A multi-pad electrode based functional electrical stimulation system for restoration of grasp
© Malešević et al.; licensee BioMed Central Ltd. 2012
Received: 25 January 2012
Accepted: 19 September 2012
Published: 25 September 2012
Functional electrical stimulation (FES) applied via transcutaneous electrodes is a common rehabilitation technique for assisting grasp in patients with central nervous system lesions. To improve the stimulation effectiveness of conventional FES, we introduce multi-pad electrodes and a new stimulation paradigm.
The new FES system comprises an electrode composed of small pads that can be activated individually. This electrode allows the targeting of motoneurons that activate synergistic muscles and produce a functional movement. The new stimulation paradigm allows asynchronous activation of motoneurons and provides controlled spatial distribution of the electrical charge that is delivered to the motoneurons. We developed an automated technique for the determination of the preferred electrode based on a cost function that considers the required movement of the fingers and the stabilization of the wrist joint. The data used within the cost function come from a sensorized garment that is easy to implement and does not require calibration. The design of the system also includes the possibility for fine-tuning and adaptation with a manually controllable interface.
The device was tested on three stroke patients. The results show that the multi-pad electrodes provide the desired level of selectivity and can be used for generating a functional grasp. The results also show that the procedure, when performed on a specific user, results in the preferred electrode configuration characteristics for that patient. The findings from this study are of importance for the application of transcutaneous stimulation in the clinical and home environments.
Functional Electrical Stimulation (FES) provides control signals to peripheral motor systems that are compromised after a central nervous system lesion (stroke, spinal cord injury, etc.). The use of FES combined with a splint for grasp control was introduced by Long and Masciarelli to assist patients with tetraplegia. A group from the University of Ljubljana, Slovenia[2, 3] suggested the use of two-channel electrical stimulation to control hand opening and closing in patients with tetraplegia. This research led to the first therapeutic application of FES, which resulted in significant carryover effects. The results of these studies and many others that followed led to the development of commercial products for therapy such as the Ness H200, Bionic Glove and other similar systems[5, 6].
More recently, the carryover effects of electrical stimulation have been recognized by clinicians, and several therapeutic methods have evolved. Intensive exercise augmented by FES, termed functional electrical therapy (FET), provided evidence that there are rehabilitation benefits from effective electrical augmentation of movement during the period when this function is not achievable[8, 9]. A therapeutic FES system comprised four pairs of relatively large surface electrodes as an interface between the electronic stimulator and tissues.
The common problems in this scenario were 1) discomfort, 2) insufficient selectivity, and 3) fast fatigue. All three of these problems could be partly eliminated by the use of multi-pad electrodes instead of large single-pad electrodes. The discomfort comes from the activation of pain receptors as well as from the strong synchronous activation of all motor units within a single muscle. When large electrodes are applied, many motoneurons are activated, which results in the coactivation of various non-synergistic muscles. In short, if the electrodes are not positioned appropriately, then the desired movement will not be generated and function will not be achieved.
Finally, the firing rate of a motor unit under natural control is typically low (≈5 Hz), and a fused contraction results from the asynchronous contraction of several motor units. In contrast, FES activates many motor units simultaneously, and non-physiological frequencies (>20 Hz) are required for fused contraction, leading to a much faster occurrence of muscle fatigue.
The suggestion to use multi-pad surface electrodes was introduced in parallel with the development of multichannel electronic stimulators for arm and hand control in patients with tetraplegia[11, 12]. Detailed maps for the stimulation of forearm muscles were presented by Nathan. This study showed that there is major inter-subject variability.
Several research groups were involved in the design of multi-pad electrode stimulation systems along with control algorithms for various applications. The first practical system was reported by Fujii. Elsaify followed this principle and suggested the use of muscle twitch response for shaping an optimal electrode. A group from Eidgenössische Technische Hochschule (ETH) Zurich designed textile multi-pad electrodes with the appropriate control algorithm[16, 17]. O’Dwyer incorporated wearable feedback sensors in his system and defined control based on sensory inputs. The development of the "Actitrode" system was part of the effort to facilitate the application of multi-pad electrodes. Detailed research led to a methodology for the minimization of non-desired movement (wrist interference when grasping) in patients with tetraplegia. The results when applying the multi pad electrodes show that the required electrical charge can be delivered at current densities that do not cause more discomfort to the user than the application of large surface electrodes.
