Enhanced EEG-Based Tracking of DMN Activity Using ICA and EEG Source Localization for a VR Psychedelic Environment

A brain-computer interface within a virtual reality environment can modulate Default Mode Network activity by targeting the Posterior Cingulate Cortex. An EEG processing pipeline combining Independent Component Analysis with source localization derived a subject-specific spatial filter from resting-state data, applied during a virtual psychedelic experience. Compared to traditional parietal-EEG channel selection, the PCC-filtered method revealed significantly enhanced spectral differentiation between baseline and VR states, including a greater decrease in alpha power and increase in delta power during the virtual hallucination, and a larger dynamic range in both alpha and delta power modulations in response to VR scene changes. This approach offers a more sensitive measure of DMN dynamics for closed-loop BCI systems.

Study at a glance

Abstract

The therapeutic potential of psychedelic substances faces limitations due to safety and legal concerns. This study explores a non-pharmacological alternative using a brain-computer interface (BCI) within a Virtual Reality (VR) environment to modulate Default Mode Network (DMN) activity, specifically targeting the Posterior Cingulate Cortex (PCC). We present an offline validation of an EEG processing pipeline that combines Independent Component Analysis (ICA) with exact Low Resolution Electromagnetic Tomography (eLORETA) source localization to derive a subject-specific PCC spatial filter from resting-state data, which is then applied to EEG data recorded during a virtual psychedelic experience. Compared to traditional parietal-EEG channel selection, our PCC-filtered method revealed significantly enhanced spectral differentiation between baseline and VR states, particularly a greater decrease in alpha power (p=0.0295) and increase in delta power (p=0.0108) during the virtual hallucination relative to baseline. Furthermore, our PCC-filtered approach showed a significantly larger dynamic range in both alpha (p=0.0240) and delta (p=0.0070) power modulations in response to VR scene changes. These findings suggest that an ICA-based, source-informed approach targeting the PCC offers a more robust and sensitive measure of DMN dynamics, laying the groundwork for more precise closed-loop BCI systems designed to guide brain states in VR environments.