The clinical relevance of the tendon reflex, particularly the Patellar and Achilles tendon, is based on the ability to evaluate the functional disturbance for nominal or augmented reflex arc and the assessment for the motor system. The tendon reflex is useful for the assessment of several neurological/neuromuscular disorders. The experimental significance of the tendon reflex response is its respective sensitivity for inhibiting and facilitating influences; such as during the acute, subacute, and chronic stages from upper motor neuron syndrome. The tendon reflex can be used for assessing the effect of therapy for altering the reflex arc; also the effects of training and aging can be ascertained. Two parameters are representative of the severity of the pathophysiological mechanisms resultant from lesion; these parameters are latency and amplitude.1
Assessing the characteristic of the deep tendon reflex is an essential component to a neurological examination. The intent of developing myotatic reflex scales is to provide medical professionals with clinical data, which can be conveyed among different doctors and evaluate patient status. The National Institute of Neurological Disorders and Stroke developed a standardized reflex scale in an attempt to homogenize deep tendon reflex assessment.
During the assessment of the NINDS Myotatic Reflex Scale, 80 subjects were evaluated. The experiment ascertained that intraobserver reliability was substantial to near perfect agreement for the NINDS Myotatic Reflex Scale, and moderate to substantial agreement interobserver. The study implies that the neurologist’s evaluation using the NINDS Myotatic Reflex Scale were independent of their techniques. As observed in previous reports, the lower limb reflex evaluation demonstrated better reproducibility than upper limb reflex evaluation. Also positioning the upper limb for evoking a reflex is more difficult than for the lower limb. The authors of the study claim the NINDS Myotatic Reflex Scale is sufficiently reliable to be accepted as a universal scale.2
However in a study by Manshot and colleagues contradicts the claim that the NINDS Myotatic Reflex Scale demonstrates even moderate agreement. Also noted is the fact that the NINDS Myotatic Reflex Scale only consists of a five component ordinal scale. The intent of the study by Manschot and colleagues was to determine the observer reliability of two standard scales for assessing tendon reflexes: the Mayo Clinic scale and the NINDS scale. The interobserver agreement was analyzed by κ statistics. However the agreement between doctors was never greater the “fair” for both scales, and the highest κ value was 0.35. The study by suggested a verbal description as opposed to a codified scale to improve communication. During 1989 a survey between neurologists indicated that at the institution of Manschot and colleagues 20 different reflex scales were used.3
The solution is a fully quantified evaluation system of the myotatic stretch reflex. A patellar hammer’s force input will be based on original potential energy. A MEMS accelerometer will quantify the output. The MEMS accelerometer is attached to a set anchor point near the ankle. The reflex response can be temporally averaged by integrating acceleration from initial time to final time, or by simply obtaining the maximum acceleration of the reflex response. The quantified data collected from MEMS accelerometers is transmitted by a portable computer
The global utility of the Mednode device is its characteristics are suitable to be used as a wearable device. The Mednode device is minimally restrictive, since it is wireless also. The device is designed to maximize reproducibility. The two essential components of the device are a swing arm for quantifying input and a Mednode 3D accelerometer for measuring reflex response.
The design of the swing arm is intended to quantify input based on settings of variable potential energy. The end of the swing arm is fastened to a standard neuro-reflex hammer for eliciting tendon reflexes of the patella. The swing arm can be first targeted to strike a given section of the tendon, such as the middle portion of the patella tendon. Subsequently, the swing arm can be pulled back to a given angle, such as 15, 30, or 45 degrees. Since the potential energy of the swing arm consistently retracted to 30 degrees is constant the input of the reflex hammer is constant for the tendon strike. Also since the swing arm can be varied for potential energy of input, studies on reflex modulation of variable input intensity can be conducted.
The 3D accelerometer Mednode provides the means to quantify the intensity of the reflex response. The Mednode should be placed consistently at a given anatomical anchor, such as the lateral component of the ankle joint. Given the wireless nature and minimal mass, the 3D accelerometer Mednode provides assessment of the reflex response with minimal intrusion. Also by measuring the motor response aspect of the reflex the neuromuscular characteristics of the reflex response are quantified. The signal of the 3D accelerometer Mednode is conveyed wirelessly to a local portable personal computer for subsequent post-data processing.4
2. General neuroanatomy of the reflex
The tendon reflex is considered to be the mechanical counterpart of the H-reflex. The tendon reflex involved the tapping of a distal tendon, and the muscles spindle 1a afferents are stimulated. These afferent impulses travel up to the spinal cord. In the spinal cord the 1a afferents synapse with alpha motor neurons, which are conveyed to their respective muscle inducing a short contraction. Originally, the reflex arc was proposed to be monosynaptic, however oligosynaptic contributions to the reflex arc have been found.