We compared the force generated via multi-pad electrodes vs. large electrodes (conventional method). The results showed that the fatigue induced by stimulation was postponed when using four cathodes instead of a single cathode over the Quadriceps m. of patients with chronic tetraplegia[21, 22]. A charge delivered through four electrodes at low frequency asynchronously activated different motoneurons and resulted in a fused contraction. This force was similar to the force resulting from high-frequency activation via a conventional large electrode. This finding was also observed for Triceps Surae m. by Nguyen et al..
The application of the small pads will result in a localized electrical field that decreases spreading of currents to motoneurons of adjacent muscles or even to the same muscle. In an electrical sense, this feature enables controlled charge distribution to selected motoneurons.
If the activation of individual pads is asynchronous with short delays, then there will be a superposition of muscle forces but not action potentials. The delay between the pulses must be short enough to allow the summation of contractions. This allows the use of a lower stimulation frequency on each site and the activation of the muscle in a more physiological order, potentially postponing fatigue.
The stimulation of different motor units of a single muscle or synergistic muscles asynchronously via several pads at a high frequency f. This stimulation is achieved with the bursts of n pulses sent to n pads during a single refractory period Tp < 5 ms. The observed effect is the same as that of synchronous stimulation of several pads because very short Tp and IPI impede the occurrence of consecutive action potentials on the same motor unit. This protocol requires stimulation frequencies with ω between 20–50 Hz (τ = 50–20 ms).
The stimulation of different motor units of a single muscle or synergistic muscles asynchronously via several pads at lower frequency (f) with an IPI between 5 and 10 ms. In this regime, when various pads activate the same motor unit, the net effect is an “an-let” that produces a shorter muscle force-time interval and higher force compared with a single pulse stimulation of the same motor unit [28–31]. This protocol requires stimulation frequencies with ω between 20–50 Hz (τ = 50–20 ms).
Stimulation of different motor units of a single muscle or synergistic muscles asynchronously via several pads at low frequency ω < 20 Hz and T + IPI = 1/nω. The stimulation of each individual pad therefore contributes to the overall generated force, which can be maintained at the required level of strong fused contraction, while the frequency of action potential generation on each motor unit can be lowered to reduce muscle fatigue [21–23].
Stimulator technical specifications
115 × 72 × 25 mm
7.4 V 750 mAh, integrated rechargeable Li-Ion
> 3 hours continuous use
Battery charger input
12 VDC, >500 mA, wall adapter
Battery charging time
< 3 hours
Number of channels
Number of multi-pad electrodes
Number of fields per electrodes
Biphasic charge compensated, constant current pulses
0 – 50 mA (default 10 mA)
1 - 50 pulses per second (default 50 pulses per second)
50 – 1000 μs (default 250 μs)
Maximal stimulation voltage
Permitted output impedance
0 – 1.5 KΩ
Type of analog inputs
Range of analog inputs
0 – 3.3 V
Number of analog inputs
Sensor power supply output
3.3 V, 100 mA
Bluetooth 2.0 + EDR
Microcontroller. We use an 8-bit ATmega1284P (Atmel, California) operating at 11 MHz with 128 k bytes of In-System Self-Programmable Flash memory (Figure3). This microcontroller has two wire interface (I2C) modules capable of a data transfer speed of up to 400 kHz, two 16-bit timers and two 8-bit timers, an SPI interface and two universal synchronous/asynchronous receiver/transmitters (USARTs). The microcontroller controls all peripherals, including the multi-pad electrodes, and it acts as a slave device to a host controller (PC, Smartphone, etc.).
Power supply. The INTFES device is battery powered, which protects it from hazardous current loops with other devices connected to the grid. The power supply is based on an incorporated Li-ion rechargeable 7.4 V, 750 mAh battery and includes a monitoring chip for battery capacity level. The device has two isolated parts: 1) control logic with high voltage source and 2) slave logic for controlling of multi-pad electrodes and analog inputs. This is a safety feature that eliminates the possibility of current loops between the high voltage source and an electrode printed circuit board (PCB) or malfunctioning sensor.
Charging of INTEFES occurs by an external 12 V voltage charger via a battery charging chip that provides optimal charging stages for prolonged battery life. To protect the user, the device cannot be turned ON during charging.
