Abstracts

Rationale: Our previous study identified that automated spike detection software algorithms have a sensitivity less than 0.50 when using a standard bipolar montage. Spike detection software, using state-dependent spike detection, is montage dependent. This study addresses whether any specific montage increases sensitivity of automated spike detection.Methods: All patients admitted to our seizure monitoring unit over a two month period were included in the study, however patients with a confirmed diagnosis of psychogenic non-epileptic spells were excluded. Two hours worth of data from sleep (0400-0600 hours) were evaluated using the most sensitive automated detection parameters we previously identified, including state-dependent spike detection with an amplitude threshold of 4 within the Stellate Harmonie software. Four montages were chosen for this analysis (i.e., average, bipolar, combination, and transverse). The automated spike detections were compared to visual review of the entire two hour file by an experienced epileptologist and then compared across different montages to determine any effect on sensitivity of spike detection. One-way repeated measures ANOVA and pairwise comparisons were used to identify differences between each montage selection.Results: The mean spikes detected for each montage was maximally 0.552, which is higher than previously reported. The one-way repeated measures ANOVA demonstrated significant differences in sensitivity between montages, F (3, 59) = 6.215, p < .01. Furthermore, paired comparisons demonstrated significant differences in sensitivity between all montages except bipolar and transverse montages and combination and average montages. Bipolar montage had the lowest sensitivity, while combination montage had the highest.Conclusions: These pilot data suggest that montage selection, among other parameters such as state dependent spike rejection and amplitude threshold, is important during automated spike detection to increase sensitivity.