Changes in perceived peripersonal space following the rubber hand illusion

Participants

In total, 65 undergraduate and graduate students participated in this study. The Embodiment Questionnaire²² was used to define to what extent participants experienced the illusion in both the synchronous stroking group (SG) and the asynchronous stroking group (AG). The synchronous group only received synchronous visuo-tactile stimulation when exposed to the rubber hand. In contrast, the asynchronous group only received asynchronous visuo-tactile stimulation (see Table 1 for demographic details of the two groups).

All participants were right-handed by self-report. Participants received course credits or 6 euros as a compensation for their time. They were naïve to the purpose of the study and a written informed consent was obtained from all individual participants prior to the experiment. This study was conducted in accordance with the standards of the declaration of Helsinki and was approved by the FETC of the Faculty of Social and Behavioural Sciences at Utrecht University.

Design

We conducted a pre-post between subjects design with type of stroking (i.e. either synchronously or asynchronously) as the between subjects factor. All participants completed several pre- and post-illusion measures (see task/stimuli below for detailed information for all the measures) in the exact same order (Fig. 1). Our primary outcome measure was an estimate of the point of subjective equality (PSE, see Analyses for details) generated from the pre- and post-Landmark tests. The test procedure started with the pre-measure of straight ahead pointing (SAP) followed by the pre-measure of the proprioceptive drift and the pre-session of the landmark-task (LM). Thereafter the RHI was induced (i.e. either synchronously or asynchronously depending on the group); the post-session of the proprioceptive drift, the landmark and straight ahead pointing followed the illusion respectively. In order to check whether the illusion was well executed, the proprioceptive drift (e.g. behavioral measure) measure followed the RHI immediately instead of the post-test of the Landmark. Ultimately, the embodiment questionnaire was administered.

Task/stimuli/procedure

Prior to the experiment participants were asked to remove all jewelry (i.e. rings and watches). The participant then seated oneself at the long end of a table, in front of a large horizontal screen (55inch) (see Fig. 2). The participant’s head was stabilized with a chinrest. For an overview of the set-up, see Fig. 2.

Straight ahead pointing

The first experimental task was the straight ahead pointing (hereafter SAP, reflecting the subjective body midline), the participants placed their lower arms on the screen, in front of their body (wrists at 32.5 and 102.5 cm from the left side of the tablet, the wrists were equidistant from the body midline which was located at 67.5 cm from the left side of the tablet). The participants were instructed to point, with eyes closed to prevent feedback, with either their left or right hand at their own body midline and then in front of them where this midline would be on the monitor. This procedure was then repeated for the other hand. These locations were measured in cm from the left side of the tablet with a tape measure. For analyses, the SAP’s (in cm) were only used from the right hand as the left hand could have been influenced by the RHI, and the pre- and post-measures (cm) were included in the analysis. The duration of this task was about 3–5 min.

Proprioceptive drift

After the SAP the pre-measure of the proprioceptive drift was conducted. A black occluder was placed over both hands to make them invisible, see Fig. 2b. The experimenter moved a stick from left to right (or right to left, random sequence) alongside the long end of the table. Participants were instructed to say stop when the stick was at the felt location of left or right index fingertip (again random sequence). The experimenter documented the exact location (by means of a tape measure) of the reported felt position. De left finger was always located 32.5 cm from the left edge of the box. The right index was 102.5 cm from the left edge of the box, see Fig. 2a. Thereafter, the occluder was removed. The pre- and post-measure (cm) were included in the analysis. The duration of this task was about 3–5 min.