Similar to the H-reflex the major parameters of the tendon reflex are latency and amplitude. Latency is the elapsed time between the tendon tap stimulus and the first deflection from the recorded signal. The latency is representative of the summation of afferent and efferent impulse conduction time and also the synaptic transmission time in the spinal cord. For EMG assessed reflexes the amplitude is calculated as the difference between positive and negative peak in terms of mV.
The lower and upper extremity consists of several tendons which are amenable for evoking reflexes in healthy subjects. Generally the reflex is evoked commonly from the lower limb, specifically the knee (Patellar tendon reflex) and ankle (Achilles tendon reflex).1
3. Previous concepts
Previous concepts have been assessed as similar to the present concept of quantified deep tendon reflexes using Mednodes. However although also novel the presented concept is established as having greater utility, both is terms of scalability and non-intrusiveness. One similar device is presented by the Italian research group of Pagliaro and Zamparo.
Pagliaro and Zamparo’s research group quantifies both input and output. The input of the tendon reflex is produced by an instrumented hammer from PCB Piezotronics Inc., USA. The output of the tendon reflex response is assessed by connecting the ankle to a load cell via an inextensible cable. The load cell and cable were aligned, so the measurement of the force component would be in the direction normal to the leg.5
Although the input is reliably quantified by the instrumentation of the tendon hammer, two aspects introduce variability in the system. The clinician would naturally have variation of intensity input and variation of accuracy of input impact on the patellar tendon. Since the load cell involves connecting the ankle to a cable, the measurement of the output involves restrictive devices, which do not characterize the full response. The load cell can also only accurately measure the maximal force of the reflex response not the full temporal nature of the reflex response.
Another device developed by Cozens and colleagues involved activation of the biceps brachii stretch reflex by a servo-positioned tendon hammer. Antagonist inhibition was also evoked by vibration of the triceps. The amplitude of biceps stretch reflex was measured using surface EMG. Cerebral dysfunction due to brain injury is demonstrated by abnormal myotatic reflexes, and the biceps brachii stretch reflex can be assessed at bedside during the acute phases of brain injury. Such reflex measurements can augment monitoring head injured patients.
The study investigates a myotatic reflex technique to assist with GCS observations. Such a method involves supraspinal centers of strong descending control for spinal reflex circuitry. Postulated is the concept that cerebral dysfunction due to brain injury is illustrated by abnormal myotatic reflexes consequential of disordered descending influence. The study envisions specifically the stretch reflex is disorganized in proportion to the degree of cerebral dysfunction.
The study consisted study consisted of 36 patients. GCS score was contrasted against EMG response. The data supports the hypothesis that cerebral dysfunction subsequent to brain injury is represented by myotatic reflex abnormalities. The study by Cozens concludes that abnormalities of myotatic reflexes observed in head injured patients is a consequence due to disordered descending supraspinal control of the circuitry associated with spinal reflexes. There is also a continuum of abnormality, which is inversely correlated with GCS score. The described technique for quantifying the stretch reflex of the biceps brachii has been demonstrated as a means for quantifying reflex abnormalities, which is capable of supplementing conventional clinical monitoring techniques. For further advancement of this concept, reproducibility must be established.6
The device developed by Cozens and colleagues bears a resemblance to the present device, which quantifies deep tendon reflexes. However the most significant contrast is the intended subject group. For the device by Cozens, the intended subject group is acute brain injured subjects, generally in intensive care and comatose. Given these acute conditions the GCS score is relevant to assessing the quality of the recovery. Rather than measuring the patellar tendon reflex, the biceps brachii tendon reflex is measured. The biceps brachii tendon reflex would be a preferred tendon reflex measurement for bed confined subjects. The study by Cozens and colleagues notably correlated the GCS score inversely with the abnormality of the tendon reflex response.
