3.2.3. Phase Coordination Index (PCI)

The PCI and bilateral stepping phase symmetry were calculated to evaluate gait coordination of hemiparetic patients and healthy adults as illustrated in Figures 12 and 13, respectively. The average value of PCI in hemiparetic patients was 19.50 ± 13.86% and the ϕ\_CV value was 0.20 ± 0.17%, the ϕ\_ABS was 41.09 ± 37.83◦ and the ϕ average was 180.88 ± 56.78◦. The PCI value of healthy subjects was 5.62 ± 5.06%, the ϕ\_CV was 0.12 ± 0.11%, the ϕ\_CV was 9.89 ± 8.99◦, and the ϕ average was 176.64 ± 13.14◦ as summarized in Table 15.

**Figure 11.** Average stride time CV of each subject group. (RH: right hemiparesis, LH: left hemiparesis, BH: bilateral hemiparesis, HA: healthy adults).

**Figure 12.** The results of PCI value comparison between hemiparetic patients and healthy adults.

Figure 13 shows a sample comparison result in a scatter plot of bilateral stepping phase symmetry between a hemiparetic (subject #14; left hemi) patient and a healthy adult (subject #34). The hemiparetic patient's bilateral stepping phase was observed to be widely scattered from the standard normal value of 180◦ compared to that of the healthy subject.

**Table 15.** The results of the phase coordination index (PCI) detection.


**Figure 13.** Sample bilateral stepping phase symmetry values plotted based on the data of, (**a**) left hemiparetic subject number 14; (**b**) healthy subject number 34.

## **4. Discussion**

With the aim of producing a practical and reliable device to quantitatively analyze the human gait, a 10-channeled TCPSI and a real-time monitoring system were developed and applied in this study for hemiparetic gait pattern analysis in comparison with healthy adults. The results of this study showed that our sensor is capable of detecting and distinguishing the differences in plantar pressure (PPD), the step count, the stride time, the coefficient of variation, and the phase coordination index (PCI) between paretic and healthy limbs. These parameters were selected for analysis, because a stroke survivors' gait abnormality is characterized by a pronounced gait asymmetry [25], decreased gait speed and stance phase, shorter stride length, and prolonged swing phase of the paretic limb [39]. These gait abnormalities along with muscle imbalance and weakness leads to a high risk of falls [40,41].

In this study, the subjects with hemiparesis showed the highest PPD in the toe area where the sensor numbers 1 and 2 were located (right hemi: 13.43%, left hemi: 28.27%). The closer to 0% the PPD is, the less the differential pressure of the feet, which means that the feet are well balanced [17]. Sanghan et al. (2015) reported that the PPD of hemiparetic patients was 3 times higher than that of the healthy group [17]. Normally in healthy subjects, the heel contacts first, followed by the midfoot, the forefoot, and then toe pressure. However, since hemiplegic patients do not pressurize the paralyzed feet properly when walking, it may have showed higher pressure in the toe area and in general compared to the healthy adult group. This result was congruent with the results in Perry's research (1992) [39]. Based on this, our proposed sensor was confirmed that it can detect the dynamic pressure difference of the two feet and be analyzed. In gait analysis, plantar pressure distribution is an important parameter for evaluating the balancing capability in terms of weight-shifting, balance strategies, and a risk of falls.

The plantar pressure distribution comparison analysis showed that the plantar pressure moves towards the lateral side of the affected foot in hemiparetic patients. This phenomenon occurs as a form of compensation to avoid falling. The patients shift their bodyweight towards the unaffected side for more secured stability [42]. The plantar dynamics of the involved leg exhibited a transfer of initial contact from the hind to the forefoot, increased lateral plantar support, limited roll-over, and reduced or absent push-off at a terminal stance. There was a tendency for heel support to disappear on the paretic side with reduced functional abilities.

Field et al. (2013) reported that the patients with a stroke reduce daily stepping activities by 27% compared to a healthy adult [43]. This was also found in our study results. The step count of hemiparetic patients was higher than that of healthy adults.

