journal of advanced mechanical design, systems, and

Journal of Advanced Mechanical Design, Systems, and

Manufacturing

Vol. 7, No. 1, 2013

82

Development of a Whole Body Motion Support Type Mobile Suit and Evaluation of Cerebral Activity Corresponding to

the Cortical Motor Areas*

Eiichirou TANAKA**, Shozo SAEGUSA*** and Louis YUGE**** ** Department of Machinery & Control Systems, Shibaura Institute of Technology

307 Fukasaku, Minuma-ku, Saitama, Japan E-mail: [email protected]

*** Center for Collaborative Research & Community Cooperation, Hiroshima University 2-313, Kagami-yama, Higashi-Hiroshima, Hiroshima, Japan

**** Graduate School of Biomedical & Health Sciences, Hiroshima University 1-2-3, Kasumi, Minami-ku, Hiroshima, Japan

Abstract We developed a new whole body motion support type mobile suit. This suit can be used separately for supporting the upper and/or lower limbs, for assisting in ADL (Activities of Daily Living). We also developed a mobile lifter system which can bear both the equipped person and the suit. This suit and the lifter can be used by motor palsy patients, people who have suffered a stroke, spinal-cord-injury patients, and people with central nerve disorders. Using this device, these patients can recover normal gait with no risk of falling. In this paper, the cerebral activity during walking using the suit and normal gait without the suit are compared. According to multiple trials with the suit on a treadmill, the activities of the premotor area sensory motor cortex decreased. Especially, while walking using the suit for supporting lower limbs without swinging arms, the cerebral activities of most of the areas decreased. This data shows that it bcomes ineffective for patients accustomed to the suit in rehabilitation. However, on the contrary, by walking while swinging bilateral arms (even though these arms were assisted by the suit), the activity in the supplementary motor area increased (this area of the brain is related with memory of motion). Furthermore, the cerebral activities while walking on a treadmill and while walking in a corridor with an outside view were compared using NIRS (Near-Infrared Spectroscopy). From the results of this experiment, we found it is most effective for gait training to actually walk and not stay fixed in one location. We also found it is important for patients to swing their arms during gait training in rehabilitation.

Key words: Neuro-Rehabilitation, Gait Training, Whole Body Motion Support Type Mobile Suit

1. Introduction

Recently, Neuro-Rehabilitation, which repairs the neural circuit by directly inputting walking movement to the legs has attracted attention, and numerous training machines have been developed. Neuro-Rehabilitation is neuroscience based rehabilitation. In this paper, we focused on Neuro-Rehabilitation for stroke patients to obtain the normal gait. When the ideal leg gait motion for the motor palsy patient is input by using a gait assistance device,

*Received 23 July, 2012 (No. 12-0302) [DOI: 10.1299/jamdsm.7.82]

