. 2011 Apr 6;6(4):e18262. doi: 10.1371/journal.pone.0018262

Table 1. Results of the multivariate linear model.

Dependent	Block	R²	Independents and partial coefficients
			FNum	Φ	SKW	KUR	AGV	SAP	PKV
Fielder	VB	0.864¹	−0.045¹	0.007²			−0.001²
	AB	0.703¹		−0.009¹	0.001¹	0.026²
a	VB	0.894¹	0.166¹					−0.037³
	AB	0.790¹	0.208¹		−0.288³
b	AB	0.309¹	−0.002¹		0.035³			−0.002²	2e-5³
c	VB	0.902¹	0.010¹				0.0001²
	AB	0.827¹		0.003¹		−0.006¹
β ₀	AB	0.822¹	0.240¹		0.354³
β ₁	VB	0.798¹		−0.246¹
	AB	0.889¹	−0.877¹		−1.090³
β ₃	VB	0.690¹		0.067¹
	AB	0.822¹	0.240¹		0.354³
β ₄	AB	0.111¹			0.598¹

The Fiedler and spectral embedded metrics of scan path were predicted by main scan path parameters in a stepwise multivariate linear regression model. VB, the valence block; AB, the arousal block; R², the adjusted coefficient of determination; FNum, the fixation number; Φ, the scan path diameter; AGV, the average velocity of saccade; SKW, the weight skewness; KUR, the weight kurtosis; SAP, saccade amplitude; PKV, the peak velocity of saccade.

p<0.001, by F-test for the significance of the regression model or by t-test for the significance of partial coefficients.

p<0.01, by t-test for the significance of partial coefficients.

p<0.05, by t-test for the significance of partial coefficients.