Entropy Could Quantify Brain Activation Induced by Mechanical Impedance-Restrained Active Arm Motion: A Functional NIRS Study.

Brain activation has been used to understand brain-level events associated with cognitive tasks or physical tasks. As a quantitative measure for brain activation, we propose entropy in place of signal amplitude and beta value, which are widely used, but sometimes criticized for their limitations and shortcomings as such measures. To investigate the relevance of our proposition, we provided 22 subjects with physical stimuli through elbow extension-flexion motions by using our exoskeleton robot, measured brain activation in terms of entropy, signal amplitude, and beta value; and compared entropy with the other two. The results show that entropy is superior, in that its change appeared in limited, well established, motor areas, while signal amplitude and beta value changes appeared in a widespread fashion, contradicting the modularity theory. Entropy can predict increase in brain activation with task duration, while the other two cannot. When stimuli shifted from the rest state to the task state, entropy exhibited a similar increase as the other two did. Although entropy showed only a part of the phenomenon induced by task strength, it showed superiority by showing a decrease in brain activation that the other two did not show. Moreover, entropy was capable of identifying the physiologically important location.

There is No test-retest reliability of brain activation induced by robotic passive hand movement: A functional NIRS study.

INTRODUCTION: The basic paradigm of rehabilitation is based on the brain plasticity, and for promoting it, test-retest reliability (TRR) of brain activation in which certain area of the brain is repeatedly activated is required. In this study, we investigated whether the robotic passive movement has the TRR of brain activation. While active training has been shown to have TRR, but there still have been arguments over the TRR by passive movement. METHODS: In order to test TRR, 10 repetitive sessions and various intervals (1 day, 3 days, 7 days, 23 days, 15 min, and 6 hr) were applied to five subjects, which had the same statistical power as applying two sessions to 50 subjects. In each session, three robot speeds (0.25, 0.5, and 0.75 Hz) were applied to provide passive movement using the robot. The fNIRS signal (oxy-Hb) generated in the primary sensorimotor area (SM1) was measured on a total of 29 channels. At this time, we used activation maps and intraclass correlation coefficient (ICC) values to examine the TRR and the effect of robot speeds and intervals on TRR. RESULTS: As a result, activation maps showed prominent variation regardless of robot speeds and interval, and the ICC value (=0.002) showed no TRR of brain activation for robotic passive movement. CONCLUSION: The brain activation induced by the robotic passive movement alone has very poor TRR, suggesting that further enhancement is required to strengthen the TRR by complementing active user engagements.

A pilot study on the optimal speeds for passive wrist movements by a rehabilitation robot of stroke patients: A functional NIRS study.

The optimal conditions inducing proper brain activation during performance of rehabilitation robots should be examined to enhance the efficiency of robot rehabilitation based on the concept of brain plasticity. In this study, we attempted to investigate differences in cortical activation according to the speeds of passive wrist movements performed by a rehabilitation robot for stroke patients. 9 stroke patients with right hemiparesis participated in this study. Passive movements of the affected wrist were performed by the rehabilitation robot at three different speeds: 0.25 Hz; slow, 0.5Hz; moderate and 0.75 Hz; fast. We used functional near-infrared spectroscopy to measure the brain activity during the passive movements performed by a robot. Group-average activation map and the relative changes in oxy-hemoglobin (ΔOxyHb) in two regions of interest: the primary sensory-motor cortex (SM1); premotor area (PMA) and region of all channels were measured. In the result of group-averaged activation map, the contralateral SM1, PMA and somatosensory association cortex (SAC) showed the greatest significant activation according to the movements at 0.75 Hz, while there is no significantly activated area at 0.5 Hz. Regarding ΔOxyHb, no significant diiference was observed among three speeds regardless of region. In conclusion, the contralateral SM1, PMA and SAC showed the greatest activation by a fast speed (0.75 Hz) rather than slow (0.25 Hz) and moderate (0. 5 Hz) speed. Our results suggest an optimal speed for execution of the wrist rehabilitation robot. Therefore, we believe that our findings might point to several promising applications for future research regarding useful and empirically-based robot rehabilitation therapy.

