Objects exhibit body model like shape distortions

Journal of Experimental Brain Research

Aurelie Saulton1, Trevor J. Dodds1, Heinrich H. Bülthoff1,2, Stephan de la Rosa1

1 Human Perception, Cognition and Action,

Max Planck Institute for Biological Cybernetics,

Tübingen, Germany

2Department of Brain and Cognitive Engineering,

Korea University,

Seoul, South Korea

Correspondence should be addressed to Aurelie Saulton or Heinrich H. Buelthoff:
Aurelie Saulton
Max Planck Institute for Biological Cybernetics
Spemannstr. 38
72076 Tübingen
Germany
Email:
Tel: +49 (0) 7071 601645

Heinrich H. Bülthoff
Max Planck Institute for Biological Cybernetics
Spemannstr. 38
72076 Tübingen
Germany
Email:
Tel: +49 (0) 7071 601 601

Supplementary material

Hand / Rake / Post-it / Box
Upright / Rotated / Upright / Rotated / Upright / Rotated / Upright / Rotated
t(15)=11.81
p<.001
r=0.95 / t(15)=5.9
p<.001
r=0.84 / t(15)=4.28
p=.0020
r=0.74 / t(15)=3.17
p=.012
r=0.63 / t(15)=3.84
p=.0032
r=0.70 / t(15)=-0.75
p=.46
r=0.19 / t(15)=2.89
p=.011
r=0.60 / t(15)=-5.61
p<.001
r=0.82

S1:Comparison between width and length dimensions.The table cells provide the t statistic, p value and the effect size measure r. P-values are Holm-corrected. Positive t values indicate that width estimates were larger than length.

Rake / Post-it / Box
Hand / t(15)=2.77
p=.014
r=.58 / t(15)=3.66
p=.0047
r=.69 / t(15)=4.21
p=.0023
r=.74
Rake / - / t(15)=2.27
p=.042
r=.51 / t(15)=2.58
p=.042
r=.55
Post-it / - / - / t(15)=1.42
p=.18
r=.34

S2: Differences between items in the localization task.The table cells provide the t statistic, p value and the effect size measure r. P values are Holm-corrected and based on planned comparisons:namely, the three comparisons resulting from comparing the hand to the three objects as well as the two comparisonsresulting from comparing the rake to the post-it and the Box. Significant p values highlighted with a grey cell background.

Item
Dimensions / Hand / Rake / Post-it / Box
Length
upright / M= -38.46%
t(15) = -9.79 p<.001
r=.93 / M= -19.69%
t(15) = -3.47
p=.0034
r=.67 / M= -29.93%
t(15) = -9.18
p<.001
r=.92 / M= -32.49%
t(15) =-13.59
p<.001
r=.96
Length
rotated / M= -38.46%
t(15) = -9.62 p<.001
r=.93 / M= -23.03%
t(15) = -4.08
p=.002
r=.72 / M= -18.68%
t(15) = -7.18
p<.001
r=.88 / M= -24.66%
t(15) = -12.06
p<.001 r=.95
Width
upright / M= 19.52%
t(15) = 3.86 p=.0031
r=.70 / M= 14.37%
t(15) = 2.46
p=.053
r=.54 / M= -18.17%
t(15) = -5.50
p<.001
r=.82 / M= -24.75%
t(15) =-6.95
p<.001
r=.87
Width
rotated / M= 2.3%
t(15) = .48 p=.64
r=.12 / M= 6.87%
t(15) = .94
p=.36
r=.24 / M= -20.74%
t(15) = -8.26
p<.001
r=.90 / M= -34.51%
t(15) = -19.69
p<.001
r=.98

S3: Holm corrected paired t-tests comparing percent overestimation of each item’s length and width against baseline (baseline corresponding to the actual size of the item) and for each orientation (upright and rotated). The table cells provide the mean value, t statistic, p value and the effect size measure r.Positive mean values are highlighted with a grey cell background.

Comparison
Dimensions / Orientation / Hand vs. Rake / Hand vs. Post-it / Hand vs. Box
Length / Upright / t(15)=-3.32
p=.014
r=.65 / t(15)=-1.91
p=.15
r=.44 / t(15)=1.42
p=.17
r=.34
Rotated / t(15)=-3.33
p=.0046
r=.65 / t(15)=-4.67
p<.001
r=.77 / t(15)=3.66
p=.0046
r=.69
Width / Upright / t(15)=.96
p=.35
r=.24 / t(15)=7.77
p<.001
r=.89 / t(15)=-8.97
p<.001
r=.91
Rotated / t(15)=-.65
p=.52
r=.17 / t(15)=-4.67
p<.001
r=.77 / t(15)=-7.78
p<.001
r=.89

S4:Holm corrected paired t-tests comparing the magnitude of the distortions (percent overestimation) between hand and other items (Rake, Post-it, Box). This was done for each dimension separately (length and width) within each orientation (upright and rotated). The difference between the hand and rake distortion is highlighted in grey. Interestingly, within each orientation, the hand appears more distorted than the rake along the length but not the width. The table cells provide the t statistic, p value and the effect size measure r.

Correlation between upright & rotated orientation within the length and width of rake and hand / Hand / Rake
Width / r= .41
p= .12 / r= .60
p= .024
Length
/ r= .87
p<.001 / r= .52
p= .040

S5:Pearson’s correlations between distortions of upright and rotated orientations shown for the rake and the hand (along columns) and dimension (along rows).The Holm-correction was carried out for each item separately.

S6: Percent overestimation (e.g.100*[judged finger length-actual finger length]/actual finger length) of finger length and spacing between pairs of knuckles. Consistent with previous experiments we observed an overall decrease in the magnitude of finger underestimation from the thumb to the little finger and a slight overestimation of the spacing between the knuckles. As mentioned in the discussion, the difference in results between previous work and ours could come from a variation in finger position. From the line drawn from the knuckle of the little finger to the knuckle of the index finger, the mean angles to the thumb, index, middle, ring and little fingers in our study were 39.33º, 67.27º, 75.39º, 83.34º and 99.47º respectively, giving a mean difference of 15.04º, compared to 17.3º in Longo et al., 2014.

1