The development of rehabilitation techniques modulating the sense of verticality is a major challenge in balance disorders related to lateropulsion or retropulsion (Stroke, Parkinson, etc.). Through this pilot study in healthy participants, we tested two techniques that are likely to modulate verticality perception: immersion in tilted virtual environment (visual cues) and walking with body weight support (somaesthesic cues).

Among 20 recruited healthy participants, 12 were studied (53.3±7 years old) for the following criteria: normal visual vertical (VV) in baseline, sensitive to an immersive virtual tilted room (VTR) without dizziness. Their VV (8 trials) was tested in 3 postural conditions (sitting, treadmill walking and treadmill walking with 30% of body weight suspended by a vertical cable (BWSW), combined with 2 visual conditions (darkness and VTR). VTR was produced by an Oculus Rift^® head mounted display (HMD), and tilted 18° clockwise. The study design aimed at introducing a VV bias through the VTR, then testing the immediate and after-effects of postural conditions. Data were analyzed with ANOVA.

Regarding VV it was found a main effect of VTR (11.0±4.4; F(1,10)=158.5; P<0.001), with an after-effect of several minutes recorded in darkness. It was also found a strong effect of the postural setting (F(2,20)=5.4, P<0.05). Post-hoc analysis showed that VV was 0.6° more vertical in the condition BWSW than in sitting (P<0.05).

Being immerged in a tilted environment induces a powerful bias in verticality perception, stronger than that reported by studies using non-immersive tilted clues. Overall walking on a treadmill being supported by a vertical cable is a mean to recalibrate a biased vertical. This modulation of the internal model of verticality is likely due partly to a reweighting of somaesthesic input, and to an implicit direction of the Earth vertical given by the cable which supports the body. This might open an interesting track for the rehabilitation of patients with postural disorders due to a biased internal model of verticality.