Design of a clinically relevant upper-limb exoskeleton robot for stroke patients with spasticity.

The purpose of this research is to propose a design of a clinically relevant upper-limb (hand, wrist, and elbow) exoskeleton that meets the clinical requirements. At first, the proposed robot was designed to have sufficient torque for passive exercise therapy and spasticity measurement of post-stroke patients with spasticity (grade 3 or lower in Modified Ashworth Scale). Because the therapy of patients with high level spasticity could be laborious for therapists by increased muscle tone, and the patients tend not to get enough rehabilitation treatment. Secondly, this robot was designed to have user friendly features like as modularity, so that users have easy approach to assemble and disassemble for practical use. Thirdly, this robot system was designed to guarantee the safety for robot-aided passive training of patients with spasticity. As a result, we were able to see the usability of the robot system, even though it was a pilot test. This shows the possibility of measuring and classifying the spasticity.

The Optimal Speed for Cortical Activation of Passive Wrist Movements Performed by a Rehabilitation Robot: A Functional NIRS Study.

Objectives: To advance development of rehabilitation robots, the conditions to induce appropriate brain activation during rehabilitation performed by robots should be optimized, based on the concept of brain plasticity. In this study, we examined differences in cortical activation according to the speed of passive wrist movements performed by a rehabilitation robot. Methods: Twenty three normal subjects participated in this study. Passive movements of the right wrist were performed by the wrist rehabilitation robot at three different speeds: 0.25 Hz; slow, 0.5 Hz; moderate and 0.75 Hz; fast. We used functional near-infrared spectroscopy to measure the brain activity accompanying the passive movements performed by a robot. The relative changes in oxy-hemoglobin (HbO) were measured in two regions of interest (ROI): the primary sensory-motor cortex (SM1) and premotor area (PMA). Results: In the left SM1 the HbO value was significantly higher at 0.5 Hz, compared with movements performed at 0.25 Hz and 0.75 Hz (p < 0.05), while no significant differences were observed in the left PMA (p > 0.05). In the group analysis, the left SM1 was activated during passive movements at three speeds (uncorrected p < 0.05) and the greatest activation in the SM1 was observed at 0.5 Hz. Conclusions: In conclusion, the contralateral SM1 showed the greatest activation by a moderate speed (0.5 Hz) rather than slow (0.25 Hz) and fast (0.75 Hz) speed. Our results suggest an ideal speed for execution of the wrist rehabilitation robot. Therefore, our results might provide useful data for more effective and empirically-based robot rehabilitation therapy.

In Vivo Estimation of Human Forearm and Wrist Dynamic Properties.

It is important to estimate the 3 degree-of-freedom (DOF) impedance of human forearm and wrist (i.e., forearm prono-supination, and wrist flexion-extension and radial-ulnar deviation) in motor control and in the diagnosis of altered mechanical resistance following stroke. There is, however, a lack of methods to characterize 3 DOF impedance. Thus, we developed a reliable and accurate impedance estimation method, the distal internal model based impedance control (dIMBIC)-based method, to characterize the 3 DOF impedance, including cross-coupled terms between DOFs, for the first time. Its accuracy and reliability were experimentally validated using a robot with substantial nonlinear joint friction. The 3 DOF human forearm and wrist impedance of eight healthy subjects was reliably characterized, and its linear behavior was verified. Thus, the dIMBIC-based method can provide us with 3 DOF forearm and wrist impedance regardless of nonlinear robot joint friction. It is expected that, with the proposed method, the 3 DOF impedance estimation can promote motor control studies and complement the diagnosis of altered wrist and forearm resistance post-stroke by providing objective impedance estimates, including cross-coupled terms.

Cortical activation change induced by neuromuscular electrical stimulation during hand movements: a functional NIRS study.

