 Hello and welcome back to the Sports Biomechanics lecture series. As always, supported by the International Society of Biomechanics in Sports and kindly sponsored by Vicon. I'm Stuart McCurley-Nailer from the University of Suffolk and today I'm joined by Dr Stephen Lindley who is European Director at Delsis. So following on from the two brilliant talks that Vicon presented with a bit of a demo or tutorial into 3D motion capture. I was delighted that Stephen and Delsis offered a kind of similar talk really into EMG presenting an overview of a number of EMG related topics and Stephen's told me a little bit about what he plans on presenting and it sounds excellent so I'm really looking forward to it and hopefully you all enjoy it too. If anybody has any questions while the talk's going on then if you use the comment section on YouTube then we'll make sure that Stephen and Delsis can get an answer back to you after the talk is finished. Yeah thank you very much. Over to you Stephen. Thank you Stuart and thank you to ISPS for supporting a wonderful informative series of lectures so far. We're proud to be able to play our part in supporting the community during these challenging times. My name is Stephen Lindley and I've been working with Delsis for the past nine years primarily supporting the European community. My background is in sports science and rehabilitation with a PhD in the area of neuromuscular biomechanics. Today I've been asked to present on the area of surface EMG in sports biomechanics and like most speakers in this series so far we could all speak for weeks about the topics we've been asked to speak of and so today's session is intended to provide the viewers with some brief insights into what is EMG and where does it come from, the challenges and some recommendations associated with measuring EMG, live demonstration, some discussion around how we can analyze those EMG signals to extract useful information and how else surface EMG could be used through some exciting new ventures into EMG decomposition for studying motor unit control. We all often find ourselves asking these big and important questions which can be seen as difficult to answer but the reward we are offered when we can present an understanding of how the brain controls the muscles is suddenly in a magnificent one. We ask ourselves what are the basic motor control properties that govern muscle contractions? Does fatigue affect specific muscles differently? And what factors influence muscle performance? Understanding how the brain controls muscles can help scientists better understand and treat motor disorders, it helps therapists improve rehabilitation treatment and trainers develop more effective training protocols or engineers to improve prosthetic control and so how do we measure this? Electromyography or EMG could be defined as the study of muscle function through the inquiry of the electrical signal the muscles emanate and EMG signals presented in the top right is the electrical manifestation of the neuromuscular activation associated with a contracting muscle. This mesmerizing time series data presents the attractive notion of insight into nervous system and neural control so let us briefly revisit the origin of these EMG signals. The brain controls muscles by sending electrical signals along motor neurons which originate in the ventral horn of the spinal cord. Now each motor neuron innovates a group of muscle fibers together this is called a motor unit. Our motor unit is the single smallest controllable muscular unit we can measure. Muscles are composed of many motor units and each motor unit contracts when it receives a signal from the brain and can be presented as a motor unit firing instance. Different motor units can fire varying firing rates. The contraction of a single motor unit creates an electrical signature called a motor unit action potential also known as a MUAP. Each MUAP provides varying amplitudes of force which response. And motor units are recruited in a relatively fixed order whereby the first recruited motor unit tends to be the fastest firing and smallest motor unit compared to the last recruited motor unit which is firing the slowest and being the largest. And the EMG signal that we see here on the screen at the top right is a composition of all of the active motor unit action potentials within the muscle and is what's commonly used to provide valuable insights into human movement. Within context of sports by mechanics the EMG can be seen as a tool that provides easy access to physiological processes that are responsible for generating force and producing movement. Therefore it can help scientists, practitioners, athletes and engineers to answer key questions around the wide and varied field of sports by mechanics. And so who are we? Does us to the world leader in designing, manufacturing and supporting high precision electromagraphic technologies since 1993. Founded by Professor Carlo De Luca we've been long navigating the field of EMG through a unique combination of research and innovation alongside the experience of our product development. Professor De Luca instilled upon us all the importance of community support and with an amplified belief in the future generations. And so we've been giving back to the community over the years through Deltas and the De Luca