Spinal cord lesions shrink peripersonal space around the feet, passive mobilization of paraplegic limbs restores it

Peripersonal space (PPS) is the space surrounding us within which we interact with objects. PPS may be modulated by actions (e.g. when using tools) or sense of ownership (e.g. over a rubber hand). Indeed, intense and/or prolonged use of a tool may induce a sense of ownership over it. Conversely, inducing ownership over a rubber hand may activate brain regions involved in motor control. However, the extent to which PPS is modulated by action-dependent or ownership-dependent mechanisms remains unclear. Here, we explored the PPS around the feet and the sense of ownership over lower limbs in people with Paraplegia following Complete spinal cord Lesions (PCL) and in healthy subjects. PCL people can move their upper body but have lost all sensory-motor functions in their lower body (e.g. lower limbs). We tested whether PPS alterations reflect the topographical representations of various body parts. We found that the PPS around the feet was impaired in PCL who however had a normal representation of the PPS around the hands. Significantly, passive mobilization of paraplegic limbs restored the PPS around the feet suggesting that activating action representations in PCL brings about short-term changes of PPS that may thus be more plastic than previously believed.


Supplemental Figures S1 The Space product method for CCE data
By means of a transdimensional hierarchical Bayesian linear regression, the Space Product Method (SPM) was applied directly to CCE indexes as follows: (1) = −

Using this index, if a PPS representation is present it should be greater than 0 and if a PPS
representation is not present it should be equal to 0. In order to set credible values, we selected the means and standard deviations using the actual data of the experiments.
In these models, the intercept and an estimation of the condition for each experimental cell were used as Fixed factors, while an estimation of the contribution of the participant was included as a Random factor.
Two categorical distributions were used as hyperpriors (prior distributions of prior distributions). One distribution was used to select the parameter values that corresponded to the various different hypotheses for the Control group. The other categorical distribution was used to select the hypothesis for the Paraplegic Group.

Figure S1
The

S3 Frequentist analyses
For the CCE analyses, a mixed-effect ANOVA was applied to the data by means of the afex package 1 (between-subject factor: group, within-subject factors: condition and left/right congruency). Pairwise t tests with Bonferroni correction were then used to investigate the data. Bootstrap versions of the tests were further applied.
For the questionnaire data in Experiment 1, ordered logistic regressions 2 were used. These are particularly suited to the analysis of ordinal multinomial data as they take into account their ordinal nature but also allow for a complex analysis such as an ANOVA. The analysis was computed by means of the rms   Bonferroni corrected post-hoc t-tests showed that the only significant difference between Ipsilateral and Contralateral trials was in the PPS condition (p = .0049). Bonferroni corrected post-hoc t-tests showed that the only significant difference between Ipsilateral and Contralateral trials was in the Motion group (p = .048, one-tail).

S4.1. Apparatus
The apparatus was designed by MS. The instrumentation for the cross-modal paradigm was composed of the following: • Two vibrating tactile stimulators for the hands that vibrate on the thumb and index finger • a wooden frame for the feet and the visual stimuli (six LEDs) • a microcontroller to control the tactile stimulators and the visual distractors A wooden frame was prepared with two feet-compartments (each 15 × 8 cm). The inner borders of the compartments were 22 cm apart. Four white LEDs to be used as distractor stimuli were mounted near the inner corners of the feet-compartments. (Figure 1A, 1B, 1C). One red and one yellow LED were positioned at the centre of the wooden frame in order to provide a fixation point and a control light, respectively. The apparatus was covered by a black rubber cloth to ensure that the LED lights contrasted with the background. The wooden frame was designed to be placed on the floor and tilted so that the feetcompartments were at an angle of 50 degrees. This design allowed the feet-compartments and the LED lights to be clearly visible.
Both the LED lights on the wooden frame and the tactile stimulators were connected to a microcontroller (Arduino Uno TM , www.arduino.cc). This was programmed using the "Processing" programming language (www.processing.org) which controlled the individual experimental trials.
The microcontroller was connected to a Notebook computer (ASUS X53S) that recorded the data in Microsoft Visual C# ™ (www.microsoft.com), correctly randomized the individual trials and recorded the vocal responses via the microphone. The tactile stimulators produced a small auditory tone which was masked by white noise on an mp3 player (played into headphones) in order to avoid a confounding effect.
The tactile stimulators were attached to flexible metal bands and connected to a comfortable plastic chair.

S4.2. Sense of Ownership Questionnaire
Sense of Embodiment