Feasibility of online seizure detection with continuous EEG monitoring in the intensive care unit.

INTRODUCTION: Continuous EEG (cEEG) is of great interest in view of the reported high prevalence of non-convulsive seizures on intensive care units (ICUs). Here, we describe our experiences applying a seizure warning system using cEEG monitoring. METHODS: Fifty comatose ICU patients were included prospectively and monitored. Twenty-eight patients had post-anoxic encephalopathy (PAE) and 22 had focal brain lesions. A measure of neuronal interactions, synchronization likelihood, was calculated online over 10s EEG epochs and instances when the synchronization likelihood exceeded a threshold where marked as seizures. RESULTS: Five patients developed seizures. Our method detected seizures in three patients, in the other patients seizures were missed because of they were non-convulsive and had a focal character. The average false positive rate was 0.676/h. DISCUSSION: This is our first attempt to implement online seizure detection in the ICU. Despite problems with artifacts and that we missed focally oriented seizures, we succeeded in monitoring patients online. Given the relatively high occurrence of seizures, online seizure detection with cEEG merits further development for use in ICUs.

Motor evoked potential: a reliable and objective measure to document the functional consequences of multiple sclerosis? Relation to disability and MRI.

OBJECTIVE: In an attempt to analyze whether MEP can serve as a valid measure for evaluating neurological dysfunction in multiple sclerosis (MS), we related MEP to clinical and MRI measures. METHODS: Transcranial magnetic stimulation was applied in 52 MS patients to determine the central motor conduction time (CMCT) to the extremities. We calculated Z-scores for each CMCT (Zcmct) corrected for height. All patients underwent two clinical measurements and a MRI scan, of which T1 and T2 brain lesion volumes, brain volume, spinal cord volume and the number of T2 spinal cord lesions were derived. RESULTS: The clinical measurements correlated significantly with various Zcmct (Spearman correlation coefficients ranged from 0.29 to 0.53; p<0.05). The number of spinal cord lesions, brain T1 and T2 lesion volume and spinal cord volume correlated with various Zcmct (r=0.31-0.53; p<0.05). Linear regression analysis revealed that the clinical measurements were explained by Zcmct left leg and T1 lesion volume (adjusted R(2)=0.38). For one clinical measurement the number of spinal cord lesions was also included (adjusted R(2)=0.43). CONCLUSION: We found a relation between MEP, brain and spinal cord MRI measures, and two clinical measures. Moreover, a model for explaining disability in MS revealed that MEP measures provide information in addition to MRI measures. SIGNIFICANCE: This study suggests that MEP is a measure that might adequately reflect pathology and neurological dysfunction in MS.

Neonatal seizure monitoring using non-linear EEG analysis.

Birth asphyxia is a major concern in neonatal care. Epileptic seizures are associated with subsequent neurodevelopmental deficits. Eighty-five percent of these seizures remain subclinical and therefore an on-line monitoring device is needed. In an earlier study we showed that the synchronization likelihood was able to distinguish between neonatal EEG epochs with and without epileptic seizures. In this study we investigated whether the synchronization likelihood can be used in complete EEGs, without artifact removal. Twenty complete EEGs from 20 neonatal patients were studied. The synchronization likelihood was calculated and correlated with the visual scoring done by 3 experts. In addition, we determined the influence of seizure length on the likelihood of detection. Using the raw unfiltered EEG data we found a sensitivity of 65.9 % and a specificity of 89.8 % for the detection of seizure activity in each epoch. In addition, the seizure detection rate was 100 % when the seizures lasted for 100 seconds or more. The synchronization likelihood seems to be a useful tool in the automatic monitoring of epileptic seizures in infants on the neonatal ward. Due to the retrospective nature of our study, the consequences for clinical intervention cannot yet be determined and prospective studies are needed. Therefore, we will conduct a prospective study on the neonatal intensive care unit with a recently developed on-line version of the synchronization likelihood analysis.

Predictive value of EEG in neonates with periventricular leukomalacia.

The aim of this study was to evaluate whether EEG (i.e. positive Rolandic sharp waves) can be used to predict neurodevelopment in newborn infants with periventricular leukomalacia and compare the predictive value with that of MRI. A sequential cohort of neonates (n=45; 33 males, 12 females; mean gestational age 31.2 weeks, SD 2.7, range 27 to 37.8 weeks; mean birthweight 1592 g, SD 601 g) with periventricular hyperechogenicities on cranial ultrasound was recruited for this study. EEGs were analyzed for positive Rolandic sharp waves. Neurodevelopment was evaluated at the ages of 12 and 18 months. In the whole group the probability of a poor outcome was 24% and the probability of any impairment was 33%. If the number of positive Rolandic sharp waves was no more than 0.1 per minute, the probability of a poor outcome was reduced to 9% (95% confidence interval [95%CI] 2 to 27%) and the probability of any impairment was reduced to 13% (95%CI 4 to 32%). In all infants with more than 0.1 positive Rolandic sharp waves per minute the probability of a poor outcome was 41% (95%CI 23 to 61%) and of any impairment was 55% (95%CI 34 to 73%). In these infants MRI identified infants with a poor outcome with a sensitivity of 1.00 (95%CI 0.70 to 1.00) and a specificity of 0.92 (95%CI 0.67 to 0.99), and infants with any impairment with a sensitivity of 0.83 (95%CI 0.55 to 0.95) and a specificity of 1.00 (95%CI 0.72 to 1.00). Results suggest that if an EEG of an infant with periventricular leukomalacia contains no more than 0.1 positive Rolandic sharp waves per minute the probability of a normal or mildly delayed development is high (0.91, 95%CI 0.73 to 0.98). MRI enhances the accuracy of the outcome prediction slightly; however, owing to a wide confidence interval, this advantage is negligible. However, if the frequency of the positive Rolandic sharp waves exceeds 0.1per minute, MRI can significantly enhance the precision of the prediction of outcome.