Understanding of human motor control is central to neuroscience with strong implications in the fields of medicine, robotics and evolution. It is thus surprising that the vast majority of motor control studies have focussed on human movement in the laboratory while neglecting behaviour in natural environments. We developed an experimental paradigm to quantify human behaviour in high resolution over extended periods of time in ecologically relevant environments. This allows us to discover novel insights and contradictory evidence to well-established findings obtained in controlled laboratory conditions. Using our data, we map the statistics of natural human movement and their variability between people. The variability and complexity of the data recorded in these settings required us to develop new tools to extract meaningful information in an objective, data-driven fashion. Moving from descriptive statistics to structure, we identify stable structures of movement coordination, particularly within the arm-hand area. Combining our data with numerous published findings, we argue that current hypotheses that the brain simplifies motor control problems by dimensionality reduction are too reductionist. We propose an alternative hypothesis derived from sparse coding theory, a concept which has been successfully applied to the sensory system. To investigate this idea, we develop an algorithm for unsupervised identification of sparse structures in natural movement data. Our method outperforms state-of-the-art algorithms for accuracy and data-efficiency. Applying this method to hand data reveals a dictionary of \emph{sparse eigenmotions} (SEMs) which are well preserved across multiple subjects. These are highly efficient and invariant representation of natural movement, and suggest a potential higher-order grammatical structure or ''movement language''. Our findings make a number of testable predictions about neural coding of movement in the cortex. This has direct consequences for advancing research on dextrous prosthetics and robotics, and has profound implications for our understanding of how the brain controls our body.

Along with playing a vital role in immune processes, the lymphatic system is a vascular network that transports extravasated fluid from the interstitial spaces to the venous return and helps maintain fluid homeostasis within the body. Its dysfunction can result in lymphoedema, or swelling of fluid within the interstitial spaces due to lymphatic impairment. There is a clear and present lack of knowledge of the pumping and transport mechanisms of the lymphatics. If physicians and scientists had a better understanding of these, more effective treatments for lymphoedema and other diseases of the lymphatic system could be derived. Bi-leaflet valves are frequent throughout the lymphatic vessels and serve to prevent retrograde flow. These valves play a vital role in fluid transport and pumping within the lymphatic system, however they have been largely understudied. Through a series of computational studies, the transport of nitric oxide (NO) within the lymphatic valve, the geometrical effect of the leaflets on the valve fluid dynamics and the interaction between the leaflets and lymphatic fluid have been studied. Computational investigations of NO transport indicated high levels of NO within the valve region as a result of flow stagnation. While the valves serve to prevent retrograde flow, they also add resistance, which is balanced out by the expansion of the bulbous sinus that encapsulates the leaflets. Consistent with experimental findings, the valve showed hysteretic opening and closing resistance values. These results strongly suggest the vital role of the valve within the lymphatic network of vessels. The research documented herein is the first recorded series of three-dimensional computational studies of flow and transport within the lymphatic valve and significantly contributes to the body of knowledge of the lymphatic vasculature.

Heterotopic ossification (HO) is the formation of mature lamellar bone in tissues that are not usually osseous. It is a significant risk in blast-related injuries and is seen in approximately 60% of UK military blast-related amputees. HO can result in reduced range of motion, pain, nerve impingement and can affect prosthesis fitting. The causes of HO are due to a mix of mechanical and biological, local and systemic factors. However, as with normal bone formation and remodelling, it may be expected that heterotopic bone also responds to mechanical stimuli to an extent. Understanding this relationship further can give insight into possible ways to manipulate the progression of HO and prevent complications in the future. The objective of this research is to create a mechanically driven physiological computational model of HO in the residual limb of a trans-femoral amputee. The current work involved expanding upon previously proposed finite element bone remodelling algorithms in the literature for the application of heterotopic remodelling in soft tissue. This study introduced an extra factor to represent the tendency for soft tissue to calcify. This factor increases in magnitude with proximity to a specified wound site and with tissue strain. Initially, soft tissue is modelled as hypoelastic with a much lower stiffness to bone. If the density in a soft tissue element exceeds a certain threshold, the material properties of the element are redefined to become proportional to the density alike bone. These newly recruited bone elements make up the projected heterotopic bone geometry. The different parameters within the algorithm were adjusted to examine their effect on the final outcome of heterotopic bone geometry. With consideration to their effects, loading was found to significantly alter the geometry of HO. Certain characteristic appearances of HO outlined in the literature were reproduced by adjusting the loading environment.

