Table 1.

Summary of studies using RAT in pwGBSs.

Author (Year) Country [Ref.]	Study Design	Sample N (pwGBSs), Gender, Mean Age ± SD, (Range)	Procedure N Duration (Session × Week)	Outcome	Results	Conclusions
Céspedes (2020) Colombia [20]	No randomized controlled study (feasibility study)	RAT + NAO robot group 4 (1) 3 M NR (20–60) Ctrl group 4 (1) 3 M NR (20–60)	RAT + NAO robot NR (1 × 1) RAT NR (1 × 1). Each patient performed two RAT sessions on Lokomat: one control and one assisted by the social NAO robot, carried out on different days of the week. (1) Therapy time should be the same during the sessions; (2) each patient should perform a unique session per condition; and (3) all Lokomat features must be the same across therapy sessions.	Cardiovascular parameters (Zephyr HxM BT), spinal posture parameters (cervical and thoracic IMU), and Borg scale.	The results of the study showed a positive and well-received effect of the robot regarding postural behavior in the cervical and thoracic area, companionship, and social interaction. In the robot-assisted scenario, patients relied more on the feedback and paid more attention.	High motivation was provided by the robot to perform the therapy with good posture. Therapists and patients agreed that the robot had been helpful in the session.
de Crignis (2023) Germany [19]	No randomized controlled study (feasibility study)	RAT group 11 (1) 9 M 64.4 ± 11.2 (47–85) Ctrl group 11 (0) 9 M 64.3 ± 9.1 (49–79)	RAT group NR (4–5 × NR). Patients performed four to five sessions of upper limb rehabilitation with a RobExReha device. Ctrl group, 4–90 min NR. Trained with ArmeoPower or ArmeoSpring, both with an average of 15 (±25 min) therapy sessions completed.	Patients: QUEST, RTLX, HoloLens, and pAR questionnaire. Therapists: completed subscale of QUEST, SUS and UEQ-short to evaluate perceptions of the device’s usability.	The patients’ usability ratings were significantly higher in the Reference Group for two items of the QUEST: reliability and ease of use. Workload (RTLX) ratings did not differ significantly between the groups. Nearly all patients using the RobExReha system perceived the gaming scenario in AR as functioning adequately despite eight patients having impairments in stereoscopic vision. The therapists valued the system’s approach as interesting and inventive.	Therapy with the RobExReha system was safe and feasible for patients and therapists, with no serious adverse events being reported. This structured approach allowed researchers to systematically evaluate the role of social robots in enhancing rehabilitation outcomes for individuals with mobility impairments, providing valuable insights into future applications of robotics in therapy.
Yabuki (2024) Japan [17]	Case report	GBS (1) 78	ABAB design Phase A 10–40 min (6 × 10). Conventional gait program included level walking and treadmill gait training. Phase B 20–30 min (5 × 6). Gait trainer HWA-01 included HGT during physical therapy (walk and climb stairs).	CWS, stride length, and cadence at the start of training in all phases. GBS disability score, MRC sum, 6MWT, motor-FIM, and ONLS.	The study reported significant improvements (p < 0.05) in the patient’s functional mobility following the use of the exoskeleton. Specific metrics of mobility enhancement were documented, indicating that robotic assistance could play a vital role in rehabilitation for pwGBSs.	Exoskeleton robots could be beneficial for patients with GBS experiencing ongoing gait disturbances. This case report highlighted the potential for such technologies to augment traditional rehabilitation methods, suggesting further research into larger cohorts and long-term effects.
Chen (2023) Taiwan [18]	Case report	GBS (1) 75	RAT with Lokomat add-on rehabilitation program 54 min (Two sessions) First session included 80% reduction in body weight, treadmill speed of 1.2 km/h, distance of 421 m. The second included 70% reduction in body weight, treadmill speed of 1.3 km/h, distance of 623 m.	BI, CFS, IADLs, CGA, SPPB, SACR-F, GDS, CAM, MMSE.	The findings revealed that robot-assisted gait training led to significant improvements in multiple areas. Participants showed increased lower limb muscle strength following the intervention. Additionally, there was a notable enhancement in functional independence, as reflected in the improved BI scores, indicating better performance in daily activities. Furthermore, the CFS results demonstrated a reduction in frailty levels, suggesting an overall improvement in the patients’ functional status.	RAGT is a feasible and safe intervention for improving ambulation and functional capabilities in geriatric pwGBSs. The robotic approach seems to have many potential benefits in rehabilitation programs for this patient population.

Key: Ctrl, control group; RAT, robot-assisted therapy; IMU, inertial measurement unit; QUEST, Quebec User Evaluation of Satisfaction with Assistive Technology; RTLX, row task load index; pAR, “presence in augmented reality”; SUS, system usability scale; UEQ-short, user experience questionnaire; HWA, hip-wearable exoskeleton; CWS, comfortable walking speed; motor-FIM, motor-functional independence measure; 6MWT, six-meter walking test; ONLS, overall neuropathy limitations scale; BI, Barthel Index; CFS, clinical frailty scale; IADLS, Instrumental Activities of Daily Living Scale; CGA, comprehensive geriatric assessment; SPPB, short physical performance battery; SARC-F, Strength, Assistance in Walking, Rise from a Chair, Climb Stairs, and Falls questionnaire; GDS, geriatric depression scale; CAM, confusion assessment method; MMSE, mini mental state examination.