Intrinsic neural activity predisposes susceptibility to a body illusion

Abstract

Susceptibility to the rubber hand illusion (RHI) varies. To date, however, there is no consensus explanation of this variability. Previous studies, focused on the role of multisensory integration, have searched for neural correlates of the illusion. But those studies have failed to identify a sufficient set of functionally specific neural correlates. Because some evidence suggests that frontal α power is one means of tracking neural instantiations of self, we hypothesized that the higher the frontal α power during the eyes-closed resting state, the more stable the self. As a corollary, we infer that the more stable the self, the less susceptible are participants to a blurring of boundaries-to feeling that the rubber hand belongs to them. Indeed, we found that frontal α amplitude oscillations negatively correlate with susceptibility. Moreover, since lower frequencies often modulate higher frequencies, we explored the possibility that this might be the case for the RHI. Indeed, some evidence suggests that high frontal α power observed in low-RHI participants is modulated by δ frequency oscillations. We conclude that while neural correlates of multisensory integration might be necessary for the RHI, sufficient explanation involves variable intrinsic neural activity that modulates how the brain responds to incompatible sensory stimuli.

Keywords: EEG; intrinsic neural activity; rubber hand illusion; self-relatedness; α band power.

Fig. 1

Questionnaire scores for synchronous and asynchronous conditions. Items #1–#3 were probed in order to investigate the ownership experience, with special attention given to #3, which explicitly concerns ownership.

Fig. 2

Ground averaged power spectrums during resting states. Power spectra were plotted as a function of electrodes and, from left to right, panels are depicted for: EO, EC, and the difference between EC and EO. Bottom panels depict topography of α frequency power for each condition. The α frequency power differences are stronger, progressively, from frontal to posterior electrodes.

Fig. 3

Correlation between total difference scores with EO/EC α power. The left panel is EO and the right panel is EC condition. All topography of the EEG showed rho values between the scores across electrodes and power across frequency bands. δ, θ, and α power are listed from top to bottom. Take note of (f): A significant negative correlation between total difference scores and α power was only evinced for frontal and LLP electrode sites (white dots, q < 0.05), in the EC condition. To depict the negative correlation between alpha power and the total difference score, we combined alpha power from those significant electrodes under the EC condition and correlated with the total difference score (f). As for (a) to (e), no significant correlation was observed. All electrode sites were averaged in order to calculate relationships to total difference scores.

Fig. 4

Correlation between item #1–3 difference scores with EO/EC α power. The left panel is EO and the right panel is EC condition. All topography of the EEG showed rho values between the difference scores for each item across electrodes and power across frequency bands. Items #1–#3 are listed from top to bottom. Take note of (f): A significant negative correlation between item #3 difference scores and α power was only evinced for frontal and LLP electrode sites (white dots, q < 0.05) in the EC condition. To depict the negative correlation between the α power and the item #3 difference score, we combined α power from those significant electrodes under the EC condition and correlated with the total difference score (f). As for (a) to (e), no significant correlation was observed. All electrode sites were averaged in order to calculate relationships to total difference scores.

Fig. 5

The δ phase and α amplitude coupling for low- and high-RHI performers in the EC condition. The top panel depicts topography of α frequency power in the EC condition. Left topography depicts low-RHI subjects; right topography, high-RHI subjects. Averaged PACz in low-RHI subjects was significantly larger than Z-score 1.96 (bottom left), whereas averaged PACz in high-RHI subjects was not (bottom right).