Simple regression–linear model: Independent variable, RNFL; dependent variables, NAA, N-acetyl-aspartate; Cho, choline; Cr, creatine; LL per BV, lesionload per Brain Volume, LL AR, lesion load along anterior right visual pathway; LL AL, lesion load along anterior left visual pathway; LL PR, load along posterior right visual pathway; LL PL, load along posterior left visual pathway; Evan’s Index; CHI; CMI; BCI; the maximum width of the 3rd ventricle; the maximum width of the 4th ventricle; MIF, the maximum width of the anterior interhemispheric fissure; MFSS, the maximum frontal subarachnoid space; MSF, the maximum width of the Sylvian fissure; DD, disease duration; EDSS, expanded disability severity scale. 1
st
row: all right eyes (n = 28; with and without ON) of all included MS patient. 2
nd
row: all left eyes (n = 28; with and without ON) of all included MS patient. 3
rd
row: all right eyes of MS patients who never experienced an ON (neither on their right nor on their left eye; RRMS, n = 17, SPMS, n = 2; Table 1). 4
th
row: all left eyes of MS patients who never experienced an ON (neither on their left nor on their right eye; RRMS, n = 17, SPMS, n = 2; Table 1). Patients are the same as in the 3rd row. 5
th
row: right eyes of 6 MS patients who experienced an ON on their right eyes (note, 4 out of 6 experienced ON on both eyes, 2 only on their right eyes; Table 1). 6
th
row: left eyes of 6 MS patients who experienced an ON on their left eyes (note, 4 out of 7 experienced ON on both eyes, 3 only on their left eyes Table 1). For each analysis the correlation coefficient (corr. coeff.), R-squared (percent), the standard error of estimate (STE of Est.) and the p-value (analysis of variance, ANOVA) is given. Since the p-value in the ANOVA table is less than 0.01, there is a statistically significant relationship between the maximum width of the 4th ventricle and the RNFL (for all patients’ right eyes, n = 28, 1st row and for all patient’s left eyes, who never experienced ON, n = 17, 4th row) at 99% confidence level. However, the low correlation coefficient indicates that there is only a weak relationship between the variables. R-squared statistic indicates that the simple/linear regression explains only 24.92% (1st row) or 28.68% (4th row) of the variability of the independent variable. In all other analyses presented here (and performed for the six OCT-Sectors, see material and methods or Table 5) no statistically significant correlation could be found (data not shown).