EEG motor imagery decoding: a framework for comparative analysis with channel attention mechanisms.

Objective.The objective of this study is to investigate the application of various channel attention mechanisms within the domain of brain-computer interface (BCI) for motor imagery decoding. Channel attention mechanisms can be seen as a powerful evolution of spatial filters traditionally used for motor imagery decoding. This study systematically compares such mechanisms by integrating them into a lightweight architecture framework to evaluate their impact.Approach.We carefully construct a straightforward and lightweight baseline architecture designed to seamlessly integrate different channel attention mechanisms. This approach is contrary to previous works which only investigate one attention mechanism and usually build a very complex, sometimes nested architecture. Our framework allows us to evaluate and compare the impact of different attention mechanisms under the same circumstances. The easy integration of different channel attention mechanisms as well as the low computational complexity enables us to conduct a wide range of experiments on four datasets to thoroughly assess the effectiveness of the baseline model and the attention mechanisms.Results.Our experiments demonstrate the strength and generalizability of our architecture framework as well as how channel attention mechanisms can improve the performance while maintaining the small memory footprint and low computational complexity of our baseline architecture.Significance.Our architecture emphasizes simplicity, offering easy integration of channel attention mechanisms, while maintaining a high degree of generalizability across datasets, making it a versatile and efficient solution for electroencephalogram motor imagery decoding within BCIs.

A System for Reproducible 3D Ultrasound Measurements of Skeletal Muscles.

In 3D freehand ultrasound imaging, operator dependent variations in applied forces and movements can lead to errors in the reconstructed images. In this paper, we introduce an automated 3D ultrasound system, which enables acquisitions with controlled movement trajectories by using motors, which electrically move the probe. Due to integrated encoders there is no need of position sensors. An included force control mechanism ensures a constant contact force to the skin. We conducted 8 trials with the automated 3D ultrasound system on 2 different phantoms with 3 force settings and 10 trials on a human tibialis anterior muscle with 2 force settings. For comparison, we also conducted 8 freehand 3D ultrasound scans from 2 operators (4 force settings) on one phantom and 10 with one operator on the tibialis anterior muscle. Both freehand and automated trials showed small errors in volume and length computations of the reconstructions, however the freehand trials showed larger standard deviations. We also computed the thickness of the phantom and the tibialis anterior muscle. We found significant differences in force settings for the operators and higher coefficients of variation for the freehand trials. Overall, the automated 3D ultrasound system shows a high accuracy in reconstruction. Due to the smaller coefficients of variation, the automated 3D ultrasound system enables more reproducible ultrasound examinations than the freehand scanning. Therefore, the automated 3D ultrasound system is a reliable tool for 3D investigations of skeletal muscle.

Predicting EEG Responses to Attended Speech via Deep Neural Networks for Speech.

Attending to the speech stream of interest in multi-talker environments can be a challenging task, particularly for listeners with hearing impairment. Research suggests that neural responses assessed with electroencephalography (EEG) are modulated by listener's auditory attention, revealing selective neural tracking (NT) of the attended speech. NT methods mostly rely on hand-engineered acoustic and linguistic speech features to predict the neural response. Only recently, deep neural network (DNN) models without specific linguistic information have been used to extract speech features for NT, demonstrating that speech features in hierarchical DNN layers can predict neural responses throughout the auditory pathway. In this study, we go one step further to investigate the suitability of similar DNN models for speech to predict neural responses to competing speech observed in EEG. We recorded EEG data using a 64-channel acquisition system from 17 listeners with normal hearing instructed to attend to one of two competing talkers. Our data revealed that EEG responses are significantly better predicted by DNN-extracted speech features than by hand-engineered acoustic features. Furthermore, analysis of hierarchical DNN layers showed that early layers yielded the highest predictions. Moreover, we found a significant increase in auditory attention classification accuracies with the use of DNN-extracted speech features over the use of hand-engineered acoustic features. These findings open a new avenue for development of new NT measures to evaluate and further advance hearing technology.

Optimal Identification of Muscle Synergies From Typical Sit-to-Stand Clinical Tests.