Foundation. Especially so during these challenging times we are committed to continue that support. Now at Deltas we join you the scientists in the same starting point which is the scientific question or the research question. It's where a lot of our work requires conceptualization of new EMG technologies which then allows us to develop an EMG product and thus supporting the ongoing desire to advance our understanding of human movement together with the scientific community and into the real world. And as you see this process loops around our understanding will further our next scientific research questions and our next EMG products along the way which will then just further increase the research applications that we can consider. Following the award of the Walton Weiler Memorial Lecture at ISPS in 1993 in Paris, Professor De Luca wrote the paper the 1997 paper about the use of surface EMG and biomechanics and I would encourage readers viewers to study this and many of the materials like it in more depth to really get a fuller understanding of some of the topics we're going to study today. Now in this paper he now famously stated to his detriment, electromography is too easy to use and consequently too easy to abuse. This statement could be applied for all measurement technologies but it's especially present in the context of EMG. Now there are numerous factors we need to consider when measuring EMG. These selected intrinsic factors we cannot always control but it's important to be aware of when recording and analyzing and ultimately interpreting your EMG data whereas these extrinsic factors can be controlled through rigorous experimental design and the use of quality technology. One of the cornerstones in the ability to capture the beauty of the neural control system is to select and use high fidelity recording technologies. Choose a technology that offers you the full and flexible band with signals that can assure synchronized signals and smart electro design to reduce or suppress the introduction of sources of unwanted noise or artifact. Even with the best of technologies at hand the user needs to consider other critical components in collecting high fidelity EMG data. Where another cornerstone is the sensor location. The best location of the sensor on the muscle is generally on the midline of the muscle far from the tendon origins and the innovation zones. The amplitude and frequency spectrum of the EMG signal is affected by the location of the sensor or electrode. The figure on the left shows the effects of the EMG signal amplitude and frequency spectrum when the sensor or electrode is placed with respect to the innovation zone. On the top graph with the my tenderness junction on the bottom graph or on the lateral edge of the muscle in the middle plot. Accurate sensor placement can maximize the EMG signals amplitude and minimize muscle cross torque from neighboring muscles. And so we implore you to avoid where possible the innovation zones or origin insertion locations in the edges of the muscle. Aim for that midline of the muscle belly place the sensor or electrode in orientation of the muscle fiber. And if you follow these guidelines you will have an increase in the signal to noise ratio and a reduction in muscle cross torque. Another critical step in capturing valuable EMG data is to ensure the signals are of good quality. Unfortunately a number of noise sources can contaminate the recording of EMG signals and may not be easily recognized by visual inspection even to the most trained eye. Therefore here we present a tool that offers the user insight into some key metrics. First line interference from power lines. So this will be represented as 50 or 60 hertz noise. Typically introduced from power lines or fluorescent lights or rather electrical devices and they originate from the electrical magnetic radiation that is pervasive in all environments. While this is generally not a concern with modern technology, line interference may in some cases contaminate EMG recordings. Sensor detachment or excessive EMG signal amplitude may cause saturation of the signal, commonly referred to as clipping. If this occurs the contact between the sensor and the skin should be secured. The amplifier gain should be reduced if possible or the location of the sensor and the muscle should be moved to reduce the signal amplitude. The baseline noise is the composition of the electronics noise from the EMG sensor and the skin electrode interface. Accurate sensor placement and skin preparation ensures a low baseline noise but focus should be given on skin preparation of a subject. Often a simple cleaning of the application side with an alcohol wipe will suffice but users may require to remove excessive hair and or hydrate the skin. The signal to noise ratio or also known as SNR is arguably the best measure of quality of the EMG signal. It indicates the ratio of the EMG signal amplitude during a muscle contraction versus the unwanted electrical signal recorded when the muscle is at rest, i.e. the baseline noise. The higher the signal to noise ratio the more reliable the discrimination of EMG data for the underlying baseline noises. So now we've very quickly and briefly covered an introduction to EMG and some cornerstone recommendations around the measurement of an accurate signal. We should now show