OBJECTIVES: Neuromuscular electrical stimulation (NMES) has been used in the field of rehabilitation for a long time. Previous studies on NMES have focused on the peripheral effect, in contrast, relatively little is known about the effect on the cerebral cortex. In the current study, we attempted to investigate the change of cortical activation pattern induced by NMES during execution of hand movements in normal subjects, using functional near infrared spectroscopy (fNIRS). METHODS: Twelve healthy normal subjects were randomly assigned to the NMES group (six subjects) and the sham group (six subjects). We measured oxy-hemoglobin (HbO) in six regions of interest (ROI) during pre-NMES and post-NMES motor phase; the left dorsolateral and ventrolateral prefrontal cortex, premotor cortex, primary sensory-motor cortex (SM1), hand somatotopic area of SM1, and posterior parietal cortex. Between the pre-NMES and the post-NMES motor phases, real or sham NMES was applied on finger and wrist extensors of all subjects during a period of 5 minutes. RESULTS: In all groups, during the pre-NMES motor phase, the HbO value in the hand somatotopic area of the left SM1 was higher than those of other ROIs. In the NMES group, during the post-NMES motor phase, HbO value variation in the hand somatotopic area of the left SM1 showed a significant decrease, compared with that of sham group (p < 0.05). However, in the sham group, similar aspect of results in HbO values of all ROIs was observed between pre-NMES and post-NMES motor phases (p > 0.05). CONCLUSIONS: Results of this study showed that NMES induced a decrease of cortical activation during execution of hand movements. This finding appears to indicate that application of NMES can increase the efficiency of the cerebral cortex during execution of motor tasks.

The cortical activation pattern by a rehabilitation robotic hand: a functional NIRS study.

INTRODUCTION: Clarification of the relationship between external stimuli and brain response has been an important topic in neuroscience and brain rehabilitation. In the current study, using functional near infrared spectroscopy (fNIRS), we attempted to investigate cortical activation patterns generated during execution of a rehabilitation robotic hand. METHODS: Ten normal subjects were recruited for this study. Passive movements of the right fingers were performed using a rehabilitation robotic hand at a frequency of 0.5 Hz. We measured values of oxy-hemoglobin (HbO), deoxy-hemoglobin (HbR) and total-hemoglobin (HbT) in five regions of interest: the primary sensory-motor cortex (SM1), hand somatotopy of the contralateral SM1, supplementary motor area (SMA), premotor cortex (PMC), and prefrontal cortex (PFC). RESULTS: HbO and HbT values indicated significant activation in the left SM1, left SMA, left PMC, and left PFC during execution of the rehabilitation robotic hand (uncorrected, p < 0.01). By contrast, HbR value indicated significant activation only in the hand somatotopic area of the left SM1 (uncorrected, p < 0.01). CONCLUSIONS: Our results appear to indicate that execution of the rehabilitation robotic hand could induce cortical activation.

Anatomical location of the frontopontine fibers in the internal capsule in the human brain: a diffusion tensor tractography study.

The frontopontine fibers (FPFs) originate from the frontal lobe and end in the pontine nuclei. Many neuroanatomy textbooks have described the FPFs as descending through the anterior limb of the internal capsule. However, several studies have reported controversial results. In this study, using diffusion tensor tractography, we investigated the anatomical location of the FPFs in the internal capsule in the human brain. We recruited 53 healthy volunteers for this study. For reconstruction of the FPFs, the seed region of interest was given in the medial cerebral peduncle of the reconstructed corticospinal tract. The target regions of interest were placed in the three cerebral cortices, respectively: Brodmann's area (BA) BA 6, BA 8, and BA 9. The anatomical locations of the FPFs were evaluated using the highest probabilistic location in the internal capsule. We measured the relative distance of the FPFs from the middle point at the genu of the internal capsule to the most posterior point of the lenticular nucleus. The relative mean distances of the highest probabilistic location for the FPFs from BA 9, 8, and 6 were 18.18, 32.08, and 43.83% from the middle point of the genu of the internal capsule, respectively. By contrast, the highest probabilistic location for the corticospinal tract was 74.18%. According to our findings, the FPFs were located at the anterior half of the posterior limb in the internal capsule, in the following order, from the anterior direction: the FPFs from BA 9, BA 8, and BA 6.

