Table 1 Data distribution categories defined after data reduction and clustering solution inspection.
From: Unsupervised EEG preictal interval identification in patients with drug-resistant epilepsy
Category | Definition |
|---|---|
Category 1 | There is no evidence of a preictal structure. There might be a clear separation into a smaller cluster, but that either comprises samples separated in time or comprises samples strictly located previously to the 120 min before seizure onset |
Category 2 | Separation into two evenly distributed clusters that might indicate some external interference, such as the transition of the sleep-wake cycle |
Category 3 | Clear separation into two differently sized clusters, the smaller one resembling a preictal interval located within the 120 min before seizure onset |
Category 4 | Data distribution indicating progression over time, with samples following a temporal trajectory |
Category 5 | It seems that a smaller cluster can be identified, but it would be difficult to isolate it in a cluster using clustering methods |
Category 6 | Category assigned when the clustering solution comprises more than two clusters that may indicate the existence of brain multistates and even progression over time. The preictal interval is represented by the cluster located within 120 min before seizure onset and nearest to the onset. It might be possible to observe evidence of sleep stage transition, preictal interval aside |
