Measuring transmembrane water exchange can provide potential biomarkers for tumors and brain disorders. Diffusion Magnetic Resonance Imaging (dMRI) is a well-established tool that can non-invasively measure water exchange across cell membranes. Diffusion Exchange Spectroscopy (DEXSY) is one of the dMRI-based frameworks used to estimate exchange. DEXSY provides a detailed picture of multi-site exchange processes but requires a large quantity of data. Several models based on the DEXSY framework have been proposed to reduce the acquisition time. Filter Exchange Imaging (FEXI) and curvature models are two of them that only require certain samples of the DEXSY dataset. Diffusion-Exchange Weighted (DEW) Imaging model is another data reduction method accounting for restricted diffusion within cells and can use a specific subset of the DEXSY dataset to measure exchange. Furthermore, a more general expression of the DEXSY signal, referred to as the general model, can theoretically analyze the full space or reduced DEXSY datasets and estimate exchange. However, the results of the subsampling schemes and the data reduction models have not been compared to the full space estimation.  Therefore, this thesis aims to experimentally explore the feasibility of estimating exchange using these four models (the general, FEXI, curvature and DEW models) with the data acquired using a low-field benchtop MR scanner, and compare the estimates from the general model with different subsampling schemes and the data reduction models to the full space estimation. For this purpose, a double diffusion encoding (DDE) sequence was modified from an existing sequence on the benchtop MR scanner and a DEXSY experiment was conducted on this MR scanner and a yeast phantom to acquire a full space dataset. The exchange parameters estimated from the full space dataset using the general model were used as "ground truths" to evaluate the estimates from the reduced datasets analyzed using the general, FEXI and curvature models. Moreover, two alternative subsampling schemes named the shifted DEW and new trajectory schemes were proposed and employed to measure exchange. The results indicate that all the methods except the curvature sampling scheme employed with both the general and curvature models provided comparable estimates to the "ground truths". The shifted DEW and new trajectory sampling schemes performed better over others in terms of consistency with the "ground truths" and low variations between voxels, suggesting the theoretical and experimental optimization of these two subsampling schemes can be further studied and developed.

The knowledge of blood vessel biomechanics is used for understanding and developing treatments for cardiovascular disease. The objective of this project was to develop an experimental protocol, for education and research, that separates active diameter response from the passive, as a function of the intraluminal pressure in a pressure myograph. The design process was performed in three steps. First the protocol was designed for an artificial vessel and then expanded to include passive properties of blood vessels, finally further developments needed to analyze active blood vessels were suggested. The system was built as a pressure myograph containing a vessel chamber where the vessel was mounted on two cannulas, two pressure sensors for calculating the intraluminal pressure, and one microscope equipped with a camera for diameter observations. Reference data for the artificial vessel material was acquired from a uniaxial tensile test. The results was in the form of stress-stretch relations. Both the results from the artificial vessel and the passive blood vessel was in a acceptable reference range. The results indicate that the experimental protocol can be used for testing passive properties of both artificial vessels and small blood vessels. No results were obtained for active blood vessels, thus the experimental protocol can not be used for separating the active response to diameter change of blood vessels. However, further developments of the experimental protocol are discussed. / Kunskap om blodkärlens biomekanik används för att förstå och utveckla behandlingsmetoder mot hjärt- och kärlsjukdomar. Syftet med det här projektet var att utveckla ett experiment som mäter små blodkärls diameter och vätsketryck in vitro. Experimentet skulle sedan separera det aktiva muskelbidraget till diameterförändring från det passiva bidraget. Genom att göra detta kan kunskapen om biomekaniken hos blodkärl utvecklas inom såväl forskning som utbildning. Experimentet utvecklades i tre steg. Först utvecklades det för artificiella blodkärl och anpassades sedan för passiva blodkärl. Slutligen diskuteras vidareutvecklingar av experimentet gällande de aktiva egenskaperna för blodkärl. Experimentet utformades genom att ett kärl placerades i en kammare. Blodkärlet fästes i vardera ände på varsin kanyl som var kopplade till trycksensorer. För att registrera diametern placerades kammaren under ett kameramikroskop. Resultatet består av spänning-sträcknings diagram. Både resultatet för det artificiella blodkärlet och det passiva blodkärlet var inom ett godkänt referensintervall. Dessa resultat stödjer antagandet att experimentet kan användas för att studera passiva egenskaper av artificiella och verkliga blodkärl med storlek mellan 1.9-4.4 mm i ytterdiameter. Ingen mätdata från aktiva blodkärl kunde samlas in, utan utvecklingskrav på systemet för hantering av aktiva blodkärl har föreslagits.