you a signal live and collect some data. So I'm going to switch across to EMG works acquisition. Okay, so on my other screen I already had the software laid it up. So what we're using today is the Dallas EMG works acquisition software platform to be able to collect our rule EMG data and give you a bit of a demonstration. What we're going to be using today is the Dallas Trinio Avanti sensors. So the Dallas Trinio Avanti sensors here on the webcam you better see is a single unit weighing 14.7 grams, which on the rear of the sensors where we house our EMG electrodes in these parallel bar formats. With inside the sensor, we also house an IMU and a natural measurement unit. But for today's purpose, we're going to be focusing just on the EMG element to be able to show you some EMG signals live. So firstly, I want to sort of just cover some of those very briefly the cornerstones you mentioned in the in the presentation. So firstly, what we're going to study is the biceps muscle, simply for ease of use for the webcam here. And what I firstly want to establish is the sensor location. So as per our previous guidelines, the simplest form is aimed for the middle of the muscle belly. So in larger muscle groups, prime movers like the biceps, then it's usually relatively straightforward to find that middle of the muscle belly, we have a large area to try and locate. But again, we're trying to avoid the origin insertion locations of the my tenderness junctions and the edges of the muscle or known areas of innovation zones. So for the biceps, we are just going to place in the midline and fix our EMG sensor down to that location in orientation with the muscle fibers. Before we do that, we need to go through skin preparation. So for the minimum, as we recommend, if a skin preparation would be a simple wipe with an alcohol wipe, doesn't have to be aggressive. Just a simple wipe up and down. And this alcohol will wicker away and open up the pools to increase the EMG signal amplitude coming through, but as well as removing the naturally occurring debris or oils that are present on the skin. Again, depending on the application side of the subjects you're working with, and your research question, you may need to remove excessive hair or perform tape peels or hydrate the skin before you go ahead and collect data. With the Delta sensors, we have a dry electrode system. So there's no electrolytes or pre-gelged electrodes that are required. We simply use a double sided adhesive onto the rear of the sensor, and then we fix this adhesive then to the subject skin. So what's often is quite handy when looking for the midline of the muscle belly. Of course, we can use our anatomy knowledge, and we can mark the area where we want to place the EMG sensor before we go ahead. Also quite common is to ask the subject to perform an isometric or some form of contraction, so you can maybe quite clearly see or palpate the muscle belly. So I've now fixed the sensor on the bicep. I'm just giving it a good push down to make sure that a D-server's got good contact with the skin. At this point, we are ready to go ahead and collect some data. The Delta Trinium Avanti sensor here is going to communicate wirelessly from the sensor that's my receiving base station, which is just next to me here on the table. Communicate via USB to the computer whereby the data is then streamed in digitally into EMG works acquisition. So firstly, I'm going to zoom in on the signal because the amplitude is certainly going to be quite low. And I'm just going to perform some very simple low level bicep curls with no weight. And so immediately we're presented with this interference pattern, this EMG signal as we saw briefly into the presentation. What you'll also notice is that as I contract harder, we'll see the larger signal. As we do smaller contractions, we see a smaller signal. You see when we're completely relaxed, we have a nice flat line compared to we have a contraction, we see all of the modians firing at the same time. So hold on to some of this information as we go into the next part of our lecture where we talk a bit more about analysis. What I want to look at now is revisit some of the qualities we spoke about about what is a good or a bad signal. So importantly, we went through the correct steps to be able to get to this point. We prepared the skin, we found a good sense of location, and we used high precision technology. But what we should really evaluate now is is it really in the correct location? And is that signal of good enough quality for me to go ahead and continue with my data collection? Now the experienced EMG user has seen many, many EMG signals in their life would be able to look at this raw signal and know just by eyeballing whether that is a good or a bad signal. But as we mentioned in the lecture in the slides, we now present a real time tool to assess the signal quality. So just as per the slides we saw, we have these three different dials that gives information on the line interference, the clipping noise, our baseline noise activation, and estimated signal to noise ratios. So if a line noise, we can see we're right down at the bottom in the green. There is no line interference here. I don't have too many fluorescent lights. I am on a desk with computers and screens and mobile phones. We can see that's not affecting