A Dutch research team consisting of Van de Crommert, Faist, Berger, and Duysens integrated a quantified input device and EMG measurements of the output reflex response. The tendon reflex input was produced by a reflex hammer driven by a magnetic motor. The input device weight was approximately one kilogram. The input device was affixed to the back of the lower leg for eliciting biceps femoris tendon reflexes.7 In a later study by Faist, Ertel, Berger, and Dietz the input device was affixed to the anterior section of the calf. The hammer can be accelerated to a constant velocity by the magnetic motor. A potentiometer in the motor could determine movement of the hammer and impact.8
Although the quantified input of the device is similar to the quantified deep tendon reflex device using Mednodes; the functional intent and inherent characteristics of the device are different. The motorized input device is intended for eliciting tendon reflexes during gait cycle. In contrast the quantified deep tendon reflex device using Mednodes is intended to evoke patellar tendon reflexes for a static position. Given the functional disparity for both input devices, the motorized input device must be affixed to the lower limb of the subject. The motorized device imparts a one kilogram load on the calf during the tendon reflex response. Therefore the reflex response is perturbed given a one kilogram load. The subject’s natural tendon reflex is preserved given the quantified deep tendon reflex device using Mednodes, since the device has minimal mass intrusion. As previously discussed the EMG readings lack the global information provided by the accelerational load assessed by the Mednode 3D accelerometer, such as the quality of the neuromuscular junction.
Another study, which attempted to quantify reflexes, was conducted by Lebiedowska and colleagues. The input device was quantified by a sweep-triggering hammer, which is manually operated, and equipped with a strain-gauge accelerometer. The response of the patellar tendon reflex was measured by a strain gauge beam attached to the ankle joint. Notably different is that the experiment involved instructing the subject, via biofeedback, to exert a certain load, which was parameterized as a percentage of maximum voluntary contraction. The study precipitated an interesting finding. The curves of reflex response as a function of percentage of maximum voluntary contraction are disparate for nominal neurology subjects relative to brain injured subjects.9
The study by Lebiedowska and colleagues is a fundamental step toward the quantification of reflexes. With respect to quantified input, the sweep triggered hammer lacks the variability of the swing arm potential energy variable hammer input of the quantified deep tendon reflex device. Also the Mednode 3D accelerometer can provide temporal characteristics of the complete reflex response. The simplicity of the quantified deep tendon reflex device allows for variable hammer strike inputs, rather than having the subject comply with biofeedback requirements.
4. Quantified deep tendon reflex device
The device is designed to maximize reproducibility. The two essential components of the device are a swing arm for quantifying input and a Mednode 3D accelerometer for measuring reflex response. The following picture characterizes the quantified deep tendon reflex device.
Figure 1. Prototype of the device for quantifying reflex response and latency
The design of the swing arm allows for quantified input based on settings of variable potential energy. The end of the swing arm is fastened to a standard neuro-reflex hammer for eliciting tendon reflexes of the patella. The swing arm position can be fixed to strike a specific point on the tendon. The swing arm is raised to a given angle (15, 30, or 45 degrees), and allowed to drop. The hammer strikes the patellar tendon with a known force. The design allows for studies of variable input intensity.
The 3D accelerometer Mednode provides the means to quantify the intensity of the reflex response. The Mednode is placed in a constant anatomical location to improve the reproducibility of the measurement. The signal of the 3D accelerometer Mednode is conveyed wirelessly to a local portable personal computer for data storage and processing. The Mednode incorporates battery-powered 3D accelerometers with integrated processing units. The Mednodes are software programmable, which allows great flexibility for this application. Data is processed in the PC.4
The system requires no placement of EMG electrodes or tethering of the leg. To test the feasibility of the design proposed in this application we evaluated the performance of the device in 3 subjects. Instead of the complexity of placing EMG electrodes, the reflex response is measured simply by attaching a 3D accelerometer Mednode to the lower leg. The swing arm can be targeted to a contact point to ensure input accuracy. To test the feasibility of the design we evaluated the device in three subjects.
The initial test and evaluation for the device study consisted of three subjects; two subjects had nominal neurology and one subject was a chronic hemiplegic. The tests were performed using the following protocol:
- Place the MEMS accelerometer with an elastic strap on the lateral aspect of the leg to the medial maleolus.
- Aim the patellar tendon hammer at the level of the tibial tubercle.
- Pull back the swing arm to a predetermined angle from an initial position, such as 15, 30 and 45 degrees.
- Release the swing arm.
- Record the 3D MEMS accelerometer data.
- Include a minimum ten second delay before the next patellar tendon hammer strike.6
- Repeat the assessment protocol to gather six measurements at each input level.
The experiment consisted of three sets of six measurements at each input level. Each set was conducted on a separate day. The six trials consisted of inputs of 45, 30, 15, 15, 30, and 45 degrees. Given three subjects being tested with each leg for three sets of six measurements, a total of 108 measurements were obtained. The intent of the experiment is to demonstrate engineering proof of concept. The initial study was intended to ascertain the reliability of the fully quantified deep tendon reflex device to demonstrate the ability to quantify the deep tendon reflex.