Two-Channel Transparency-Optimized Control Architectures in Bilateral Teleoperation With Time Delay.

This paper introduces transparency-optimized control architectures (TOCAs) using two communication channels. Two classes of two-channel TOCAs are found, thereby showing that two channels are sufficient to achieve transparency. These TOCAs achieve a greater level of transparency but poorer stability than three-channel TOCAs and four-channel TOCAs. Stability of the two-channel TOCAs has been enhanced while minimizing transparency degradation by adding a filter; and a combined use of the two classes of two-channel TOCAs is proposed for both free space and constrained motion, which involve switching between two TOCAs for transition between free space and constrained motions. The stability condition of the switched teleoperation system is derived for practical applications. Through the one degree-of-freedom (DOF) experiment, the proposed two-channel TOCAs were shown to operate stably, while achieving better transparency under time delay than the other TOCAs.

The cortical activation differences between proximal and distal joint movements of the upper extremities: a functional NIRS study.

OBJECTIVES: Motor control of musculature of proximal and distal joints has been suggested to differ. However, no study comparing patterns of activation between movements of proximal and distal joints has been conducted. In this study, using functional near infrared spectroscopy (fNIRS), we attempted to compare patterns of cortical activation generated during movements of shoulder and hand. METHODS: Nine normal subjects were recruited. fNIRS was performed using a fNIRS system with 49 channels. Flexion-extension movements of the right shoulder or hand were performed. We measured values of oxy-hemoglobin (HbO) and total-hemoglobin (HbT) in three regions of interest: the primary sensory-motor cortex (SM1), the premotor cortex (PMC), and the prefrontal cortex (PFC). The relative activation ratio of HbO and HbT was estimated for comparison of the relative activity of the left PMC or PFC to the left SM1. RESULTS: Mean values of HbO and HbT of the left SM1, PMC, and PFC were higher during movements of the right shoulder, compared with movements of the left hand. Relative activation ratios for movements of the right shoulder (HbO: PMC-104.5%, PFC-110.9%; HbT: PMC-136.3%, PFC-200.1%) were greater than 100%, and, by contrast, less than 100% (HbO: PMC-57.5%, PFC-84.8%; HbT: PMC-88.9%, PFC-70.3%) for movements of the right hand. CONCLUSIONS: Our results appear to indicate that movements of the shoulder require greater neural recruitment than movements of the hand. In addition, the PMC and PFC appeared to have greater involvement than the SM1 in movements of the shoulder; by contrast, the SM1 appears to have greater involvement than the PMC and PFC in movements of the hand.

The ascending reticular activating system from pontine reticular formation to the thalamus in the human brain.

INTRODUCTION: Action of the ascending reticular activating system (ARAS) on the cerebral cortex is responsible for achievement of consciousness. In this study, we attempted to reconstruct the lower single component of the ARAS from the reticular formation (RF) to the thalamus in the normal human brain using diffusion tensor imaging (DTI). METHODS: Twenty six normal healthy subjects were recruited for this study. A 1.5-T scanner was used for scanning of diffusion tensor images, and the lower single component of the ARAS was reconstructed using FMRIB software. We utilized two ROIs for reconstruction of the lower single component of the ARAS: the seed ROI - the RF of the pons at the level of the trigeminal nerve entry zone, the target ROI - the intralaminar nuclei of the thalamus at the level of the commissural plane. RESULTS: The reconstructed ARAS originated from the pontine RF, ascended through the mesencephalic tegmentum just posterior to the red nucleus, and then terminated on the intralaminar nuclei of the thalamus. No significant differences in fractional anisotropy, mean diffusivity, and tract number were observed between hemispheres (p > 0.05). CONCLUSION: We reconstructed the lower single component of the ARAS from the RF to the thalamus in the human brain using DTI. The results of this study might be of value for the diagnosis and prognosis of patients with impaired consciousness.

Differences of the frontal activation patterns by finger and toe movements: a functional MRI study.

