Design exploration predicts designer creativity: a deep learning approach.

Area of Science:

• Neuroscience
• Cognitive Science
• Design Studies

Background:

• Understanding the neural basis of creativity is crucial for optimizing design processes.
• Graphic designers' brain activity during creative tasks offers insights into innovation.

Purpose of the Study:

• To investigate brain activation patterns in graphic designers during pictorial stimulation and exploration tasks.
• To determine if design exploration can predict designer creativity using deep learning.

Main Methods:

• Examined brain activation in graphic designers using pictorial stimuli during exploration tasks.
• Employed deep learning models, including shallow and deep architectures, Long Short-Term Memory (LSTM) networks, Convolutional Neural Networks (CNNs), and Directed Acyclic Graph (DAG) networks.
• Classified participants into high-creativity and low-creativity groups based on performance.

Main Results:

• Shallow neural network architectures demonstrated higher prediction accuracy than deeper ones.
• Shallow Long Short-Term Memory (LSTM) networks outperformed Convolutional Neural Networks (CNNs).
• Bandpower analysis and shallow LSTM networks with specific power spectra improved prediction accuracy, outperforming other deep learning methods. Design exploration effectively predicted designer creativity.

Conclusions:

• Design exploration is a significant predictor of creativity in graphic designers.
• Shallow neural network architectures, particularly shallow LSTM networks with bandpower analysis, are effective for predicting creativity.
• The findings suggest novel approaches for assessing and fostering creativity in design professionals.