Skip to main content
. 2011 Aug 22;5:99. doi: 10.3389/fnins.2011.00099

Table 1.

Subject-specific data and different measures of spelling performance.

VPnv VPnw VPnx VPny VPnz VPmg VPoa VPob VPoc VPod VPja VPoe Avg
Sex m m w w m M m m m m m w
Age 26 21 25 23 34 23 23 24 24 25 29 24 25.1
Binary cl. acc. 77.0 74.8 52.7 75.0 74.4 60.2 72.2 78.6 79.6 70.0 82.0 73.2 72.5
|Diff| in counts 23 13 18 35 5 33 16 6 14 10 1 26 16.7
#Selectionsa 29 31 23 29 38 26 28 23 28.4
#False sel.a 2 4 0 2 5 1 3 0 2.1
Time (min)a 25.1 23.0 15.4 21.7 26.9 18.1 19.1 17.9 20.9
#Selectionsb 63 53 97 51 49 45x 61 49 48 57.3
#False selb 7 4 27 5 3 1x 8 3 4 6.9
Time (min)b 47.1 36.5 76.7 36.8 39.5 30.9x 48.9 36.2 38.6 43.5

Results based on offline data have a gray background and binary classification accuracy is given in classwised balanced accuracy. Each subject spelled the same two sentences: a short (18 char) sentencea and a long (36 char) sentenceb. Spelling performance is quantified by the time (min), required to spell the complete sentence. Individual differences arise due to false selections. For participants VPnv, VPoc, and VPoe, the second sentence was canceled or not even started. One spelling runx was stopped after the 45th trial.