It is well-known that physical exercise can affect cognition and the frontal lobe is an important structure involved in motor function and cognition. Furthermore, many functional neuroimaging studies have demonstrated that cortical activation patterns of hand and leg movements differ. However, no study has been undertaken to identify differences between the frontal activation patterns generated by hand and leg movements. In the present study, the frontal activation patterns associated with finger and toe movements, as visualized by functional MRI, were investigated and compared. Twelve healthy volunteers were recruited. Functional MRI was performed using a 1.5 T Philips Gyroscan Intera. Flexion-extension movements of fingers or toes were performed in one extremity. Regions of interest (ROIs) were set at the primary sensory-motor cortex (SM1: Brodmann area [BA] 1, 2, 3, 4), the premotor area (PMA: BA 6), and the prefrontal cortex (PFC: BA 8, 9, 10, 11, 46). In SM1, finger movements (10,809) induced more activation than toe movements (5349). On the other hand, in the PMA and PFC, toe movements (PMA: 4201, PFC: 921) induced more activation than finger movements (PMA: 2887, PFC: 912) respectively. In the analysis of relative voxel counts in the PMA and PFC versus the SM1, toe movements generated more activation in the PMA and PFC than finger movements. The PMA and PFC were more activated by toe than finger movements, although the SM1 was more activated by finger movements.

Neural tracts injuries in patients with hypoxic ischemic brain injury: diffusion tensor imaging study.

Many studies have reported on vulnerable areas of the brain in hypoxic ischemic brain injury (HI-BI). However, little is known about the involvement of neural tracts following HI-BI. We investigated neural tract injuries in adult patients with HI-BI, using diffusion tensor tractography (DTT). Twelve consecutive patients with HI-BI and 12 control subjects were recruited for this study. We classified the patients into two subgroups according to the preservation of alertness: subgroup A-5 patients who had intact alertness and subgroup B-7 patients who had impaired alertness. DTI-Studio software was used for evaluation of seven neural tracts: corticospinal, cingulum, fornix, superior longitudinal fasciculus, inferior longitudinal fasciculus, inferior fronto-occipital fasciculus, and optic radiation. We measured the DTT parameters (fractional anisotropy, apparent diffusion coefficient and voxel number) of each neural tract. In the individual analysis, all 12 patients showed injuries in all 24 neural tracts in terms of both DTT parameters and integrity, except for the corticospinal tract (75.0% injury). In the group analysis, the patient group showed neural injuries in all 24 neural tracts. In comparison of subgroups A and B, subgroup B showed more severe injuries: subgroup B showed a higher rate of disruption (39.8%) than subgroup A (12.9%) on individual DTTs and subgroup B had more severe injuries in both the cingulum and superior longitudinal fasciculus. In conclusion, we found that extensive injuries in the neural tracts were accompanied by HI-BI. Patients with impaired alertness appeared to show more severe injuries of neural tracts.

Anatomical location of the corticospinal tract according to somatotopies in the centrum semiovale.

Little is known about the somatotopic location of the corticospinal tract (CST) in the centrum semiovale (CS). We investigated the somatotopic location of the CST in the CS in the human brain using diffusion tensor tractography (DTT). Fifty-two healthy volunteers were recruited for this study. Diffusion tensor images (DTIs) were obtained at 1.5T, and CSTs for the hand and leg were obtained using FMRIB software. Normalized DTT images were reconstructed using the Montreal Neurological Institute echo-planar imaging template supplied with the SPM. Individual DTI data were calculated as number of pixels in the CS. In the mediolateral direction, average distances of the highest probabilistic locations for hand and leg somatotopies were 25.57 mm and 21.72 mm from the midline between the right and left hemispheres, respectively. For the anteroposterior direction, the average distance of the highest probabilistic locations for hand and leg somatotopies were 0.4 mm and 5.2 mm behind the horizontal line between the medial end of the central sulcus and midline, respectively. In conclusion, hand somatotopy of the CST was found to be located at about 26 mm lateral to the midline almost along the horizon line between the medial end of central sulcus and midline, and leg somatotopy of the CST was found to be located medioposteriorly to the hand somatotopy of the CST.