Copyright © 2013 by JSME

Journal of Advanced Mechanical Design,Systems, and Manufacturing

Vol. 7, No. 1, 2013

83

the subject regains his/her motor function, and the motion activates the subject’s brain on the undamaged area. Then the neural network on the brain is reconstructed. Miyai et al. said NIRS (Near-Infrared Spectroscopy) was useful to evaluate cerebral activity during walking and hypothesized that the alternating leg movements seen during gait are controlled by the leg regions of the primary motor cortex and other related motor areas [1]. Suzuki et al. showed that the prefrontal and premotor cortices are involved in adapting walking and running speed on the treadmill using NIRS [2]. They also indicated that preparation for walking cued by a verbal instruction enhanced frontal activations both during the preparation and execution of walking as well as walking performance [3]. Furthermore, Yano et al. developed a footpad-type locomotion interface and by using NIRS, they confirmed it activated user’s brain more effectively than walking on a treadmill [4]. Therefore, it is more effective to use the motion assistance device than normal training. However, most of these machines require the motor palsy patient to remain stationary during use (e.g., [5]). To address the needs and problems noted above, we developed a new whole body motion support type mobile suit. This suit can be used separately for supporting the upper and/or lower limbs, for assisting in ADL (Activities of Daily Living). Many conventional devices [6][7][8] are fixed to arms and legs tightly. In contrast, this suit is restraint free of user’s arms and legs for reducing the feeling of oppression and discomfort by the slight difference of the points between the suit and user’s joints and it can be easy to use for the patients even if whose arms or legs are distorted. We also developed a mobile lifter system which can bear both the equipped person and the suit. This suit and the lifter can be used by motor palsy patients, people who suffered a stroke, patients with spinal-cord-injury, and people with central nerve disorders. Using this device, these patients can recover normal gait with no risk of falling. Therefore, the purpose of this research is to study and find the method of effective gait training using the suit, and to grasp the inclination of the cerebral pattern for able bodied persons at the first stage. Previously, we compared the cerebral activity during walking on a treadmill using the suit and normal gait without the suit using NIRS [9]. However, according to the subjects accustomed to the suit, the cerebral activity decreased. Therefore, in this paper, the cerebral contour maps between un-accustomed and accustomed subjects were compared and found the difference of the activated area. And the influence of the difference of the training place (on a treadmill or in a corridor) to the cerebral activity was also compared. From these results, We found it is important for patients to swing their arms and actually walk during gait training from the aspect of motivation in Neuro-Rehabilitation.

2. Development of a whole body motion support type mobile suit

We developed a whole body motion support type mobile suit as shown in Fig. 1. This suit can be used separately for supporting the upper and/or lower limbs as shown in Figs. 2 and 5. The assistance apparatus for upper limbs (Fig. 2) is designed for the patients who can control their finger but they cannot lift up their arms themselves, especially myopathy patients. The mechanism of the assistance is utilized the differential gears to lose the weight and volume of the mechanical arm, and that enabled to configure three motors to drive two DOFs (Degrees of freedom) for the shoulder and one DOF for the elbow around the root of the mechanical arm as shown in Fig. 3. The arm can lift up to 4 [kgf] at the tip of the extended arm, and each angle velocity is 1.57 [rad/sec]. This arm has two support trays, for wrist and upper arm. Each tray is equipped a pressure sensor at the contact point to user’s arm, and by using the measured result of these four sensors, the control computer can be obtain the state of the user’s arms. Furthermore, to realize to do some ADL (activities of daily living) motions (for instance, eating, writing, making up, wiping his/her face, and so on) them selves, we proposed to control the device using the targeted posture map for the


Vol. 7, No. 1, 2013

84

mechanical arm as shown in Fig. 4. At first, various desired or necessary postures of the assisted arms for the equipped person are determined, and each posture is defined as a control target. Next, each target (which is relatively close) is mutually connected, and the map can be accomplished. To be able to choose the appropriate input way for each patient, various input interfaces, for example, joy-stick, push buttons, sensor glove using bending sensors, and so on, are equipped as shown in Fig. 5. In this paper, this device is combined to lower limb assistance device and used for assisting arm swinging during walking. The center of Fig. 2 shows the example for assisting to play the violin inputted with the sensor glove, and the right of the Fig. 2 shows the example for the motion of drinking inputted with the joy stick. This target mapping system is for patient safety and feedbacked from patients’ opinion.

We also developed the lower limb assistance device for supporting walking as shown in Fig. 6 [10]. We have developed this walking assistance apparatus and confirmed the effectiveness for an able bodied person by using it. For instance, muscle activity during walking decreased [11], and the mean power of frequency of electromyography during walking, i.e., the subject might be able to prevent fatigue [12]. These walking experiments for able bodied persons using our apparatus were already obtained the permission with Shibaura Institute of Technology Bioethical Committee.

