Table 3.

Summary of the statistical parameters of the models C vs non-responder (NR), C vs responder (R), and R vs NR.

	Orthogonal partial least square discriminant analysis models					Permutationd
Groups	Componentsa	R2Xcumb	R2Ycumb	Q2cumc	p	R2 intercept	Q2 intercept
Controls vs NR	1P + 1O	0.518	0.850	0.824	<0.001	0.140	−0.360
Controls vs R	1P + 1O	0.545	0.739	0.566	<0.001	0.316	−0.490
Responders vs NR	1P + 1O	0.508	0.631	0.467	<0.001	0.278	−0.395

^aThe number of predictive and orthogonal components used to create the statistical models.

^bR²X and R²Y indicated the cumulative explained fraction of the variation of the X block and Y block for the extracted components.

^cQ²cum values indicated cumulative predicted fraction of the variation of the Y block for the extracted components.

^dR² and Q² intercept values are indicative of the validity of the model. The permutation test was evaluated on the corresponding partial least square discriminant analysis model.