A Low-Cost Balance Training Platform Using Augmented Reality in Neurorehabilitation: a Usability Study

1. INTERVENTION The first phase of the intervention will last four weeks and will be done in the hospital. After this intervention, if the patients meet the inclusion criteria, they will be enrolled in the second phase, which will be done at the patient's home for two additional weeks. Three assessments will be performed during the study: The first one ("Day 0") will be assessed before starting the treatment, the second one will be done ("Week 4") directly after the intervention at a hospital setting, and the last one ("Week 6"), after the home intervention. The intervention will be conducted in 1-hour sessions three times a week for four weeks, performing different exercises using an augmented reality platform accessed through RGSweb. If the patients continue the intervention at home, a three-session training for two weeks will be conducted in the presence of a patient's relative, who will accept to be the training partner.
2. EQUIPMENT AND TECHNICAL SPECIFICATIONS. RGSweb can be used with any computer or device with a camera. However, our target users are patients who might have visibility issues due to old age. Therefore, bigger screens will be used to enhance the patients'; visibility. Additional screen support will be required to hold the monitor.

There are three protocols in RGSweb:

Twister Buddy: It has been developed for balance training, which involves a series of levels intended to strengthen the muscles crucial for maintaining upright posture, notably those in the legs. The aim is to improve stability and help prevent falls.

Levels The user will have to shift their weight from one leg to the other while doing different exercises, such as touching targets with their hands and feet or taking steps.

All levels will start with a calibration phase to ensure the user is in the correct position.

Level 1 -Balance: In this level, two colored columns will be displayed on the screen, while user motion trackers will also be reflected on the screen. The goal is to align the trackers with the same color column.

Level 2 - Balance + Hands: In addition to the balance movement from Level 1, a target will be displayed on the screen. The user must maintain balance on the highlighted column while reaching the target with their hand.

Level 3 - Side Steps: To initiate, the user must have their feet on a square-like target at the bottom center of the screen. After that, a target will appear adjacent to the square. The user's task is to adjust their foot position so that both the toe and heel are inside this target area. Once achieved, the user must return their foot to the original position within the square, prompting the appearance of the next target for the opposite foot.

Level 4 - Front steps: Similar to level 4, the user begins by positioning both feet on the square target. Then, an arrow will appear above one of the user's feet, indicating the direction of the step (either forward or backward). Once the user moves the designated foot past a predefined threshold value, currently indicated by a slider, the arrow will disappear, prompting the user to return both feet to the square. Subsequently, an arrow will appear above the other foot, indicating the direction for the next step. The direction of the steps is randomized for each interaction.

Level 5 -- Balance + Hands + Feet: This is the next step from level 4. It starts with Balance, then the target for the hand, and finally, the target for the opposite foot. Like level 2, the user must maintain only the bottom part (hip and under) inside the column while reaching the target with their hand and foot.

Costume Party: This protocol's main focus is sit-to-stand. The levels of this exercise follow different steps to stand up from a chair, starting with leaning forward.

Levels All levels will start with a calibration phase to ensure the user is in the correct position.

The objective is to dress up for a costume party using a reference shown at the bottom right of the screen. Various costume parts will appear on a bar and move across the screen. The user needs to catch the right parts for their costume and put them on. Once the costume is complete, there will be a few seconds to admire it before the process starts over with a new costume reference.

Level 1-Lean Forward: In this level, the costume elements move from one side of the screen to another at the user's eye level.

To grab the element, they must lean forward enough to surpass a threshold value defined by a slider.

Level 2-Flex Knees: The user starts the motion of standing up by flexing the knees and rising slightly from the chair.

The elements will appear close to the user's head, and leaning forward is unnecessary.

Level 3-Stand Up: This level is similar to level 2, but the elements are positioned higher, requiring the user to stand up to reach them.

Gentle Giant: This exercise targets less-impaired patients who can perform sit-to-stand and balance training exercises.

Levels:

The aim is to allow the elements to move from one platform to another following color coding. To do so, the user will act as a bridge between the different floors.

All levels will start with a calibration phase to ensure the user is in the correct position.

Level 1-Squats: In this level, the user must move the elements from one platform to another by moving the bridge up and down, that is, by squatting to go down and standing again to come up.

Level 2-Tip-Toe: In this level, one platform will be higher than the user's height. The aim is similar, though: The user must stand in a tip-toe position to move the bridge higher and connect the platforms.