The ISEK Congress will include a series of pre-Congress workshops, scheduled to take place on Friday, June 29th, 2018. Please see the various descriptions below. You can sign up for a workshop when registering for the 2018 Congress.
You can register to take part in the Workshop day as an add on when you register for the full ISEK Congress. The fee is €90, which includes access to all workshops, lunch and a coffee break.
Want to add a workshop to your existing registration? Click here to login to your registration.
Methods for estimating synaptic potentials in human muscular system
Organisers: Kemal S. Türker, Gizem Yilmaz, Gorkem Ozyurt, Betilay Topkara and B.Selen Senocak, Koç University School of Medicine, İstanbul, Turkey
Since it is not possible to record directly from human neurons, indirect methods have been developed to study the synaptic potentials. An indirect estimate of a synaptic potential can be obtained by recording the activity developed in muscles in response to stimulating a set of afferent fibres. Classically, these indirect methods use surface or intramuscular electromyography (EMG) to represent the responses of motoneurons to stimuli. The most common classical techniques are rectification and averaging of the EMG around the time of stimulation and compiling peristimulus time histograms (PSTH) from the single motor unit records. The limitations of these classical techniques in estimating synaptic potentials were recognized and reports have claimed that they contain significant errors in estimating the underlying potentials . We have studied this problem in regularly discharging motoneurons in rat brain slice preparations . In these studies, we have illustrated that the classical methods for estimating pathways in the central nervous system do in fact contain significant errors and that these errors are minimized when the same discharge information is used in a peristimulus frequencygram (PSF). This workshop aims to highlight the differences between the classical and the novel methods via a formal talk followed by a demonstration to illustrate the errors for estimating the synaptic potentials when classical methods are used for estimation. For demonstration, we will: Use a simple reflex circuitry; Apply electrical stimuli on index finger and record surface EMG and single motor unit potentials from the APB muscle of the thumb; Analyse the stimulus evoked reflex responses in surface EMG using rectification and averaging around the time of the stimulus; Analyse the same single motor unit discharges using the PSTH and PSF methods to indicate the differences between them. By the end of the workshop the participants will appreciate that there are many ways of estimating synaptic potentials in human neuromuscular system and also recognize the pros and cons of each of the analyses methods. Since profile of the synaptic potentials indicate the functional wiring diagram between the stimulated afferent system and the motor neuron pool, it is extremely important that it is estimated correctly since neurologists use this information to diagnose and treat their patients.  TÜRKER, K.S. and CHENG, H.B. (1994) Motor-unit firing frequency can be used for the estimation of synaptic potentials in human motoneurones. Journal of Neuroscience Methods, 53:225-234. TÜRKER, K.S. and POWERS, R.K. (2005) Black box revisited: A technique for estimating postsynaptic potentials in neurones. Trends in Neuroscience, 28:379-386.
Extraction of information from high-density EMG: recent developments and perspectives
Organisers: Dario Farina, Imperial College London and Ales Holobar, University of Maribor
High density electromyography (hdEMG) is fast developing research methodology that led to many physiological discoveries in the last decade. Although the use of surface arrays is relatively simple, the interpretation of acquired signals remains to be challenging, especially in the case of pathology and in dynamic muscle contractions. Analysis of hdEMG is often nontrivial and errors often lead to misunderstanding of addressed physiological mechanisms. Different central and peripheral nervous system properties that may be assessed from hdEMG differ substantially in their sensitivity to intrinsic experimental and physiological factors, such as electrode-skin contact, thickness of interposed tissue, depth of muscle fibers, electrical properties of uptake electrodes, muscle crosstalk, fatigue, muscle geometry and anatomy of muscle fibers. These factors cannot be fully controlled, even in laboratory environments. It is, thus, of paramount importance that we understand the limitations and sensitivities of different information extraction techniques before we interpret the hdEMG signals. This workshop will systematically explain the main methodological steps needed for accurate and efficient interpretation of hdEMG signals. We will present and discuss large set of methodologies from amplitude and spectral estimators to non-negative matrix factorization and motor unit identification techniques. We will also present methodologies for robust assessment of neural codes and muscle activation primitives out of decomposed hdEMG signals. Methodological explanations will be accompanied with practical (hands on) demonstrations.
Wearable Technologies – Challenges & Advances
Organisers: Brian Caulfield, University College Dublin
If we look back 10 years we had just begun to get to grips with the new smart phone phenomenon and companies like Fitbit were still in start-up mode. Since then the fields of wearable sensing and mobile computing have expanded to the extent that these are now ubiquitous technologies. We now have the capacity to digitize human performance and behaviour on a 24-7 basis using a variety of relatively inexpensive technology platforms.
