The extrastriate body area is involved in illusory limb ownership

Highlights

• We induce the Rubber Hand Illusion using a novel, fully automated fMRI-setup.

• We use a somatotopical design to match RHI and control in visuo-tactile synchrony.

• We functionally localize the body-part selective extrastriate body area and MT +.

• Brain activity in EBA and anterior insulae is stronger during RHI versus control.

• Reported intensity of illusory ownership is positively correlated with EBA activity.

Abstract

The Rubber Hand Illusion (RHI) is an established paradigm for studying body ownership, and several studies have implicated premotor and temporo-parietal brain regions in its neuronal foundation. Here we used an automated setup to induce a novel multi-site version of the RHI in healthy human participants inside an MR-scanner, with a RHI and control condition that were matched in terms of synchrony of visual and tactile stimulation. Importantly, as previous research has shown that most of the ownership-related brain areas also respond to observed human actions and touch, or body parts of others, here such potential effects of the experimenter were eliminated by the automated procedure. The RHI condition induced a strong ownership illusion; we found correspondingly stronger brain activity during the RHI versus control condition in contralateral middle occipital gyrus (mOCG) and bilateral anterior insula, which have previously been related to illusory body ownership. Using independent functional localizers, we confirmed that the activity in mOCG was located within the body-part selective extrastriate body area (EBA). Crucially, activity differences in participants' peak voxels within the left EBA correlated strongly positively with their behavioral illusion scores. Thus EBA activity also reflected interindividual differences in the experienced intensity of illusory limb ownership. Moreover, psychophysiological interaction analyses (PPI) revealed that contralateral primary somatosensory cortex had stronger brain connectivity with EBA during the RHI versus control condition, while EBA was more strongly interacting with temporo-parietal multisensory regions. In sum, our findings demonstrate a direct involvement of EBA in limb ownership.

Introduction

To be oneself among others, one needs to identify with a particular body (Blanke and Metzinger, 2009, Gallagher, 2000, Jeannerod, 2007). Most accounts of body ownership have emphasized multimodal information integration in hierarchical cortical networks as a fundamental mechanism underlying a coherent self-representation (Apps and Tsakiris, 2013, Blanke, 2012, Hohwy, 2007, Hohwy, 2010, Petkova et al., 2011, Seth et al., 2011, Tsakiris, 2010). These theories are supported by recent neuroimaging experiments that have provided novel insights into how the brain self-attributes body parts based on such integration of visual, tactile, and proprioceptive information. In the Rubber Hand Illusion (RHI; Botvinick and Cohen, 1998), synchronous stroking of a dummy body part together with one's own corresponding body part typically misleads the brain to self-attribute the dummy limb (Botvinick and Cohen, 1998, Ehrsson et al., 2004, Tsakiris and Haggard, 2005) or even a whole body (Ehrsson, 2007, Lenggenhager et al., 2007). The experience of (illusory) body ownership has been linked to activity in frontal brain regions, predominantly the ventral premotor cortex (PMv; Ehrsson et al., 2004, Ehrsson et al., 2005, Petkova et al., 2011), but also posterior regions like the right temporo-parietal junction (rTPJ; Blanke et al., 2002, Blanke et al., 2005, Ionta et al., 2011, Tsakiris et al., 2008), posterior parietal cortex and intraparietal sulcus (PPC/IPS; Brozzoli et al., 2012, Ehrsson et al., 2004, Gentile et al., 2011, Petkova et al., 2011, Shimada et al., 2005; Tsakiris, 2010), and occipito-temporal regions like the body part-selective extrastriate body area (EBA; Arzy et al., 2006, Blanke and Mohr, 2005, Downing et al., 2001, Ionta et al., 2011). Primary somatosensory cortex (SI; Kanayama et al., 2007, Kanayama et al., 2009, Lenggenhager et al., 2011, Tsakiris et al., 2007) and the anterior insula (AI; Ehrsson et al., 2007) have also been associated with body ownership. Activity in these regions has been interpreted as reflecting the degree of illusory self-attribution or “incorporation” of the fake limb or body (Blanke, 2012, Ehrsson et al., 2004, Holmes and Spence, 2004, Petkova et al., 2011, Tsakiris, 2010).

Here, we used a fully automated setup to induce a novel, multi-site version of the RHI inside an fMRI scanner with high spatial resolution, addressing two potential caveats of the procedures typically used to evoke the illusion. First, we matched visual and tactile stimuli of both RHI and control condition in temporal synchrony, in contrast to the typically used asynchronous stroking control condition where observed touch on the dummy hand and felt touch on the own hand are presented serially. In our control condition, observed and felt touch were presented synchronously at spatially incongruent locations (palm and forearm). This synchronous stimulation countered potential problems associated with a serial, isolated presentation of observed and felt touch: For example, premotor cortex has been shown to be engaged in (serial) sensory predictions even in tasks using abstract, nonbiological stimuli (Schubotz and von Cramon, 2002, Schubotz and von Cramon, 2003, Schubotz and von Cramon, 2004), and the presentation of observed touch before felt touch at the same location could potentially be influenced by effects of anticipation of touch (see e.g. Carlsson et al., 2000, Keysers et al., 2010, Kuehn et al., 2012). Moreover, the resulting design enabled us to calculate a joint contrast comparing two RHI and control conditions, in which spatiotemporal differences between stimuli in the conditions were averaged out, and thus the resulting effects were attributable to the experienced illusion only. Second, by fully automating our experimental setup, we eliminated the human experimenter from the procedure. The induction of the RHI by touch from another person may interfere with self-related information processing, as many brain regions associated with body ownership (e.g., EBA, insula, PMv, and SI) also respond to observed human actions and touch, or mere vision of bodies of others (Bernhardt and Singer, 2012, Blanke, 2012, Ebisch et al., 2008, Keysers et al., 2010, Peelen and Downing, 2007, Zaki and Ochsner, 2012). Therefore, we aimed to isolate body ownership mechanisms from effects introduced by social interaction. The RHI has been induced automatically in one PET study (Tsakiris et al., 2007), but to our knowledge no automated MR-compatible RHI setup has been reported to date. We tested for BOLD signal differences between the RHI versus control condition within the ownership-related regions identified in previously published studies, expecting effects in regions whose response to the illusion is not influenced by receiving human touch. Moreover, we tested whether activity in those regions would reflect individual differences in the experienced intensity of the ownership illusion (Ehrsson et al., 2004, Petkova et al., 2011, Tsakiris et al., 2007).