The neural connectivity of the inferior olivary nucleus in the human brain: a diffusion tensor tractography study.

OBJECTIVES: Many animal studies have reported on the neural connectivity of the inferior olivary nucleus (ION). However, the neural connectivity of the ION has not been clearly elucidated in the human brain. In this study, the neural connectivity of the ION in the human brain was investigated by diffusion tensor imaging (DTI). METHODS: Forty healthy subjects were recruited. DTIs were acquired using a sensitivity-encoding head coil at 1.5T. Connectivity was defined as the incidence of connection between the ION and regions of interest (ROIs) in the brain. RESULTS: In these subjects, the ION showed higher connectivity to the reticular formation (100%), the posterior limb of internal capsule (100%), the red nucleus (93.75%), the cerebral peduncle of midbrain (91.25%), the primary motor cortex (86.25%), the primary somatosensory cortex (85%), the periaqueductal gray mater (81.25%), the globus pallidus (81.25%), the anterior limb of internal capsule (62.5%), the pontine basis (62.5%), and the posterior parietal cortex (60%). CONCLUSIONS: The ION shows high connectivity with motor function-related areas, such as, the posterior limb of internal capsule, the red nucleus, the cerebral peduncle of midbrain, the primary motor cortex, and the pontine basis. These results indicate that the ION is closely related to motor function in the human brain.

Telepresence index for bilateral teleoperations.

This paper proposes a performance index called telepresence index for bilateral teleoperation, which can be used both for the performance evaluation of bilateral control architectures and for design purposes. This index is intended to represent a comprehensive performance objective consisting of transparency and kinematic correspondence, which are two major performance objectives of bilateral teleoperation. In order to quantify the performance objective, telepresence index has employed the error vector magnitude, which enables a seamless combination of magnitude and phase errors and the accommodation of time delay. In comparison with existing performance indices, it was observed that telepresence index possesses the comprehensiveness of performance objectives, magnitude/phase integrity, and the capacity to include time delay, which the others lack in one way or another. The index was applied to evaluate the performances of two widely known control architectures: PD-type bilateral control and Ueda's ideal control. In all cases, telepresence index has been compared favorably with the other indices in terms of clarity, convenience, and accuracy, thereby demonstrating its superiority.

Development of a one-body optical torque sensor for rehabilitation robotic systems.

This paper proposes a one-body optical torque sensor. The proposed sensor has advantages of anti-slip and low cost due to the simple one-body structure. Simulations for stress and strains analysis are accurately performed. To demonstrate the performance of proposed design, experiments were also carried out to compare it with a commercial force/torque sensor (Mini45, ATI).

Stochastic estimation of human shoulder impedance with robots: an experimental design.

Previous studies assumed the shoulder as a hinge joint during human arm impedance measurement. This is obviously a vast simplification since the shoulder is a complex of several joints with multiple degrees of freedom. In the present work, a practical methodology for more general and realistic estimation of human shoulder impedance is proposed and validated with a spring array. It includes a gravity compensation scheme, which is developed and used for the experiments with a spatial three degrees of freedom PUMA-type robot. The experimental results were accurate and reliable, and thus it has shown a strong potential of the proposed methodology in the estimation of human shoulder impedance.

Shoulder mechanism design of an exoskeleton robot for stroke patient rehabilitation.

Shoulder girdle movement is critical for stabilizing and orientating the arm during daily activities. During robotic arm rehabilitation with stroke patients, the robot must assist movements of the shoulder girdle. Shoulder girdle movement is characterized by a highly nonlinear function of the humeral orientation, which is different for each person. Hence it is improper to use pre-calculated shoulder girdle movement. If an exoskeleton robot cannot mimic the patient's shoulder girdle movement well, the robot axes will not coincide with the patient's, which brings reduced range of motion (ROM) and discomfort to the patients. A number of exoskeleton robots have been developed to assist shoulder girdle movement. The shoulder mechanism of these robots, along with the advantages and disadvantages, are introduced. In this paper, a novel shoulder mechanism design of exoskeleton robot is proposed, which can fully mimic the patient's shoulder girdle movement in real time.