Imaging Motor Unit Recruitment Patterns

Within skeletal muscles the fibres are grouped into distinct functional units, motor units (MUs), with groups of fibres innervated by a single motoneuron located in the spinal chord. The size of MUs (number of innervated fibres) and their physiological properties varies within and between muscles. To complete any movement the nervous system must activate the appropriate number and combination of MUs to produce the varying levels of force required throughout the task.

Quantifying patterns of MU activation provides a means of i) studying how MUs are controlled by the nervous system to produce force; ii) measuring dysfunction caused by diseases such as cerebral palsy or Parkinson's disease; iii) diagnosing neurodegenerative diseases e.g. motor neurone disease; iv) estimating changes in MU number which occur due to disease, injury or ageing. Currently such studies are completed using electromyography (EMG), a method of recording electrochemical changes in the muscle fibre which occur when it is activated. However, the techniques available to analyse EMG signals and provide information on MU recruitment patterns limit the range of movement tasks that can be studied. So it is not currently possible to study patterns of MU recruitment during the majority of movements completed during acts of daily living (e.g. reaching, grasping) or more demanding tasks (e.g. running, negotiating different terrains or inclines).

Activation of MUs causes muscle tissue to move. The number and distribution of the fibres of a MU in a muscle will influence: i) the area of the muscle over which tissue movement spreads and ii) the timing of the spread of the movement which occurs when the MU is activated. Activation of MUs with different properties should therefore cause different 'signature' muscle tissue movement patterns to occur. The relationship between MU activation and tissue movement patterns has not yet been fully explored, but could be revealed using ultrasound imaging. These types of images have already been shown to reveal surprisingly small contractions (e.g. activation of single muscle fibres) and can be easily collected during different movements e.g. locomotion, reaching and grasping. We therefore propose developing novel ultrasound image analysis techniques, to provide a new way of studying MU recruitment patterns and extending the conditions under which MU properties can be studied.

We will develop algorithms to provide: i) high frame rate (>1000 frames per second) ultrasound imaging of skeletal muscle in 3D; ii) a means of defining muscle movement templates, identifying the 'signature' muscle tissue movement patterns associated with activation of different MUs; iii) a means of using the signature movement patterns to analyse image sequences and identify activation patterns of different MUs.

This work will therefore provide a new method of studying MU recruitment during a wide range of different movement tasks. In addition, our work will provide a means of: i) collecting experimental evidence to underpin the development of more accurate and valid mathematical models of muscle; ii) evaluating and monitoring changes in MU properties which occur as a result of ageing, injury, rehabilitation, disease and different forms of treatment aimed at improving physical health and well-being; iii) non-invasive investigation of MU properties which could be applied in humans and animals and provide novel data from the same or smaller numbers of animals.

We will adapt our current algorithms controlling a fully configurable high frame rate (>1000fps) ultrasound imaging device to enable simultaneous control of 2 ultrasound probes. The known geometric association between the probes and electronically controlled sweeping of the ultrasound beam will be used to provide high frame rate 3D images of muscles. Images will be recorded from custom built phantom devices and human muscle, with the accuracy of constructed images evaluated against known phantom properties and MRI of scanned muscles.

We will use the developed algorithms to collect image sequences of muscles during: i) evoked muscle twitches and the resulting reflex response, when MUs with known properties are activated and ii) low level voluntary activation when different populations of MUs will be recruited. Image sequences will therefore contain information on the spatial and temporal characteristics associated with activation of different MUs and MU populations. Multi-channel surface and intramuscular EMG will be simultaneously recorded during image collection. These data, coupled with the experimental condition, will provide independent information on the MUs activated during image recording. Data will be collected from biceps brachii and medial gastrocnemius which have different fascicle geometries.

We will use collected images to develop algorithms which discriminate between activation of different MUs based on information of the spatial and temporal tissue activation-displacement patterns. This will provide statistical models, or signature tissue movement patterns, from which activation of different MUs can be defined. In the final stage of the project we will develop machine learning techniques, applying the developed statistical models, to analyse previously unseen image sequences and map spatial and temporal activation characteristics within the imaged muscle volume to provide novel information on patterns of MU recruitment.

The potential impacts of the proposed work are broad with the following communities likely to benefit:

UK's International Profile: Building further specialist expertise and knowledge within our group would create a valuable resource for industrial partners and help maintain the UK's leading position in biomedical imaging research and translation. In addition, this research could enhance point of care imaging technologies which would be applied in clinical environments across the world strengthening the UK's reputation in this area.

Academics: Providing a means of quantifying and classifying temporal and spatial features of muscle tissue displacement in relation to MU recruitment patterns will provide a new, non-invasive and objective technique with which to study the neuromusculoskeletal system. For the first time it will be possible to begin studying the integration between neural drive and 3D anatomical muscle properties and provide experimental data crucial for the development of novel computer models of muscle. The proposed computational developments could therefore benefit scientific researchers across a multi-disciplinary community including: basic, applied and clinical medical and veterinary sciences, mathematical modelling, engineering and computer vision.

Industry: The proposed system has the potential to open up new uses of existing ultrasound imaging technology, with possible commercial exploitation by SMEs and global enterprises. Spin-out companies offering image analysis training services could arise contributing to job creation and increased skills within research and health service communities.

Clinical Impact: In the short-term we will provide a new method of mapping temporal and spatial MU activation properties within skeletal muscle. Such a tool could be of significant value to the clinical community for identification and monitoring of neurodegeneration which occurs as a result of pathology (e.g. Motor Neurone Disease, Multiple Sclerosis) or ageing (e.g. sarcopenia). The high spatiotemporal resolution and objective nature of the approach provides the potential to detail disease aetiology and changes to activation characteristics, improving the understanding of disease pathogenesis and quantitative evaluation of current and newly developed therapeutic interventions.
In addition, the newly developed computational methods could improve the value of ultrasound technology currently used by public and private healthcare providers. Extending current technology to facilitate new, objective methods of assessing the neuromuscular system could facilitate quicker, more sensitive disease diagnosis, reducing the number of referrals for re-examination and contributing to more efficient and cost-effective diagnostic services. Such application in medical care of humans could also be replicated in veterinary medicine with care of companion animals (e.g. cats, dogs) and performance animals (e.g. horses, greyhounds) potentially benefiting. In the longer term mapping temporal and spatial MU activation properties could lead to the development of techniques integrating measures of neural control (MU activation), mechanical tissue displacement (tissue strains) and the resulting force production which would provide even more powerful diagnostic and monitoring tools.

Teaching institutions: Revealing principles of MU recruitment and muscle anatomy through the use of 3D imaging techniques will facilitate the development of a number of tools which could be applied to teaching students at a range of educational levels from A-level biology through to undergraduate medical and veterinary sciences. Visualisation of different body systems is a powerful means of encouraging learning and interest in health and well-being and could also contribute to the development of novel means of promoting interest in anatomy and physiology which could be easily distributed to many members of the general public.