*2.4. Force Assessment System Based on Virtual Reality Games*

A Force Assessment System based on VR games was developed to assess the child's improvements in grip force control and modulation. The system was composed of force sensors positioned at the fingertips of the Gloreha Sensor Glove and of custom-designed VR games in the form of tracking tasks purposely developed to assess different aspects of force control and modulation.

Previous works in the literature proposed tracking tasks for the assessment and training of grip force control in mildly to severely affected hemiparetic stroke patients. In the study of Kurillo et al. [22], the authors presented and tested a grip force training system that enabled the improvement of grip force control in 8 out of 10 post-stroke patients. In the study of Lindberg et al. of 2012 [23], the authors proved that the power grip force tracking tasks developed to assess grip force modulation capability were feasible to quantify the accuracy of grip force control.

The force sensors to be embedded in the Gloreha Sensor Glove were chosen according to the following technical specifications: (i) maximum dimension of 2 × <sup>10</sup> <sup>−</sup><sup>2</sup> m diameter; (ii) thickness lower than 5 × <sup>10</sup> <sup>−</sup><sup>4</sup> m; (iii) force range comparable to the one detected during hand rehabilitation training (0–25 N) [24]. The piezoresistive sensor FSR® Model 402 Short Tail (Interlink Electronics) was selected to meet the specifications. It has a diameter of 1.83 × <sup>10</sup> <sup>−</sup><sup>2</sup> m, a thickness of 4 × <sup>10</sup> <sup>−</sup><sup>4</sup> m, a force sensitivity range of 0.2–20 N and continuous force resolution. Piezoresistive sensors were selected because they are low cost, suitable for wearable applications, require a very simple conditioning electronics and have a good shock resistance.

Three sensors were embedded, respectively, in the fingertips of the first, second and third finger of the Gloreha Sensor Glove, i.e., the fingers mostly involved in daily life grasps, as shown in Figure 2. The interface between the sensor and the glove was modified by fixing a thin 3D printed PLA plate with double-sided tape (1 × <sup>10</sup> <sup>−</sup><sup>3</sup> m thickness and <sup>2</sup> × <sup>10</sup> <sup>2</sup> <sup>m</sup> diameter) to uniformly distribute the force on the sensitive area of the sensor. A 3D-printed PLA support was positioned to interlock the plastic components of the glove that allow connecting Gloreha Sinfonia flexible transmissions to the glove itself. The sensors are secured on the fingertips by means of an elastic band fastened at the top of the PLA support.

The sensors with the developed PLA plate were calibrated using the Instron® testing machine to relate the force and output voltage.

The VR game was developed with Unity, using Microsoft Visual Studio as the pre-set editor and C# as the programming language. Attention was paid to provide an intuitive, clear and engaging visual feedback to the user. In fact, providing the patient with biofeedback can improve the outcome of the treatment and promote neuroplasticity [22].

The aim of the VR game in the form of tracking tasks was to assess the patient capability to finely control the sub-maximal forces exerted when grasping real objects, the ability to balance and release the grip and the general accuracy of force control.

The average force exerted by the three fingers was given in input to the VR game to move the avatar according to the force exerted on the object grasped by the participant (i.e., a wood parallelepiped 1.35 × <sup>10</sup>−<sup>1</sup> <sup>m</sup> × 6.5 × <sup>10</sup>−<sup>2</sup> <sup>m</sup> × <sup>4</sup> × <sup>10</sup>−<sup>2</sup> m). In the proposed VR game, shown in Figure 3, the patient was asked to move the avatar of the game vertically, according to the exerted force, in order to track the three proposed waveforms moving on the screen. The three waveforms were a "Ramp", a "Square Wave" and a "Sinusoidal Wave" [25]. Each waveform was developed to assess different aspects of grip force control: the Ramp aimed at assessing the capability to gradually increase and decrease the grip force; the Square Wave aimed at assessing the capability to exert discrete force levels and stabilize the force; and the Sinusoidal Wave aimed at assessing the overall force modulation capability. "Ramp" and "Square Wave" were composed of 10 discrete force levels to be reached (and held for the "Square Wave") and are uniformly distributed between the maximum and minimum forces recorded at the beginning of the trial, whereas the "Sinusoidal Wave" had a peak-to-peak amplitude that corresponded to the range between the minimum and maximum forces.