Physical forces regulate cellular behaviour and function during all stages of life. Mechanotransduction, the process by which cells convert mechanical stimuli into biochemical signalling events is central to a number of physiological and pathological processes. The first part of this work focuses on the effect of retinoid therapy on the mechanobiology of pancreatic cancer. Pancreatic cancer is characterised by a persistent activation of stromal fibroblasts, known as pancreatic stellate cells (PSCs), which can perturb the biomechanical homeostasis of the tumour microenvironment to favour tumour invasion. Using biophysical and biological techniques, we report that all-trans retinoic acid (ATRA), an active vitamin A metabolite restores mechanical quiescence in PSCs via a mechanism involving a retinoic acid receptor beta (RAR-β)-mediated downregulation of actomyosin (MLC-2) contracility. We show that ATRA reduces the ability of PSCs to generate high traction forces and adapt to extracellular mechanical cues (mechanosensing), as well as suppresses force-mediated extracellular matrix remodelling to inhibit local cancer cell invasion in 3D organotypic models. We thus suggest that ATRA may serve as a stroma reprogramming agent for the treatment of pancreatic cancer. In the second part of this work, we focus on syndecan-4 (Syn-4) - a ubiquitous transmembrane proteoglycan receptor. We identify Syn-4 as a cellular mechanotransducer that tunes cell mechanics by eliciting a global mechanosignalling response. We outline a mechanotransduction model whereby localised tension on Syn-4 triggers a synergistic cell-wide activation of β1 integrins, in a PI3K-dependent manner, to subsequently activate the RhoA pathway and induce adaptive cell stiffening. Furthermore, syndecan-4 mediated mechanosensing is required for YAP activation and downstream changes in gene expression. We propose that this newly identified mechanotransductive ability of Syn-4 should have direct implications for the field of mechanobiology.

This thesis focuses on designing and developing both wired and wireless multimodal clinical monitoring electronic instruments for traumatic brain injury (TBI) patients. Potentiostats using both a transimpedance resistive design and a novel switched-capacitor based design have been developed to support amperometric sensors and biosensors in detecting concentrations of important biomolecules, such as glucose and lactate in the injured brain tissue. The performance and relevant advantages/disadvantages of these two potentiostat designs have been also discussed. A high-resolution low-noise bio-potential acquisition system has been developed, to monitor the electroencephalography and electrocorticography of TBI patients. To optimise the monitoring performance, the system has been separated into two functional parts: a battery-powered head-stage box that records, amplifies, and digitalises the biopotentials; a bed-side digital unit that integrates with a new patient monitoring instrument. A commercial device has been used as a comparison, and the new design shows better electrical interference elimination. I have also designed an intelligent multimodal TBI instrumentation box to simultaneously monitor chemical, electrical, and physical parameters from a TBI patient. The instrumentation box has been carefully designed and evaluated to meet IEC60601 (3rd edition) safety regulation and it has passed local hospital test for clinical uses. The thesis then concentrates on developing different wireless solutions for the TBI instruments, and aims at eliminating clinical cable congestion and practical issues in intensive care units. By integrating with TBI-related electronics, different wireless modules have been evaluated and compared. Finally, to enable the device to display real-time tissue concentrations to the clinical team, a wireless control of a sensor auto-calibration system has been developed. Relevant graphic user interfaces have been designed on both PC and android-based smart phones or tablets.

