Estimation and quantification of vigilance using ERPs and eye blink rate with a fuzzy model-based approach

Shabnam Samima, Monalisa Sarma, Debasis Samanta, Girijesh Prasad

Research output: Contribution to journalArticlepeer-review

11 Citations (Scopus)
357 Downloads (Pure)


Vigilance, also known as sustained attention, is defined as the ability to maintain concentrated attention over prolonged time periods. Many methods for vigilance detection, based on biological and behavioural characteristics, have been proposed in the literature. In general, the existing approaches do not provide any solution to measure vigilance level quantitatively and adopt costly equipment. This paper utilizes a portable electroencephalography (EEG) device and presents a new method for estimation of vigilance level of an individual by utilizing event-related potentials (P300 and N100) of EEG signals and eye blink rate. Here, we propose a fuzzy rule-based system using amplitude and time variations of the N100 and P300 components and blink variability to establish the correlation among N100, P300, eye blink and the vigilance activity. We have shown, with the help of our proposed fuzzy model, we can efficiently calculate and quantify the vigilance level, and thereby obtain a numerical value of vigilance instead of its mere presence or absence. To validate the results obtained from our fuzzy model, we performed subjective analysis (for assessing the mood and stress level of participants), reaction time analysis and compared the vigilance values with target detection accuracy. The obtained results prove the efficacy of our proposal.
Original languageEnglish
Pages (from-to)517-533
Number of pages17
JournalCognition, Technology and Work
Early online date2 Nov 2018
Publication statusPublished (in print/issue) - 1 Aug 2019


  • Electroencephalography (EEG)
  • Event Related Potentials
  • Eye blink rate
  • Fuzzy model
  • Vigilance detection


Dive into the research topics of 'Estimation and quantification of vigilance using ERPs and eye blink rate with a fuzzy model-based approach'. Together they form a unique fingerprint.

Cite this