it largely due to the wireless nature of these EMG sensors, but also due to the referencing guide, referencing electrode system on the rear of our sensor where we have two bars measuring EMG and two bars being reference electrodes. But we couldn't perhaps introduce this by touching something that's got a high leakage, electrical leakage, such as an electrical power supply, a laptop power supply is typically quite noisy. So here I'm just holding on to my laptop power supply and we can see the dial is moving up towards the red. We have started to get some line interference, but it's not of large enough magnitude to go into the red. Now if I just paused and zoomed into the signal, we should also see that characteristic line noise. So we can see here in the signal, this cyclical noise. And again, if we looked at the frequency content of this, it would be in the 50 hertz range. And so that again backs up with what our dials are telling us, but also if we just want to look at this manually to ensure that it was correctness calculation, we can certainly see the cyclical nature of a signal coming down here. There's not just motor units firing. So we'll return to the real time plot. The other step would be the clipping interference. So going beyond the measurement capability of the sensor. Now the input signal here, my biceps activity isn't going to exceed the maximum range of our EMG sensor. And the sensor is well fixed down into the skin. If the sensor wasn't fixed down to the skin or was partly hanging off or had fallen off, we may see clipping interference coming in. But we don't see that either within the raw signal or in the indicator. The middle dial indicates information about the baseline noise. So remember, this is the composition of the electronics noise of the sensor, which is very low, but is mainly due to the skin electrode interface and the skin preparation and the subcutaneous tissues and all the different tissue layers before we actually get to this. So we're aiming for a smaller baseline noise as possible. So we can see on the dial, we're around about one microvolt RMS. And again, we can look at the raw EMG signal zoom all the way in plus or minus 10 microvolts here. So it's a very small baseline noise activity, which then allows us to have a very high signal to noise ratio, the next dial along. So again, if you remember from the slides, this is a ratio between the amplitude of the EMG signal and of the baseline noise. So as I perform a contraction, we can see a large EMG signal next to a very small baseline noise. That allows us to discern the important information about when a muscle is on set, when he's on, when the muscle is off and ensure that we have got a good quality signal, we can run a lot of complex analysis with afterwards. If you're starting your EMG JD collection with line noise clipping interference, a high baseline or poor signal to noise ratio, there's only limited amounts we can do post processing with those signals to be able to answer the questions you want. And so as always, we encourage users to spend the time at this point of study conception, study design and pilot studies, and also now with these handy tools for each subject and for each muscle, actually go ahead and investigate what this signal looks like. So at this stage, we're going to move away from the live signal acquisition and we're going to head back to the presentation to be able to take it a step further into looking at how we can, after we've captured these type of bicep data from bicep curls here, what can we do with that data and how can we quiz this data to provide us useful and relevant information. Okay, so now we've just switched back to the lecture notes now whereby we're going to take a step into data analysis. Now, analyzing and interpreting an EMG signal is a much studied topic and one which can fill many lectures over. Today, I aim to provide a brief and simplified view of the most common areas for analysis, which of course can evolve into ever complicated methodologies, but I want to demonstrate the simplified view to excite and peak the interest in EMG and sports by mechanics, not to blind users with maths or complexity. With this in mind, I've presented a simple model for guiding the analysis of EMG data. You can see here we have three categories, physiological effect, EMG signal characteristic and an example computation. Although I suspect we will understand the analysis of any data is more involved in this simplification, I present here, there is a novelty in presenting an understanding complicated data in very simple ways. So with that in mind, firstly we have muscle effort. So this is where the user would desire to understand the physiological effect of how a muscle or multiple muscles are working, how hard they're working. And the EMG signal can be characterized and defined by its amplitude. So you could use a simple moving average window filter such as a Rootman squared or an RMS and then find the peak to compute that amplitude. So as you increase the force or torque around a joint, the EMG signal demonstrates increase in the signal amplitude. However, this observation only provides a qualitative indication of the relationship. And this is a very attractive notion of around about force and EMG and I would sincerely encourage users to read more in detail around