The output stage (current control). The output stage of the stimulator is limited to 94 V. This voltage is generated by a step-up DC/DC converter MAX773 (Maxim Integrated Products, Sunnyvale, CA, U.S.A.). The output current is generated in an H-bridge configuration that generates symmetrical biphasic pulses; its amplitude is defined using an 8-bit D/A convertor, and its duration is defined by two 16-bit timer modules in a microcontroller responsible for the direct and compensatory pulse.
Analog inputs and multi-pad electrode control. For safety reasons, analog inputs and multi-pad electrode control logic are isolated from high voltage circuitry with a digital circuit ADUM1201 (Analog Devices, Norwood, MA, U.S.A.) and a DC/DC power supply ISF0505A. We used MCP3208 (Microchip, Chandler, Arizona, U.S.A.) with 6 12-bit channels.
Communications. A Bluetooth module WT12-A-AI (Bluegiga Technologies Inc., Espoo, Finland) maintains communication between the INTFES device and the host controller. Bluetooth communication provides isolation from the grid and remote control of the device using any device with a standard Bluetooth module. Using this communication, the host can set all of the stimulation parameters, initiate a stimulation protocol, activate the pads and read analog signals and locally calculated parameters. The guaranteed delay in any read/write is 15 ms, which, for a relatively slow physiological system, permits real-time control in the INTFES stimulator or within the host controller. Thus, an algorithm can be executed by the microcontroller, but if its complexity becomes too high for the 8-bit processor, the host can co-opt some of the decision process and return command to the stimulator. In the case of recovery assessment or centralized database logging, it is possible to establish a link between the host controller application and a remote service. The Bluetooth link also allows the change of stimulator firmware.
To minimize the number of leads connecting the stimulator unit to the electrode, multiplexing of the stimulation pulses is performed in the electrode connector (Figure4). The INTFES connector also contains signal conditioning electronics for the pressure sensor layer that detect pressed pads based on a defined threshold. There are seven total leads (2 for I2C protocol, 2 for power supply, sensory output, stimulation and enable) for any number of pads within the multi-pad electrode.
Control software. The control software is an application running on a host controller. The communication protocol is supported by several programming languages (C, C#, JAVA, LabView, MATLAB), enabling custom-made control software to be developed for a preferred platform. We developed user applications in a .NET environment for Windows and JAVA for Android (versions 2.3 and newer) for the setting of the stimulation parameters, and our optimization algorithm was developed in LabView (National Instruments, U.S.A.). Stimulation can be initiated or terminated promptly using control software, and it has the highest priority in the INTFES system.
From the example in Figure6, one can derive a rule-based logic that can be implemented in the grasp control algorithm. Pad 5 does not produce any response, and it should not be selected as a part of the active surface. Pads 2, 3 and 4 (Intensity 1) all move the same finger, A; therefore, these pads can be activated asynchronously with lower stimulation frequency on each pad, which can possibly postpone fatigue. Pad 1 activates only the wrist. Lower current intensities on this pad can be used to stabilize the wrist during finger extension. Pad 3 activates more than one finger, and Pad 4 produces additional activation of the wrist with higher current intensities. These pads can be included when the activation of multiple joints is required. Based on the proposed logic, after testing muscle responses for each of the electrode pads, an automated algorithm must be used to determine the preferred electrode number and position of active pads (stimulation active surface) and current intensities that produce the desired movement. The term “preferred” is dependent on the application, and in the case of grasp generation, it is used to describe finger flexion/extension and a stable wrist. “Stable wrist” is a term describing the minimal excursion of the wrist during finger opening and closing. The selection of a preferred active surface can be static, meaning one set of pads is used for each type of movement (i.e., different types of grasps), or dynamic, meaning several sets of pads are used depending on the position of the forearm. Dynamic selection (dynamic compensation) is necessary in tasks that involve movement of the skin with respect to the underlying tissue, i.e., during forearm supination/pronation. Once defined, all of the stimulation patterns can be activated by a real-time controller, reducing the number of controllable parameters in the open loop system. In closed loop systems, the derived look-up table represents the initial state of the controllable parameters, which can be further adjusted to achieve a defined goal by increasing/decreasing stimulation intensities or activating additional pads if necessary.