Goal: The goal of this manuscript is to investigate the optimal methods for extracting muscle synergies from a sit-to-stand test; in particular, the performance in identifying the modular structures from signals of different length is characterized. Methods: Surface electromyography signals have been recorded from instrumented sit-to-stand trials. Muscle synergies have then been extracted from signals of different duration (i.e. 5 times sit to stand and 30 seconds sit to stand) from different portions of a complete sit-to-stand-to-sit cycle. Performance have then been characterized using cross-validation procedures. Moreover, an optimal method based on a modified Akaike Information Criterion measure is applied on the signal for selecting the correct number of synergies from each trial. Results: Results show that it is possible to identify correctly muscle synergies from relatively short signals in a sit-to-stand experiment. Moreover, the information about motor control structures is identified with a higher consistency when only the sit-to-stand phase of the complete cycle is considered. Conclusions: Defining a set of optimal methods for the extraction of muscle synergies from a clnical test such as the sit-to-stand is of key relevance to ensure the applicability of any synergy-related analysis in the clinical practice, without requiring knowledge of the technical signal processing methods and the underlying features of the signal.

Predictive simulation of sit-to-stand based on reflexive-controllers.

Sit-to-stand can be defined as a set of movements that allow humans to rise from a sitting position to a bipedal standing pose. These movements, often categorized as four distinct kinematic phases, must be coordinated for assuring personal autonomy and can be compromised by ageing or physical impairments. To solve this, rehabilitation techniques and assistive devices demand proper description of the principles that lead to the correct completion of this motor task. While the muscular dynamics of the sit-to-stand task have been analysed, the underlying neural activity remains unknown and largely inaccessible for conventional measurement systems. Predictive simulations can propose motor controllers whose plausibility is evaluated through the comparison between simulated and experimental kinematics. In the present work, we modelled an array of reflexes that originate muscle activations as a function of proprioceptive and vestibular feedback. This feedback encodes torso position, displacement velocity and acceleration of a modelled human body with 7 segments, 9 degrees of freedom, and 50 actuators. We implemented two controllers: a four-phases controller where the reflex gains and composition vary depending on the kinematic phase, and a simpler two-phases controller, where three of the kinematic phases share the same reflex gains. Gains were optimized using Covariance Matrix Adaptation. The results of the simulations reveal, for both controllers, human-like sit-to-stand movement, with joint angles and muscular activity comparable to experimental data. The results obtained with the simplified two-phases controller indicate that a simple set of reflexes could be sufficient to drive this motor task.

Dynamic 3D Ultrasound Imaging of the Tibialis Anterior Muscle.

Skeletal muscle volume has been mainly investigated under static conditions, i.e. isometric contractions. The aim of our study is to use ultrasound imaging to determine muscle deformation during movement. We used a custom-designed scanning rig to obtain 3D ultrasound images of a subject moving the foot from plantarflexion to dorsiflexion at constant velocity. Using motion capture, we computed the respective angle of the ankle for each frame and collected them in bins based on the measured angle (rounded on the next normal number). For each degree, we used Stradwin for the 3D reconstruction of the respective volume. We found increasing cross-sectional areas for increasing dorsiflexion angles. The proposed method is a promising approach for determining muscle volume during movement. Future studies aim at collecting more data to compute muscle volume and length during contraction and compare the results to isometric measurements.

Investigating the spatial resolution of EMG and MMG based on a systemic multi-scale model.

While electromyography (EMG) and magnetomyography (MMG) are both methods to measure the electrical activity of skeletal muscles, no systematic comparison between both signals exists. Within this work, we propose a novel in silico model for EMG and MMG and test the hypothesis that MMG surpasses EMG in terms of spatial selectivity, i.e. the ability to distinguish spatially shifted sources. The results show that MMG provides a slightly better spatial selectivity than EMG when recorded directly on the muscle surface. However, there is a remarkable difference in spatial selectivity for non-invasive surface measurements. The spatial selectivity of the MMG components aligned with the muscle fibres and normal to the body surface outperforms the spatial selectivity of surface EMG. Particularly, for the MMG's normal-to-the-surface component the influence of subcutaneous fat is minimal. Further, for the first time, we analyse the contribution of different structural components, i.e. muscle fibres from different motor units and the extracellular space, to the measurable biomagnetic field. Notably, the simulations show that for the normal-to-the-surface MMG component, the contribution from volume currents in the extracellular space and in surrounding inactive tissues, is negligible. Further, our model predicts a surprisingly high contribution of the passive muscle fibres to the observable magnetic field.

Optimizing NV magnetometry for Magnetoneurography and Magnetomyography applications.

