**4.Discussion**

**Table 2.** Summary of pinch force, pulling force,

In 274 trials (91.3% of total trials), the volunteers could adjust the pinch force in proportion to the changes in pulling force. However, the remaining trials were excluded prior to analysis due to improper adjustment of pinch force to the pulling force. This is considered as human errors such that the volunteers might not follow the experiment's procedures and instructions probably due to delay in reaction, missing the instructions, or absence of visual feedback. According to the UE-FMA protocol, the subject should resist a gentle pull to avoid slip. This means that the pincer object slips when the subject is no longer able to exert enough pinch force. In this study, it was initially hypothesized that the slip onset was expected at the maximum pinch force just before the point where the pincer object totally slipped. Interestingly, it was found that the slip onset occurred much earlier and was subjected to a sub-maximal pinch force. This result opens up the possibility of using the slip onset as a measurement of pinch improvement such that the patient would develop a stronger pinch force over rehabilitation time at the slip onset. Furthermore, the continuous changes of slip onset would improve the responsiveness of the pinch assessment such that the scoring is not subjected only to score 1 and 2 as in the current UE-FMA scoring criteria.

The pinch–force relationship was found to be linear prior to the slip onset as the static COF between human skin and copper alloy was constant. However, there was some variability in pinch and pulling force, as indicated by peaks approximately at 7.5, 12, and 14.5 N, as depicted in Figure 12. The highest variability was observed at 7.5 N pinch force at the loading phase in which the linear electric actuator started to pull. This is consistent with previous studies stated that the subjects generally show stronger pinch forces at the loading phase [59]. Besides, the pinch force rate (N/s) reached the highest at the loading phase when the volunteer was subjected to unpredictable loading [27]. The study conducted by Takamuku et al. [60] showed that pinch force variability is lower when visual feedback is provided to the subject. In addition, the pinch force control is greatly influenced by the safety margin factor during the unpredictable loading [61]. The safety margin is an additional amount of pinch force applied to guard against slip, which is di fferent from volunteer to volunteer [62]. The other two peaks at 12 and 14.5 N can be interpreted such that the volunteer should adapt his pinch force against the continuous automatic pull without visual feedback; hence, the safety margin is continuously changed to adapt the continuous increase in pulling force. Therefore, it is possible that the volunteer may unconsciously fail to adapt his pinch force to the increase in pulling force at some events. The safety margin is also influenced by the static COF [63], such that it will be high in the case of slippery object as in copper alloy used in our experiment. The pinch–pulling force relationship can be used as an assessment for pinch force control against unpredictable loading for stroke patients.

The CV value (represented by standard deviation divided by the mean) indicates how well the mean values of pinch and pulling forces summarizes the whole dataset. Interpretation of CV may vary; the rule of thumb can be used, which states the CV value of under 1 shows a small variance [64]. In this study, the CV values were 0.248 and 0.35 for pinch and pulling forces of the right hands, respectively. The CV values were 0.241 and 0.31 for pinch and pulling forces of left hands, respectively. Hence, the mean pinch and pulling force values are a reasonable representation of healthy adults in Malaysia. The results also indicate that the pinch and pulling force vary among the volunteers at the slip onset, which is expected as each volunteer has a di fferent body size, hand size, and amount of safety margin. For instance, a bigger fingertip size leads to a larger contact area between the fingertip and the object's surface. Consequently, larger contact area would lead to larger pinch force, as reported in a related study by Derler et al. [65], who showed that there is a positive association between contact area and force exerted by the index finger. In addition, each volunteer has di fferent skin characteristics such as oiliness, moisture, roughness, and age that may a ffect the friction between fingertip skin and pincer object's surface. In this study, the mean static COFs for right and left hands were computed as 0.518 ± 0.146 and 0.517 ± 0.145, respectively. This is relatively consistent with previous findings related to measuring COF between human skin and copper material, reported by Sivamani et al. [66] (mean COF = 0.55) and Li et al. [67] (mean COF = 0.58).

