Body representation changes after repeated lower-limb exoskeleton use in patients with a chronic spinal cord injury.

Exoskeleton use in patients with spinal cord injury may bring health benefits but little information is available about its impact on body perception and cognition.Aim of the study is to explore how repeated use of a powered lower-limb exoskeleton modulates the user’s body representation. Methods: Exo group: 13 SCIpatients (all male, mean age 43 ± 14.54y; time since lesion: 11 ± 6.34 years). Lesions ranged from thoracic (T3) to lumbar (L2): 7 participants had complete lesions(ASIA Impairment Scale [AIS] A), and 6 patients featured sensory and motor incomplete (AIS B) lesions. This group followed an exoskeleton assisted trainingprogram (8 sessions in 4 weeks). SCI control group: the group was selected to match the experimental group (mean age 44 ± 14; years since lesion: 8 ± 8.22years; all p’s > 0.4). Lesions range was T3-L1 with 8 AIS A and 5 AIS B. Control activities consist in passive movements, standing, and strengthening the upper partof the body. Healthy controls: similar age to SCI groups, 8 session wheelchair training program. The body image task (BIT) - administered before (T0) and after(T1) the training period- collects quantitative measures of perceived body spatial organization. On a touchscreen, the participants indicate the location of targetbody part using an anchor image (head or feet) as a reference. The trials includ 14 targets (e.g. left elbow or right knee) for a total of 4 blocks of 52 trials each.Body targets are presented on the screen in written form. Then, an anchor image appears randomly in one of four locations. The participants touch the screen torecord the perceived position of the body target. Results: Measurements are expressed as the difference between the perceived length, width, or height and truebody sizes as a proportion of the true size (distortion index). To compare the body distortion patterns of the rehabilitation programs, a mixed-model analysis ofvariance (ANOVA) with a T0/T1 condition, body part sizes as within-subject factor, and group as between-subject factor was used. The exoskeleton groupperceived hips as narrower (difference T0-T1 ratio = 0.39 to 0.33; p = < 0.017), lower arms longer (difference T0-T1 ratio = -0.21 to -0.10; p = < 0.0001), andupper legs longer (difference T0-T1 ratio = -0.38 to -0.25; p = < 0.0008). The changes go in the direction of real body proportions. Conclusion: exoskeleton usecan modulate perception of specific body parts. Active control of the exoskeleton (agency), jointly with postural and sensorimotor changes, may facilitatewheelchair dis-embodiment. These events may promote the emergence of body representations formed before the injury.