Objective evaluation of fatigue by EEG spectral analysis in steady-state visual evoked potential-based brain-computer interfaces

• DOI: 10.1186/1475-925x-13-28
• Publication Date: 2014-03-12T16:05:46Z
• Authors: Teng Cao; Feng Wan; Chi Man Wong; Janir Nuno da Cruz; Yong Hu
• Keywords: Adult; Analysis of Variance; Epilepsy; Time Factors; Electroencephalography - methods; Research; 0206 medical engineering; Biomedical Engineering; Electroencephalography; Signal Processing, Computer-Assisted; 02 engineering and technology; Signal-To-Noise Ratio; Young Adult; Fatigue - diagnosis; Brain-Computer Interfaces; Surveys and Questionnaires; Evoked Potentials, Visual; Humans; Visual; Evoked Potentials; Fatigue; Photic Stimulation; Signal Transduction

The fatigue that users suffer when using steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) can cause a number of serious problems such as signal quality degradation and system performance deterioration, users' discomfort and even risk of photosensitive epileptic seizures, posing heavy restrictions on the applications of SSVEP-based BCIs. Towards alleviating the fatigue, a fundamental step is to measure and evaluate it but most existing works adopt self-reported questionnaire methods which are subjective, offline and memory dependent. This paper proposes an objective and real-time approach based on electroencephalography (EEG) spectral analysis to evaluate the fatigue in SSVEP-based BCIs.How the EEG indices (amplitudes in δ, θ, α and β frequency bands), the selected ratio indices (θ/α and (θ + α)/β), and SSVEP properties (amplitude and signal-to-noise ratio (SNR)) changes with the increasing fatigue level are investigated through two elaborate SSVEP-based BCI experiments, one validates mainly the effectiveness and another considers more practical situations. Meanwhile, a self-reported fatigue questionnaire is used to provide a subjective reference. ANOVA is employed to test the significance of the difference between the alert state and the fatigue state for each index.Consistent results are obtained in two experiments: the significant increases in α and (θ + α)/β, as well as the decrease in θ/α are found associated with the increasing fatigue level, indicating that EEG spectral analysis can provide robust objective evaluation of the fatigue in SSVEP-based BCIs. Moreover, the results show that the amplitude and SNR of the elicited SSVEP are significantly affected by users' fatigue.The experiment results demonstrate the feasibility and effectiveness of the proposed method as an objective and real-time evaluation of the fatigue in SSVEP-based BCIs. This method would be helpful in understanding the fatigue problem and optimizing the system design to alleviate the fatigue in SSVEP-based BCIs.