The average stride time difference between the right and left foot of left hemiparesis was 0.14±0.04 s (Table 4), those of the right hemiparesis was 0.11 ± 0.04 s, and of the bilateral hemiparesis was 0.04 ± 0.04 s, respectively (Tables 5 and 6). The temporal aspects of hemiplegic gait are characterized by increased stride times [44]. Hemiplegic patients usually have reduced joint excursion and insufficient forward propulsion, which may lead to an asymmetrical and unstable walking pattern [5]. Since the single support time of the affected limb is significantly shorter than that of the unaffected limb, the

unaffected limb's step length is shorter than the affected limb's step length [4,6]. Likely, the stride time of the affected foot was longer than the unaffected side in our study.

As presented in Table 6, the time difference in the average stride time in healthy adults was 0.02 ± 0.01 s and the average stride time difference was three times lower than the hemiparetic patients. Mackenzie et al. (2006) and Sartini et al. (2010) reported that the stride time was related to falling and an abnormal gait pattern causes a risk of falls [45,46]. The elderly who have experienced falls are associated with a decrease in walking speed, stepping and an increase in step symmetry [47]. Our proposed sensor system can also detect and decrease the risk of falls by real-time gait monitoring.

The CV is used to compare the magnitude of the change regardless of the data dimension. In this study, because the step length and the stride time may be different depending on the subject and the degree of the disease, the CV was calculated and the change amount was examined. The CV of stride time with heathy adults was 5 times higher than hemiparetic patients as shown in Table 8 and Figure 9. This means that hemiparetic patients cannot walk with a constant gait speed and thus walk irregularly. It was confirmed that our sensor can analyze various gait parameters.

This study performed the PCI comparison between hemiparetic patients and healthy adults because hemiparetic gait is characterized by mild to severe asymmetric patterns. Hemiparetic patients have increased stance phase and double support duration. In addition, the stride length and step length decrease, and the paralyzed side has longer step [48]. The PCI is an indicator for evaluating the balance of a pair of feet. A value closer to 0% refers to a higher balance between the two feet [23–25,48]. The PCI was originally developed to evaluate the asymmetry during walking and many studies evaluated the gait asymmetry of patients with Parkinson's disease and a stroke [23–25,48]. In the previous studies evaluating gait asymmetry in patients with stroke patients [48], the PCI value of stroke patients was 19.5% ± 2.3% and that of healthy subjects was 6.2% ± 1.0%. The PCI value was approximately 3 times higher in the patient group than in the healthy group.

In this study, the PCI value of stroke patients was 19.5% ± 13.9% and that of healthy subjects was 5.6% ± 5.1%, showing similar results to the previous studies as summarized in Table 7. Our results showed a higher standard deviation of the PCI value than that of previous studies. This may be because the age range of the subjects participating in this study was larger (63.3 ± 8.6 yrs.) The ages of the subjects (number 2, 6, 8, 10, and 14) were 65, 63, 63, 60, and 68, respectively, and were diagnosed with hemiparesis. The PCI values of the five subjects were 4.07%, 4.00%, 4.64%, 0.98%, 4.03%, respectively, which seemed to be similar to those of healthy adults. According to the previous studies, aging is associated with decreased stride time, velocity, step length during gait. Slowed walking speed leads to larger gait variability [37]. Other previous studies explored age-related changes by analyzing gait speed and PCI values, and found out that the average gait speed decreases and the PCI value increases every decade from the age of 70 years [49–51].

Study limitations: Generalizing the results of this study may be difficult because the number of subjects was small. In our previous study [35], a soft-material-based pressure insole and its performance feasibility test for performing a clinical experiment was developed and analyzed. The study results showed that our pressure insole sensor was confirmed to have a high correlation with F-scan (R<sup>2</sup> <sup>≥</sup> 0.90, *p* < 0.05). Therefore, the sensor performance analysis was not repeated in this study.
