16 sessions of 45 minutes of reeducation with robot of the walk and the procedure to follow with the same patient for the same phase of rehabilitation of the balance and the walk or of several of them (Annex 3b), at the rate of 3 at 5 sessions per week, for 4 weeks and according to predefined exercises

16 sessions of 45 minutes of reeducation without robot of the walk and the procedure to follow with the same patient for the same phase of rehabilitation of the balance and the walk or of several of them (Annex 3b), at the rate of 3 at 5 sessions per week, for 4 weeks and according to predefined exercises

A quality of life assessment: SIS questionnaire completed by the patient, the results of which will be reported in the observation book by the coordinating physiotherapist who carried out the evaluations described above. (D0, D30 and D90)

An assessment of patient acceptance of robotic rehabilitation (for patients in the robot group): self-questionnaire to be filled in by the patient (before carrying out the gait tests in order to avoid influence the results) (D30 ans D90)

An evaluation of cognitive functions (between D0 and D30): research dysexual cognitive-behavioural syndrome, hemineglect visio-spatial, memory disorders. Done by a doctor evaluator.

An evaluation of the robot's modalities of use (for patients in the robot group): time of use, distance travelled, walking speed, exercise duration, percentage of regulated suspension, number of breaks, average length of breaks, total length of breaks. (D30)

The motricity index, Motricity Index, is an ordinal method of measuring muscle strength developed by Demeurisse in 1980. In the lower limb, 3 movements are tested and performed in a sitting position: hip flexion, knee extension and dorsiflexion. For each movement, the score is 0.9, 14.19, 25 or 33 (same items as the MRC classification on 5). The maximum total score in the lower limb is therefore 99. The evaluation is made on the side of the paretic limb. This scale has excellent fidelity (Collin and Wade 1990, Fayazi 2012) and construct validity (Collin and Wade 1990). The susceptibility to change in the motor index is average, and for example the TDM10m is more sensitive to change than the motor index (Vos-Vromans 2005).

Stroke Impact Scale (SIS) is a quality of life assessment tool. Its latest version, SIS 3.0, contains 59 items (64 for version 2.0) measuring eight domains: strength (4 items), hand functionality (5 items), ADL / IADL (10 items), mobility ( 9 items), communication (7 items), emotions (9 items), memory / thinking (7 items), and social participation (8 items). A final item assessing overall recovery from stroke is requested from the patient (score out of 100).The items are scored using a Likert type scale with five answer options. Patients are asked to complete the instrument according to the difficulty they perceived during the last week. The scores of 5, 4, 3, 2 and 1 correspond respectively to: '' not difficult at all '', '' a little difficult '', '' rather difficult '', '' very difficult '', and '' extremely difficult ''. The scores are then processed by computer to obtain a total score out of 100, 100 corresponding to the absence of difficulty.

Postural Assesment Scale:

The PASS, Postural Assesment Scale, assesses balance in a supine, sitting and standing position. this scale has been designed specifically for stroke patients and is suitable for all individuals regardless of postural performance (Benaim 1999). It is an adaptation of the Fugl-Meyer evaluation equilibrium subscale. Initially developed in French, it was later translated into English. This scale consists of 12 items of increasing difficulty.

It includes items not evaluated by the Berg / BBS scale (Blum and Korner-Bitensky, 2008) and demonstrates better psychometric properties than the Berg / BBS scale and the modified balance subscale of the evaluation. Fugl-Meyer (Mao 2002, Hsieh 2002, Wang 2004).

The Wade test, or 10-meter walk test (TDM10m), is designed to evaluate walking speed per second over a short distance. The patient walks for 10 meters, and the time required to cover this distance is measured, and the number of steps is calculated. Most often the patient is started, that is to say he starts a few meters before and he starts the test at constant speed. Three in-line tests must be performed, and the average over the three trials is calculated. The test can be done at comfortable speed or fast speed (must be specified), with or without technical assistance (walking stick, walker), with or without orthosis (foot lifter, orthopedic shoes). The Standard Error of Measurement (SEM) was evaluated at 0.04 m / sec (Perera 2006), and the Minimum Detectable Change-MDC at 0.16 m / sec (Tilson 2010). The cut-off for safe outdoor independent walking has been set at 0.8 m / s (Bowden 2008).

TUG, Timed Up and Go Test (Podsiadlo and Richardson, 1991), is an instrument for the overall assessment of mobility, balance, walking ability and risk of falling in the elderly. The patient should stand up from a chair, walk a distance of three meters, turn around, then return to the chair while walking and sitting down. These tasks must be performed at comfortable speed. The measured time begins with the verbal instruction "go" and stops when the client returns to sitting position. It is permissible to use a walking aid, but its use must be indicated. No physical assistance is given. TUG performance is rated on a scale of 1 to 5, where 1 indicates "normal functioning" and 5 indicates "severe abnormal behavior", depending on the individual's perception of risk of falling. The score is the time taken to complete the test, in seconds.

Walking test of 6 minutes:

The six-minute walk test (6MWT / TDM6min) (Butland, Pang, Gross, Woodcock and Geddes, 1982) is a functional walk test in which the distance traveled by the patient in six minutes is evaluated. This test has been validated in different populations including stroke patients (Kosak and Smith, 2005). It is a simple test that requires an indoor, straight, flat, 30m long corridor. Mean normative data vary between 408 and 422 m (Wevers 2011).

