Association between low-level geometric features of central gaze-guiding patterns and SSVEP responses in mixed reality head-mounted display environments

Abstract

This study investigated the correlation between the low-order geometric properties of centrally fixated guide patterns and steady-state visual evoked potential (SSVEP) responses in a mixed reality (MR) head-mounted display environment. An MR-SSVEP experimental environment was established using Apple Vision Pro. For 22 types of central gaze-guiding patterns, computational vision methods were employed to quantify four categories of low-order geometric features: directionality, centrality, radiation index and complexity. Correlation analysis, regression analysis and time-frequency spatial response analysis were conducted to identify the relationships between these geometric properties and SSVEP response characteristics. At the pattern-mean level, centrality was positively associated with SNR ( r = 0.5407, p = 0.0094) and the CCA coefficient ( r = 0.4452, p = 0.0379), but not with classification accuracy (ACC; r = 0.2966, p = 0.1801). Directionality, radiation index, and complexity did not show statistically significant associations with these outcomes. Descriptive analysis within the sample revealed that centrally located gaze-guided patterns with high centrality and low complexity exhibited superior overall SSVEP performance under the current experimental conditions, accompanied by a more concentrated distribution of energy at the target frequency and a more stable occipital response pattern. This study preliminarily reveals a potential association between visual-spatial organisation characteristics and neural responses in a near-eye mixed reality display environment, demonstrating the reference value of low-order geometric feature quantification methods in the design of mixed reality brain-computer interface systems.