Skip to main content
. 2021 Jun 30;3:688122. doi: 10.3389/fdgth.2021.688122

Table 5.

Resulting confusion matrices of (A) condition Fpz+EOG test-retest (missing participant number 5, since not included in Fpz+EOG 1st Rating), (B) condition Fpz+EOG vs. cEEGrid and (C) condition Fpz+EOG vs. cEEGrid+EOG.

(A) Fpz+EOG test-retest Art W N1 N2 N3 REM
Art 63 19 1 1 0 5
W 4 927 96 18 2 8
N1 0 58 370 260 0 53
N2 0 117 168 2,870 226 23
N3 1 7 2 39 1,495 0
REM 0 102 91 104 1 1,562
(B) Fpz+EOG vs. cEEGrid Art W N1 N2 N3 REM
Art 100 8 2 19 1 2
W 63 769 47 135 10 98
N1 27 86 257 306 3 60
N2 135 71 133 2,958 244 82
N3 2 2 1 198 1,753 1
REM 83 23 70 399 2 1,191
(C) Fpz+EOG vs cEEGrid+EOG Art W N1 N2 N3 REM
Art 70 13 21 14 0 14
W 84 714 85 133 3 103
N1 29 51 300 295 1 63
N2 63 11 96 3,288 121 44
N3 6 3 1 241 1,705 1
REM 35 14 75 126 0 1,518

Each count refers to the annotation of a single 30 second segment. The respective hypnograms of all participants were concatenated to create the confusion matrices. Only the first 5 h and 20 min of rating results of participant number 1 entered the analyses of (B, C), since several cEEGrid electrodes separated from the skin during the second half of the night.