The early years of wearable sensing research were focused on developing our capability to sense human performance and generate data, and we are now well into the next phase of this field – developing our capacity to process and analyze the resultant data and use it in meaningful applications in health, wellness and sport. This work is associated with a number of important challenges; understanding the optimal strategies for sensing target data in an unobtrusive manner, extracting accurate biomechanical and physiological variables for the signals, turning this into actionable information that can effect the desired effect in different application contexts, and enhancing the human-machine experience. In this workshop we will bring together researchers from many different fields to present the different challenges and ongoing efforts to address them in leading research centres.
Talks will address topics such as:
- Selecting the right sensors for the job.
- Classification of human performance using wearables
- From sensing to behaviour change
- Actigraphy – moving beyond step counts
- From sensing to actuation – interfacing with robotics
- Remote monitoring in chronic disease
How good is my robot? The increasing importance of benchmarking in wearable robotics research
Organisers: Diego Torricelli, Jan Veneman, Jose Luis Pons, Cajal Institute Spanish National Research Council
Wearable robots for gait assistance and rehabilitation, such as exoskeletons and prostheses, are becoming increasingly available in the market. These robotic systems are currently tested according to self-defined procedures and metrics, which can hardly be transferred across different laboratories and platforms. This makes it very difficult, if not impossible, to compare similar systems to each other. Benchmarking is a powerful tool that can help researchers and developers to overcome this problem. Recent efforts, mostly based on competitions (e.g. Cybathlon), have raised the attention for this topic. However, a consolidated benchmarking methodology for Robotics is still not available. The definition of standardized metrics and protocols to evaluate the effectiveness of wearable robotics technology is a crucial yet very complex issue, which involves multiple perspectives at technical, clinical and usability aspects. The recently funded H2020 project EUROBENCH is working on the creation of a unified methodology for the assessment of wearable robotics performance, and will soon offer third party funding opportunities for research groups working on the following topics:
- Methods and protocols to test human and human-like locomotion performance in unstructured environments
- Equipment and test beds to allow replicability of benchmarking protocols across labs
- Simulation and modelling approaches to measure human-robot interactions
- Creation and use of databases to quantify human and robot performance in complex environments
- Testing safety of wearable robots
- Testing physiological and subjective impact of robotic use
This workshop will explore and discuss the main challenges behind robotic benchmarking as a follow-up of previous discussions workshops at WeRob2014, ICORR 2015, and WeRob2016, ICORR2017 conferences. The workshop is promoted by the H2020 Project EUROBENCH and supported by the benchmarking bipedal locomotion hub (www.benchmarkinglocomotion.org) and the COST Action CA16116 “Wearable Robots for Augmentation, Assistance or Substitution of Human Motor Functions” http://www.cost.eu/COST_Actions/ca/CA16116.
Surface EMG alive – an update of the current lines of application in real life settings
Organisers: Catherine Disselhorst Klug, Aachan University and Karen Sogaard , University of Southern Denmark
Developing a subject-specific tri-dimensional model of human muscles, from experimental anatomical and physiological data
Organisers: Leonardo Gizzi, Oliver Röhrle, Ekin Altan, Okan Avci, Filiz Ates, University of Stuttgart, Germany
Neuromuscular modelling is a powerful tool to complement and extend the information obtained from experimental data. If properly implemented, it can predict experimental results and provide information in cases where experiments would be unethical, impractical or even unfeasible. Musculoskeletal modelling allows investigating neuromechanical parameters that defy contemporary experimental techniques and overcome their intrinsic limitations. The aim of this workshop is to present a novel concept of neuromusculoskeletal modelling that, starting from anatomical (i.e. Magnetic resonance imaging – MRI and diffusion tensor imaging – DTI) and neuromechanical (i.e. high-density EMG, torque) data, provides a physiologically sound neuromechanical output and is able to account for those variables that are currently impossible to evaluate. The four main challenges of this process are addressed during the workshop 1) Conceive a fast and reliable pipeline to segment MRI and DTI Data and create a subject-specific finite elements (FE) model, comprehensive of fibres orientation, for a given anatomical district. 2) Determine the mechanical load and the cumulative neural drive to each muscle (and each sub-compartment) given a certain force/torque trace 3) From the specific neural drive to a muscle, create a physiologically rigorous population of motor neurons and populate the FE model. 4) Compute physiologically-sound action potential trains for each synthetic motor unit 5) Combine optimally the action potentials from the artificial motor units to generate synthetic force traces. The main advantages of this approach reside in its relatively short and straightforward pipeline, in the possibility to quickly solve the inverse dynamics problem for several muscles and to generate detailed external and internal interaction forces and a realistic neural activation patterns. The resulting force output can be used as a validation of the model which can, afterwards, be driven in open-loop (by prescribing motor units action potentials), or in a closed loop (by generating the neural drive to the muscle starting from a given force trace).