Musculoskeletal models and simulations allow for the estimation of forces acting on muscles and joints during human movement and athletic performance. In order to improve the accuracy of these models for a specific application, knowledge about subject-specific in vivo properties of human muscle and tendon is needed. This study presents a method for estimating in vivo mechanical properties of human tendon in the lower leg, using a combination of ultrasound imaging and motion capture. Key mechanical parameters — such as tendon stiffness, moment arm, slack length and force-strain relationship — and the contribution of tendon elongation to ankle mobility of the medial gastrocnemius (MG) and soleus (SOL) aspects of the Achilles tendon were obtained in vivo in 8 typically-developed adults, and the applicability of the method on the tibialis anterior (TA) tendon was investigated. In contrast to previous studies using a comparable method, variable tendon moment arm lengths during passive movement of the ankle joint was taken into consideration. As a novelty, the passive mechanical properties of the Achilles tendon were obtained in vivo in 4 hemiplegic post-stroke subjects and compared to the 8 typically-developed subjects. The estimated mechanical parameters of the MG and SOL aspects of the Achilles tendon were consistent with findings in the literature. In order to estimate stiffness of the TA tendon, it was shown that a larger range of motion (ROM) of the foot during the passive rotation experiments is needed. The comparison between typically-developed and hemiplegic post-stroke subjects revealed significantly lower tendon stiffness and slack angle, and significantly higher contribution of tendon elongation to ankle mobility in the post-stroke group. The developed method enables estimation of in vivo mechanical properties of tendon in the lower leg and contributes to improving the accuracy of subject-specific musculoskeletal models and simulations. / Person-specific biomechanical models

Factors affecting the risk of ribcage injury in adult subjects during CPR were investigated using the torso region of the THUMS model, a full human body FE-model, representing an average adult male. The thoracic dynamic response of the model was compared to experimental PMHS hub loading impact data and live-subject CPR data found in the literature. The model was then used to study the risk of obtaining injuries in various simulated CPR conditions, also varying the stiffness of the costal cartilage. Parameters that are known to predict induced injuries were extracted from the model simulations, i.e. chest deflections, and maximum 1st principal strain and von-Mises stress in the ribs and sternum, as well as the pressure in the heart muscle. These were compared with values that have been reported to have the potential to cause injury. The predictions were compared to experimental findings of the probability of CPR resulting in fractures of the ribs and sternum. The previously mentioned parameters did not reach high enough values to predict fracture occurrences, but interesting trends were highlighted with regards to the different loading conditions investigated. It was demonstrated that human body FE-model simulation studies can be useful for investigating the influence of different CPR related loading conditions on the risk of occurrences of rib and sternal fractures.

