Table 4.

Characterization of the Distortion of the Phosphene Maps Relative to the Retinotopic Electrode Locations Using Procrustes Analysis

			Translation (° Visual Arc)
Subject	Task	Scale Factor	Horizontal	Vertical	Rotation (°)
S1	Gaze-based (nonimplanted eye)	0.72**	0.0	4.4	−5.5
	Gaze-based (implanted eye)	0.57**	0.0	2.5	2.2
	Finger-based (early time-point)	0.37**	0.0	11.7	−29.4
	Finger-based (touchscreen)	0.63**	0.3	4.8	−6.0
S2	Gaze-based (nonimplanted eye)	0.99*	0.0	0.2	−7.3
	Gaze-based (implanted eye)	0.87***	0.0	−0.2	−2.2
	Finger-based (early time-point)	0.67**	2.0	1.0	−8.8
S3	Gaze-based (nonimplanted eye)	0.07	16.9	10.2	22.4
	Gaze-based (implanted eye)	0.06	15.7	10.0	15.5
	Finger-based (early time-point)	0.83*	3.9	5.7	−1.7

For each phosphene map, the linear transformation (translation and scaling) of the electrode locations that minimized the mean squared error between the phosphene and electrode locations was computed. Scale factors were tested for significance against a bootstrap distribution derived from a random permutation of the electrode locations. For S3 gaze-based mapping, the nonsignificant scale factor near zero indicates that no linear transform mapping the electrode locations to the measured phosphene locations was found.

^*P < 0.05.

^** P < 0.01.

^*** P < 0.001.