what this means. Although it is an attractive notion to demonstrate the increase in EMG signal and increase in force, seemingly being a linear relationship and causative, then there are a lot of more technicalities involved in this. And so I would encourage people to learn more around this for some of your research applications. However, we can simply demonstrate this quantitative, a qualitative view by, if we look on the right-hand side here of the graph, the slide where the bicep curl is executed with two different weights, 10 kilograms and 20 kilograms. And you can clearly see these plots at the same scale, you can clearly see the heavier weight at the bottom results in a larger EMG signal. Now, this area of exploration may provide the user with the ability to study the relative load across muscles, between tasks or across subject groups, many more. The next physiological effect commonly studied is the activation timing of muscles, or also known as the on-off. The user would be looking at the signal when the amplitude goes above a determined threshold, and thus judged on being on or off. And a typical approach would be to filter the EMG signal, again perhaps with a moving average window filter, and use a threshold calculation, perhaps being based on the number of standard deviations above the baseline noise level, to determine whether it's an on or off state. In the quantification of the muscle on or off state is arguably the most common, and therefore, unfortunately, the most variance for analysis. Now this example here we see on the right hand side, uses three standard deviations above the baseline noise level, during some repetitive bicep-carl action. So the same data we just did in the live demonstration, just pre-recorded this to demonstrate this analysis. The red square square wave pulses that are over the top of the blue raw EMG signal indicate when the muscle is on and when the muscle is off. So zero when it's off, one when it's on. And the bottom graph here shows how when you can analyze multiple muscles, you can determine the time when muscles are contracting at the same time, i.e. co-contraction, which is again a certainly well studied area, and can provide some valuable insight into biomechanical efficiencies or sporting performance or rehabilitation states. Also for an EMG signal, we can provide an estimate or an index to the fatigue of the muscle through the inquiry of the frequency domain of the signal. Interestingly, the signal presents with an increase in EMG amplitude over time, and a decrease in the mean or median frequency. Again, the example computation is typically a frequency based calculation, and most commonly has been the median density frequency or median frequency. Now, there are many methods to study fatigue where physiologists would tend to monitor the failure point. We in biomechanics are interested perhaps in the progression of the fatigue over time, and we could extrapolate from an EMG signal a fatigue index. So in this example, on the right hand side, we've again used the biceps brachii and we've held a same 10 kilogram weight at a fixed joint position. So the elbows and flex 90 degrees position tucked into the side of your trunk, and this was held for 210 seconds before failure point was reached. And you can see in the raw EMG plot, the one at the top, we have an increase in EMG amplitude over time, as the muscle is working harder to maintain the weight in this fixed joint position. Then we follow on to the graph underneath the raw plot whereby it shows the output after the signal has been processed with the median frequency calculation, being the blue line, and the red line is a line of best fit, just demonstrating the decrease in frequency as the contraction progress through time. Again, although this is an attractive line of inquiry, we would encourage readers to seek more information about the requirements of data collection and analysis configurations to offer a valid interpretation. For example, the most commonly explored area for muscle fatigue for data collection is that we're studying an isometric contraction state whereby the joint position and therefore the muscle length is remaining fixed, which is a method here of how we're using the median frequency to analyze that signal. So it's useful to explore further around all of these analysis topics so you can understand the nuances around them. Now finally, there is certainly a popular application of EMG in the field of biofeedback, whereby users would work with their subjects or their patients, their athletes, to show them in real time an EMG signal as a form of feedback on the action or task that they're performing, and biofeedback can be represented in many visualizations through simply plotting of a raw or smoothed EMG data, or active bar graphs from singular multiple muscles, or a more modern approach with muscle mapping type concepts. Now although again this was a whirlwind tour of some of the options of how you could utilize the EMG signal, it was targeted to be simplified and without lengthy discussion on the nuances of each method and their options, and again we would encourage viewers to read the provided materials, if you look in the links below the video on YouTube, and the steward's going to upload the Dallas's website and lots of other educational tools, or contact