File:Fneur-12-750667-g001.jpg

English: Figure caption quoted from Frohlich et al. 2021: "EEG burst-suppression positively predicts outcomes at chronic follow up [...] Outcomes were assessed using the Glasgow Outcome Scale extended (GOSe) and were divided at discharge and chronic timepoints using median splits (dotted lines, GOSe scores are scattered with jitter to avoid overlap). Patients with burst-suppression (red circles) were overrepresented in the upper half of each median split (blue circles = no burst-suppression; circle size is proportional to patient's highest Glasgow Coma Scale score). Note that 2 patients missing chronic GOSe scores (Patients 4 and 10, Table 1) are not shown in the scatter plot but are represented in the discharge GOSe histogram. GOSe scores at discharge showed low spread, with most patients either scoring 2 (vegetative state) or 3 (low severe disability). Similarly, chronic GOSe scores showed a bimodal distribution with peaks flanking a large dip at 4 (upper severe disability). Owing to this fact, we opted for a median split on scores (discharge: 1-2 vs. 3-4; chronic: 1-4 vs. 5-8) to create dichotomous outcome variables for logistic regression. In the upper half of both the discharge and chronic GOSe median split, 40% of patients displayed BSR > 0.5, vs. 0% (discharge) and 6.7% (chronic) of patients in the lower half of the GOSe distribution."