Eric Mah

Presenter: Eric Mah

Title: Measuring individuals’ mental representations

Abstract:
Mental representations of categories and concepts play a fundamental role in most aspects of cognition, including perception, decision-making, learning and memory. However, measuring these mental representations–particularly at the individual level–is challenging. Historically, researchers have used techniques like multidimensional scaling (MDS), which produces an interpretable visualization of representational space using behavioural similarity judgements. Although a useful technique for creating aggregate representational spaces, MDS has a number of limitations. I will present PsiZ, a novel machine-learning approach that shows promise for obtaining rich individual-level representational spaces, and discuss several preliminary experiments applying this approach to individual differences in knowledge and expertise. I hope to demonstrate the utility of this method for researchers interested in individuals’ mental representations, and invite discussion and suggestions for future research and applications.