Skip to main content
. 2021 Apr 7;38(5):2513–2531. doi: 10.1007/s12325-021-01718-6

Table 2.

Diagnostic performance of two versions of the Cognalyzer® algorithm after participants with missing EEG recordings or who had electrode placement errors and were removed from the population (PP2, n = 62)

Diagnostic performance of cognalyzer®
Post-inhalation Pre-inhalation Total Sensitivity (95% CI) Specificity (95% CI) Accuracy (95% CI) Percent false positive (95% CI) Percent false negative (95% CI) PPV (95% CI) NPV (95% CI) p value
Cognalyzer®—algorithm version 1
 Cognalyzer® version 1—post-inhalation 54 10 64 87.1 (76.1–94.3) 83.9 (72.3–92) 85.5 (78–91.2) 16.1 (8–27.7) 12.9 (5.7–23.9) 84.4% (73.1–92.2) 86.7% (75.4–94.1) 0.637
 Cognalyzer® version 1—pre-inhalation 8 52 60
 Total 62 62 124
Cognalyzer®—algorithm version 2
 Cognalyzer® version 2—post-inhalation 55 13 68 88.7 (78.1–95.3) 79.0 (66.8–88.3) 83.9 (76.2–89.9) 21 (11.7–33.2) 11.3 (4.7–21.9) 80.9% (69.5–89.4) 87.5% (75.9–94.8) 0.180
 Cognalyzer® version 2—pre-inhalation 7 49 56
 Total 62 62 124

McNemar’s chi-squared test was used to compare the difference in performance between the diagnostic test being evaluated and the comparison standard