Magnetometers based on color centers in diamond are setting new frontiers for sensing capabilities due to their combined extraordinary performances in sensitivity, bandwidth, dynamic range, and spatial resolution, with stable operability in a wide range of conditions ranging from room to low temperatures. This has allowed for its wide range of applications, from biology and chemical studies to industrial applications. Among the many, sensing of bio-magnetic fields from muscular and neurophysiology has been one of the most attractive applications for NV magnetometry due to its compact and proximal sensing capability. Although SQUID magnetometers and optically pumped magnetometers (OPM) have made huge progress in Magnetomyography (MMG) and Magnetoneurography (MNG), exploring the same with NV magnetometry is scant at best. Given the room temperature operability and gradiometric applications of the NV magnetometer, it could be highly sensitive in the pT / Hz -range even without magnetic shielding, bringing it close to industrial applications. The presented work here elaborates on the performance metrics of these magnetometers to the state-of-the-art techniques by analyzing the sensitivity, dynamic range, and bandwidth, and discusses the potential benefits of using NV magnetometers for MMG and MNG applications.

Variations in Muscle Activity and Exerted Torque During Temporary Blood Flow Restriction in Healthy Individuals.

Recent studies suggest that transitory blood flow restriction (BFR) may improve the outcomes of training from anatomical (hypertrophy) and neural control perspectives. Whilst the chronic consequences of BFR on local metabolism and tissue adaptation have been extensively investigated, its acute effects on motor control are not yet fully understood. In this study, we compared the neuromechanical effects of continuous BFR against non-restricted circulation (atmospheric pressure-AP), during isometric elbow flexions. BFR was achieved applying external pressure either between systolic and diastolic (lower pressure-LP) or 1.3 times the systolic pressure (higher pressure-HP). Three levels of torque (15, 30, and 50% of the maximal voluntary contraction-MVC) were combined with the three levels of pressure for a total of 9 (randomized) test cases. Each condition was repeated 3 times. The protocol was administered to 12 healthy young adults. Neuromechanical measurements (torque and high-density electromyography-HDEMG) and reported discomfort were used to investigate the response of the central nervous system to BFR. The investigated variables were: root mean square (RMS), and area under the curve in the frequency domain-for the torque, and average RMS, median frequency and average muscle fibres conduction velocity-for the EMG. The discomfort caused by BFR was exacerbated by the level of torque and accumulated over time. The torque RMS value did not change across conditions and repetitions. Its spectral content, however, revealed a decrease in power at the tremor band (alpha-band, 5-15 Hz) which was enhanced by the level of pressure and the repetition number. The EMG amplitude showed no differences whilst the median frequency and the conduction velocity decreased over time and across trials, but only for the highest levels of torque and pressure. Taken together, our results show strong yet transitory effects of BFR that are compatible with a motor neuron pool inhibition caused by increased activity of type III and IV afferences, and a decreased activity of spindle afferents. We speculate that a compensation of the central drive may be necessary to maintain the mechanical output unchanged, despite disturbances in the afferent volley to the motor neuron pool.

A Systematic Review and Meta-Analysis on the Longitudinal Effects of Unilateral Knee Extension Exercise on Muscle Strength.

The aim of the study was to investigate the time-dependent increase in the knee extensors' isometric strength as a response to voluntary, unilateral, isometric knee extension exercise (UIKEE). To do so, a systematic review was carried out to obtain data for a Bayesian longitudinal model-based meta-analysis (BLMBMA). For the systematic review, PubMed, Web of Science, SCOPUS, Chochrane Library were used as databases. The systematic review included only studies that reported on healthy, young individuals performing UIKEE. Studies utilizing a bilateral training protocol were excluded as the focus of this review lied on unilateral training. Out of the 3,870 studies, which were reviewed, 20 studies fulfilled the selected inclusion criteria. These 20 studies were included in the BLMBMA to investigate the time-dependent effects of UIKEE. If compared to the baseline strength of the trained limb, these data reveal that UKIEE can increase the isometric strength by up to 46%. A meta-analysis based on the last time-point of each available study was employed to support further investigations into UIKEE-induced strength increase. A sensitivity analysis showed that intensity of training (%MVC), fraction of male subjects and the average age of the subject had no significant influence on the strength gain. Convergence of BLMBMA revealed that the peak strength increase is reached after ~4 weeks of UIKEE training.

Modelling the electrical activity of skeletal muscle tissue using a multi-domain approach.

