpha, and lower beta frequency bands. In these frequency bands, females show higher activation in the right dorsolateral prefrontal cortex, right occipitotemporal cortex, secondary visual cortex (i.e., part of the occipital cortex), and prefrontal cortex on both tasks. Both genders show higher power in the theta and alpha one band in the right hemisphere and higher power in the alpha two band and the beta band across hemispheres during the open design sketching task. Subsequently, they found significant differences between constrained and open designs for the EEG power in beta bands at the earliest reaction stage immediately after reading the request. Then, the alpha two band and beta band were found to have significant differences between the two tasks in the stage of sketching to externalize the solution. Despite numerous studies on sketching-based creativity tests and some efforts in sketching for design, there remains a significant gap in research focusing on neurocognitive behavior during sketching and visual creativity within design contexts. The current literature lacks sufficient exploration of how brain activity during sketching varies across different scenarios also raises the question of whether findings from creativity tests can be reliably used to interpret design cognitive behavior through EEG measurements.

This difficulty in generalizing findings becomes even more pronounced when considering the differences in context, objectives, and cognitive demands between structured creativity tests and real-world design tasks. The contrast between sketching in creativity tests and design is consistent with dual-process in cognitive psychology, which distinguishes human cognition as being governed by two systems (i.e., system one: fast, intuitive and effortless thinking; system two: slower, analytic thinking). Sketching in a creativity test such as TTCT-IF is open-ended and primarily divergent. It emphasizes intuitive associative processing (system one), imposes no functional requirements on the drawing and does not set a specific outcome goal for the sketch. The constraints are task-based, arising only from the fragments provided for elaboration. In contrast, design sketching is goal-oriented and context-dependent. To propose a novel design concept by means of sketching, the designer should integrate multiple partial solutions into a coherent whole that satisfies explicit functional requirements under practical constraints. This process involves analytic processing (system two), iterative evaluation of trade-offs, and a strong balance between divergent and convergent thinking. Compared with creativity tests, these additional constraints increase cognitive load by introducing higher demands on problem-solving, visuospatial reasoning and decision making. Furthermore, it also aligns with design frameworks that describe design as complex, iterative activities of idea generation and evaluation rather than a pure intuitive thinking process. Therefore, it is crucial to investigate sketching activities in both contexts to better understand the cognitive processes involved and to determine the applicability of creativity test findings to real-world design tasks. By analyzing the neurophysiological data, such as EEG in this study, we intend to explore cognitive differences in design activities that utilize the same idea expression modalities as those used in creativity tests (e.g., sketching). This exploration may also shed light on the ongoing debate about the differences between design creativity and creativity measured by divergent thinking tests.

In light of the previously identified research gaps, the present study aims to address the use of inconsistent baselines, non-comprehensive frequency band analysis, participant background diversity, and challenges in generalizing results. Specifically, it focuses on distinguishing and comparing EEG behaviors observed during sketching tasks performed by the same group of engineering students. By employing an eyes-open condition as the baseline and analyzing EEG data across a broader range of frequency bands, this study aims to provide a more comprehensive understanding of cognitive processes in sketching for different purposes. Through this approach, the study seeks to answer the following research questions:

Research question one: What is the de/synchronization brain behavior of engineering students when sketching in TTCT-IF with reference to the different eye conditions used as a baseline for TRP estimation?

Research question two: How does the brain behavior of engineering students differ when sketching in a design task and in the TTCT-IF test?

The methodology used to address these research questions is presented in the following section.