Search

We established a direct communication channel between Microsoft HoloLens 2 and 3D Slicer to exchange transform and image messages between the platforms in real time. This allows us to seamlessly display a CT reslice of a patient in the AR world. Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. Research supported by projects PI122/00601 and AC20/00102 (Ministerio de Ciencia, Innovación y Universidades, Instituto de Salud Carlos III, Asociación Española Contra el Cáncer and European Regional Development Fund “Una manera de hacer Europa”), project PerPlanRT (ERA Permed), TED2021-129392B-I00 and TED2021-132200B-I00 (MCIN/AEI/10.13039/501100011033 and European Union “NextGenerationEU”/PRTR) and EU Horizon 2020 research and innovation programme Conex plus UC3M (grant agreement 801538). APC funded by Universidad Carlos III de Madrid (Read & Publish Agreement CRUE-CSIC 2023).

Memorias en formato póster de los proyectos que han obtenido reconocimiento en la 19ª Convocatoria de Apoyo a Experiencias de Innovación Docente. Curso 2021-2022. Presentadas en la Jornada de Innovación docente 2022 celebrada en Getafe el 23 de junio de 2022.

Breast cancer is the most common invasive cancer in women worldwide. Many women with breast cancer have their malignant tumors detected before the lesions become clinically palpable. Occult lesions must be marked for the surgeon to ensure that they can be effectively resected. Image-guided wire localization (WGL) is the current standard of care for the excision of non-palpable carcinomas during breast conserving surgery. The integration of the information from multimodal imaging may be especially relevant in surgical planning as complement or an alternative to WGL. The combination of information from images in different positions is especially difficult due to large breast deformation. This work presents a system based on surface registration to localize the lesion in the operative position, starting from a prone MRI study and a surface of the patient in the supine positon. The pre-operative surface from the MRI is registered to the surface obtained in a supine position similar to the intraoperative setting. Triangular meshes have been used to model breast surface in both positions and surfaces are aligned using a Laplacian deformation with fiducials automatically obtained from 3 anatomical references. The evaluation of the methodology has been carried out in 13 cases in which a supine- CT was available achieving an average localization error of 6.7 mm

Purpose: To present a method for spatiotemporal alignment of in-utero magnetic resonance imaging (MRI) time series acquired during maternal hyperoxia for enabling improved quantitative tracking of blood oxygen level-dependent (BOLD) signal changes that characterize oxygen transport through the placenta to fetal organs. Materials and Methods: The proposed pipeline for spatiotemporal alignment of images acquired with a single-shot gradient echo echo-planar imaging includes 1) signal nonuniformity correction, 2) intravolume motion correction based on nonrigid registration, 3) correction of motion and nonrigid deformations across volumes, and 4) detection of the outlier volumes to be discarded from subsequent analysis. BOLD MRI time series collected from 10 pregnant women during 3T scans were analyzed using this pipeline. To assess pipeline performance, signal fluctuations between consecutive timepoints were examined. In addition, volume overlap and distance between manual region of interest (ROI) delineations in a subset of frames and the delineations obtained through propagation of the ROIs from the reference frame were used to quantify alignment accuracy. A previously demonstrated rigid registration approach was used for comparison. Results: The proposed pipeline improved anatomical alignment of placenta and fetal organs over the state-of-the-art rigid motion correction methods. In particular, unexpected temporal signal fluctuations during the first normoxia period were significantly decreased (P < 0.01) and volume overlap and distance between region boundaries measures were significantly improved (P < 0.01). Conclusion: The proposed approach to align MRI time series enables more accurate quantitative studies of placental function by improving spatiotemporal alignment across placenta and fetal organs. National Institutes of Health (NIH) . Grant Numbers: U01 HD087211 , R01 EB017337 Consejeria de Educacion, Juventud y Deporte de la Comunidad de Madrid (Spain) through the Madrid-MIT M+Vision Consortium.

El procedimiento actual para cirugías de cáncer de mama incluye en la mayoría de los casos el empleo de un sistema aguja-arpón como herramienta de guiado durante la intervención. Aunque su uso está ampliamente extendido, presenta ciertas desventajas como el aumento en coste y tiempo, la incomodidad para el paciente o los posibles resultados estéticos. Por ello existen diferentes alternativas, entre las cuales se encuentra el uso de imágenes preoperatorias como herramienta de apoyo durante la cirugía. Sin embargo, su adquisición se lleva a cabo con el paciente en diferentes posiciones a la encontrada en la intervención, dificultando su interpretación. En este trabajo se presenta un procedimiento para la adquisición de la superficie del paciente durante la cirugía a partir de un escáner de luz estructurada combinado con el uso de un sistema de posicionamiento electromagnético para la localización del tumor. La información proporcionada por ambas técnicas se fusiona mediante un registro, empleando para ello unos marcadores diseñados específicamente para ser localizables por ambas. Tras una evaluación de la precisión, los resultados demuestran la viabilidad del procedimiento con errores inferiores a 1 mm.

Actas de: XXIX Congreso Anual de la Sociedad Espñaola de Ingeniería Biomédica (CASEIB 2011). Cáceres, 16-18 Noviembre 2011. La radioterapia intraoperatoria (RIO) es una técnica que combina cirugía y radioterapia, que se aplica en pacientes con tumores para los que se ha indicado su resección y con alto riesgo recidivante. El oncólogo radioterapeuta planifica esta intervención sobre una imagen TAC del paciente. Sin embargo, los datos del escenario real encontrado en la sala de tratamiento (posición y orientación del aplicador respecto del paciente y energía del haz) se deben recoger de forma manual para actualizar la planificación. Esta información es muy valiosa para la documentación y posterior seguimiento del procedimiento realizado. En este trabajo se evalúa la viabilidad de la integración de un sistema de posicionamiento óptico en el entorno RIO, y su precisión para localizar el aplicador sobre la imagen de planificación. Este trabajo ha sido financiado por el Minsiterio de Ciencia e Innovación (PI09/90568 IPT-3000000-2010-3, TEC2010-21619-C04-01), la Comunidad de Madrid (ARTEMIS S2009/DPI-1802) y fondos FEDER. Publicado

Proceeding of: The 72nd Annual Scientific Meeting of the College on Problems of Drug Dependence, Scottsdale, Arizona, June 12-17, 2010. MICINN (SAF2007-064890); Ministerio de Sanidad y Consumo (RD06/001/0029 y CP08/00017 del ISC III; Plan Nacional sobre Drogas 2004-2007 y 2008-2010); Fundación Mutua Madrileña; and CAM (S-SAL/0261/2006).

