NONLINEAR APPROXIMATE ENTROPY ANALYSIS OF BRAIN ELECTRICAL ACTIVITY IN A GENERALIZED EPILEPSY ANIMAL MODEL
Abstract number :
3.074
Submission category :
Year :
2002
Submission ID :
1289
Source :
www.aesnet.org
Presentation date :
12/7/2002 12:00:00 AM
Published date :
Dec 1, 2002, 06:00 AM
Authors :
Deng-Shan Shiau, J. Chris Sackellares, Leon D. Iasemidis, Meadow F. Maze, Paul R. Carney. Neuroscience, University of Florida, Gainesville, FL; Pediatric Neurology, University of Florida, Gainesville, FL; Biomedical Engineering and Neurology, University o
RATIONALE: Approximate entropy (ApEn), a statistic for quantifying the system regularity/complexity of a time series, has been applied on medical data analysis, such as in heart rate analysis (PNAS 1991; 88: 2297). This methodology allows for distinguishing normal from abnormal data. In theory, the smaller the ApEn value, the less ordered is the system. In this study, we apply ApEn on continuous video-EEG recordings in an animal model of generalized epilepsy in which the H218/AGR16/edg-5/LPB2 sphingosine 1-phosphate gene has been disrupted and a littermate control mouse. The aim of this study was to test the hypothesis that H218 mice and littermate controls can be distinguished by both the mean values and the variability of ApEn. After reviewing this abstract, the participants should be able to distinguish an animal model of generalized epilepsy and control mice by the complexity (or the changes of complexity) of their EEG signals.
METHODS: Continuous daily bifrontal and bilateral hippocampal intracranial video-EEG recordings were obtained in postnatal days (P) 18-25 H218 mice (n = 8) and in littermate controls (n = 2). Seizures were characterized by generalized spike-and-wave discharges (SWD) with categorical behavioral changes. ApEn were estimated in every 10 seconds non-overlapping window of EEG signals for each electrode site in three postnatal age intervals of H218 mice (P18; seizure vulnerable period, P21 and P25; seizure free periods) and one postnatal age interval in an age-matched P18 littermate control. At each age (P18, P21, and P25), mean value and the variance of ApEn were compared with a P18 control by employing the standard two-sample t test and F test for mean and variance equality, respectively.
RESULTS: Mean value and variance of ApEn were 1.582 and 0.016, respectively in littermate controls. Mean values were 1.418 at P18 during the seizure vulnerable period, 1.372 at P21 and 1.328 at P25 in H218 mice. Variances were 0.042 at P18 during the seizure vulnerable period, 0.035 at P21, and 0.043 at P25 in H218 mice. For mean values and variances for P18-25 H218 mice were significantly different (p-value [lt] 0.0001) from age-matched littermate controls.
CONCLUSIONS: (1) Mean ApEn values were significantly lower in H218 mice than in littermate controls. (2) Variances of ApEn values were significantly higher in H218 mice than in littermate control. These results suggest that brain dynamics is more ordered (or less complex) in H218 seizure-prone mice, and the complexity of the brain dynamics has less variability in littermate controls. These findings may help to develop a simple algorithm to distinguish the normal and abnormal states in epileptic patients.
[Supported by: NIH/NINDS NS 039687
Veterans Affairs
University of Florida Division of Sponsored Research
Children[ssquote]s Miracle Network]