Landmark task

After the administration of the pre-test for proprioception, the pre-landmark task started. Participants were instructed to determine whether a landmark (i.e. vertical transection mark) was either left or right from the center of a grey line (Fig. 2c). In order to prevent feedback from previous landmark positions, each trial started with a static dot (500 ms) (Fig. 2c) appearing either on the left and right (alternately) from the center of the screen. Participants were instructed to look at these dots. Then a horizontal (dark grey) line across the whole width of the monitor (light grey background) was presented (Fig. 2a), followed by a 750 ms presentation of a vertical line (126 mm (200 pixels)) (i.e. the landmark; at a different location across the horizontal line in each trial; Fig. 2c). Locations of the vertical line were at (− 25.2, − 6.3, − 3.1, − 2.5, − 1.8, − 1.2, − 0.6, 0 (center), 0.6, 1.2, 1.8, 2.5, 3.1, 6.3, 25.2 mm; − 40, − 10, − 5, − 4, − 3, − 2, − 1 0, 1, 2, 3, 4, 5, 10, 40 pixels). To avoid aftereffects, a mask consisting of vertical lines was shown immediately after landmark presentation until the end of the trial. The participant indicated verbally whether the vertical line was located either to the left or right from the center of the screen. The experimenter pressed ‘A’ if the answer was ‘left’ and ‘L’ if the answer was right. Then the next trial started, for a total of 60 trials. Each of the 15 locations was presented 4 times in random order. Each participant’s data were fitted with a cumulative normal distribution function to generate estimates of the point of subjective equality (PSE) in Matlab; the location of the landmark where the participant was equally likely to determine it ‘left’ or ‘right’. The PSE, our primary outcome measure, was included in the analyses. The shift in PSE (before versus after the RHI) reflected the shift in space, hence peripersonal space. The duration of this task was about 10–15 min.

Rubber hand illusion

After the first landmark session, the left arm was covered up by the occluder and the RHI was set up, see Fig. 2d. While the participant had his eyes closed, the rubber hand was placed in an anatomical congruent position 15 cm to the right of the real left hand²³ and therefore 20 cm to the left of the body midline. To optimize the illusion, a cloth was placed over the shoulder of the participant. In the experimental condition, the illusion was established by stroking the index finger of the real and rubber hand simultaneously with a soft brush for 90 s, while the participant was visually focusing on the rubber hand. In the other group, the asynchronous group, the stroking was asynchronous: first the rubber hand was touched and then the real hand. Location and velocity of stroking were held constant.

After inducing the illusion, the rubber hand was removed and both real hands were covered by the occluder (Fig. 2b). The proprioceptive drift was now measured for the second time. This procedure was identical to pre-session. The occluder was then removed, so both hands were visible again, as in starting position (Fig. 2a). Then the landmark task started for the second time with the exact same procedure as in the pre-illusion session. Thereafter, the straight-ahead pointing task was performed and again the procedure was identical to the pre-illusion session.

Embodiment questionnaire

To conclude the experiment, the participant filled out the Embodiment Questionnaire²². This questionnaire contained 10 items to measure the experience of the rubber hand illusion. The first three items measured the illusion experience, for example: ‘It seemed like the rubber hand was my own’, while the remaining seven items were control questions e.g. ‘It seemed like I had more than two hands’. The participant responded on a 11-point Likert scale with 0 = strongly disagree and 10 = strongly agree. The overall duration of the experiment was about 40 min.

Analysis

For all our outcome-measures we used a Mixed ANOVA with time (pre-test versus post-test) as the within subject factor and synchrony (synchronous stroking versus asynchronous stroking) as between subjects factor. All p-values from post-hoc analyses were Bonferroni-corrected.

In addition, we also analyzed our data with a Bayesian mixed ANOVA, which uses a linear mixed model. We used Cauchy (uninformative) priors on effect size^24,25. Thus, next to the frequentist approach we report Bayes factors which yields the probability of a model given the data (i.e., a certain combination of effects) relative to a null model (i.e., no effects), that is, values larger than 1 are in favor of H1. Bayes Factors (BF) that provide evidence in favor of the null model are abbreviated as BF₀₁, Bayes Factors that provide evidence in favor of a difference are abbreviated as BF₁₀. Since the Bayesian approach can quantify evidence for both directions (e.g. evidence for H1 and evidence for H0), it allows evaluating null effects, which is not the case in the classical frequentist approach²⁴.