Glaucoma is the leading cause of irreversible blindness worldwide, and is characterized by elevated intraocular pressure (IOP) caused by increased resistance to aqueous humor outflow. The majority of outflow resistance is generated near the inner wall endothelium of Schlemm’s canal (SC). The inner wall experiences a basal-to-apical directed flow as aqueous humor crosses the outflow pathway. As IOP increases, the outflow pathway responds in a pressure-dependent manner, resulting in the expansion of the trabecular meshwork (TM) and the collapse of SC. This effectively reduces the cross-sectional area of the SC lumen and increases the shear stress experienced by SC cells, reaching levels known to activate endothelial nitric oxide synthase (eNOS) in vascular endothelia. Our central hypothesis examines the role of eNOS as part of a dynamic mechano-regulatory feedback system to regulate the outflow resistance sites through nitric oxide (NO) production to maintain IOP homeostasis. We firstly demonstrated the physiological role of NO and eNOS in regulating aqueous humor outflow through the use of NO-donor and NOS-inhibitors. We also demonstrated that spatial variations in eNOS expression in the SC correlates with regions of greater outflow in the TM. Furthermore, we developed NO-sensitive biosensors to detect changes in NO production in response to elevated IOP, showing that NO production was pressure-dependent. Finally, we demonstrated that targeted delivery of NO to the outflow resistance sites in the TM results in a ~3-fold increase in outflow facility. Taken together, these studies reveal that eNOS plays a crucial regulatory role in conventional outflow physiology by modulating outflow resistance through NO production. This mechano-regulatory feedback mechanism appears to be altered in glaucoma, and thus leads to ocular hypertension and pathogenesis of the disease. Therefore, targeting the NO-regulatory machinery within the outflow pathway may provide a promising therapeutic target for treating glaucoma.

Total shoulder arthroplasty (TSA) is used to treat patients with glenohumeral joint osteoarthritis. Despite high success rates, complications occur and many patients undergo revision surgeries. The top three most frequent complications are: instability/dislocation of the joint, glenoid loosening and rotator cuff insufficiency. It is believed that these complications occur in most part due to alignment of the implant. In this thesis a two part investigation was conducted to test the effects of joint replacement humeral head version (-15° to 15°) and tilt (-10° to 10°) and glenoid version and tilt (-15° to 15°). The first part investigates the effect of the humeral head alignment on range of motion and activities of daily living using a collision detection modelling method. The second part investigates the effect of both humeral head and glenoid variations on the joint reaction and muscle forces to describe the risk of the three most frequent complications using the United Kingdom National Shoulder Model. This thesis shows that increasing humeral head posterior version decreased the ability to perform activities of daily living (up to 32% at 15°) mostly due to bone-implant collision, increased the risk of the rocking-horse mechanism (by up to 37% at 15°) and increased subscapularis activity (by up to 14% at 15°). Similarly, increasing inferior tilt of the glenoid to 10° produced the best outcomes: vertical rocking-horse mechanism decreased by 19% and no significant differences in muscle forces were observed. In conclusion, normal alignment of the humeral head following surgical guidelines is recommended to increase the chances of implant survival. Posterior versions of the humeral head should be avoided more so than other small mal-alignments. Increasing glenoid inferior tilt to 10° produced favourable results but after combining the results from this thesis and from the literature, it is concluded that all glenoid mal-alignments should be avoided and highlights the need for more effective surgical tools to accurately position the shoulder replacement.

This thesis is a monograph on Markov chains and deterministic approximation schemes that enable the quantitative analysis thereof. We present schemes that yield approximations of the time-varying law of a Markov chain, of its stationary distributions, and of the exit distributions and occupation measures associated with its exit times. In practice, our schemes reduce to solving systems of linear ordinary differential equations, linear programs, and semidefinite pro- grams. We focus on the theoretical aspects of these schemes, proving convergence and providing computable error bounds for most of them. To a lesser extent, we study their practical use, applying them to a variety of examples and discussing the numerical issues that arise in their implementation.