By combining these devices, the whole body motion support type mobile suit was developed. This suit can be used without fixation of user’s arms and legs, different from many conventional suits [6][7][8], so it can be easy to use for motor palsy patients even if their arms and legs are deformed. We also developed a mobile lifter system which can bear both the equipped person and the suit as shown in Fig. 7 [10], and can take into account the vertical fluctuation during walking. This suit and the lifter can be used by motor palsy patients, people who have suffered a stroke, patients with spinal-cord-injury, and people with central nerve disorders. Using this device, these patients can recover normal gait with no risk of falling. This suit is equipped with 10 motors, and the total power is 950 [W]. Total weight is 25 [kg]. The control system of the suit is shown in Fig. 8. In this system, the self-contained computer (LEPRACAUN-CPU, General Robotics Inc.) controls 10 motors in a lump and control interval is 1 [msec]. Four pressure sensors which are located under soles (thenar eminence and heel) can be grasped the timing of heel off (start of swing phase) as shown in Fig. 9. Furthermore, four ultrasonic sensors attached on the top and heel of the flat steps measure the distance to stairs, and the computer controls to avoid the collision with the flat steps and stairs. The gyro sensor measures the angle of the slope where the equipped person is standing, and compensates the trajectory of the flat steps during walking.

Walking assistance control method is as follows: According to the estimated walking phase, both flat steps under the soles are controlled along the trajectory which was calculated by using the data of the equipped person’s legs and a variation of data from hip, knee, and ankle joints as written in medical books [13]. Simultaneously, both upper arms are swinging by the range between -20 [deg] and 30[deg] for shoulders, and between 0 [deg] and 20[deg] for elbows (data was determined from observing able bodied walking).

3. Measurement of cerebral activity with NIRS during walking

It is important to activate the brain extensively, not only PFC (prefrontal cortex), because the areas of PSC (primary somatosensory cortex), PMA (premotor area), PMC (primary motor cortex) , and SMA (supplementary motor area) are used during motor learning [14]. To verify the effectiveness of our developed suit for Neuro-Rehabilitation, the cerebral activity during walking using the suit and normal gait without the suit are measured with NIRS and compared. We used ETG-4000 (Hitachi Medical Corporation) and measured the variation of Oxy-Hb with 44 probes at tetragonal intervals of 3 [mm]. The measured


Vol. 7, No. 1, 2013

85

areas are PFC (associated with consideration), PMA (associated with generate motion with sense), SMA (associated with generate motion with memory), PMC (associated with the output signal of motion), and PSC (associated with the received signal of somatesthesia) as shown in Fig. 10. The experimental block design is as follows:

Fig. 1 Whole body motion support type mobile suit

Fig. 2 ADL assistance apparatus for upper limbs

Fig.3 Mechanism of the assistance device for upper limbs


Vol. 7, No. 1, 2013

86

Fig. 4 Example of the targeted posture map for the assistance device for upper limbs

Fig. 5 Input interfaces

Fig. 6 Walking assistance apparatus for lower limbs


Vol. 7, No. 1, 2013

87

Fig. 7 Mechanism of the weight bearing lift

Fig. 8 Control system of the suit

Fig. 9 Measurement of walking phase

Detachable

Operation PC

Gyro Sensor

Ultra SonicSensor ×4

Upper limbs ×

×8

Encoders ×10

DC motors ×10 (42W ×4, 90W ×2, 150W ×4

5V Battery

24V Battery


Vol. 7, No. 1, 2013

88

Fig.10 Configuration of each probe

{Rest (Pre) 20[sec], Task 40 [sec], Rest (Post) 40 [sec]}*5 [sets].

Each group of eight tasks is shown in Table 1. By comparing the results of Task 1, 2, 3,

4, the influence of the difference of walking velocity and swing arms or not are estimated. The walking velocity of each task was determined in response to the value which each subject felt comfortably during walking previously. By comparing the results of Task 3, 4, 5, 6, the influence of assistance or not and swing arms or not are estimated. By comparing the results of Task 7 and 8, the influence of assistance or not for swing arms are estimated. In Task 6-1, subjects walked as well as grasping the grip which was attached the treadmill. In Task 6-2, subjects crossed arms during walking. Each task was carried out five times respectively and averaged. Each measured data was processed as the baseline, which was connected mean data while 10-20 [sec] and 90-100 [sec]. Furthermore, to compare each area and subject, Effect size was calculated by using Eq. (1).