The aim of this research was to produce a reliable finite element model of a helmet, that could be used to simulate approval tests as well as impacts to investigate the safety offered. A 2D and 3D mesh was generated from the CAD file of an Easton Synergy 380 with HyperWorks, and then checked referring to standard parameter values. A few specimens cut from the liner were tested with the Instron Electropuls E3000 (Instron, High Wycombe, Great Britain) machine to determine Young’s modulus, Poisson’s ratio and the density of the EPP. The numerical model was characterised with appropriate materials with Ls-PrePost, such as ABS for the shell, EPP for the liner and steel for the impact anvil. The foam was implemented both with the *063_CRUSHABLE_FOAM and the *126_MODIFIED_HONEYCOMB card, in two different configurations. The helmet model was coupled with a finite element model of a HIII head form and three impact scenarios were set up. Backward, lateral and pitched impact were simulated and results were compared with those obtained from the experimental tests carried on at the MIPS. The two configurations were tested in all the three scenarios. The correlation between numerical and experimental results was evaluated by analysing the linear and rotational acceleration, and the rotational velocity, recorded by the accelerometer positioned inside the HIII headform. The parameters used were the Pearson correlation coefficient, the peak linear acceleration score, the shape of the curves, the time occurrence of peaks and the percentage of the difference between them. The first configuration showed good correlation scores (>85%) for the backward and lateral impact, for the rotational velocity and acceleration, while lower values were recorded for the pitched impact simulation. Lower values (70.88% and 77.76%) were obtained for the peak linear acceleration score, which stress the need for modifications of the contact definition in Ls-PrePost or a more detailed material testing. Worse results were recorded for the second configuration, but the smaller computational time required suggests that more attempts should be done in this direction.

WebAssembly is a new binary code specification, which was initially designed to complement JavaScript in web applications. WebAssembly is inherently portable and small, designed for multiplatform usage. Therefore, WebAssembly modules can be created to support embedded system-based biomedical sensor operation. However, WebAssembly has its own limitations to compensate with its portability. In this thesis, we show how WebAssembly modules can be applied to the basic biomedical modalities of body temperature, heart rate, and breathing pattern.  We show how the implementation performed, and what challenges were met during the development. It is concluded that WebAssembly can be applied for achieving safe and effective biomedical sensor devices, although with some limitations. / WebAssembly är ett nytt binärt maskinkodsformat, ursprungligen skapat för att komplettera JavaScript i webbapplikationer.  WebAssemblys kod är liten och kan lätt användas på flera plattformar. Därför kan WebAssembly-moduler skapas för att stödja inbyggda system för biomedicinska sensorer. WebAssembly har dock sina egna begränsningar på grund av sin portabilitet.  I denna avhandling visar vi hur WebAssembly-moduler kan användas på enkla biomedicinska mätningar av kroppstemperatur, hjärtfrekvens och andningsmönster. Vi visar hur implementeringen genomfördes och vilka utmaningar som möttes under utvecklingen. Slutsatsen är att WebAssembly kan tillämpas för att skapa säkra och effektiva biomedicinska sensorenheter, även om det finns en del begränsningar.

As upper-limb rehabilitation devices development continue to grow, lower limb rehabilitation devices have limited development. Thus a device that measures weight bearing of heel and forefoot during gait and also produces auditory feedback at certain thresholds were made in this project. The device was constructed based on an ARM-Core microcontroller and was supported by 5 force sensing resistor sensors for each foot. Communications between the modules and the computer as the GUI were established using an UART interface. For testing, sensors were placed beneath the foot using a stretchable cloth so each different subjects’ pressure points were correctly placed. The biofeedback used for this device was by enabling a buzzer and LED embedded to the designed board that turns on when heel strike or push off reached acertain threshold. Testing were divided into phase one and two. From phase one, it was concluded that both push off and heel strike event can have a reading exceeding 8.5% of the body weight. It was then used as a threshold for the biofeedback event. During the biofeedback tests, improvements in the subjects’ walking pattern and reading were noted. Results suggest that biofeedback can change the behavior of the subject. / I takt med att utvecklingen av rehabilitationsapparater i övre extremiteterna fortsätter att växa, har rehabiliteringsanordningar för nedre extremiteterna begränsad utveckling. Således gjordes en anordning som mäter viktbearbetning av häl och framfot under gång och också ger hörselåterkoppling vid vissa trösklar i detta projekt. Enheten var konstruerad baserad på en ARM-Core mikrokontroller och stöds av 5 kraftsavkännande motståndssensorer för varje fot. Kommunikation mellan modulerna och datorn som GUI upprättades medhjälp av ett UART-gränssnitt. För testning placerades sensorer under foten meden töjbar duk så att varje olika försöks tryckpunkter placerades korrekt. Den biofeedback som användes för den här enheten var genom att aktivera en summer och LED inbäddad i det konstruerade kortet som slås på när hälslaget eller skjutstoppet nådde en viss tröskel. Testningen delades in i fas en och två. Frånfas ett drogs slutsatsen att både push off och hälstreffhändelse kan ha en avläsningsom överstiger 8.5% av kroppsvikt. Den användes sedan som en tröskelför biofeedback-händelsen. Under biofeedback-testen noterades förbättringar i försökspersonernas gångmönster och läsning. Resultaten tyder på att biofeedback kan förändra beteendet hos ämnet.