Actas de: XXVIII Congreso Anual de la Sociedad Española de Ingeniería Biomédica (CASEIB 2010). Madrid, 24-26 de noviembre de 2010. El análisis de la imagen de miocardio en cardiología nuclear es una herramienta muy importante para la valoración de la enfermedad coronaria, dado que permite la cuantificación objetiva de la extensión y de la intensidad del infarto de miocardio. En este trabajo se presenta una herramienta que permite realizar dicha cuantificación mediante el uso de mapas polares tanto estáticos como dinámicos, resultantes del análisis de la imagen PET que se viene realizando en los pacientes. Esta aproximación es una alternativa que permite la visualización integral del infarto en vez de requerir una segmentación (manual o asistida) de todas las rodajas tomográficas que componen el estudio. Este trabajo ha sido financiado por los proyectos del Ministerio de Ciencia e Innovación, TEC2007-64731, TEC 2008-06715-C02-1, la RETIC-RECAVA del Ministerio de Sanidad y Consumo, y el programa ARTEMIS S2009/DPI-1802 de la Comunidad de Madrid. Publicado

[Poster] 4th European Molecular Imaging Meeting, Barcelona, Spain, May 27 - 30, 2009 Non-uniform intensity artifacts confound the quantitative analysis of magnetic resonance (Mr) images of animal studies, particularly when using surface coils and high-field magnets. The use of correction methods proposed and validated on human brain images such as the n3 algorithm has previously been reported only on mouse images acquired with a volume coil. here, we evaluate the performance of n3 specifically on Mr rat brain images acquired with a surface coil CdTeaM (CeniT-ingenio 2010), Ministerio de Ciencia e innovación, Ciber Cb07/09/0031 CiberSaM, Ministerio de Sanidad y Consumo Publicado

The Medical Imaging Laboratory is a research group within the Hospital General Universitario Gregorio Marañón. The main research line of the group is focused towards the development and exploitation of medical imaging techniques, including the development of new processing tools for image analysis in clinical and preclinical research. The group has a multi-disciplinary profile and a priority for translational research topics, derived from real problems faced by the clinical specialists. One of the main research areas is the development of technologies for molecular imaging, some of which have been transferred to the industry and are now among the top products of the market. These systems include high-resolution PET, CT and PET-CT. Over the last years the group has developed several software tools to enable quantification of multimodal brain images using morphometric and functional data. Some research applications of these hardware and software tools are illustrated in the paper “Ministerio de Sanidad y Consumo”, CDTEAM Programa CÉNIT, Ministerio de Industria. FIS PI052271 “Ministerio de Sanidad y Consumo”. CIBER Salud Mental (CIBERSAM) “Ministerio de Sanidad y Consumo” and Fundación Mutua Madrileña Publicado

[Abstract] The second International conference of the European Society for Molecular Imaging in Naples, Italy June 14-15, 2007 Publicado

of: AMI Annual Conference 2006, March 25-29 Chronic exposure to opiates has been shown to influence neural activity in brain regions related to the rewarding process. It also induces neuroadaptations which lead to addiction. We have found in previous works that morphine self-administration produces neuroadaptative changes in brain areas of Fischer-344 rats. The aim of this study is to examine the effect of chronic self-administration of morphine on cerebral glucose metabolism Publicado

of: AMI Annual Conference 2006, March 25-29 Small animal studies are not always acquired in perfectly controlled conditions. This leads to reconstructed images that sometimes are ill-positioned: the animal is rotated or translated from the expected position. The study must be reoriented using a manual reformatting tool until the result has the desired orientation or position. To avoid this manual step, we propose a totally automatic procedure making use of the image principal axes and centroid, combined with an auto-registration process. The result is a correctly positioned study Publicado

Resumen Esta Tesis Doctoral profundiza en los métodos de registro de estudios multimodales (anatómicos y funcionales) de animales de laboratorio mediante técnicas basadas en teoría de la información. Se realiza un análisis de la influencia de los parámetros del algoritmo de registro, así como una validación sistemática de los resultados. Para poder evaluar el funcionamiento de los métodos de registro basados en teoría de la información, se presentan una serie de estudios patrón que se registran gracias a un conjunto de marcadores externos. En estos estudios se calcula la precisión del registro que se consigue por medio de los marcadores, validando el diseño y configuración de éstos. Se evalúa la convergencia y precisión de los algoritmos de registro, y su dependencia de diversos factores (desalineamiento inicial y parámetros del algoritmo), por medio de un diseño experimental que permite una validación sistemática. Los resultados permiten cuantificar la influencia de los factores en el resultado. Dado que los métodos descritos en esta Tesis Doctoral tienen como principal limitación el desalineamiento inicial máximo que son capaces de resolver, se propone un método basado en los ejes principales y los centros de masas de las imágenes, para prealinear los estudios. Utilizando este método se consigue mejorar el comportamiento de los algoritmos automáticos, consiguiendo un porcentaje de éxitos para las transformaciones evaluadas cercano al 100%. Finalmente, todos los algoritmos desarrollados y utilizados en esta Tesis Doctoral se encuentran integrados dentro de una interfaz de adquisición, reconstrucción, visualización y análisis denominada MMWKS, específicamente diseñada para estudios de animales de laboratorio. De esta forma los usuarios de este tipo de dispositivos pueden integrar la información proveniente de distintas modalidades con facilidad. Abstract The main topic of this PhD Thesis is multimodality (anatomical and functional) image registration of small animal studies, using information theory based techniques. The influence of algorithm parameters in the result and a systematic validation of these methods are also addressed. A set of reference studies are presented in order to evaluate the behavior of information theory based registration algorithms. These studies are registered using external references. The accuracy of the registration calculated from these markers is obtained, and allows validating their design and configuration. Since the methods described in this PhD Thesis have the initial misalignment they are able to solve as principal limitation, a method to pre-align the studies is proposed. This method is based in the principal axes transformation. By using this previous step, the results form automatic algorithms are improved, reaching a percentage of successful registrations near to 100%. Finally, algorithms developed and used in this PhD Thesis are part of a specific user interface named MMWKS. This software is designed to acquire, reconstruct, visualize and analyze small animal studies. By using the software developed, the users of small animal imaging devices can integrate information from different modalities easily.