**Figure 2.** Force sensors positioned at the fingertips of Gloreha Sensor Glove.

To assess the child's improvements in grip force control and modulation, she was instructed to move the avatar (i.e., a turtle) and to follow the proposed waveform pattern in order to collect the maximum number of "bubbles" of the VR game with the avatar. In the first assessment session (T0), she performed 5 one-minute repetitions of the "Ramp" and 3 one-minute repetitions of the "Square Wave" and "Sinusoidal Wave" exercises, respectively. In the second assessment session (T1), she performed 3 one-minute repetitions of the "Ramp" and 2 one-minute repetitions of the "Square Wave" and "Sinusoidal Wave" exercises. Before each session, the maximum and minimum forces applicable by the child were recorded with a custom-made graphical user interface to set the force range in which the avatar could be moved. Then, the maximum force to be reached was set at 90% of the maximum recorded value.

**Figure 3.** VR game for force assessment.

#### *2.5. Intervention*

The Gloreha treatment lasted from January to April 2021, and the commitment was once a week for a total of 10 weeks; each session lasted 60 min. During each session, the child performed passive, active-assisted and active movements with gradually increasing complexity exercises supported and stimulated by sensory feedback. In addition, the child was also offered exercises in bi-dexterity to improve both the quality of the recovery of the paretic limb and its coordination and functionality in daily activities.

#### *2.6. Outcome Measures*

The child was evaluated with functional scales, with upper limb kinematic analysis and with an ad hoc developed force assessment tool (FSR model 402 Short Tail, Interlink Electronics) at pre-treatment (T0) and at the end of the treatment (T1).

She was evaluated with Fugl-Meyer Assessment-Upper Extremity (FMA-UE) for upper-extremity motor impairment and the Visual Analogic Scale (VAS) for pain intensity and activities and participation of daily life with ADL scale [26].

Improvements in force control and modulation capability were assessed by means of the Force Assessment System. Two main performance indicators were computed for each repetition of the three VR exercises: the Root Mean Square Error (RMSE) (N) between the target force pattern and the exerted force pattern and the Peak Performance (%), computed as the percentage of reached force peaks, for "Ramp" and "Sinusoid", and the percentage of reached and held discrete force levels, for the "Square Wave". A statistical analysis was performed to evaluate improvements of the child from one session to the other. The non-parametric one-way ANOVA test (i.e., Kruskal–Wallis test) was conducted between RMSE and Peak Performance indicators in the two sessions. The significance level was set at 5%.

#### **3. Results**

Results are reported in Table 1. At the end of treatment (T1), the patient improved in functional scales: FMA-UE had a percentage variation (Δ%) of 44% from T0 (34/66) to T1 (49/66). Moreover, at the end of treatment, her quality of life was better for greater involvement in some little but significative activities of daily living such as food, dress and undress (ADL scale: 4/6 at T0 vs. 6/6 at T1) and had greater interest in the surrounding world with less fear of feeling different from others.

**Table 1.** Results of outcome measurement.


ADL: Activity of Daily Living; FMA-UE: Fugl-Meyer Assessment for Upper Limb; VAS: Visual Analogic Scale.

It is important to highlight that the child had never complained of pain (VAS scale 0/10) during both evaluations.

The Kruskal–Wallis test performed on RMSE and Peak Performance indicators between sessions at T0 and T1, allowed the assessment of improvements in force control and modulation. RMSE was significantly reduced from one session to the other (*p* value = 0.0018), meaning that training with the Gloreha improved the child's capacity to follow target force patterns. Peak Performance significantly increased from the first to the second evaluation session (*p* value = 0.0120). Boxplots for the two performance indicators are shown in Figure 4, where the central mark is the median and the box edges are the 25th and 75th percentiles.

No adverse events occurred during the entire treatment.

**Figure 4.** Boxplots of RMSE and Peak Performance for session 1 (T0) and session 2 (T1).

#### **4. Discussion**

The aim of this study is to describe the effects on the range of motion, muscle tone and functionality of the paretic upper limb, particularly the hand, when Gloreha Sinfonia combined with conventional therapy was used for rehabilitation treatments in a pediatric stroke patient. The results revealed an improvement in FMA-UE, a significant reduction in RMSE and a significant improvement in Peak Performance.

To the best of our knowledge, this is the first study using Gloreha Sinfonia in pediatric strokes. Previous rehabilitation studies on pediatric stroke patients were mostly dedicated to the robotic treatment of the proximal portion of the upper limb, while there are still very few studies on the distal extremity, especially the hand. Gloreha Sinfonia allows the performance of tasks that combine the activity of the entire upper limb. In fact, unlike other robots designed exclusively for shoulder and elbow movements, Gloreha permits the improvement of distal control through the implementation of exercises focused on the use of hands combined with the involvement of the entire upper limb, reproducing activities of daily living. As other studies have previously mentioned, when a patient uses the distal part of the paretic upper limb, at the same time, the proximal segment is also trained, albeit the upper arm is supported or restrained in the distal group [27]. Furthermore, Gloreha Sinfonia is well suited for use in combination with other traditional rehabilitation activities because it can integrate rehabilitation treatments with highly stimulating and interactive exercises for patients.

