Neural correlates of learning in 360° virtual laboratory environments: A data-driven approach

The adoption of 360° virtual laboratory environments in educational settings has gained significant traction, which can be attributed to improved learning experiences, accessibility of learning, and scalability. In this study, we investigated how learning a laboratory technique within a virtual environment affects cortical activity. Two groups of participants studied the materials, either using a virtual laboratory or without it (control). During the lessons, EEG data were recorded, and non-learning tasks were randomly intercalated. Machine learning algorithms were used to decode the learning cognitive state, utilizing sub-band (delta, theta, alpha, beta, and gamma) relative wavelet energy (RWE) values as features. Learning patterns were reconstructed to examine differences across learning environments. Changes in accuracy levels over time and the quantification of times when learning was decoded in all frequency bands were used to examine the evolution of the learning process over time. Results showed decreased decoded learning in the theta band and increased decoded learning in the alpha band in the virtual laboratory condition, which are associated with lower cognitive load and attentional levels, respectively. Despite significant differences in neural decoding of learning, no differences in learning performance were found. This study proposes a data-driven approach to studying neural patterns of learning in virtual learning environments.