of: AMI Annual Conference 2006, Orlando, Florida, March 25-29 Small animal studies of different modalities require certain image processing tools to be properly visualized and analyzed. We present a multimodality workstation that performs these tasks for different image modalities: positron emission tomography (PET) and computed tomography (CT) Publicado

of: AMI Annual Conference 2006, March 25-29 The use of small animal studies of different modalities isnowadays widespread, since functional studies like positron emission tomography (PET) or single photon emission computed tomography (SPECT) are better analyzed when mapped to the underlying anatomical structure coming from magnetic resonance imaging (MRI) or computed tomography (CT). The different positioning of the animals in the scanners results in the need for automated image registration methods. One important drawback of the existing algorithms is their limited capture range: if case of large initial misregistration (large translations and rotations) the optimization may not converge to the right solution, and the images do not become correctly coregistered. To avoid this problem we propose the use of a pre-alignment step based on the principal axes transformation. Publicado

[Abstract] 11th Annual Meeting of the Organization for Human Brain Mapping, June 2005, Toronto, Canada The aim of this work is to measure the repeatability of different segmentation strategies on regional volume estimates. We choose to focus on three possible alternatives of the most widely used segmentation method, the EM algoritIun. Thus, we have obtained regional volumetric data using raw EM, EM with PVE modeling, and EM using anatomical templates. To account for the most common sources of error, tIlese methods have been evaluated under different conditions of patient positioning, MR scarmer, and bias field correction Supported in part by: TIC (2001-3697-C03-03), FIS (02/1178,0213095), Red TeITk'ltica 1M3 G031185 Publicado

La posibilidad de visualizar y cuantificar la función de un determinado órgano en animales de laboratorio es una herramienta científica de gran importancia en el estudio de modelos de enfermedades humanas, en el descubrimiento y desarrollo de nuevos medicamentos y en la caracterización del fenotipo de animales transgénicos y noqueados. La tomografía por emisión de positrones (PET) y la imagen proyectiva de radiofármacos marcados con emisores de positrones proporcionan una herramienta genérica y no invasiva para la investigación científica en los casos anteriormente expuestos. En este trabajo se describe un sistema mutimodal PET/CT basado en cámaras de coincidencia y detectores de estado sólido para los rayos X, diseñado para adquirir tanto imágenes tomográficas como de proyección con alta resolución temporal, con capacidad para visualizar ratas y ratones, estos últimos en cuerpo completo en un solo campo de visión. Publicado

[Abstract] AMI Annual Conference 2004, March 27-31, Orlando, Florida This contribution reports preliminary results of a high-resolution small animal positron emission tomography (PET) based on pairs of opposed scintillation rotating detectors working in time coincidence Publicado

This work presents a new algorithm (nonuniform intensity correction; NIC) for correction of intensity inhomogeneities in T1-weighted magnetic resonance (MR) images. The bias field and a bias-free image are obtained through an iterative process that uses brain tissue segmentation. The algorithm was validated by means of realistic phantom images and a set of 24 real images. The first evaluation phase was based on a public domain phantom dataset, used previously to assess bias field correction algorithms. NIC performed similar to previously described methods in removing the bias field from phantom images, without introduction of degradation in the absence of intensity inhomogeneity. The real image dataset was used to compare the performance of this new algorithm to that of other widely used methods (N3, SPM'99, and SPM2). This dataset included both low and high bias field images from two different MR scanners of low (0.5 T) and medium (1.5 T) static fields. Using standard quality criteria for determining the goodness of the different methods, NIC achieved the best results, correcting the images of the real MR dataset, enabling its systematic use in images from both low and medium static field MR scanners. A limitation of our method is that it might fail if the bias field is so high that the initial histogram does not show bimodal distribution for white and gray matter Publicado

[Abstract] The 10th International Conference on Functional Mapping of the Human Brain, June 13-17, 2004, Budapest, Hungary This work presents an alternative method for reporting negative results in statistical parametric maps, consisting in estimating the maximum effect size that the test would not detect as significant with a certain probability Publicado

[Abstract] The 10th International Conference on Functional Mapping of the Human Brain, June 13-17, 2004, Budapest, Hungary The aim of this work is to study the repeatability of brain tissue volume quantification achieved by different MRI segmentation methods Publicado

[Abstract] AMI Annual Conference 2004, March 27-31, Orlando, Florida Statistical parametric mapping (SPM) is used to detect subtle activity changes in brain not requiring a priori assumptions about the expected activations. We adapt the methodology for analyzing rat brain positron emission tomography (PET) scans Publicado

[Abstract] AMI International Conference 2003, September 21 - 27, Madrid, Spain: "High Resolution Molecular Imaging: from Basic Science to Clinical Applications" The Advanced Technology Laboratory Animal Scanner (ATLAS) is a small animal positron emission tomography (PET) scanner with depth-of-interaction (DOI) capability designed to image animals the size of mice and rats. We report performance measurements and show animal studies that suggest that ATLAS offers advantages over systems without DOI capability Publicado

