Abstracts

Rationale: Reliable computerized detection of epileptiform transients (ETs), characterized by interictal spikes and sharp waves in the EEG, is a useful goal since this would assist neurologists in interpreting EEGs. To demonstrate the need for such automated systems, we collected inter-rater scoring data from a group of academic clinical neurophysiologists (ACNs) and private practice neurologists Methods: Two hundred 30-second routine scalp EEG segments from 200 different patients were included in the study. EEG segments were divided as follows: 100 EEG segments containing subtle ETs or benign paroxysmal activity such as exaggerated alpha activity, wicket spikes, and small sharp spikes that may be misinterpreted as abnormal; 50 EEG segments retrieved from consecutive normal EEG recordings; and 50 EEG segments retrieved from consecutive abnormal EEGs interpreted as containing ETs. Level of fellowship training, years of practice, and board certification status were collected for all participants. EEG scoring was performed using EEGnet, a distributed web-based platform for the analysis of scalp EEG recordings. Participants were asked to mark the location of all ETs in the recording on the channel on which they thought it was best represented. All marked ETs were clustered to determine common ET markings between scorers. For inter-rater analysis, we examined linearly weighted Gwet's agreement coefficient (AC2) and Matthew's correlation coefficient (MCC). The AC2 agreement between each scorer and events marked by at least 20% of other scorers was assessed. All members of the Critical Care EEG Monitoring and Research Consortium and all private practice neurologists in the states of SC, NC, GA, and parts of FL were invited to participate. Results: 19 academic neurophysiologists and 16 private practice physicians participated. Each scorer marked an average of 68 ETs (range 6 ?" 212). For all events, AC2 was 0.81 and for events marked by at least two scorers was 0.63. Participants who were board certified by the American Board of Clinical Neurophysiology (ABCN) had a higher AC2 than those not certified (0.61 versus 0.46, p < 0.05). MCC for scorers with ABCN certification was higher than for those without (0.16 versus 0.10, p < 0.02). AC2 performance did not correlate with length of fellowship training. MCC was higher for scorers with at least 1.5 years of fellowship training (0.16 versus 0.09, p < 0.01).Length of fellowship training correlated moderately with performance based on MCC (Spearman rho = 0.42, p < 0.01). Conclusions: Inter-rater agreement for labeling ETs shows substantial variability among neurologists and is overall moderate. Neurologists with board certification by the ABCN and with neurophysiology fellowship training are more proficient at labelling ETs. Funding: None