Microbubble-mediated focused ultrasound therapies such as blood-brain barrier opening, sonothrombolysis, and sonoporation can non-invasively deliver drugs to a targeted region. Despite the potential impact this technology could have in the clinic, there are currently concerns regarding its efficacy, uniformity and safety. These challenges ultimately stem from a limited ability to control the microbubble dynamics during ultrasound exposure. In the current thesis, we sought to design ultrasound exposure sequences and develop monitoring techniques that promote the desired acoustic cavitation activity and suppress unwanted stimuli that do not produce a safe therapeutic bioeffect. For example, violent cavitation activity could cause irreversible damage within the treatment area. The behaviour of microbubble populations exposed to low-power therapeutic ultrasound was first qualitatively studied using high-speed microscopy. Microbubbles were found to form large clusters within milliseconds of exposure and collectively coalesce into larger bubbles. Based on these observations and findings in the literature, new therapeutic sequences were designed and tested. Rapid short-pulse sonication consisted of μs-long pulses separated by short off-times. When compared to conventional ultrasound sequences, this pulse sequence enhanced the lifetime and mobility of cavitation nuclei and resulted in more uniform acoustic activity distributions. To better monitor ultrasound treatment as it evolves, we developed a method that passively measures microbubble velocities via the Doppler effect emerging in the microbubble acoustic emissions. Using standard passive cavitation detection techniques in one and two dimensions, we estimated microbubble velocities on the order of m s-1 during ultrasound exposure. Finally, we tested our new therapeutic design in a mouse model in order to improve the safety of blood-brain barrier opening. We achieved drug delivery with a similar magnitude but with a better uniformity compared to conventional sequences, thus demonstrating evidence of favourable microbubble dynamics within the targeted region. Taken together, our contribution in ultrasonic stimulation using new sequences and monitoring using passive acoustic detection techniques improves our control of microbubble dynamics in ultrasound therapy and has the potential to promote treatment efficacy and suppress unwanted damage.

Understanding the functional principles of the mammalian cortical circuit is a major challenge in neuroscience. To make progress towards this understanding, one needs to be able to assess the behavioural dynamics of individual neuronal elements of this circuit. Manual whole-cell recording (WCR) in vivo is recognised as the “gold standard” method for electrophysiological interrogation of individual neurons. It allows subthreshold and suprathreshold signals to be recorded, perturbations to be applied through current injection, and DNA vectors to be directly delivered into the patched cells as part of the pipette internal solution. Unfortunately, the WCR technique for in vivo application is a “blind” procedure and has a low-throughput. In addition, the genetic and morphological identity of the recorded neuron often cannot be accounted for. The inherent cell-type non selectivity of this technique can be overcome by combining WCR with two-photon laser scanning microscopy, and targeting recordings to specifically labelled individual cells or cell classes. Targeted electrophysiological interrogation allows one to examine the properties of both single cells and neuronal assemblies and, additionally, the role of cell type-specific proteins in orchestrating neuronal responses. Targeted recordings in vivo may enable to test a wide range of hypotheses related to information processing in the cortical circuits. However, probing and studying the properties of individual cells in live animal preparations remains a challenge in neuroscience. In particular, precise vision-guided control of patch pipette motion and viewpoint generation of microscope objective for targeted single-cell electrophysiological interrogation is problematic. It requires specialised skills acquired through extensive practice and training by individual operators. Although automatic patch clamp technology has been in use for some years exclusively for cell culture-based paradigms, only recently has Kodandaramaiah et al. demonstrated a “blind” automated patch clamp system for in vivo recordings. However, a fully automatic method for in vivo WCR targeting specific cells or cell-types has not been implemented in any robotic system so far. In this study an automated two-photon targeted whole-cell patch clamping algorithm is demonstrated as a workable solution. The aim of this work was to develop a robotic integrated targeted autopatcher that minimised labour intensive procedures and increased the throughput both in blind and two-photon targeted WCR in live animals. The system automatically controls a micromanipulator, a microelectrode signal amplifier a two-photon microscope and a custom made regulator for controlling the internal pressure of the pipette. The two-photon microscope acquires images of fluorescently labelled cells, and cell-targets for patch clamp are selected via a point-and-click graphical user interface. Optical coordinates are initially converted to the micromanipulator coordinate system and a suitable path calculated to guide the patch pipette towards the target. This platform allows to compensate for brain tissue deformation and subsequent neuronal target movement caused by pipette insertion. As proof-of-concept, the system was tested in both “blind” and two-photon targeted paradigms and achieved performances comparable to human operators, in terms of yield, recording quality and operational speed. Hit rate for “blind” WCR was 51.4% (n=18, 35 attempts across 5 mice). RITA was also calibrated and tested for targeting specific cell types in the cortex of intact mouse brain labelled via fluorescent dye loading (e.g. Oregon Green BAPTA-1 and/or Sulforhodamine 101). Pipettes were automatically guided to the target cells and recordings obtained from visually identified neurons as well as astrocytes. These results prove the feasibility of robotic targeted WCR patch clamp in vivo and establish this system as a powerful tool for automated electrophysiological experiments in the brain.