[Abstract] AMI International Conference 2003, September 21 - 27, Madrid, Spain: High Resolution Molecular Imaging: from Basic Science to Clinical Applications Statistical parametric mapping (SPM) is an analysis technique long been used in clinical research to detect subtle activity changes in brain; it is an excellent exploratory tool as it does not require a priori assumptions about the expected brain region activations. Research in animal imaging may also take benefit from this technique, if properly adapted to the new scenario. This is the case of brain activation studies in murine models using PET tracers and dedicated imaging devices. This work proposes the use of an SPM methodology adapted to the analysis of 2-deoxy-2-[18F] fluoro-D-Glucose (FDG) positron emission tomography (PET) scans of rat brains. Advantages over conventional region of interest (ROI) based analysis were assessed in an experiment addressing the detection of brain activation in of rats which underwent three different visual stimulation paradigms Publicado

AMI International Conference 2003, September 21 - 27, Madrid, Spain: "High Resolution Molecular Imaging: from Basic Science to Clinical Applications" A common problem in quantitative studies involving multi-subject analysis of brain functional images is the requirement of a spatial normalization step to provide a standardized reference space. Since SPM is one of the most widely used analytical approaches for functional studies, our aim is to investigate differences in metabolic activity measurements in positron emission tomography (PET) images with or without the spatial normalization performed by SPM. The possible effect of the normalization will be evaluated in a comparison of metabolic activity between schizophrenic patients and controls Publicado

Proceeding of: 2002 IEEE International Symposium On Biomedical Imaging, Washington, D.C., USA, July 7-10, 2002 ATLAS (Advanced Technology Laboratory Animal Scanner), a small animal PET scanner designed to image animals the size of rats and mice, is about to enter service on the NIH campus in Bethesda, Maryland. This system is the first small animal PET scanner with a depth-ofinteraction capability and the first to use iterative resolution recovery algorithms, rather than conventional filtered back projection, for "production" image reconstruction. ATLAS is also proximate to, and co-axial with, a high resolution small animal CT scanner. When fully integrated, spatially registered PET and CT images of each animal will be used to correct the emission data for radiation attenuation and to aid in target identification. In this report we describe some of the technical and functional features of this system and illustrate how these features are used in an actual small animal imaging study This work was supported in part by projects FIS 00/0036 and 11I PRICIT Comunidad de Madrid, Spain.

[Abstract] AMI Annual Conference 2002, October 23 - 27, San Diego, California Publicado

[Abstract] AMI Annual Conference 2002, October 23 - 27 Publicado

Mención Internacional en el título de doctor The role of technology inside the operating room is constantly increasing, allowing surgical procedures previously considered impossible or too risky due to their complexity or limited access. These reliable tools have improved surgical efficiency and safety. Cancer treatment is one of the surgical specialties that has benefited most from these techniques due to its high incidence and the accuracy required for tumor resections with conservative approaches and clear margins. However, in many cases, introducing these technologies into surgical scenarios is expensive and entails complex setups that are obtrusive, invasive, and increase the operative time. In this thesis, we proposed convenient, accessible, reliable, and non-invasive solutions for two highly complex regions for tumor resection surgeries: pelvis and head and neck. We explored how the introduction of 3D printing, surgical navigation, and augmented reality in these scenarios provided high intraoperative precision. First, we presented a less invasive setup for osteotomy guidance in pelvic tumor resections based on small patient-specific instruments (PSIs) fabricated with a desktop 3D printer at a low cost. We evaluated their accuracy in a cadaveric study, following a realistic workflow, and obtained similar results to previous studies with more invasive setups. We also identified the ilium as the region more prone to errors. Then, we proposed surgical navigation using these small PSIs for image-to-patient registration. Artificial landmarks included in the PSIs substitute the anatomical landmarks and the bone surface commonly used for this step, which require additional bone exposure and is, therefore, more invasive. We also presented an alternative and more convenient installation of the dynamic reference frame used to track the patient movements in surgical navigation. The reference frame is inserted in a socket included in the PSIs and can be attached and detached without losing precision and simplifying the installation. We validated the setup in a cadaveric study, evaluating the accuracy and finding the optimal PSI configuration in the three most common scenarios for pelvic tumor resection. The results demonstrated high accuracy, where the main source of error was again incorrect placements of PSIs in regular and homogeneous regions such as the ilium. The main limitation of PSIs is the guidance error resulting from incorrect placements. To overcome this issue, we proposed augmented reality as a tool to guide PSI installation in the patient’s bone. We developed an application for smartphones and HoloLens 2 that displays the correct position intraoperatively. We measured the placement errors in a conventional and a realistic phantom, including a silicone layer to simulate tissue. The results demonstrated a significant reduction of errors with augmented reality compared to freehand placement, ensuring an installation of the PSI close to the target area. Finally, we proposed three setups for surgical navigation in palate tumor resections, using optical trackers and augmented reality. The tracking tools for the patient and surgical instruments were fabricated with low-cost desktop 3D printers and designed to provide less invasive setups compared to previous solutions. All setups presented similar results with high accuracy when tested in a 3D-printed patient-specific phantom. They were then validated in the real surgical case, and one of the solutions was applied for intraoperative guidance. Postoperative results demonstrated high navigation accuracy, obtaining optimal surgical outcomes. The proposed solution enabled a conservative surgical approach with a less invasive navigation setup. To conclude, in this thesis we have proposed new setups for intraoperative navigation in two complex surgical scenarios for tumor resection. We analyzed their navigation precision, defining the optimal configurations to ensure accuracy. With this, we have demonstrated that computer-assisted surgery techniques can be integrated into the surgical workflow with accessible and non-invasive setups. These results are a step further towards optimizing the procedures and continue improving surgical outcomes in complex surgical scenarios. Programa de Doctorado en Ciencia y Tecnología Biomédica por la Universidad Carlos III de Madrid Presidente: Raúl San José Estépar.- Secretario: Alba González Álvarez.- Vocal: Simon Drouin