Setup of a preferred electrode for patients
qf represents the flexion quality factor of the i-th ST, Wf represents the weight factor for each finger, Af represents the flexion amplitude of a finger, Ww represents the weight factor for wrist movement, and Aw represents the bending amplitude of the wrist. The presented cost function takes into account neighboring ST-s if they originate from the same active pad, favoring pads to which an increase of current intensity gradually increases all flexions. If it is the first or last current amplitude of an active pad, the SC indexes become (i + 2) instead of (i-1) for the first current amplitude and (i-2) instead of (i + 1) for the last amplitude. The gradient is based on the quality factor; all pads and their current amplitudes are rated, and the pads with the largest quality factor are selected as preferred. Application of the quality factor produces a list of pads and current intensities for functional movement (as in Figure6). A similar strategy was used for grasp optimization by Popović and Popović.
Testing of intfes in stroke patients
The goal of this test was to estimate this system’s effectiveness and simplicity in a clinical environment and to test the effect of the preferred electrode selection procedure for hand opening and closing as part of the grasp/release procedure.
Basic data for hemiplegic subjects
Period between stroke and testing of the INTFES (months)
Modified Ashworth scale
The average time required to position the whole system, including the data glove, rings and stimulation electrodes, was less than 4 minutes. The results from one patient are presented in the following figures.
The range of motion during evaluation testing is dependent on the patient’s neutral finger and wrist position, which varied from patient to patient. As shown in Figure12, both the smallest flexion and extension were achieved for the index finger because the electrode size was not large enough to cover all of the hand muscle motoneurons.
The results presented in Figure12 are for one of the patients (A). For the other two patients, we obtained similar twitch measurement results in terms of selectivity. The preferred electrode was found to be different among patients due to intersubject physiological variability (Figure12). This finding indicates that the optimization algorithm must be applied at least one time for each patient. Repeated tests in all three patients showed that the preferred electrodes were the same if the electrode was positioned at the same position on the forearm.
This pilot study in three patients clearly shows that INTFES allows wrist stabilization and selective activation of muscles that are required for hand opening and closing. Importantly, the multi-pad electrode application enables asynchronous operation.
The feedback coming from the sensorized garment effectively provided necessary information for the algorithm used in the tests, and the system was shown to be robust. The automated algorithm for selection of preferred grasp has been tested successfully; it used the twitch response (assessed by the flex sensors) and, in short time, automatically generated the preferred electrode configuration.
The size and shape of the electrode, in line with previous results, were found to vary substantially from patient to patient but remained the same from day to day in each patient. These results suggest that once the electrode size and shape are determined with respect to the anatomical features of the forearm, the INTFES system can be set up on the patient in less than 2 minutes by a naïve user. The results demonstrate that INTFES eliminates the tedious search process and electrode deterioration due to multiple attachments and detachments during the placement of a single pad electrode.
The preferred electrodes do not have regular shapes; they typically have a branching shape.
The INTFES system is operational and is currently being used in clinical studies at the Dr. Miroslav Zotović Institute for Rehabilitation, Belgrade, Serbia. The device was well accepted by the therapists, who learned to operate the system in sessions lasting less than 30 minutes. The same system can be applied for other therapeutic modalities (e.g., foot drop correction, lower back pain, and fitness).
This work was partially supported by the Ministry for Science and Technological Development of Serbia (projects: OI175016 and TR35040). We would like to acknowledge valuable suggestions by Prof Laszlo Schwirtlich, M.D. and Aleksandra Dragin, M.D. and assistance in clinical tests by Anđelka Pjanović, P.T. and Jelena Milovanović, P.T., Institute for Rehabilitation “Dr Miroslav Zotović”, Belgrade, Serbia. In part financed by the CONSOLIDER INGENIO 2010 program (grant CSD2009-00067) by the Spanish Science Ministry.
