Automatic detection of non-convulsive seizures using AR modeling

Author

Khan, Yusuf Uzzaman ; Farooq, Omar ; Tripathi, Meenakshi ; Sharma, Parmanand ; Alam, P.

Author_Institution

Biosignal Res. Div., Aligarh Muslim Univ., Aligarh, India

fYear

2012

fDate

17-19 Dec. 2012

Firstpage

1

Lastpage

4

Abstract

This paper proposes an algorithm for the detection of non-convulsive seizures (NCSz), based on the Autoregressive (AR) modeling of EEG over the frequency band of 0-31 Hz. Simple linear classifier was used for the classification between the normal and seizure EEG. The algorithm was tested on the scalp EEG database of 5 patients, collected at All India Institute of Medical Sciences (AIIMS), New Delhi. The database consists of 13 seizures of different duration. The results were reported for 11 seizures. Overall sensitivity and specificity achieved by the method was 86.8% and 96.9% respectively.

Keywords

autoregressive processes; electroencephalography; medical signal detection; signal classification; AIIMS; AR modeling; All India Institute of Medical Sciences; NCS; New Delhi; automatic detection; autoregressive modeling; frequency 0 Hz to 31 Hz; linear classifier; nonconvulsive seizures; normal EEG; scalp EEG database; seizure EEG; AR modeling; EEG; Linear classifier; Linear predictor coefficients (LPC); Non-convulsive seizures; Reflection coefficients;

fLanguage

English

Publisher

ieee

Conference_Titel

Power, Control and Embedded Systems (ICPCES), 2012 2nd International Conference on

Conference_Location

Allahabad

Print_ISBN

978-1-4673-1047-5

Type

conf

DOI

10.1109/ICPCES.2012.6508099

Filename

6508099