Mención Internacional en el título de doctor In the last twenty years, a new way of practicing medicine has been focusing on the problems and needs of each patient as an individual thanks to the significant advances in healthcare technology, the so-called personalized medicine. In surgical treatments, personalization has been possible thanks to key technologies adapted to the specific anatomy of each patient and the needs of the physicians. Tracking systems, augmented reality (AR), three-dimensional (3D) printing and artificial intelligence (AI) have previously supported this individualized medicine in many ways. However, their independent contributions show several limitations in terms of patient-to-image registration, lack of flexibility to adapt to the requirements of each case, large preoperative planning times, and navigation complexity. The main objective of this thesis is to increase patient personalization in surgical treatments by combining these technologies to bring surgical navigation to new complex cases by developing new patient registration methods, designing patient-specific tools, facilitating access to augmented reality by the medical community, and automating surgical workflows. In the first part of this dissertation, we present a novel framework for acral tumor resection combining intraoperative open-source navigation software, based on an optical tracking system, and desktop 3D printing. We used additive manufacturing to create a patient-specific mold that maintained the same position of the distal extremity during image-guided surgery as in the preoperative images. The feasibility of the proposed workflow was evaluated in two clinical cases (soft-tissue sarcomas in hand and foot). We achieved an overall accuracy of the system of 1.88 mm evaluated on the patient-specific 3D printed phantoms. Surgical navigation was feasible during both surgeries, allowing surgeons to verify the tumor resection margin. Then, we propose and augmented reality navigation system that uses 3D printed surgical guides with a tracking pattern enabling automatic patient-to-image registration in orthopedic oncology. This specific tool fits on the patient only in a pre-designed location, in this case bone tissue. This solution has been developed as a software application running on Microsoft HoloLens. The workflow was validated on a 3D printed phantom replicating the anatomy of a patient presenting an extraosseous Ewing’s sarcoma, and then tested during the actual surgical intervention. The results showed that the surgical guide with the reference marker can be placed precisely with an accuracy of 2 mm and a visualization error lower than 3 mm. The application allowed physicians to visualize the skin, bone, tumor and medical images overlaid on the phantom and patient. To enable the use of AR and 3D printing by inexperienced users without broad technical knowledge, we designed a step-by-step methodology. The proposed protocol describes how to develop an AR smartphone application that allows superimposing any patient-based 3D model onto a real-world environment using a 3D printed marker tracked by the smartphone camera. Our solution brings AR solutions closer to the final clinical user, combining free and open-source software with an open-access protocol. The proposed guide is already helping to accelerate the adoption of these technologies by medical professionals and researchers. In the next section of the thesis, we wanted to show the benefits of combining these technologies during different stages of the surgical workflow in orthopedic oncology. We designed a novel AR-based smartphone application that can display the patient’s anatomy and the tumor’s location. A 3D printed reference marker, designed to fit in a unique position of the affected bone tissue, enables automatic registration. The system has been evaluated in terms of visualization accuracy and usability during the whole surgical workflow on six realistic phantoms achieving a visualization error below 3 mm. The AR system was tested in two clinical cases during surgical planning, patient communication, and surgical intervention. These results and the positive feedback obtained from surgeons and patients suggest that the combination of AR and 3D printing can improve efficacy, accuracy, and patients’ experience In the final section, two surgical navigation systems have been developed and evaluated to guide electrode placement in sacral neurostimulation procedures based on optical tracking and augmented reality. Our results show that both systems could minimize patient discomfort and improve surgical outcomes by reducing needle insertion time and number of punctures. Additionally, we proposed a feasible clinical workflow for guiding SNS interventions with both navigation methodologies, including automatically creating sacral virtual 3D models for trajectory definition using artificial intelligence and intraoperative patient-to-image registration. To conclude, in this thesis we have demonstrated that the combination of technologies such as tracking systems, augmented reality, 3D printing, and artificial intelligence overcomes many current limitations in surgical treatments. Our results encourage the medical community to combine these technologies to improve surgical workflows and outcomes in more clinical scenarios. Programa de Doctorado en Ciencia y Tecnología Biomédica por la Universidad Carlos III de Madrid Presidenta: María Jesús Ledesma Carbayo.- Secretaria: María Arrate Muñoz Barrutia.- Vocal: Csaba Pinter

Mención Internacional en el título de doctor Craniosynostosis is a congenital defect defined as the premature fusion of one or more cranial sutures. This fusion leads to growth restriction and deformation of the cranium, caused by compensatory expansion parallel to the fused sutures. Surgical correction is the preferred treatment in most cases to excise the fused sutures and to normalize cranial shape. Although multiple technological advancements have arisen in the surgical management of craniosynostosis, interventional planning and surgical correction are still highly dependent on the subjective assessment and artistic judgment of craniofacial surgeons. Therefore, there is a high variability in individual surgeon performance and, thus, in the surgical outcomes. The main objective of this thesis was to explore different approaches to improve the surgical management of craniosynostosis by reducing subjectivity in all stages of the process, from the preoperative virtual planning phase to the intraoperative performance. First, we developed a novel framework for automatic planning of craniosynostosis surgery that enables: calculating a patient-specific normative reference shape to target, estimating optimal bone fragments for remodeling, and computing the most appropriate configuration of fragments in order to achieve the desired target cranial shape. Our results showed that automatic plans were accurate and achieved adequate overcorrection with respect to normative morphology. Surgeons’ feedback indicated that the integration of this technology could increase the accuracy and reduce the duration of the preoperative planning phase. Second, we validated the use of hand-held 3D photography for intraoperative evaluation of the surgical outcome. The accuracy of this technology for 3D modeling and morphology quantification was evaluated using computed tomography imaging as gold-standard. Our results demonstrated that 3D photography could be used to perform accurate 3D reconstructions of the anatomy during surgical interventions and to measure morphological metrics to provide feedback to the surgical team. This technology presents a valuable alternative to computed tomography imaging and can be easily integrated into the current surgical workflow to assist during the intervention. Also, we developed an intraoperative navigation system to provide real-time guidance during craniosynostosis surgeries. This system, based on optical tracking, enables to record the positions of remodeled bone fragments and compare them with the target virtual surgical plan. Our navigation system is based on patient-specific surgical guides, which fit into the patient’s anatomy, to perform patient-to-image registration. In addition, our workflow does not rely on patient’s head immobilization or invasive attachment of dynamic reference frames. After testing our system in five craniosynostosis surgeries, our results demonstrated a high navigation accuracy and optimal surgical outcomes in all cases. Furthermore, the use of navigation did not substantially increase the operative time. Finally, we investigated the use of augmented reality technology as an alternative to navigation for surgical guidance in craniosynostosis surgery. We developed an augmented reality application to visualize the virtual surgical plan overlaid on the surgical field, indicating the predefined osteotomy locations and target bone fragment positions. Our results demonstrated that augmented reality provides sub-millimetric accuracy when guiding both osteotomy and remodeling phases during open cranial vault remodeling. Surgeons’ feedback indicated that this technology could be integrated into the current surgical workflow for the treatment of craniosynostosis. To conclude, in this thesis we evaluated multiple technological advancements to improve the surgical management of craniosynostosis. The integration of these developments into the surgical workflow of craniosynostosis will positively impact the surgical outcomes, increase the efficiency of surgical interventions, and reduce the variability between surgeons and institutions. Programa de Doctorado en Ciencia y Tecnología Biomédica por la Universidad Carlos III de Madrid Presidente: Norberto Antonio Malpica González.- Secretario: María Arrate Muñoz Barrutia.- Vocal: Tamas Ungi

The importance of resolution is crucial when working with medical images. The possibility to visualize details lead to a more accurate diagnosis and makes segmentation easier. However, obtention of high-resolution medical images requires of long acquisition times. In clinical environments, lack of time leads to the acquisition of low-resolution images. Super Resolution (SR) consist in post-processing images in order to enhance its resolution. During the last years, a branch of SR is getting promising results. This branch focuses in the application of Convolutional Neural Networks (CNNs) to the images. This project is intended to create a network able to enhance resolution of knee MR stored in DICOM format. Different networks are proposed, and evaluation is made by computing Peak Signal-to-Noise Ratio (PSNR) and normalized Cross-Correlation. One of the networks proposed, SR-DCNN, presented better results than the conventional method, bicubic interpolation. Finally, visual comparison of the SR-DCNN and bicubic interpolation also showed that the network proposed outperforms the conventional methods. Ingeniería Biomédica

This thesis was developed with the aim of building and implementing new informatic tools that could facilitate research and international benchmarking in paediatric transplantation. Two different studies were created for that purpose. The first one consisted on an analysis of the current paediatric transplantation activity situation in the European Union, utilizing Google My Maps to obtain a general geographical distribution per countries, and Data Studio to complement the previous one and get more complete results. The second method was the construction of a patient registry for TransplantChild, the European Reference Network for paediatric transplantation, so that all the transplanted children in the member hospitals are registered, and their information is stored together to be analysed. The results obtained give an accurate vision of the unequal transplantation distribution across the continent and allowed the identification of the most expert and specialized centres in Europe. Also, it was possible to recognize potential cases in which the child needed to be moved to receive a transplant, and proposed solutions for them. On the other hand, two platforms, one for data collection and one for data exploitation, were integrated together to build up the patient registry. The latter was developed from scratch in this project using Python and Flask. At last, it was concluded that by implementing the mentioned tools it is possible to promote paediatric transplantation research and perform a benchmarking across the countries and hospitals. Hence, this turns out to be an indirect way to improve transplants success rate and, in the end, patients’ survival and quality of life. Ingeniería Biomédica

