Multisensory temporal processing in own-body contexts: plausibility of hand ownership does not improve visuo-tactile asynchrony detection

Abstract

Tracking one's own body is essential for environmental interaction, and involves integrating multisensory cues with stored information about the body's typical features. Exactly how multisensory information is integrated in own-body perception is still unclear. For example, Ide and Hidaka (Exp Brain Res 228:43-50, 2013) found that participants made less precise visuo-tactile temporal order judgments (TOJ) when viewing hands in a plausible orientation (upright; typical for one's own hand) compared to an implausible orientation (rotated 180°). This suggests that viewing one's own body relaxes the precision for perceived visuo-tactile synchrony. In contrast, visuo-proprioceptive research shows improvements for multisensory temporal perception near one's own body in asynchrony detection tasks, implying an increase in precision. Hence, it is unclear whether viewed hand orientation generally modulates the ability to detect small asynchronies between vision and touch, or if this effect is specific to TOJ tasks. We investigated whether viewed hand orientation affects detection of visuo-tactile asynchrony. In two experiments, participants viewed model hands in anatomically plausible or implausible orientations. In one experiment, we stroked the hands to induce the rubber hand illusion. Participants were asked to detect short delays (40-280 ms) between vision (an LED flash on the model hand) and touch (a tap to fingertip of the participant's hidden hand) in a two-interval forced-choice task. Bayesian analyses show that our data provide strong evidence that viewed hand orientation does not affect visuo-tactile asynchrony detection. This study suggests the mechanisms for fine-grained time perception differ between visuo-tactile and visuo-proprioceptive contexts.

Visual body form and orientation cues do not modulate visuo-tactile temporal integration

“…Hoover and Harris [34] and Zopf et al [35] demonstrated that visual form and orientation cues led to greater sensitivity to detect small temporal delays, which suggests that visual plausibility leads to relatively more segregation of visual and proprioceptive temporal inputs. Our data, however, together with Keys et al [28], show that such visual cues do not modulate visuo-tactile integration. Thus, it appears that different mechanisms could underlie the temporal processing of visuo-tactile and visuoproprioceptive bodily inputs.…”

“…As Bayesian causal inference models propose that the degree of multisensory integration is influenced by a causal inference process that determines the relative probability that signals belong together or not [19][20][21], a realistic pairing between a visual and tactile touch stimulus might be important to observe modulations in multisensory integration. A study might therefore fail to show an effect if the stimuli are normally unlikely to co-occur, such as in the Keys et al [28] study. In that study, the visual stimulus was a light presented next to a 3D plaster hand, paired with a tactile stimulus on the participant's own hand.…”

“…Previous studies that investigated the effect of visual cues on the degree of visuotactile temporal integration have reported conflicting findings [26][27][28]. Keys et al [28] used an asynchrony detection task, and found no effects. In contrast, two studies using a TOJ task reported an effect [26,27].…”

“…The conflicting findings may also be influenced by the relative sample sizes. Both the current study and the Keys et al [28] study used relatively large sample sizes combined with a clearly defined sampling plan to ensure sensitive data for hypothesis evaluation and a procedure to check for outliers. Studies with high power and a complete analysis strategy are typically more robust and therefore less easily influenced by outliers and analytical flexibility [30][31][32].…”

“…This could have affected whether the visual and tactile signals were perceived to 'belong together' and hence the degree of temporal integration. Finally, Keys et al [28] used relatively large samples of 30 (Experiment 1) and 31 participants (Experiment 2) whereas both Ide and Hidaka [26] and Maselli et al [27] used much smaller samples (12including the two authors -and 14, respectively) and it is unclear if these studies used a procedure to check for outliers. Current focus on methodological issues in experimental psychology and other sciences have emphasised the potential for false positive effects when using small sample sizes [30][31][32][33].…”