Modeling the Edge-Cloud Continuum: A Brain-Computer Interface Case Study

The availability of low-cost EEG headsets and advancements in signal processing have expanded the potential of Brain-Computer Interface (BCI) systems, particularly in cognitive buildings for enhanced accessibility and automation. However, a structured approach for integrating EEG-based applications within an edge-cloud continuum is lacking. This paper proposes a modeling approach for designing such applications and evaluates its effectiveness through a case study on steadystate visual evoked potential (SSVEP) recognition. Experiments conducted on four subjects show that SSVEP recognition is subject-dependent and influenced by electrode configuration, with classification accuracies ranging from 95.56 % to 100 % for individuals and 95.33 % to 89.60 % for aggregated data, with a Random Forest classifier. The proposed methodology lays the foundation for scalable, intelligent applications that leverage EEG signals to infer user preferences.