Electromyography (EMG) can be used to study the behaviour of the motor neurons and thus provides insights into the physiology of the central nervous system. However, due to the high complexity of neuromuscular control, EMG signals are challenging to interpret. While the exact knowledge of the excitation patterns of a specific muscle within an in vivo experimental setting remains elusive, simulations allow to systematically investigate EMG signals in a controlled environment. Within this context, simulations can provide virtual EMG data, which, for example, can be used to validate and optimise signal analysis methods that aim to estimate the relationship between EMG signals and the output of motor neuron pools. However, since existing methods, which are employed to compute EMG signals, exhibit deficiencies with respect to the physical model itself as well as with respect to numerical aspects, we propose a novel homogenised continuum model that closely resolves the electro-physiological behaviour of skeletal muscle tissue. The proposed model is based on an extension of the well-established bidomain model and includes a biophysically detailed description of the electrical activity within the tissue, which is due to the depolarisation of the muscle fibre membranes. In contrast to all other published EMG models, which assume that the electrical potential field for each muscle fibre can be calculated independently, the proposed model assumes that the electrical potential in the muscle fibres is coupled to the electrical potential in the extracellular space. We show that the newly proposed model is able to simulate realistic EMG signals and demonstrate the potential to employ the predicted virtual EMG signal in order to evaluate the goodness of automated decomposition algorithms.

Myofascial trigger points alter the modular control during the execution of a reaching task: a pilot study.

Myofascial trigger points (TP) constitute a conundrum in research and clinical practice as their etiopathogenesis is debated. Several studies investigating one or few muscles have shown that both active and latent TP causes an increased muscle activity, however the influence of TP on modular motor control during a reaching task is still unclear. Electromyographic signals, recorded from the muscles of the shoulder girdle and upper arm during a reaching task, were decomposed with Non-Negative Matrix Factorization algorithm. The extracted matrices of motor modules and activation signals were used to label the muscles condition as dominant or non-dominant. The presence of latent and active TP was detected in each muscle with manual examination. Despite a similar muscle activity was observed, we found that muscles with active TP had increased weighting coefficients when labeled in the dominant condition. No influences were found when muscles were in the non-dominant condition. These findings suggest that TP altered the motor control without co-contraction patterns. As a preliminary evidence, the present results suggest that the increased weighting coefficients in presence of TPs are associated with an alteration of the modular motor control without affecting the dimensionality of motor modules for each individual and reciprocal inhibition.

People with low back pain show reduced movement complexity during their most active daily tasks.

BACKGROUND: Actigraphy is a quantitative method for the investigation of human physical activity and is normally based on accelerometric and/or kinematic data. METHODS: A multichannel actigraphy system, able to record both acceleration and spine angles, was employed in this study to measure the quality of movement in 17 individuals with chronic low back pain (LBP) and 18 healthy individuals during unrestricted daily activities. An indication of movement complexity was computed by means of non-negative matrix factorization throughout the 24 hr period and in the 60 min of highest activity. RESULTS: Movement complexity differed only when the 60 min of highest activity was taken into account, with the LBP group showing reduced complexity (e.g., for dimensionality = 8, over 90% of the comparisons showed a significant reduction in the LBP group). CONCLUSIONS: The results are compatible with the hypothesis that pain induces a reduction in the available kinematic trajectories and degrees of freedom during natural movements, which becomes more evident when more demanding tasks are performed. A reduced movement complexity suggests a persistent alteration of the descending neural pathways and/or a disrupted somatosensory information processing, which could be possibly contrasted by administering highly variable motor tasks. SIGNIFICANCE: People with chronic pain move differently. Movement quality is difficult to evaluate during daily activities, yet it may prove more informative than quantitative measurements. We proposed a new approach for computing movement complexity and found out that patients' movements get more stereotyped when higher spinal acceleration is required.

Age-related changes in trunk muscle activity and spinal and lower limb kinematics during gait.