Our results show an important enhancement in FMA-UE, with values that proceed from 34/66 to 49/66 between T0 and T1; they indicate an overall improvement of the upper limb due to a clinical advance both in the shoulder district, implemented thanks to reaching exercises, and to the hand control, achieved thanks to manipulation exercises. These results were significantly higher than the minimal clinically important difference (MCID) seen in adults by Page et al. [28], but it is not possible to make a comparison with a pediatric population, because their MCID is not present in the literature. In addition, FMA-UE results are in line with ADL records, which report improvements in activities of daily life due to a better use of the upper limb.

From Figure 4, it is evident that the child significantly improved her capacity of reaching force peaks and holding force levels. Furthermore, the results show that force control and modulation capability significantly increased after treatment.

We know from the girl's parents that she has reported a psycho-emotional improvement and presented new interest in socializing with other children without fear of being judged and perceived less pronounced motor deficits and less disability at the end of therapy. This statement can be considered as another positive effect of the combined (robot + conventional rehab therapy) and continuative therapy. It is important to underline that the girl never complained of pain, demonstrating a good ad safe adaptability of the robot with respect to the child.

Another aspect that makes this robot an interesting tool in pediatric applications is the possibility of performing specific repetitive task-oriented exercises, placed in the form of playing a videogame; as already reported, robot-mediated therapy in children with acquired or congenital brain injury appears to be beneficial: enhancing motivation and improving perception, it incorporates the advantages of the enjoyable game-like experience [15,29]. Furthermore, the use of a virtual and motivational environment in which the patient has immerged has a positive effect to facilitate motor and sensory and cognitive relearning by stimulating the patient to challenge himself to perform better and to obtain a higher score in the game. Moreover, as recorded also by Mirkowski et al. in their systematic review [18], robotic therapy seems to significantly improve the upper extremity function and spasticity in children after strokes. This theory is also supported by the study of Wann et al., which considers motor learning theory and reports how the repetition of motor patterns is seen as a key factor in improving movement [30]. Moreover, Rizzo et al. reports how a rehabilitation treatment that uses feedback, either visual or auditory, contributes to gains made in motor learning [31]. It follows that Virtual Reality is a powerful medium for providing stimulation in the form of visual and auditory events to increase the motivation and desire to continue practicing [32].

To date, there are few studies in the literature on this specific population. Most pediatric upper extremity rehabilitation studies are aimed at patients with cerebral palsy, spinal cord injury or quadriplegia. As reported also in the case report of Colovi´ ˇ c et al. [33], there is still no clear indication on how robotic therapy can be combined with conventional therapy and for how many times; however, combining robotic and traditional rehabilitation can improve the functional motor performance of the arm involved in the chronic recovery phase after a pediatric stroke.

Our results, being relative to a single case, cannot find a generalization. Further studies on larger samples, with control and randomization groups and with adequate follow-ups, are needed in order to reach meaningful conclusions. Furthermore, we have not assessed patient's emotional changes and the level of social integration before and after the robotic treatment with specific and objective measures; the administration of scales that could quantify these changes could be useful to determine social aspects in further studies.

#### **5. Conclusions**

Given the results and given the literature evidence, Gloreha Sinfonia seems to be suitable for the treatment of post-stroke hand disabilities in the pediatric age, but further studies on larger populations, with stratification of the sample for clinical characteristics, as well as clinical scales that are more sensitive to any change are needed to support this hypothesis. It would be interesting in the future to consider how this type of technology can also support cognitive difficulties. In addition, this type of rehabilitative approach facilitates the need for personalized and easily monitored rehabilitation protocols. Clinical trials with follow-ups and on large populations could confirm the results obtained.

**Author Contributions:** Conceptualization, F.B., C.A. and D.M.; methodology, F.S. and F.C.; software, L.Z., F.C. and M.L.; formal analysis, F.S.; investigation, F.N.; data curation, L.Z. and F.C.; writing original draft preparation, F.N., B.C. and L.C.; writing—review and editing, F.B. and L.C.; supervision, S.S. and M.B.; project administration, F.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Fondazione Policlinico Campus Bio-Medico of Rome (protocol code PT-O 20.22; date of approval: 29 March 2022).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.