us at Dallas to learn more whereby we're more than happy to support you in your EMG ventures. Now I want to briefly touch upon an emerging field of interest with EMG, a one that holds great promise for further unraveling the complexities and mysteries of the nervous system. The focus on the lecture so far has been this, the EMG signal, being the composition of the active motor units. Now imagine the possibilities if we're able to take a step back and provide insight into the composition of that signal, where we could measure the motor unit firing rates and capture the motor unit action potentials and the deeper exploration of neural control. The study, the topic of studying motor unit action potentials is one that's been under exploration for decades. In 1972 Professor DeLuca published a solution of measuring motor unit activity using a quadrified neat electrode, which is an invasive method before fine wire electrodes inserted into the muscle belly through hypodermic needle. Now this work has been continued over the years by DeLuca and colleagues to further explore the concept of decomposing an EMG signal, albeit from within the muscle, to provide information on the motor units themselves. Now we fast forward to 2006, whereby DeLuca and colleagues presented a solution that enabled researchers to capture a surface EMG signal. Again remember previous attempts at this have all been invasive with the quadrified needle or fine wire on needle EMG electrode. So here they showed a solution that captured a surface EMG signal and were able to accurately decompose that signal into high yields of concurrently active motor units. This was a huge breakthrough in the field, mainly due to the advanced methods of analysis, rather than advanced hardware technologies. But you will notice here on the slide whereby the EMG electrode certainly differs to the one we showed you, you may have been previously exposed to, whereby we have five pins. These five pins provide us four channels of single differential EMG signal from a single sensor. And that is important while we have the four channels adding this very particular orientation. But as I mentioned a lot of the hard work has been placed within the decomposition algorithm which is provided with those four channels of raw surface EMG and when we expand that signal that we presented with across all four channels we're able to start recognizing some unique shapes or templates. And now these are motor units or superpositions of motor units within the surface EMG signal. Now the algorithm scans all the signals from all four channels of data up and down and identifies as many different shapes that it can find within the signal. Following on from that it works through a discriminative step whereby we look at the different shapes and how they may be different differently allocated different motor unit action potential trains. And so here now you see on the screen we've expanded out this surface EMG signal. We're able to isolate numerous different motor unit action potential or numerous shapes which then were allocated to the motor unit action potential trains whereby we could present an average shape of the motor unit action potential as you can see on the left hand side of the screen. And each one of those fire and the firing instance for each one of those motor unit action potentials. And again here we just showed a subset of motor units but we've easily demonstrated up to 50 plus motor units within an isometric contraction. So as I just mentioned the previous work was limited due to the contraction type to decomposing isometric only contractions. And this is because the signal of a dynamic contraction is highly complex with the motor unit action potential shapes changing throughout a contraction cycle as the muscle lentils or as the force increases. And therefore we also see an increase in superposition of motor units and many of the technical challenges associated with this. However in 2014 Deluca demonstrated the evolution of the previous decades worth of processing abilities to be able to decompose the cyclical dynamic contractions. And therefore offering the community the scientific community possibilities of studying motor unit control in more applied fields. And for the first time in ravelling perhaps how our nervous system can control dynamic muscle actions at a motor unit level. Now looping back to some of the original slides here and understanding how deltas operates so that original 40 years worth of scientific work took its time to come through but we're able to then leverage the understanding and scientific findings we found in the research team at deltas to create and mold this into a product over the years. And we evolved our scientific work in thinking from the research team into our wireless platform alongside our existing wireless platform called the deltas tree system which allowed us to create a new concept and method to capture data called the neuromap system. To continue the theme of bicep cars here you can we can be seen capturing EMG data with the Trinio Galileo sensors again the middle line of the muscle belly and we're capturing simple bicep cars like in our previous example like in the data we the previous example data we got from the traditional surface EMG. Now after we collect that raw EMG data again four channels of