6. Results and Discussion
6.1. Preliminary results
A prototype system was designed and built consisting of a swing arm, Mednode, and reflex hammer. Communication between the various parts of the system is wireless and does not limit or restrict movement. The goal of this system was to use the Mednode device and associated hardware to develop a reliable means of measuring and quantifying patellar tendon reflexes.
The input for the tendon reflex is based on potential energy of the swing arm. The intensity of the input strike is highly consistent, provided the joint of the swing arm’s friction is minimized and the height of the arm prior to release is at the same level. The swing arm stand’s positioning allows for reliable input impact on the same point on the patellar tendon. The hammer can be first aimed, then pulled back to the desired potential energy setting to always achieve the same energy and strike position. The output is measured by the 3D MEMS accelerometer in the Mednode device. The Mednode conveys information to a portable computer by wireless connectivity. Therefore with exception to the minimal mass of the Mednode, the device measures output nonintrusively. The 3D MEMS accelerometer of the Mednode can reliably measure the complete temporal acceleration profile of the reflex response. This should be an improvement over other existing devices. With a Mednode also attached to the swing arm the reflex latency can be obtained by the time disparity between maximum acceleration of reflex response and maximum acceleration of swing arm strike.
6.2. Test and evaluation results:
The following graph consists of the raw acceleration signal from the Mednode.
Graph 1. Test sample of one subject (Mednode 3D accelerometer signal vs. reflex input)
Subsequently the data was further processed by converting the voltage signal to g’s of gravity. The following six graphs consist of the calibrated to g’s of gravity acceleration of reflex response for subject 1, the chronic hemiplegic subject both affected and unaffected leg; subject 2 and 3 with nominal neurology left and right leg.
Both Graphs 2 through 7 characterize averaged reflex response amplitude as a function of input. Graph 2 represents the chronic hemiplegic subject’s affected leg, and Graph 3 represents the chronic hemiplegic subject’s unaffected leg. Graphs 4 though 7 represent subjects with nominal neurology.
The results are based on a total of 108 measurements. The results indicate that the relative variation of 107 measurements for the quantified amplitude of reflex response was bounded by a maximum relative variation of 10%. Only one measurement for the quantified amplitude of reflex response was bounded by a maximum relative variation of 15%. These findings suggest it is possible to quantify deep tendon reflex using the proposed device. Determination of reproducibility of the measurement requires further testing.
Given the initial results, we believe that further trials are warranted to establish the reproducibility of the measurement technique. The goal of our initial clinical trial will be to assess the reproducibility of the device for quantifying deep tendon reflexes.
The deep tendon reflex is a fundamental component of the neurological examination. However the present quantification of the reflex response, by for instance the NINDS reflex scale, has inherent issues of accuracy. The proposed quantified deep tendon reflex device provided quantified input of tendon reflex strike and quantified assessment of reflex response. The preliminary test and evaluation of the device suggests a considerable degree of accuracy and reproducibility.
7.1. Advanced implications
Upon clinical validation for reproducibility of the Mednode 3D accelerometer based quantified deep tendon reflex device, the device can be extended into the field of gait analysis. During nominal gait cycle, the modulation of reflexes is considered to be functionally significant. The reflex modulation is reduced for the affected side of patients with hemiparesis relative to healthy subjects. Gait cycle for subjects with hemiparesis is generally asymmetric.8 Given equivalent tendon reflex inputs, the reflex response is disparate for the unaffected vs. affected lower limb for hemiparetics.9 The quantified deep tendon reflex device can measure the quality of reflex modulation for the affected and unaffected leg for hemiparetics. The static and quantified evaluation of deep tendon reflex modulation based on variable quantified input could be used to assess gait disparity for hemiparetics.
In order to further substantiate the novelty of the second generation quantified deep tendon reflex device, the initial aspects of the concept were disclosed in a UCLA Neuroengineering graduate class during June 2005.10 The concept for the quantified deep tendon reflex device has been presented multiple times at the 15th International Conference on Mechanics in Medicine and Biology during 2006 and also at the 35th and 36th Society for Neuroscience.11-13
- Instrumental has been the support from the UCLA IGERT NSF fellowship.
Provisional patents filed
- UC Case No. 2006-288; “Fully quantified evaluation of myotatic stretch reflex”
- UC Case No. 2006-660; "Quantified Deep Tendon Reflex Device”.
If you are the original writer of this essay and no longer wish to have the essay published on the UK Essays website then please click on the link below to request removal: