Search

Near-infrared spectroscopy (NIRS) is a non-invasive and easy-to-use diagnostic technique that enables real-time tissue oxygenation measurements applied in various contexts and for different purposes. Continuous monitoring with NIRS of brain oxygenation, for example, in neonatal intensive care units (NICUs), is essential to prevent lifelong disabilities in newborns. Moreover, NIRS can be applied to observe brain activity associated with hemodynamic changes in blood flow due to neurovascular coupling. In the latter case, NIRS contributes to studying cognitive processes allowing to conduct experiments in natural and socially interactive contexts of everyday life. However, it is essential to measure systemic physiology and NIRS signals concurrently. The combination of brain and body signals enables to build sophisticated systems that, for example, reduce the false alarms that occur in NICUs. Furthermore, since fNIRS signals are influenced by systemic physiology, it is essential to understand how the latter impacts brain signals in functional studies. There is an interesting brain body coupling that has rarely been investigated yet. To take full advantage of these brain and body data, the aim of this thesis was to develop novel approaches to analyze these biosignals to extract the information and identify new patterns, to solve different research or clinical questions. For this the development of new methodological approaches and sophisticated data analysis is necessary, because often the identification of these patterns is challenging or not possible with traditional methods. In such cases, automatic machine learning (ML) techniques are beneficial. The first contribution of this work was to assess the known systemic physiology augmented (f)NIRS approach for clinical use and in everyday life. Based on physiological and NIRS signals of preterm infants, an ML-based classification system has been realized, able to reduce the false alarms in NICUs by providing a high sensitivity rate. In addition, the SPA-fNIRS approach was further applied in adults during a breathing task. The second contribution of this work was the advancement of the classical fNIRS hyperscanning method by adding systemic physiology measures. For this, new biosignal analyses in the time-frequency domain have been developed and tested in a simple nonverbal synchrony task between pairs of subjects. Furthermore, based on SPA-fNIRS hyperscanning data, another ML-based system was created, which is able distinguish familiar and unfamiliar pairs with high accuracy. This approach enables to determine the strength of social bonds in a wide range of social interaction contexts. In conclusion, we were the first group to perform a SPA-fNIRS hyperscanning study capturing changes in cerebral oxygenation and hemodynamics as well as systemic physiology in two subjects simultaneously. We applied new biosignals analysis methods enabling new insights into the study of social interactions. This work opens the door to many future inter-subjects fNIRS studies with the benefit of assessing the brain-to-brain, the brain-to-body, and body-to-body coupling between pairs of subjects.

It is critical to quantitatively analyse the developing human fetal brain in order to fully understand neurodevelopment in both normal fetuses and those with congenital disorders. To facilitate this analysis, automatic multi-tissue fetal brain segmentation algorithms are needed, which in turn requires open datasets of segmented fetal brains. Here we introduce a publicly available dataset of 50 manually segmented pathological and non-pathological fetal magnetic resonance brain volume reconstructions across a range of gestational ages (20 to 33 weeks) into 7 different tissue categories (external cerebrospinal fluid, grey matter, white matter, ventricles, cerebellum, deep grey matter, brainstem/spinal cord). In addition, we quantitatively evaluate the accuracy of several automatic multi-tissue segmentation algorithms of the developing human fetal brain. Four research groups participated, submitting a total of 10 algorithms, demonstrating the benefits the dataset for the development of automatic algorithms., Scientific Data, 8 (1), ISSN:2052-4463

Neurophotonics, 6 (2)

In preterm infants, the brain is the most vulnerable organ. Cerebral ischemia induces severe brain lesions that lead to neurodevelopmental impairments. Monitoring the brain’s oxygenation enables an early diagnosis and will thus help to treat ischemia. For this purpose, a non-invasive multimodal high-resolution imager is developed. This imaging system combines the advantages of two complementary imaging techniques, near-infrared optical tomography (NIROT) and optoacoustic imaging (OAI). OAI features the high spatial resolution of ultrasound imaging as well as the high contrast of optical imaging of vessels. The OAI signal is, however, spectrally distorted due to the unknown fluence distribution within the tissue. Fluence calculation based on NIROT reconstructions of the optical properties enable a correction of the OAI signal. Additionally, NIROT enables the monitoring of the tissue oxygenation of background tissue. The multimodal imager is developed in a collaborative effort between the Biomedical Optics Research Laboratory (BORL) in Zurich and the Institute of Applied Physics in Berne. This thesis presents the research and development of the NIROT part. Four studies are presented. First, the robustness of the image reconstruction to perturbations in source/detector location was examined. This was performed analytically and numerically for homogeneous phantoms and numerically for inhomogeneous phantoms. Numerical simulations were based on the NIROT software, Nirfast. The simulations yielded that reconstructed optical properties are highly sensitive to errors in source/detector location for one source-detector pair. However, small errors in the source/detector locations tend to cancel each other for more than one source-detector pair. Second, the successful development and design of the NIROT imager is presented. The NIROT sensor features highly flexible microfiber bundles as light guides. 3D printing technology was applied for inner sensor parts and to create a mold for silicone casting. Biocompatible 2-component silicone is casted to form a soft curved sensor shell that adapts to the surface structure. At the same time, it prevents light from directly travelling from a source fiber to the detector. Silicone phantoms were produced to perform a thorough phantom validation of the sensor prototype. Third, silicone dyes were characterized in terms of their optical properties to create phantoms with wavelength dependent optical properties. This enables a stable verification and validation method of NIRS/NIROT devices. These three studies lead up to the fourth study: A first laboratory experiment combining the full multimodal system, which was then used in a phantom study. Finally, it is found that NIROT serves to correct spectrally distorted OA signals, likewise OA and Ultrasound images could be used to provide NIROT reconstructions with priors. Overall, we reinforced the notion that this hybrid modality has great potential for clinical application, i.e. the early detection of cerebral ischemia in preterm infants.

The brain is a very vulnerable organ and damage to it is often followed by severe implications such as long-term disabilities and it may even lead to death. Cerebral oximetry by near-infrared spectroscopy (NIRS) has repeatedly been cited as promis- ing technology, potentially enabling clinicians to prevent these outcomes. Preterm neonates are likely to suffer from complications leading to brain damage and may thus benefit strongly from NIRS monitoring. Although a number of commercial NIRS oximeters are clinically approved, the method has not yet been widely established. Two of the major reasons for this are poor precision of instruments and that tissue oxygen haemoglobin saturation (StO 2 ) obtained from different oximeters and sensors are incomparable. This thesis addresses these two problems and provides solutions. OxyPrem was developed with the objective to provide an instrument with increased precision to clinicians. Two versions, OxyPrem v1.2 and v1.3, are introduced in this thesis and their performance is validated in vivo and in vitro. Both sensors are based on symmetric arrangement of light sources and detectors and employ a self-calibrating algorithm. OxyPrem v1.2 performed excellently in vivo in a precision assessment on the forearm of adults (repeatability = within-subject standard deviation (S w ) = 1.7 %). Repeata- bility in preterm neonates was S w = 3.3% which is still good, taking into account that S w in neonates is typically higher than in adults. An improved version of the instrument, OxyPrem v1.3, was assessed in vivo in neonates as well. The study demonstrated S w as good as 2.8 %. Simultaneous mea- surements with another OxyPrem v1.3 sensor and a pulse oximeter revealed unstable physiology in some neonates. In a second analysis without these confounding sub- jects, S w improved drastically to 1.9 % which is amongst the best precision values ever achieved for NIRS oximeters. To overcome the lack of comparability of different oximeters and sensors, we have performed several studies with liquid phantoms simulating optical properties of neona- tal brain tissue. We first conducted experiments with a simple, homogeneous phantom and then refined the set-up to model a more realistic two-layer geometry resembling skull and brain. Our studies showed substantially different StO 2 readings provided by different oximeters which, however, were mostly linearly related. With the improved set-up, we characterized a large number of commercially available oximeters and sen- sors and provided coefficients for their pairwise linear relation. The method showed good repeatability and helps establishing comparability. As neonates are a very heterogeneous group, we investigated the effect that a varia- tion in total haemoglobin concentration (c tHb ) has on StO 2 readings by NIRS oxime- ters. We found strong influence of c tHb on StO 2 while only OxyPrem v1.3 proved to be largely immune to this effect, which causes substantial uncertainty to readings of other instruments. As the presented phantom set-up is very versatile, we additionally investigated sev- eral other effects with slight adaptations. These showed that StO 2 readings were un- affected by a thin superficial layer, while sensitivity decreased substantially for a layer with 16 mm thickness. In another experiment, we did not observe a change in StO 2 readings for very thin clear layers such as oil on the skin of neonates, whereas thicker layers must be avoided. Partial placement of sensors on top of hair and birth marks may seriously flaw StO 2 of sensors without symmetric source-detector arrangement and self-calibrating algorithm. In summary, this thesis provides solutions to two of the problems mentioned most often in association with cerebral oxygenation monitoring by NIRS. We have intro- duced OxyPrem v1.2 and v1.3 and demonstrated superior precision of the instruments in vivo. OxyPrem v1.3 proved to be largely immune to variation in c tHb , which reduces this considerable uncertainty in StO 2 readings to a minimum. By a series of in vitro experiments with liquid phantoms we were able to establish comparability of different instruments and systematically assessed several types of influences to StO 2 readings.

ISBN:978-3-906327-63-1

For controlled caffeine release, light-responsive membranes were developed. It was possible to produce membranes that reduced their caffeine permeability resistance by about 97% when irradiated with UV-light compared to measurements at daylight. This was achieved by grafting polymers possessing photochromic units onto track-edged polycarbonate membranes. Covalently linked coatings on porous polycarbonate membranes were obtained by plasma activation of the membrane surface followed by plasma-induced graft polymerization. Copolymerization of spiro-compounds during the coating process as well as postmodification of preformed coatings with spiropyran resulted in photochromic membranes. For the copolymerization process, the synthesis of five photochromic methacrylic and acrylic spiropyrans and spirooxazines was successfully performed. Additionally, a spiropyran with carboxylic acid functionality was synthesized for the postmodification process. This enabled us to postmodify polymeric materials containing alcohol or amine groups to obtain photochromic materials. UV-irradiation of these light-responsive membranes resulted in a strong colouration of the membrane, in a reduction of surface tension, which resulted in a decreased caffeine permeability resistance. The membranes were characterized using XPS for the elemental composition of the coating, contact angle measurements for the surface tension, solid-state UV/VIS measurements for the determination of the kinetic and stability properties, and two-photon microscopy for the localisation of the photochromic substance in the porous membrane., RSC Advances, 3 (45), ISSN:2046-2069