Table 5.

The ten most discriminating parameters according to the Fisher (F-) coefficient and corresponding Wilcoxon test p-values.

White matter
Most Discriminative Texture Parameters on Patients	p-values (patients)	p-values (controls)	Most Discriminative Texture Parameters on Controls	p-values (controls)	p-values (patients)
LEVEL 2
S(4,4)Correlat	<0.001*	0.492	S(5,5)SumVarnc	0.002*	0.008*
S(4,4)Contrast	0.001*	0.557	S(5,5)Correlat	0.004*	0.008*
S(4,4)SumVarnc	<0.001*	0.695	S(4,4)InvDfMom	0.064	0.003*
S(4,4)InvDfMom	0.003*	0.064	WavEnLH_s-3	0.922	0.359
S(4,4)DifVarnc	0.002*	0.770	S(5,5)Contrast	0.010*	0.015*
S(5,5)Correlat	0.008*	0.004*	S(0,1)SumAverg	0.164	0.114
S(2,-2)AngScMom	<0.001*	0.396	S(1,-1)SumAverg	0.105	0.225
S(0,3)DifEntrp	0.002*	0.695	S(5,-5)SumVarnc	0.105	0.253
S(5,5)Contrast	0.015*	0.010*	S(5,-5)Correlat	0.105	0.603
S(5,5)SumVarnc	0.008*	0.002*	Teta1	0.064	0.274

The ten most discriminating parameters according to the Fisher (F-) coefficient and their corresponding Wilcoxon test p-values are calculated between hemispheres in the cerebral white matter in the level of corona radiata (level 2) for both patients and controls.

Correlat, correlation; DiffVarnc, difference variance; AngScMom, angular second moment; Teta, vectors of autoregressive model; SumAverg, sum average; InvDfMom, inverse difference moment; SumVarnc, sum variance; DifEntrp, difference entropy; WavEnLH, energy of wavelet coefficients in subband LH.

* p-values of under 0.05 are considered statistically significant