Does the integration of haptic and visual cues reduce the effect of a biased visual reference frame on the subjective head orientation?

Abstract

Background: The selection of appropriate frames of reference (FOR) is a key factor in the elaboration of spatial perception and the production of robust interaction with our environment. The extent to which we perceive the head axis orientation (subjective head orientation, SHO) with both accuracy and precision likely contributes to the efficiency of these spatial interactions. A first goal of this study was to investigate the relative contribution of both the visual and egocentric FOR (centre-of-mass) in the SHO processing. A second goal was to investigate humans' ability to process SHO in various sensory response modalities (visual, haptic and visuo-haptic), and the way they modify the reliance to either the visual or egocentric FORs. A third goal was to question whether subjects combined visual and haptic cues optimally to increase SHO certainty and to decrease the FORs disruption effect.

Methodology/principal findings: Thirteen subjects were asked to indicate their SHO while the visual and/or egocentric FORs were deviated. Four results emerged from our study. First, visual rod settings to SHO were altered by the tilted visual frame but not by the egocentric FOR alteration, whereas no haptic settings alteration was observed whether due to the egocentric FOR alteration or the tilted visual frame. These results are modulated by individual analysis. Second, visual and egocentric FOR dependency appear to be negatively correlated. Third, the response modality enrichment appears to improve SHO. Fourth, several combination rules of the visuo-haptic cues such as the Maximum Likelihood Estimation (MLE), Winner-Take-All (WTA) or Unweighted Mean (UWM) rule seem to account for SHO improvements. However, the UWM rule seems to best account for the improvement of visuo-haptic estimates, especially in situations with high FOR incongruence. Finally, the data also indicated that FOR reliance resulted from the application of UWM rule. This was observed more particularly, in the visual dependent subject.

Conclusions: Taken together, these findings emphasize the importance of identifying individual spatial FOR preferences to assess the efficiency of our interaction with the environment whilst performing spatial tasks.

Figure 1. Illustration of the head apparatus which permitted the deviation of head centre of mass.

Figure 2. Illustration of frame scene and modes of sensory adjustment:

visual (a), haptic (b) and visuo-haptic (c).

Figure 3. Effect of modalities of response according to frame tilt and deviation of head CM all subjects combined.

Figure 4. Correlation between individual visual frame effect scores in the visual modality of response in the ‘no deviation of head CM’ condition, and individual head CM effect in the haptic modality of response in the ‘no frame’ condition.

Figure 5. Mean Error according to conditions and modality of response.

V = Visual modality of response, H = Haptic modality of response, VH obs = Visuo-haptic modality of response, VH MLE = Visuo-haptic calculated with MLE, VH WTA = Visuo-haptic calculated with WTA, VH UWM = Visuo-haptic calculated with unweighted mean error, F−18 = Frame tilted at 18° to the left, WF = cylinder-shape optical tunnel, F+18 = Frame tilted at 18° to the right, CM−9 = deviation of the head centre of mass at 9° to the left right, CM0 = no deviation of the head centre of mass, CM+9 = deviation of the head centre of mass at 9° to the right.

Figure 6. Correlation between individual visual frame effect scores in visual modality of response in the ‘no deviation of head CM’ condition and slope between visuo-haptic response predicted by MLE rule and observed visuo-haptic response.

Figure 7. Correlation between individual visual frame effect scores in visual modality of response in the ‘no deviation of head CM’ condition and slope between visuo-haptic response predicted by WTA rule and observed visuo-haptic response.

Figure 8. Mean Error according to conditions and modality of response.

V FOR = mean error with only visual FOR disrupted (F±18 and CM0), E FOR = mean error with only egocentric FOR disrupted (WF and CM±9), VE FOR obs = mean error observed with visual and egocentric FOR disrupted (F±18 and CM±9), VE FOR MLE = mean error with visual and egocentric FOR disrupted calculated with MLE, VE FOR WTA = mean error with visual and egocentric FOR disrupted calculated with WTA, VE FOR UWM = mean error with visual and egocentric FOR disrupted calculated with unweighted mean error, F−18 = Frame tilted at 18° to the left, WF = cylinder-shape optical tunnel, F+18 = Frame tilted at 18° to the right, CM−9 = deviation of the head centre of mass at 9° to the left, CM0 = no deviation of the head centre of mass, CM+9 = deviation of the head centre of mass at 9° to the right.

Figure 9. Correlation between individual visual frame effect scores in visual modality of response in the ‘no deviation of head CM’ condition and slope between double disruption response predicted by UWM rule and observed double disruption response.