In the literature, many studies have reported a significant di fference in maximum pinch force between right and left hands [68]. Although our results show a slight increase in the means of pinch and pulling force for right hand compared to the left hand, there was no significant di fference (*p* > 0.05). Thus, it is not required to distinguish between right and left hands when a single-gender group is

recruited for further research. Furthermore, the una ffected hand can be used as a reference for the affected hand during the rehabilitation of post-stroke patients. The explanation of this result can be that the volunteers did not exert their maximum pinch force at the slip onset. This is consistent with a study conducted by Li and Yu [69], who reported that there is no significant di fference between left and right hands when the grip force is at 25%, 50%, and 75% of the maximum force. The results of their study indicate that there is a significant di fference only at maximum grip force.

Recently, three main components of the assessment system are required for an assessment system to be fully automated: administration, data acquisition, and rating [19]. Administration involves the instructions and procedures to evaluate the patient, while the data acquisition involves obtaining the outcome measurements from patients. Rating is the criteria to evaluate the outcome measurements of the patients, which are usually designed as an ordinal scale. In this study, only the pinch data acquisition system is automated based on UE-FMA protocol. The limitation of this study is that the administration and rating components are not ye<sup>t</sup> automated. In the future, this limitation will be addressed, such that stroke patients with pinch deficits at di fferent levels of impairment (severe, moderate, and mild) will be recruited to determine the threshold values for each level so that an automated rating process can be performed. In addition, the administration of the pinch test will be automated by including a graphical user interface to visualize the instructions and procedures.

To use the automated data acquisition system in clinical practice based on UE-FMA protocol, the design can be improved to be more compact, less wires, and portable by: (1) replacing the DC power supply with a portable DC battery to power the two amplifiers and LVDT sensor; (2) using a customized small size linear electric actuator as the with pulling force of smaller than 15 N; and (3) using a wireless connection (e.g., ZigBee module) to transmit/receive digitalized signals to/from the PC instead of signal wires. This improvement will be addressed in the future work.

## **5. Conclusions**

In this paper, an automated data acquisition system for pinch assessment based on UE-FMA protocol is presented. The automated system is able to provide objective measurements of pinch and pulling forces and detect slip onset rather than the subjective manual gentle pull and visual observation of slipping occurrence. In addition, the therapist's gentle pull has been replaced with a linear actuator sub-system exerting a consistent amount of pulling force. Right and lefts hands of 50 volunteers were recruited to investigate the pinch and pulling force measurements at the slip onset using the developed system. The pinch–pulling force relationship is linear, which is indicated by a proportional increase of pinch force against the continuous increase of pulling force prior to slip onset. In addition, the volunteers were subjected to submaximal pinch force at the slip onset. The mean pinch force values at the slip onset were 12.17 and 11.67 N for right and left hands, respectively. The mean pulling force values at the slip onset were 6.25 and 5.92 N for right and left hands, respectively. It was found that there is no significant di fference in force measurements between right and left hands. A further study can be conducted to investigate the hypothesis of considering slip onset and pinch–pulling force relationship as a pinch assessment for stroke patients.

**Author Contributions:** A.A. (Abdallah Alsayed) and R.K. designed and developed the pinch assessment system; H.R. performed the integration of sensors and actuator; A.A. (Azizan As'arry) performed CAD and finite element analysis as well as provided access to volunteers; A.A. (Abdallah Alsayed) performed data collection and results analysis; and all authors contributed to writing and editing the paper. All authors have read and agreed to the published version of the manuscript

**Funding:** This research was funded by Universiti Putra Malaysia, gran<sup>t</sup> number IPS No. 9574400.

**Acknowledgments:** We would like to thank Yayasan Khazanah Berhad, Malaysia for their technical and financial support.

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

**Data Availability:** The data and codes used in this study are available from the corresponding author upon request. *Appl. Sci.* **2020**, *10*, 3436

## **Appendix A**

**Figure A1.** Force–voltage relationship during loading and unloading.

**Figure A2.** Flowchart of experiment execution.

**Figure A3.** (**a**) Three trials recorded from one volunteer; and (**b**) displacement in decibel scale.