Standard Error of Measurement (SEM) has been estimated to be between 12.4 and 18.6 m (Eng 2004, Flansbjer 2005), and the Minimal Detectable Change-MDC at about 35 m for a population chronic stroke (Eng 2004, Flansbjer 2005) and 61 m in the acute phase (Perera 2006). Fidelity has been found by studies to be excellent (Flansbjer 2005, Fulk 2008, Wevers 2011) or adequate (Kosak and Smith 2005). Validity against excellent criteria (Flansbjer 2005, Kosak and Smith 2005, Wevers 2011).

It consists of 14 items, with a maximum score of 56. Static and dynamic balance tasks of increasing difficulty are performed, with a score per item ranging from 0 to 4, depending on the ability to perform the task evaluated.

This equilibrium scale was initially conceived as a quantitative measure of the balance and the risk of falling in non-institutionalized elderly people (Berg 1989). The psychometric qualities of this scale were then demonstrated in patients after stroke, including fidelity, validity and internal consistency (Berg 1992, Berg 1995, Mao 2002, Wang 2004, Hiengkaew 2012). Sensitivity to change was considered moderate to wide according to the studies (Wood-Dauphinee 1996, Mao 2002, Chou 2006). The Standard Error of Measurement (SEM) varies from 1.5 to 2.9 points depending on the studies (Liston and Brouwer 1996, Stevensen 2001, Flansbjer 2012, Hiengkaew 2012).

The MMSE, Mini Mental State Examination, is a tool for assessing global cognitive efficiency (Folstein and Folstein, 1975).The MMSE (Boxes 1 and 2) consists of a series of questions grouped into 7 subtests and designed in such a way that normal subjects can easily answer each question. The questions concern: orientation in time (5 points), orientation in space (5 points), immediate recall of three words (3 points), attention (5 points), delayed recall of three words (3 points), language (8 points) and constructive praxis (1 point). The maximum score is therefore 30 points.

Evaluation of patient acceptance of robotic gait rehabilitation at 1 month (D30± 5 days): Acceptance questionnaire, developed by our laboratory colleagues of Social Psychology, LP3C, University of Rennes 2.

Evaluation of patient acceptance of robotic gait rehabilitation at 3 months (D90 ± 2 days): Acceptance questionnaire, developed by our laboratory colleagues of Social Psychology, LP3C, University of Rennes 2.

Time of use Time of use

Distance travelled

Walking speed

Duration of the exercise

Regulated suspension percentage

Number of breaks

Average lengh of breaks

Total duration of breaks

BREF(Batterie Rapide d'Efficience Frontale) or FAB (Frontal Assessment Battery in english is a neuropsychological evaluation of executive functions developed by Dubois and Pillon in 2000. It is a rapid test (between 5 and 10 minutes) and is composed of 6 sub-tests: similarities (exploring conceptual elaboration), lexical evocation (exploring mental flexibility), Luria's gesture sequence (exploring programming), conflicting instructions (exploring sensitivity to interference), the Go-No Go (exploring inhibitory control) and the search for a grasping behavior (exploring behavioral autonomy with respect to the environment). The BREF is the most widely used screening tool for examining executive functions, due to its sensitivity. The cut-off score is 16 for levels 2 and 3, and 15 for level 1 (below the school leaving certificate).

The TMT B, Trail Making Test Part B (Reitan, 1958).In part B of the TMT (TMT-B), the sheet given to the subject includes 13 circles containing numbers from 1 to 13 and 12 circles containing letters from A to L, distributed in a semi-random way on the page. The subject is instructed to connect, as quickly as possible, the numbers and letters in their respective order, alternating each time between a number and a letter. Any errors are not corrected by the examiner. A pre-test containing six items to be connected is carried out beforehand, in order to make sure that the instructions have been understood. The number of correct moves and the time taken to complete the task are collected. The French norms of the TMT B were made available by the GREFEX (Groupe D'Evaluation des Fonctions Exécutives).

The Virtual Action Planning Supermarket (VAP-S) consists of a medium-sized supermarket. The task is to collect 7 items in the supermarket according to a predefined list, then to check out, pay and leave the store. The software allows the collection of the following data: the total distance traveled (in meters), the time (in seconds) needed by the subject to accomplish the task, the number of products purchased, the number of correct actions, the number of incorrect actions (intrusions and perseverations), the number of pauses (greater than 3 seconds), the total duration of the pauses (in seconds), the collisions inside the supermarket The composite score of time per object weighted by the number of objects found ((time/number of objects found)/(number of objects found/7)) has already been validated and published in several studies (Cogné et al., 2016; 2018) and has been shown to be highly correlated with executive functions assessed using paper-and-pencil tests and questionnaires.

Memory test: RL/RI-16: The RL/RI-16, or 16-item free recall/indexed recall test The word list consists of 16 words belonging to 16 semantic categories. The task starts with a categorical semantic encoding phase. Then, an immediate recall, three free/indicative recalls and a delayed free/indicative recall are performed. The aim of this procedure is to elicit effective encoding and retrieval strategies in a patient while monitoring the effective implementation of these strategies. Several studies have shown the usefulness of this procedure in distinguishing demented patients from normal elderly subjects (Christensen et al. 1991; Grober et al., 1988; Grober and Kawas, 1997; Petersen et al., 1994). This test is widely used throughout the world, and has been used extensively in various studies assessing memory in brain-damaged subjects. The duration of the encoding and immediate recall phase is about 10 minutes, and that of delayed recall 3-4 minutes.

Bell barrage test

Catherine Bergego ladder

Subjective Straight Ahead (SSA) haptique,

Subjective Visual Vertical (SVV),

Longitudinal Body Axis (LBA)