Loading...
Please wait, while we are loading the content...
Similar Documents
Assessing alertness from EEG power spectral bands
| Content Provider | Semantic Scholar |
|---|---|
| Author | Rueda, Robin Gerardo Álvarez |
| Copyright Year | 2006 |
| Abstract | The assessment of low level of alertness and drowsiness conditions of humans, while performing critical tasks, requires the development of automatic detection systems to work in real time, to be as pervasive as possible for long lasting periods of use and robust enough to cope with a wide intraand inter-individual variability. A new alertness detection procedure based on the spectral analysis of the EEG signal is proposed, mostly concerned with the provision of robust classification criteria under the working conditions depicted above. The wide interindividual variability has been reduced down to operational levels by means of a personal dependant normalization algorithm, which consists of describing the EEG spectral morphology as a function of the alpha behaviour of each subject. With this approach, drowsiness classification can be achieved by simple thresholding of the EEG spectral variable selected: the power ratio between a high frequency and an alpha bands defined for each individual. Variable that has been selected to optimize the discriminant power and its interindividual stability. The experimental results include the selection of the preferred recording sites and the demonstration of the reliability of the classification criteria along the time for each individual. The paper also analyses the time resolution of the algorithms to assure their real time operation. Technological requirements of the method proposed allow concluding that the design of a wearable one single EEG lead nonintrusive device it is feasible to reliably discriminate continuously drowsiness situations. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://bibdigital.epn.edu.ec/bitstream/15000/9872/1/2006AJIEE-22.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
