Exploring Model Design Spaces for Cognitive and Affective State Recognition in Immersive Virtual Reality
This repository contains the code, data, and analysis related to the project "Exploring Model Design Spaces for Cognitive and Affective State Recognition in Immersive Virtual Reality Using Multimodal Neuro-physiological Signals." This research aims to advance the understanding of how cognitive and affective states can be inferred from neuro-physiological signals in immersive virtual reality (iVR) environments.
In this project, we conducted a secondary analysis on a multimodal dataset collected from 10 participants during a multi-phase iVR experiment. The experiment included various phases designed to elicit different cognitive and affective states, such as rest, effortful control, stress induction, and decision-making in a virtual buffet environment.
Key aspects of the project include:
Model Design Exploration: Investigated the impact of different model design choices, such as modality, feature set, and personalization, on the accuracy of inverse inference models that map neuro-physiological signals to cognitive and affective states.
Machine Learning Evaluation: Applied machine learning techniques to assess the ability of models to discriminate between different iVR phases.
Between-Person Variation: Used non-linear manifold embedding methods to analyze individual differences, emphasizing the need for adaptive models tailored to personal characteristics.