Abstracts

Rationale: We have previously shown that almost all electroencephalographic generalized seizures after lateral fluid-percussion (FPI) induced traumatic brain injury (TBI) occur in the transition between stage III and REM sleep. Our objective was to investigate in a larger cohort of rats whether (a) the change in the characteristics of sleep spindles would serve as diagnostic biomarkers for post-traumatic epilepsy (PTE) and (b) the presence of brain abscess in a subpopulation of animals would affect the sleep spindle duration. Methods: From the total cohort of 251 adult male Sprague-Dawley rats, 110  video-EEG-monitored animals (86 TBI, 24 sham) were included in the analysis. A 3-wk video-EEG monitoring was started at 3 months post-TBI in order to assess occurrence of spontaneous seizures and sleep disturbances. For sleep analysis, a 24-h continuous EEG recording was divided into 2 947 epochs (30 sec each). Sleep stages and graphoelements (spindles) during the transition between stage III and REM sleep were identified. Duration of each individual spindle was measured. Results: Spindle duration (SD) was shorter in the TBI than the Sham group (p=0.029). Within the TBI group, rats with post-traumatic epilepsy (PTE)(n=26) had shorter SDs than rats without epilepsy (n=60)(p=0.0001). ROC analysis showed that spindle duration differentiated TBI animals with or without epilepsy with high sensitivity and specificity as the area under the curve (AUC) was 0.884 (sensitivity 81% at 97% specificity; SD 0.05). However, their seizure frequency was higher than in PTE group without abscess (p=0.002). Conclusions: Our data show that the shortening of SDs during the transition from stage III to REM sleep differentiates epileptic from non-epileptic animals after TBI. In animals with PTE, the presence of brain abscess as a dual pathology does not augment SD shortening, but does increase the frequency of seizures. Funding: No funding.