Effect Size = (Rmean - Nmean) / NSD, (1)

where Rmean [mMmm] is the mean data of each task, Nmean [mMmm] is the mean data of Task 1, NSD [mMmm] is the standard deviation of the Rest of Task 1. For fear of the fatigue by the influence of each task’s turn, the tasks were carried out in order of low fatigue for un-accustomed subjects we assumed, i.e., 1) Only swing arms (Tasks 7, 8), 2) Walking without the device (Tasks 1, 2, 3, 4), 3) Walking with the device (Tasks 5, 6). Furthermore, we asked all subjects to take an enough rest break between each task.

4. Experimental results of walking on a treadmill and considerations

The first experiment was carried out on a treadmill as shown in Fig. 11. To maximize output and to reduce expended power for the compensation of the suit’s weight, the weight


Vol. 7, No. 1, 2013

89

Table 1 Example of tables Task No.

Walking velocity [km/h]

Assist Arm swing

1 0.8-2.8 None Yes 2 0.8-2.8 None No 3 0.5-0.6 None Yes 4 0.5-0.6 None No 5 0.5-0.6 Upper and lower limbs Yes

6-1 0.5-0.6 Lower limbs No (grasp grip)6-2 0.5-0.6 Lower limbs No (cross arms)7 0 (Standing) None Yes 8 0 (Standing) Upper limbs Yes

bearing lifter was used to lift up the suit. However, to treat the condition of the floor reaction force the same for all Tasks, the lift didn’t lift up subject’s weight. Subjects were six; three un-accustomed subjects (Able bodied men, age: 21-24, height: 170-176 [cm], weight: 60-66 [kg]) and three accustomed subjects (Able bodied men, age: 22-25, height: 170-176 [cm], weight: 60-65 [kg]). Un-accustomed subjects did not have an experience to use the suit, and practice to walk with the suit for only one hour before measured. Accustomed subjects have an experience to use the suit over one year as a staff of developing the suit. The mean results of the brain activity with each task for three un-accustomed subjects and three accustomed subjects are shown in Figs. 12 and 13.

From Fig. 12, most of all results of the brain activity of unaccustomed subjects are activated than the result of Task 1. The results of Task 2 and 4 hardly changed, and the result of Task 3 slightly increased. Therefore, it is possible to influence the variation of walk velocity to cerebral activity. The results of Task 5 and 6 were also higher than the results of Task 3 and 4. PFC is especially increased, therefore the subjects might be more conscious for walking motion to keep own posture, by wearing the apparatus. However, from the result of Task 6-1 (grasp the grip) in Fig. 13, PFC is decreased. By grasping the grip and receiving lower leg’s assistance, subjects did not have to control to keep posture and consider how to behave to walk. And from the result of Task 6-2 (cross arms) in Fig. 13, the activities of PFC, PMA, PMC and PSC; the areas are concerned about his/her motion, are decreased.

The contour maps of the mean results of Task 5 and 6-2 for three un-accustomed and three accustomed subjects are shown in Figs. 14, 15, 16 and 17. Apparently, the results of un-accustomed subjects in Figs. 14 and 16 are activated extensively. Especially, the value at Ch 34 in the area of PMCr was increased. According to Penfield’s motor homunculus, it is possible to control a subject’s trunk or left arm with emphasis, because subjects are dextral. On the contrary, the result of accustomed subjects in Figs. 15 and 17 are inactivated except for the PFC area. However, only the value at Ch 38 in the area between PSCr and PSCl was activated in both figures. According to Penfield’s sensory homunculus, it is possible to increase the received signal of somatesthesia from subject’s legs, because they relied on the apparatus. Therefore, it is hard to gain effective training only while walking relying on the apparatus without swinging arms. And the results indicate the subjects accustomed the apparatus and they could walk smoothly, because PMA is concerned about the information about the sensation. In addition, from the results of our precedence research, the muscle activity of the equipped person’s leg decreased while walking and wearing the lower limbs assistance apparatus [11]. Therefore, from the results of the cerebral and muscle activities, it is effective to use the lower limbs assistance apparatus as a vehicle. However, from the result of Task 5, even though the subjects are accustomed, PFC and SMA are slightly activated. By swinging arms, inactivation of the cerebral activity is prevented, and it is


