Using artificial intelligence and machine learning to scale the application of neuro-analytics in the content design proces

The application of neuro-analytics to digital content design has the potential to drive more efficient and effective development processes. Neuro-metrics go beyond traditional user-data acquisition models (focus groups) by using attention, memory, emotion and other mind-states to help determine content’s performance. Designers predict how content will elicit neuro-responses from individuals, and then test those hypotheses with actual human subjects using EEG. While helpful, neuro-metrics also present challenges due to the barriers associated with data collection. Analyses rely on obtaining brainwave data from live subjects matching defined demographics. The data-acquisition process (recruiting subjects, EEG set-up, running content, etc.) requires time, labor and money. It is the bottleneck. To scale the benefits of neuro-design, the requirement for EEG data-acquisition must be reduced or eliminated. Machine learning and artificial intelligence (ML/AI) offer a solution. If, instead of measuring human subjects who meet pre-defined demographics, the system pulls from pre-existing neuro-datasets of similar subjects; and if, rather than testing specific content, the system uses unsupervised ML to scan previously tested content, using deep learning analytics to identify similar and matching content, then the system can output instantaneous results with sufficient validity, avoiding the need for neuro-data acquisition, and allowing for scalability and repeatability. This work presents an effort directed toward building a sufficient dataset where such ML/AI algorithms can be applied within the neuro-design process, providing immediate feedback to the designer and indicating the accuracy of his/her predictions based on the AI results. The practical applications of utilizing AI/ML in the neuro-design process will be discussed, with emphasis on how such technology accommodates the nuances inherent within the development of warfighter training content. The elements and architecture of the ML/AI components will be discussed as they apply to the NeuroDesign process, as will the precursor studies that served to motivate the application.