raw EMG data we use the neuromap software to decompose that surface EMG data using the previous mentioned steps into the motor unit action potentials in their firing instances. And the bottom plot shows the firing rate of all those motor units found within that area after being smoothed with the handing window filter and the black line going through the signal indicates the elbow joint angle. So now following the decomposition and the initial presentation of the data by presenting the motor unit action potential shapes and the firing instances and then the smoothing of those firing instances to give us a firing rate we can also perform statistical analyses of that firing rate data and the motor unit action potential amplitudes or characteristics. So here we've simply just selected the concentric phase of the bicep car across the five repetitions with inside the neuromap software and we've regressed the average firing rate versus the peak motor unit action potential amplitude and to see what it offers us in the insight into what changes across the contractions as the subjects perhaps becoming more fatigued and so we can receive from these regression lines that from the first repetition being the bottom of the graph in the purple color going to the last repetitions that blue and orange color at the top and we can see even across just a couple of repetitions how the control of the muscle differs with the increasing motor unit firing rate and change in the motor unit action potential amplitudes and so now also maybe what might be quite interesting for us to do is to move on from just starting the concentric phase or just starting the isometric behavior of this muscle during this task but instead we can now compare the concentric and eccentric phases of the contraction which previously with traditional surface cmg we simply get some onset timing or perhaps amplitude changes which are fraught with problems unfortunately and now when we isolate the content of eccentric phases within those bicep curls here we can see the results it demonstrates a simple example of smaller motor unit action potentials having higher firing rates than their larger motor unit action potentials and we can still see that there's hierarchical arrangement persists the hierarchical arrangement of the firing rates with the first recruited motor unit being the smallest motor unit firing the fastest and the last recruited motor unit being the largest forced which firing the slowest and this persists across the eccentric and concentric activities phases but also importantly that each motor unit maintains a higher firing rate during concentric than eccentric activity which for the biomechanist here is quite a fascinating concept and idea to perhaps inform us of our training or our performance gains that we're interested in now this simple example that we use the theme throughout of studying just the biceps during some simply the isometric holes or bicep curls is designed to be simple in its nature to offer as a relatively straightforward view but now imagine the possibilities of expanding more from this concept of studying perhaps the biceps and triceps and studying the co-contractions the concentric eccentric phases of that action it provides us some fascinating insight into motor unit decomposition and what that can offer is over a traditional analysis of an MG signal whereby we can ask questions of the nervous system that have never been asked before highlighting to us that there's a lot more information to find in and amongst our complex neurophysiology so I wish to thank you all for listening and watching we hope that this lecture and the materials provided useful either in introducing you to EMG or adding to your existing knowledge and perhaps further picking your interest in EMG through these novel methodologies and again we are really happy to support the community further so if you do have questions please do get in touch we'll be happy to hear from you and support you and again thank you to Stuart for organizing this event and inviting us here today and thank you to ISPS for the continued support within the community thanks Thanks Stephen that was brilliant I think as you said EMG is potentially too easy to use and too easy to abuse but you demonstrated both aspects of that really nicely where you showed us how easy it is to use but also helped us to avoid some of the common misuse that we might see and so that was really useful I really liked the demo as well especially the live signal quality demonstration yeah I thought that was brilliant and the decomposition of surface EMG signals was a really interesting avenue to end it on as well so yeah huge huge thank you to Stephen and Delsis if anybody's got any questions for Stephen then either get in touch via the mechanisms you mentioned at the end or leave a comment on YouTube and I'll be sure to pass those on and get an answer back for you and yeah all that really leaves is just take a look on the screen at what's coming up in the next few weeks building up to the ISPS conference week and don't forget to subscribe and click on the bell and you'll then receive notifications whenever things are updated to keep you informed with the series yeah thank you very much and as I say last but by no means least another huge thank you to Stephen and Delsis thank you