The influence of age on spinal muscle activation patterns and its relation to kinematics is poorly understood. We aimed at understanding age-related changes to spine and trunk muscle activity in addition to spinal and lower limb kinematics during treadmill walking under various conditions. An observational study was conducted evaluating asymptomatic young (n = 10; 3F, 7M; 26.3±2.5yrs) and older (n = 9; 3F, 6M; 67.1±4.2yrs) adults' treadmill walking at 2km/h and 4km/h, each at 0, 1, 5, and 10% inclination. Unilateral (right side) electromyography (EMG) was recorded from deep and superficial multifidus (intramuscular) and erector spinae and abdominal obliques (surface); trunk and leg kinematics were also measured. Muscle activity was characterised by peak amplitude and duration of activity, and the time-point of peak amplitude in the gait cycle (0-100%). Peak activation in older adults was lower for the superficial multifidus (p<0.0001) and higher for the thoracolumbar (p<0.001) and lumbar erector spinae (p<0.01). The duration of activation was longer in older adults for all muscles (p<0.05) except the superficial multifidus, and longer during faster walking for all participants. The time-point of peak amplitude in the gait cycle was earlier in older participants for the external obliques (p<0.05). Walking speed appeared to influence muscle activity more than inclination. Older adults used less spine, trunk and lower limb motion, except at the ankle. Age-related differences within multifidus and between paravertebral and trunk muscles were inconsistent. Walking at 4km/h at 5-10% inclination may specifically target the lumbar paravertebral muscles.

An Efficient Modelling-Simulation-Analysis Workflow to Investigate Stump-Socket Interaction Using Patient-Specific, Three-Dimensional, Continuum-Mechanical, Finite Element Residual Limb Models.

The lack of an efficient modelling-simulation-analysis workflow for creating and utilising detailed subject-specific computational models is one of the key reasons why simulation-based approaches for analysing socket-stump interaction have not yet been successfully established. Herein, we propose a novel and efficient modelling-simulation-analysis workflow that uses commercial software for generating a detailed subject-specific, three-dimensional finite element model of an entire residual limb from Diffusion Tensor MRI images in <20 min. Moreover, to complete the modelling-simulation-analysis workflow, the generated subject-specific residual limb model is used within an implicit dynamic FE simulation of bipedal stance to predict the potential sites of deep tissue injury. For this purpose, a nonlinear hyperelastic, transversely isotropic skeletal muscle constitutive law containing a deep tissue injury model was implemented in LS-DYNA. To demonstrate the feasibility of the entire modelling-simulation-analysis workflow and the fact that detailed, anatomically realistic, multi-muscle models are superior to state-of-the-art, fused-muscle models, an implicit dynamic FE analysis of 2-h bipedal stance is carried out. By analysing the potential volume of damaged muscle tissue after donning an optimally-fitted and a misfitted socket, i.e., a socket whose volume was isotropically shrunk by 10%, we were able to highlight the differences between the detailed individual- and fused-muscle models. The results of the bipedal stance simulation showed that peak stresses in the fused-muscle model were four times lower when compared to the multi-muscle model. The peak interface stress in the individual-muscle model, at the end of bipedal stance analysis, was 2.63 times lower than that in the deep tissues of the stump. At the end of the bipedal stance analysis using the misfitted socket, the fused-muscle model predicted that 7.65% of the residual limb volume was injured, while the detailed-model predicted 16.03%. The proposed approach is not only limited to modelling residual limbs but also has applications in predicting the impact of plastic surgery, for detailed forward-dynamics simulations of normal musculoskeletal systems.

People With Chronic Neck Pain Walk With a Stiffer Spine.

Study Design Controlled laboratory study, case-control design. Objective To evaluate spine kinematics and gait characteristics in people with nonspecific chronic neck pain. Background People with chronic neck pain present with a number of sensorimotor and biomechanical alterations, yet little is known about the influence of neck pain on gait and motions of the spine during gait. Methods People with chronic nonspecific neck pain and age- and sex-matched asymptomatic controls walked on a treadmill at 3 different speeds (self-selected, 3 km/h, and 5 km/h), either with their head in a neutral position or rotated 30°. Tridimensional motion capture was employed to quantify body kinematics. Neck and trunk rotations were derived from the difference between the transverse plane component of the head and thorax and thorax and pelvis angles to provide an indication of neck and trunk rotation during gait. Results Overall, the patient group showed shorter stride length compared to the control group (P<.001). Moreover, the patients with neck pain showed smaller trunk rotations (P<.001), regardless of the condition or speed. The difference in the amount of trunk rotation between groups became larger for the conditions of walking with the head rotated. Conclusion People with chronic neck pain walk with reduced trunk rotation, especially when challenged by walking with their head positioned in rotation. Reduced rotation of the trunk during gait may have long-term consequences on spinal health. J Orthop Sports Phys Ther 2017;47(4):268-277. Epub 3 Feb 2017. doi:10.2519/jospt.2017.6768.

High-density EMG Reveals Novel Evidence of Altered Masseter Muscle Activity During Symmetrical and Asymmetrical Bilateral Jaw Clenching Tasks in People With Chronic Nonspecific Neck Pain.

