**6. Muscle trajectory analysis**

In order to show the trajectory analysis, a set of box plots were used to graphically represent the scatter, symmetry, and outliers found in the muscle trajectory signal. On the graph of the healthy person (**Figure 7**), it can be seen that the upper quartile is larger than the lower one. In addition, the mean values of both sides are closer to last mentioned.

The patient that presents facial palsy (**Figure 8**), it is observed that the interquartile distribution of the injured half-face with respect to the healthy half-face is irregular. It does imply that those are not close to any particular quartile. In addition, outliers are presented in the diagram between the lower quartile and the minimum threshold; likewise, the distance between the outer quartiles and the thresholds is greater compared with the diagrams in **Figure 7**.

The outliers denoted by "+" signs outside the quartiles are neither present in the healthy person (**Figure 7**) nor in the healthy person half-face with paralysis box plot (**Figure 8a**). As the study results show, WaL is effective in quantifying the trajectory of the frontalis, zygomaticus minor, risorius, and zygomaticus major muscles. The data obtained from the tests performed on 15 healthy persons and three with facial

**Figure 7.**

*Muscular trajectory box diagrams in normal state. The names along the x–axis represent the sensor number. (a) Represents the left hemiface side with sensors 1–4 and (b) the right hemiface side with sensors 5–8. Notches were added in the diagrams of the healthy hemiface for purposes of distinction.*

paralysis suggest that the facial correlation percentage of a healthy person is in a span between 89.57% and 97.87%, with a deviation standard between 0.34822 and 0.3433.

Due to the fact that having a healthy face does not guarantee having a symmetrical face, the system does a precalibration, by showing an error coefficient in order to express the asymmetry. Under the assumption that all faces are asymmetric, the facial muscles are normal when the error coefficient is below 10.43%.

According to the graphs generated by the system (**Figure 5**), when one side of the face is in a normal state, the average response speed from the point of largest relaxation to the point of largest contraction is 2.49 seconds while for the injured hemifaces, it was 3.52 seconds. Furthermore, it was observed in the experimental data that when the correlation is high, the data dispersion marked by the standard deviation increases.

From the analyses shown in **Figures 7** and **8**, it is observed that if the contraction muscle time is larger than the relaxation time, the length of the upper quartiles increase; on the other hand, it decreases if the contraction is lower, so it can be stated that the length of the quartiles depends on the muscle contraction as well as the mean. Although the possibility of anomalous values + that appear in the box plots of normalized signals from healthy half-faces is not ruled out, in the 18 studies on which this research is based, outliers were only present in the normalized paralyzed hemifaces.

The person whose results are shown in **Figure 6** was diagnosed with peripheral facial paralysis. It shows weakness on the left side of the face, which was confirmed by the system with an average standard deviation of 0.32232 and a facial correlation coefficient of 72.63%. A value that is outside of the range previously established for healthy people, which means that the electrical potentials presented the muscles of the left side of the face were attenuated by 27.37% due to the paralysis, by making it impossible for the person to perform a symmetrical contraction affecting their muscular excursion.

#### **Figure 8.**

*Muscle trajectory box plots, those represent a person with facial palsy. (a) Shows the components of the left hemiface side muscles on the left hemiface side and (b) the right hemiface side. In addition, both hemifaces display an odd behavior (a) and (b), in contrast with a healthy person shown elsewhere (Figure 7).*

*Medical System to Evaluate the Seventh Cranial Nerve through the Main Facial Mimic Muscles DOI: http://dx.doi.org/10.5772/intechopen.107134*