Mención Internacional en el título de doctor Intraoperative electron radiation therapy (IOERT) is a technique used to deliver radiation to the surgically opened tumor bed without irradiating healthy tissue. Treatment planning systems and mobile linear accelerators enable clinicians to optimize the procedure, minimize stress in the operating room (OR) and avoid transferring the patient to a dedicated radiation room. However, placement of the radiation collimator over the tumor bed requires a validation methodology to ensure correct delivery of the dose prescribed in the treatment planning system. In this dissertation, we address three well-known limitations of IOERT: applicator positioning over the tumor bed, docking of the mobile linear accelerator gantry with the applicator and validation of the dose delivery prescribed. This thesis demonstrates that these limitations can be overcome by positioning the applicator appropriately with respect to the patient’s anatomy. The main objective of the study was to assess technological and procedural alternatives for improvement of IOERT performance and resolution of problems of uncertainty. Image-to-world registration, multicamera optical trackers, multimodal imaging techniques and mobile linear accelerator docking are addressed in the context of IOERT. IOERT is carried out by a multidisciplinary team in a highly complex environment that has special tracking needs owing to the characteristics of its working volume (i.e., large and prone to occlusions), in addition to the requisites of accuracy. The first part of this dissertation presents the validation of a commercial multicamera optical tracker in terms of accuracy, sensitivity to miscalibration, camera occlusions and detection of tools using a feasible surgical setup. It also proposes an automatic miscalibration detection protocol that satisfies the IOERT requirements of automaticity and speed. We show that the multicamera tracker is suitable for IOERT navigation and demonstrate the feasibility of the miscalibration detection protocol in clinical setups. Image-to-world registration is one of the main issues during image-guided applications where the field of interest and/or the number of possible anatomical localizations is large, such as IOERT. In the second part of this dissertation, a registration algorithm for image-guided surgery based on lineshaped fiducials (line-based registration) is proposed and validated. Line-based registration decreases acquisition time during surgery and enables better registration accuracy than other published algorithms. In the third part of this dissertation, we integrate a commercial low-cost ultrasound transducer and a cone beam CT C-arm with an optical tracker for image-guided interventions to enable surgical navigation and explore image based registration techniques for both modalities. In the fourth part of the dissertation, a navigation system based on optical tracking for the docking of the mobile linear accelerator to the radiation applicator is assessed. This system improves safety and reduces procedure time. The system tracks the prescribed collimator location to solve the movements that the linear accelerator should perform to reach the docking position and warns the user about potentially unachievable arrangements before the actual procedure. A software application was implemented to use this system in the OR, where it was also evaluated to assess the improvement in docking speed. Finally, in the last part of the dissertation, we present and assess the installation setup for a navigation system in a dedicated IOERT OR, determine the steps necessary for the IOERT process, identify workflow limitations and evaluate the feasibility of the integration of the system in a real OR. The navigation system safeguards the sterile conditions of the OR, clears the space available for surgeons and is suitable for any similar dedicated IOERT OR. La Radioterapia Intraoperatoria por electrones (RIO) consiste en la aplicación de radiación de alta energía directamente sobre el lecho tumoral, accesible durante la cirugía, evitando radiar los tejidos sanos. Hoy en día, avances como los sistemas de planificación (TPS) y la aparición de aceleradores lineales móviles permiten optimizar el procedimiento, minimizar el estrés clínico en el entorno quirúrgico y evitar el desplazamiento del paciente durante la cirugía a otra sala para ser radiado. La aplicación de la radiación se realiza mediante un colimador del haz de radiación (aplicador) que se coloca sobre el lecho tumoral de forma manual por el oncólogo radioterápico. Sin embargo, para asegurar una correcta deposición de la dosis prescrita y planificada en el TPS, es necesaria una adecuada validación de la colocación del colimador. En esta Tesis se abordan tres limitaciones conocidas del procedimiento RIO: el correcto posicionamiento del aplicador sobre el lecho tumoral, acoplamiento del acelerador lineal con el aplicador y validación de la dosis de radiación prescrita. Esta Tesis demuestra que estas limitaciones pueden ser abordadas mediante el posicionamiento del aplicador de radiación en relación con la anatomía del paciente. El objetivo principal de este trabajo es la evaluación de alternativas tecnológicas y procedimentales para la mejora de la práctica de la RIO y resolver los problemas de incertidumbre descritos anteriormente. Concretamente se revisan en el contexto de la radioterapia intraoperatoria los siguientes temas: el registro de la imagen y el paciente, sistemas de posicionamiento multicámara, técnicas de imagen multimodal y el acoplamiento del acelerador lineal móvil. El entorno complejo y multidisciplinar de la RIO precisa de necesidades especiales para el empleo de sistemas de posicionamiento como una alta precisión y un volumen de trabajo grande y propenso a las oclusiones de los sensores de posición. La primera parte de esta Tesis presenta una exhaustiva evaluación de un sistema de posicionamiento óptico multicámara comercial. Estudiamos la precisión del sistema, su sensibilidad a errores cometidos en la calibración, robustez frente a posibles oclusiones de las cámaras y precisión en el seguimiento de herramientas en un entorno quirúrgico real. Además, proponemos un protocolo para la detección automática de errores por calibración que satisface los requisitos de automaticidad y velocidad para la RIO demostrando la viabilidad del empleo de este sistema para la navegación en RIO. Uno de los problemas principales de la cirugía guiada por imagen es el correcto registro de la imagen médica y la anatomía del paciente en el quirófano. En el caso de la RIO, donde el número de posibles localizaciones anatómicas es bastante amplio, así como el campo de trabajo es grande se hace necesario abordar este problema para una correcta navegación. Por ello, en la segunda parte de esta Tesis, proponemos y validamos un nuevo algoritmo de registro (LBR) para la cirugía guiada por imagen basado en marcadores lineales. El método propuesto reduce el tiempo de la adquisición de la posición de los marcadores durante la cirugía y supera en precisión a otros algoritmos de registro establecidos y estudiados en la literatura. En la tercera parte de esta tesis, integramos un transductor de ultrasonido comercial de bajo coste, un arco en C de rayos X con haz cónico y un sistema de posicionamiento óptico para intervenciones guiadas por imagen que permite la navegación quirúrgica y exploramos técnicas de registro de imagen para ambas modalidades. En la cuarta parte de esta tesis se evalúa un navegador basado en el sistema de posicionamiento óptico para el acoplamiento del acelerador lineal móvil con aplicador de radiación, mejorando la seguridad y reduciendo el tiempo del propio acoplamiento. El sistema es capaz de localizar el colimador en el espacio y proporcionar los movimientos que el acelerador lineal debe realizar para alcanzar la posición de acoplamiento. El sistema propuesto es capaz de advertir al usuario de aquellos casos donde la posición de acoplamiento sea inalcanzable. El sistema propuesto de ayuda para el acoplamiento se integró en una aplicación software que fue evaluada para su uso final en quirófano demostrando su viabilidad y la reducción de tiempo de acoplamiento mediante su uso. Por último, presentamos y evaluamos la instalación de un sistema de navegación en un quirófano RIO dedicado, determinamos las necesidades desde el punto de vista procedimental, identificamos las limitaciones en el flujo de trabajo y evaluamos la viabilidad de la integración del sistema en un entorno quirúrgico real. El sistema propuesto demuestra ser apto para el entorno RIO manteniendo las condiciones de esterilidad y dejando despejado el campo quirúrgico además de ser adaptable a cualquier quirófano similar. Programa Oficial de Doctorado en Multimedia y Comunicaciones Presidente: Raúl San José Estépar.- Secretario: María Arrate Muñoz Barrutia.- Vocal: Carlos Ferrer Albiach