- Long C: An electrophysiologic splint for the hand. Arch Phys Med Rehabil 1963, 44: 499-503.PubMedGoogle Scholar
- Rebersek S, Vodovnik L: Proportionally controlled functional electrical stimulation of hand. Arch Phys Med Rehabil 1973, 54: 378.PubMedGoogle Scholar
- Rudel D, Bajd T, Rebersek S, Vodovnik L: FES assisted manipulation in quadriplegic patients. Popovi c, D (Ed ), Advances in External Control of Human Extremities 1984, 8: 273-282.Google Scholar
- Merletti R, Acimovic R, Grobelnik S, Cvilak G: Electrophysiological orthosis for the upper extremity in hemiplegia: feasibility study. Arch Phys Med Rehabil 1975, 56: 507.PubMedGoogle Scholar
- Broderick BJ, Breen PP, Olaighin G: Electronic stimulators for surface neural prosthesis. Journal of Automatic Control 2008, 18: 25-33. 10.2298/JAC0802025BView ArticleGoogle Scholar
- Hendricks HT, IJzerman MJ, De Kroon JR, Zilvold G: Functional electrical stimulation by means of the ‘Ness Handmaster Orthosis’ in chronic stroke patients: an exploratory study. Clin Rehabil 2001, 15: 217-220. 10.1191/026921501672937235View ArticlePubMedGoogle Scholar
- Popović DB, Sinkjær T, Popović MB: Electrical stimulation as a means for achieving recovery of function in stroke patients. NeuroRehabil 2009, 25: 45-58.Google Scholar
- Yan T, Hui-Chan CWY, Li LSW: Functional electrical stimulation improves motor recovery of the lower extremity and walking ability of subjects with first acute stroke. Stroke 2005, 36: 80-85. 10.1161/01.STR.0000149623.24906.63View ArticlePubMedGoogle Scholar
- Popović DB, Sinkjær T, Popović MB: Electrical stimulation as a means for achieving recovery of function in stroke patients. NeuroRehabil 2009, 25: 45-58.Google Scholar
- Popović DB, Popović MB: Advances in the use of electrical stimulation for the recovery of motor function. Brain Machine Interfaces: Implications for Science, Clinical Practice and Society 2011, 194: 215-225.Google Scholar
- Nathan RH, Ohry A: Upper limb functions regained in quadriplegia: a hybrid computerized neuromuscular stimulation system. Arch Phys Med Rehabil 1990, 71: 415.PubMedGoogle Scholar
- Nathan R: Handmaster NMS. Present technology and the next generation. In Proceedings of the 2nd Int. Edited by: Popovic D. Burnaby: Symposium IFESS; 1997:139-140.Google Scholar
- Nathan RH: An FNS-based system for generating upper limb function in the C4 quadriplegic. Med Biol Eng Comput 1989, 27: 549-556. 10.1007/BF02441635View ArticleGoogle Scholar
- Fujii T, Seki K, Handa Y: Development of a new FES system with trained super-multichannel surface electrodes. Bournemouth, UK: Proc. 9th Annual Conf IFESS; 21-24.Google Scholar
- Elsaify A: A self-optimising portable FES system using an electrode array and movement sensors . Leicester: University of Leicester, Department of Engineering; 2005.Google Scholar
- Keller T, Lawrence M, Kuhn A, Morari M: New multi-channel transcutaneous electrical stimulation technology for rehabilitation . New York, USA: Proc of the 28th IEEE EMBS; 2006:194-197.Google Scholar
- Kuhn A, Keller T, Micera S, Morari M: Array electrode design for transcutaneous electrical stimulation: a simulation study. Med Eng Phys 2009, 31: 945-951. 10.1016/j.medengphy.2009.05.006View ArticlePubMedGoogle Scholar
- O'Dwyer SB, O'Keeffe DT, Coote S, Lyons GM: An electrode configuration technique using an electrode matrix arrangement for FES-based upper arm rehabilitation systems. Med Eng Phys 2006, 28: 166-176. 10.1016/j.medengphy.2005.03.010View ArticlePubMedGoogle Scholar
- Popović Bijelić A, Bijelić G, Jorgovanović N, Bojanić D, Popović MB, Popović DB: Multi Field Surface Electrode for Selective Electrical Stimulation. Artif Organs 2005, 29: 448-452. 10.1111/j.1525-1594.2005.29075.xView ArticlePubMedGoogle Scholar
- Popovic DB, Popovic MB: Automatic determination of the optimal shape of a surface electrode: Selective stimulation. J Neurosci Methods 2009, 178: 174-181. 10.1016/j.jneumeth.2008.12.003View ArticlePubMedGoogle Scholar