Vol. 7, No. 1, 2013

90

effective for Neuro-Rehabilitation. From the results of Tasks 7 and 8, only swinging arms activated their brains, even

though with the upper limbs assistance and by the accustomed subjects. Especially, SMA is activated. It could also be possible to activate CPG (central pattern generator) which generates rhythmic motion.

Fig. 11 Component devices of the experiment for gait training on a treadmill

Fig. 12 Results of mean data of each task on a treadmill (Three un-accustomed subjects)


Vol. 7, No. 1, 2013

91

Fig. 13 Results of mean data of each task on a treadmill (Three accustomed subjects)

Fig. 14 Contour map of the result of mean data of Task 5 on a treadmill (Three un-accustomed subjects, at intervals of 0.5 [effect size of Oxy-Hb level])

Fig. 15 Contour map of the result of mean data of Task 5 on a treadmill (Three accustomed subjects, at intervals of 0.5 [effect size of Oxy-Hb level])


Vol. 7, No. 1, 2013

92

Fig. 16 Contour map of the result of mean data of Task 6-2 on a treadmill (Three un-accustomed subjects, at intervals of 0.5 [effect size of Oxy-Hb level])

Fig. 17 Contour map of the result of mean data of Task 6-2 on a treadmill (Three accustomed subjects, at intervals of 0.5 [effect size of Oxy-Hb level])

5. Experimental results of walking in a corridor and considerations

Next experiment was carried out in a corridor with an outside view as shown in Fig. 18. Figure 19 shows the mean result of two accustomed subjects. To distinguish the experiments between on a treadmill and in the corridor, the task numbers of the experiment in a corridor are labeled with a prime “ ’ “ symbol (for example, Task 1’). From the result of Task 1’, all area is activated except for SMA. Gait training in a corridor is apparently more effective for Neuro-Rehabilitation than the training on a treadmill. In Task 5’, all subjects said that they could walk the most comfortably in all experiments, however, the result is almost the same outcome as the result of Task 5 in Fig. 13. Therefore, if the patient cannot do gait training only on a treadmill according to his/her condition or environment, it is possible to achieve the effective rehabilitation by swinging arms (even though they are assisted with a suit). In the result of Task 6-1’ (grasp the grip), most of the areas are activated, however, the result of Task 6-2’(cross arms) was inactivate except for PFC, PMCl and PSCl. From the subject’s comments, even though he didn’t swing his arms, by grasping the grip while walking, he felt like as if he drove a kind of vehicle. Consequently, it is most effective for gate training to walk as well as moving from the aspect of motivation.


Vol. 7, No. 1, 2013

93

Fig. 18 Component devices of the experiment for gait training in a corridor

Fig. 19 Results of mean data of each task in a corridor (Two accustomed subjects)

6. Conclusions

We developed a whole body motion support type mobile suit. The cerebral activity during walking using the suit and normal gait without the suit, the difference of the training place (on a treadmill or in a corridor) are compared using NIRS. From these trials, we found it is important for patients to swing their arms and actually walk during gait training, from the aspect of motivation in Neuro-Rehabilitation. We can conclude that this suit is suitable to assist patients in gait rehabilitation.


Vol. 7, No. 1, 2013

94

References

[1] I. Miyai, et al., “Cortical mapping of gait in humans: A near-infrared spectroscopic topography study”, Neuro Image, 14, (2001) , pp1186-1192.