Car collisions are, unfortunately, not uncommon and cause 1.35 million deaths each year worldwide. Children are often occupants in cars and to ensure their safety, child restraint systems (CRSs) have been developed. However, CRSs need to be used correctly to be efficient. Several studies, such as field investigations and Q-dummy tests, have shown that a misuse of a CRS can increase the risk of injuries. Typical misuses for a forward-facing CRS and a booster seat, with two real accident parameters, were constructed and simulated using the PIPER child human body model. The kinematics of each case were compared with injury parameters of the head, neck and abdomen. Comparing the parameters to existing injury criteria showed that most of the cases end in AIS3+ head injury, even cases with no misuse. When comparing the results of misuses to the cases where the CRS was correctly used, the dominant result was that misuse resulted in being less effective to protect the child. Moreover, results of chosen misuses compared to Q-dummy tests correlated with their results. Results from this thesis illustrate how important it is for parents to restrain children and route the belt correctly.

In Sweden fall-related injuries among the elderly lead to over a thousand deaths and close to 300.0000 hospital visitations annually. Fall related injuries can in many cases lead to serious head injuries along with other fractures. The elderly are more prone to fall and sustain an injury because of conditions such as osteoporosis and weak muscles. Researchers at KTH from the Division of Neuronic Engineering have developed a special floor that absorbs energy from an impact which can reduce the severity of injury when a person falls. There has been research done on the floor focusing on hip fractures which have shown good results. The floor has been set up in an elderly home in Stureby, Stockholm, Sweden, however further testing is needed to fully develop the floor to its maximum capacity. One goal of this master thesis was to investigate certain obstacles with the floor which arose during testing in Stureby. One of the challenges was that the floor had greater rolling resistance since it is softer than a normal concrete floor. The experiment was done by using a dynamo meter to measure the force needed to move a test wagon with a fixed weight and different sizes of wheels. Another goal of this thesis was to investigate head impacts on the KTH floor. That was done in a helmet testing lab at Mips helmet company. There a dummy head was dropped from various heights and the acceleration was measured. The 1st principle strain of the brain was calculated from the collected data. Other companies have created similar floors so it was a part of the thesis to compare the KTH floor with its competitors. The results from the rolling resistance showed that by increasing the width and diameter of the wheels there was less force needed to move the test wagon. Further tests are needed to investigate the problem regarding the indentation issues seen on the floor set up in Stureby. Some research has already begun with different types of linoleum and glue. In the head impact studies, the KTH floor showed the best results of all tested floors when dropped from 60 cm. However, when dropped from 100 cm a competitor floor showed better results in the resultant translational acceleration.

MRI is favorable for brain imaging due to its excellent soft tissue contrast and absence of harmful ionizing radiation. Many have proposed supervised multimodal neural networks for automatic brain tumor segmentation and showed promising results. However, they rely on large amounts of labeled data to generalize well. The trained network is also highly specific to the task and input. Missing inputs will most likely have a detrimental effect on the network’s predictions, if it works at all. The aim of this thesis work is to implement a deep neural network that learns the general representation of multimodal MRI images in an unsupervised manner and is insensitive to missing modalities. With the latent representation, labeled data are then used for brain tumor segmentation. A variational autoencoder and an unified representation network are used for repre- sentation learning. Fine-tuning or joint training was used for segmentation task. The performances of the algorithms at the reconstruction task was evaluated using the mean- squared error and at the segmentation task using the Dice coefficient. Both networks demonstrated the possibility in learning brain MR representations, but the unified representation network was more successful at the segmentation task.