- Malešević NM, Popović LZ, Schwirtlich L, Popović DB: Distributed low frequency functional electrical stimulation delays muscle fatigue compared to conventional stimulation. Muscle Nerve 2010, 42: 556-562. 10.1002/mus.21736View ArticlePubMedGoogle Scholar
- Popovic LZ, Malesevic NM: Muscle fatigue of quadriceps in paraplegics: comparison between single vs. multi-pad electrode surface stimulation. Minneapolis, MN: Proc of IEEE EMBC; 2009:6785-6788.Google Scholar
- Nguyen R, Masani K, Micera S, Morari M, Popovic MR: Spatially Distributed Sequential Stimulation Reduces Fatigue in Paralyzed Triceps Surae Muscles: A Case Study. Artif Organs 2011, 35: 1174-1180. 10.1111/j.1525-1594.2010.01195.xView ArticlePubMedGoogle Scholar
- Bijelić G, Popović-Bijelić A, Jorgovanović N, Bojanić D, Popović DB: E Actitrode: the new selective stimulation interface for functional movements in hemiplegics patients. Serbian Journal of Electrical Engineering 2004, 1: 21-28. 10.2298/SJEE0403021BView ArticleGoogle Scholar
- Popovic LZ, Malesevic NM, Popovic MB: Optimization of multi-pad surface electrode: Selective stimulation of wrist. St. Petersburg Russia: Proc of IEEE EuroCON; 2009:142-145.Google Scholar
- Malešević N, Popović L, Bijelić G, Kvaščev G: Muscle twitch responses for shaping the multi-pad electrode for functional electrical stimulation. J Automatic Control 2010, 20: 53-57. 10.2298/JAC1001053MView ArticleGoogle Scholar
- Popović Maneski L, Jorgovanović N, Ilić V, Došen S, Keller T, Popović MB, Popović DB: Electrical stimulation for the suppression of pathological tremor. Med Biol Eng Comput 2011, 49: 1187-1193. 10.1007/s11517-011-0803-6View ArticlePubMedGoogle Scholar
- Routh GR, Durfee WK: Doublet stimulation to reduce fatigue in electrically stimulated muscle during controlled leg lifts. Engineering in Medicine and Biology Society, 2003. Proceedings of the 25th Annual International Conference of the IEEE 2003, 2: 1531-1534.Google Scholar
- Fang ZP, Mortimer JT: A method to effect physiological recruitment order in electrically activated muscle. Biomedical Engineering, IEEE Transactions on 1991, 38: 175-179. 10.1109/10.76384View ArticleGoogle Scholar
- Kebaetse MB, Lee SCK, Binder-Macleod SA: A novel stimulation pattern improves performance during repetitive dynamic contractions. Muscle Nerve 2001, 24: 744-752. 10.1002/mus.1065View ArticlePubMedGoogle Scholar
- Scott WB, Lee SCK, Johnston TE, Binkley J, Binder-Macleod SA: Effect of electrical stimulation pattern on the force responses of paralyzed human quadriceps muscles. Muscle Nerve 2007, 35: 471-478. 10.1002/mus.20717View ArticlePubMedGoogle Scholar
- Axelgaard J, Heard S: Medical electrode. 14-3-2000. US Patent 6,038,464.Google Scholar
- Keller T, Kuhn A: Electrodes for transcutaneous (surface) electrical stimulation. J Automatic Control 2008, 18: 35-45. 10.2298/JAC0802035KView ArticleGoogle Scholar
- Kuhn A, Keller T, Lawrence M, Morari M: The influence of electrode size on selectivity and comfort in transcutaneous electrical stimulation of the forearm. Neural Systems and Rehabilitation Engineering, IEEE Transactions on 2010, 18: 255-262.View ArticleGoogle Scholar
- Popovic D, Maleševic N, Keller T: Apparatus For External Activation Of Paralyzed Body Parts By Stimulation Of Peripheral Nerves. 7-7-2011. WO Patent WO/2011/079,866.Google Scholar
- Wang PT, King CE, Do AH, Nenadic Z: A durable, low-cost electrogoniometer for dynamic measurement of joint trajectories. Med Eng & Phys 2011, 33: 546-552. 10.1016/j.medengphy.2010.12.008View ArticleGoogle Scholar
- Popović Maneski L, Jevtic T, Malešvić N: Assessment Of Hand Function With Flex Sensors. Zlatibor, ME 1.3: Proceedings of ETRAN; 2012.Google Scholar
- Häger-Ross CK, Klein CS, Thomas CK: Twitch and tetanic properties of human thenar motor units paralyzed by chronic spinal cord injury. J Neurophysiol 2006, 96: 165-174. 10.1152/jn.01339.2005View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.