[2] M. Suzuki, I. Miyai, T. Ono, I. Oda, I. Konishi, T. Kochiyama, K. Kubota, “Prefrontal and premotor cortices are involved in adapting walking and running speed on the treadmill: an optical imaging study”, NeuroImage 23, (2004), 1020-1026.

[3] M. Suzuki, I. Miyai, T. Ono, K. Kubota, “Activities in the frontal cortex and gait performance are modulated by preparation. An fNIRS study”, NeuroImage 39, (2008), 600-607.

[4] H. Yano, Y. Nakajima, H. Iwata, “Evaluation of Effect of Walking Using Near-Infrared Spectroscopy”, Journal of the Virtual Reality Society of Japan, Vol. 12, No. 1, pp. 67-74, 2007. (In Japanese)

[5] Colombo, G. et al., “Treadmill training of paraplegic patients using a robotic orthosis”, Journal of Rehabilitation Research and Development, Vol. 37 No. 6, 2000, pp. 693-700.

[6] Adam B. Zoss, et al., “Biomechanical Design of the Berkeley Lower Extremity Exoskeleton (BLEEX)”, IEEE/ASME Transactions on Mechatronics, Vol. 11, No. 2, April 2006, pp. 128-138.

[7] Y. Sankai, ”HAL: Hybrid Assistive Limb Based on Cybernics”, Robotics Research, The 13th International Symposium ISRR, pp.25-34, 2010.

[8] J. Yonetake and S. Toyama, “Development of the Ultrasonic Motor-Powered Assisted Suit System”, Proc. of the 1st KSME-JSME Joint Int. Conf. on Manufacturing, Machine Design and Tribology (ICMDT), Seoul, Korea, CD-ROM, (2005).

[9] E. Tanaka, J. Setoguchi, T. Mori, S. Saegusa, L.Yuge, “Development of a Whole body Motion Support Type Mobile Suit and Evaluation of Cerebral Activity while Walking Wearing the suit”, Biomechanisms 21, Physical function: assistance and improvement research, ISBN978-4-7664-1955-9 C3047, pp. 145-156. (In Japanese)

[10] E. Tanaka, T. Ikehara, H. Yusa, Y. Sato, T.Sakurai, S. Saegusa, K. Ito, L. Yuge, “Walking-Assistance Apparatus as a Next-Generation Vehicle and Movable Neuro-Rehabilitaiton Training Appliance”, Journal of Robotics and Mechatronics, Vol. 24, No. 5, (2012), pp. 851-865.

[11] E. Tanaka, T. Ikehara, T. Omata, T. Owada, K. Nagamura, K. Ikejo, T. Sakamoto, Y. Inoue, “Development of a Walking Assist Machine Supporting Whole Legs from Their Soles”, Transactions of the Japan Society of Mechanical Engineers, Series C, Vol.72, No.724 (2006), pp. 3871-3877. (In Japanese)

[12] E. Tanaka, T. Ikehara, Y. Sato, H. Yusa, K. Ito, S. Saegusa, K. Nakagawa, Y. Aokage, L. Yuge, “Development of a Walking Assistance Apparatus without Fixation on Legs and Study on the Assistance Effectiveness with Electromyography”, Transactions of the Japan Society of Mechanical Engineers, Series C, Vol.77, No.775 (2011), pp. 1119-1132. (In Japanese)

[13] Nakamura, R. et al., Fundamental Kinesiology, sixth edition, ISBN978-4-263-21153-3, Ishiyaku p ubrishers inc, (2003), (in Japanese).

[14] H. Iwata, K. Arimura, S. Sugano, “Patient Internal State Identification based on Neural Reaction and Motor Performance for Cognitive Neuro Rehabilitation”, Proc. of the 22nd Biomechanism Symposium in Aso 2011, (2011), pp.51-58. (In Japanese)

journal of advanced mechanical design, systems, and

Documents