OBJECTIVES: To characterize the distribution of masseter muscle activity and force control during bilateral jaw clenching tasks in people with chronic nonspecific neck pain, without an associated temporomandibular disorder. METHODS: Twelve volunteers with nonspecific neck pain and 12 age-matched and sex-matched healthy individuals participated. Submaximal symmetrical and asymmetrical bilateral jaw clenching was performed with and without visual feedback of force. Force performance was assessed with indices of accuracy (mean distance, offset error) and precision (standard deviation, coefficient of variation of force). High-density, 2-dimensional, surface electromyography (EMG) was recorded to characterize bilateral masseter muscle activity. The EMG root mean square was computed for each location of the electrode grid to form a map of the EMG amplitude distribution, and the location of the center of activity was measured. RESULTS: The patient group showed a different distribution of masseter muscle activity compared with pain-free individuals during both symmetrical and asymmetrical bilateral jaw clenching. The position of the center of activity was positioned more cranial (P<0.001; right masseter only) and more anteriorly in the patient group (P<0.0001). In addition, the patients with chronic neck pain displayed higher levels of masseter muscle activation compared with the control participants regardless of the specific task performed (P<0.0001). DISCUSSION: People with chronic neck pain display increased activation and altered distribution of masseter muscle activity during a jaw-clenching coordination task. These results provide a greater appreciation of how secondary orofacial pain or temporomandibular disorders may develop in people with neck pain.

Short- and long-term effects of exercise on neck muscle function in cervical radiculopathy: A randomized clinical trial.

OBJECTIVE: To compare short- and long-term changes in neck muscle endurance, electromyography measures of neck muscle activation and fatigue and ratings of fatigue and pain after neck-specific training or physical activity in people with cervical radiculopathy. DESIGN: Randomized clinical trial. SUBJECTS/PATIENTS: Seventy-five patients with cervical radiculopathy. METHODS: Patients underwent neck-specific training in combination with a cognitive behavioural approach or prescribed physical activity over a period of 14 weeks. Immediately after the intervention and 12 months later, surface electromyography was recorded from neck flexor and extensor muscles during neck endurance tests. Time to task failure, amplitude and median frequency of the electromyography signal, and subjective fatigue and pain ratings were analysed in 50 patients who completed at least one follow-up. RESULTS: A significant increase in neck flexor endurance time was observed for both groups at 14 weeks compared with baseline and this was maintained at the 12-month follow-up (p < 0.005). No change was identified for the slope of the median frequency. For the neck-specific training group, splenius capitis was less active during neck flexion at both follow-ups (p < 0.01), indicating reduced muscle co-activation. CONCLUSION: Both specific and general exercise increased neck flexor endurance, but neck-specific training only reduced co-activation of antagonist muscles during sustained neck flexion.

Are regions of the lumbar multifidus differentially activated during walking at varied speed and inclination?

PURPOSE: Lumbar multifidus is a complex muscle with multi-fascicular morphology shown to be differentially controlled in healthy individuals during sagittal-plane motion. The normal behaviour of multifidus muscle regions during walking has only received modest attention in the literature. This study aimed to determine activation patterns for deep and superficial multifidus in young adults during walking at different speeds and inclination. METHODS: This observational cohort study evaluated ten healthy volunteers in their twenties (three women, seven men) as they walked on a treadmill in eight conditions; at 2km/h and 4km/h, each at 0, 1, 5, and 10% inclination. Intramuscular EMG was recorded from the deep and superficial multifidus unilaterally at L5. Activity was characterized by: amplitude of the peak of activation, position of peak within the gait cycle (0-100%), and duration relative to the full gait cycle. RESULTS: Across all conditions superficial multifidus showed higher normalised EMG amplitude (p<0.01); superficial multifidus peak amplitude was 232±115% higher when walking at 4km/h/10%, versus only 172±77% higher for deeper region (p<0.01). The percentage of the gait cycle where peak EMG amplitude was detected did not differ between regions (49±13%). Deep multifidus duration of activation was longer when walking at the faster vs slower speed at all inclinations (p<0.01), which was not evident for superficial multifidus (p<0.05). Thus, a significantly longer activation of deep multifidus was observed compared to superficial multifidus when walking at 4km/h (p<0.05). CONCLUSIONS: Differential activation within lumbar multifidus was shown in young adults during walking. The prolonged, more tonic activation of deep relative to superficial regions of multifidus during gait supports a postural function of deeper fibres.