Various source localization algorithms exist to perform localization with High Density (HD)-ElectroEncephaloGraphy (EEG). However, validation of these EEG source localization algorithms is lacking. The current gold standard for source localization in the brain is functional Magnetic Resonance Imaging (fMRI) by calculating the difference in hemodynamic response to different stimuli. This study aims to validate HD-EEG source localization with fMRI using an MR compatible haptic robot. Participants performed several tasks with the robot to activate cortex patches and allow localization of source activity under various circumstances. These hypothesized patches are the somatosensory, motor and visual cortex. No comparison is made with fMRI due to time constraints. As no comparison could be made, we aim to validate the experimental methodology.
Activation of the somatosensory cortex is clearly visible. The visual cortex is often localized, but lacks power in some settings. Activity during the torque task can be localized, but not conclusively to the motor cortex. In all, the experiment was a success, as it was able to induce verifiable different brain states. Hypothesized task contrasts contained different activity distributions. Improvements can be made by generating a more detailed leadfield and by applying a linear manipulator

Problem: A biomarker that accurately predicts recovery of ischemic stroke for patients with poor baseline Fugl-Meyer Assessment of the Upper Extremity (FMA-UE) is lacking. Biomarkers that predict recovery while providing functional insight into the underlying neural process are highly desired for optimal clinical care. Objectives: This explorative study aims to determine potential biomarkers from Somatosensory Evoked Potentials (SEP) using ElectroEncephaloGraphy (EEG). Brain asymmetry metrics are derived from high-density EEG recordings of five longitudinal stroke patients during the first six months of stroke recovery. In addition, EEG sources and their interactions, constrained by anatomical information from (d)MRI, are explored for a potential biomarker. Results: Subjects with low baseline FMA-UE show a trend of increased recovery as the Laterality Index (LI) of the infarcted hemisphere increases, although it is not significant within this small group. The LI of the non-infarcted hemisphere shows a significant trend of high LI values at baseline that decrease to lower values during recovery. This effect was tested on a different dataset of 17 longitudinal stroke patients (without dMRI data). The results in this second group showed strong variability between subjects and measurements. The non-infarcted hemisphere is able to significantly predict the FMA-UE for both datasets. On the other hand, estimated effects for the LI of the infarcted hemisphere did not show significant values. Biomarkers were not derived from EEG sources and their interactions. Conclusion: In this study, we have demonstrated the potential of the LI as a biomarker for stroke recovery. Small sample size and absence of controls make hard clinical conclusions impossible. However, our findings show that the LI might predict recovery of the FMA-UE, even at low baseline FMA-UE. The LI of the non-infarcted hemisphere is able to significantly predict the FMA-UE of the affected hand, a surprising discovery that might be attributed to background cortical activity or ipsilateral SEP components. Derivations from EEG sources and their interactions might lead to more sensitive metrics and novel insights into stroke rehabilitation. Significance: If the discovered trends apply to a larger sample size, asymmetry metrics from SEP’s have prognostic value in stroke recovery.

The company invented a product that can monitor vulnerable people. It is important that they know exactly where their products are located. If a situation occurred where a vulnerable person needs immediate help, first responders can use the knowledge of the location to get to the patient as fast as possible. This thesis investigates if a modular localization solution can be designed for this product. The product will always be located in an indoor environment, which is a difficult place to establish the location. A modular solution is interpreted as a system that has parts that are replaceable. Some parts can be switched out for different parts with the same functionality, with the system still working properly. In order to design this system according to the requirements, the design process was divided in hardware and software. Both were designed in a modular way, yet eventually implemented in one specific way. This was done to get a tangible prototype that could be evaluated and tested. Improvements to the system were also designed, however could not be implemented in time. Nonetheless, the resulting system works well under conditions of intended use. There were some conditions in which the system performed less than required, however this should be fixed once the improvements are implemented. The results indicate that the localization system that is needed for the company's product is possible to be implemented. The next step is to research whether parts of the modular system can be replaced by an improvement. Then the system can be tested in more distinct working conditions, which could not be done during this project. If the system passes these tests, it is ready to be integrated with the other subsystems of the product, such that the company can bring their product to the market.

This report describes the implementation of a custom streaming solution from an IP camera to a web browser. The system aims to both provide live video and Video on Demand. This will be used to monitor premature babies in incubators.