Investigating the generalizability of EEG-based Cognitive Load Estimation Across Visualizations

TL;DR

This study investigates whether EEG-based cognitive load estimation methods can reliably generalize across different visualization types during a working memory task, highlighting challenges and the need for improved machine learning approaches.

Contribution

It evaluates the generalizability of EEG-based cognitive load estimation across various visualizations using deep learning and support vector machines, revealing limitations and motivating further research.

Findings

CL estimation varies across visualization types

Current methods show limited cross-visualization generalizability

Highlights need for better ML techniques for interface usability assessment

Abstract

We examine if EEG-based cognitive load (CL) estimation is generalizable across the character, spatial pattern, bar graph and pie chart-based visualizations for the nback~task. CL is estimated via two recent approaches: (a) Deep convolutional neural network, and (b) Proximal support vector machines. Experiments reveal that CL estimation suffers across visualizations motivating the need for effective machine learning techniques to benchmark visual interface usability for a given analytic task.