EEG Feature-level integration for robust command generation in SSVEP-based brain-computer interfaces

Steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) have been widely studied for efficient command generation via spatiotemporal information integration and advanced stimulation paradigms. However, users may intentionally or unintentionally suppress SSVEP responses, leading to the SSVEP Blocking Response (SSVEP-BR). This phenomenon, characterized by attenuation or disappearance of SSVEP, has been largely overlooked despite its potential for command generation without additional stimuli.

This study aimed to exploit SSVEP-BR to enhance command generation under limited visual stimulation. A novel detection method was proposed based on alpha rhythm sensitivity during transitions between SSVEP and SSVEP-BR states. An EEG feature-level integration strategy was further developed to integrate correlation and alpha rhythm features, enabling effective characterization of their temporal evolution. An asynchronous BCI system was implemented for real-time robotic arm control.

Experiments on 11 offline subjects demonstrated that the proposed method improved SSVEP-BR detection performance. Compared with the previous SSVEP-BR detection framework, average improvements of 17.11%, 8.50%, and 5.44% were achieved in precision, recall, and specificity, respectively. Furthermore, an online robotic grasping experiment was conducted with 5 subjects, where the proposed asynchronous BCI system achieved 8-command generation using two stimulation frequencies.

Unlike prior studies focusing on algorithmic optimization or stimulation design, this work introduced SSVEP-BR as an additional controllable neural response and demonstrated its effectiveness in improving command generation without increasing stimulus complexity.

Incorporating SSVEP-BR into SSVEP-based BCI frameworks improved flexibility and robustness, enhancing the practicality of asynchronous brain computer interaction systems.