Arterial spin labeling is becoming an increasingly popular method for evaluating the cerebral blood flow. However, one of the major limitations of arterial spin labeling (and other perfusion assessment methods like PET) is the partial volume effect, by which each voxel of the image contains a mixture of tissues due to the low spatial resolution of ASL. Partial volume correction is required to retrieve the perfusion contribution of each of these tissues, which is important in the study of neurodegenerative diseases such as Alzheimer’s disease. A new partial volume correction method (3D weighted least squares) based on an existing state-of-the-art method (Asllani’s algorithm) is presented in this work. The new algorithm improves the previous algorithm by operating in a 3D way instead of a 2D way and including a weighting to the regression problem as a function of the distance between the voxels.The new method was tested over simulated cerebral perfusion images, giving better results than the Asllani’s algorithm. The algorithm was also implemented as a graphical user interface extension for the open source platform 3DSlicer. This extension automates all the correction process and allows the researchers processing the ASL images rapidly and easily. Using this extension, a real perfusion study was conducted to compare the cerebral perfusion between Alzheimer and control groups in resting state. Alzheimer group showed a significantly lower perfusion in the thalamus, caudate nucleus, hippocampus and cuneus. These regions have been reported in the literature to present atrophies in Alzheimer subjects and are involved in cognitive functions that are negatively affected by the disease. These results provide further validation for the 3DWLS as a suitable correction method and for the extension as a useful research tool. Ingeniería Biomédica

Mención Internacional en el título de doctor Esta tesis se enmarca dentro del campo de la radioterapia y trata específicamente sobre la radioterapia intraoperatoria (RIO) con electrones. Esta técnica combina la resección quirúrgica de un tumor y la radiación terapéutica directamente aplicada sobre el lecho tumoral post-resección o sobre el tumor no resecado. El haz de electrones de alta energía es colimado y conducido por un aplicador específico acoplado a un acelerador lineal. La planificación de la RIO con electrones es compleja debido a las modificaciones geométricas y anatómicas producidas por la retracción de estructuras y la eliminación de tejidos cancerosos durante la cirugía. Actualmente, no se dispone del escenario real en este tipo de tratamientos (por ejemplo, la posición/orientación del aplicador respecto a la anatomía del paciente o las irregularidades en la superficie irradiada), sólo de una estimación grosso modo del tratamiento real administrado al paciente. Las imágenes intraoperatorias del escenario real durante el tratamiento (concretamente imágenes de tomografía axial computarizada [TAC]) serían útiles no sólo para la planificación intraoperatoria, sino también para registrar y evaluar el tratamiento administrado al paciente. Esta información es esencial en estudios prospectivos. En esta tesis se evaluó en primer lugar la viabilidad de un sistema de seguimiento óptico de varias cámaras para obtener la posición/orientación del aplicador en los escenarios de RIO con electrones. Los resultados mostraron un error de posición del aplicador inferior a 2 mm (error medio del centro del bisel) y un error de orientación menor de 2º (error medio del eje del bisel y del eje longitudinal del aplicador). Estos valores están dentro del rango propuesto por el Grupo de Trabajo 147 (encargo del Comité de Terapia y del Subcomité para la Mejora de la Garantía de Calidad y Resultados de la Asociación Americana de Físicos en Medicina [AAPM] para estudiar en radioterapia externa la exactitud de la localización con métodos no radiográficos, como los sistemas infrarrojos). Una limitación importante de la solución propuesta es que el aplicador se superpone a la imagen preoperatoria del paciente. Una imagen intraoperatoria proporcionaría información anatómica actualizada y permitiría estimar la distribución tridimensional de la dosis. El segundo estudio específico de esta tesis evaluó la viabilidad de adquirir con un TAC simulador imágenes TAC intraoperatorias de escenarios reales de RIO con electrones. No hubo complicaciones en la fase de transporte del paciente utilizando la camilla y su acople para el transporte, o con la adquisición de imágenes TAC intraoperatorias en la sala del TAC simulador. Los estudios intraoperatorios adquiridos se utilizaron para evaluar la mejora obtenida en la estimación de la distribución de dosis en comparación con la obtenida a partir de imágenes TAC preoperatorias, identificando el factor dominante en esas estimaciones (la región de aire y las irregularidades en la superficie, no las heterogeneidades de los tejidos). Por último, el tercer estudio específico se centró en la evaluación de varias tecnologías TAC de kilovoltaje, aparte del TAC simulador, para adquirir imágenes intraoperatorias con las que estimar la distribución de la dosis en RIO con electrones. Estos dispositivos serían necesarios en el caso de disponer de aceleradores lineales portátiles en el quirófano ya que no se aprobaría mover al paciente a la sala del TAC simulador. Los resultados con un maniquí abdominal mostraron que un TAC portátil (BodyTom) e incluso un acelerador lineal con un TAC de haz de cónico (TrueBeam) serían adecuados para este propósito. This thesis is framed within the field of radiotherapy, specifically intraoperative electron radiotherapy (IOERT). This technique combines surgical resection of a tumour and therapeutic radiation directly applied to a post-resection tumour bed or to an unresected tumour. The high-energy electron beam is collimated and conducted by a specific applicator docked to a linear accelerator (LINAC). Dosimetry planning for IOERT is challenging owing to the geometrical and anatomical modifications produced by the retraction of structures and removal of cancerous tissues during the surgery. No data of the actual IOERT 3D scenario is available (for example, the applicator pose in relation to the patient’s anatomy or the irregularities in the irradiated surface) and consequently only a rough approximation of the actual IOERT treatment administered to the patient can be estimated. Intraoperative computed tomography (CT) images of the actual scenario during the treatment would be useful not only for intraoperative planning but also for registering and evaluating the treatment administered to the patient. This information is essential for prospective trials. In this thesis, the feasibility of using a multi-camera optical tracking system to obtain the applicator pose in IOERT scenarios was firstly assessed. Results showed that the accuracy of the applicator pose was below 2 mm in position (mean error of the bevel centre) and 2º in orientation (mean error of the bevel axis and the longitudinal axis), which are within the acceptable range proposed in the recommendation of Task Group 147 (commissioned by the Therapy Committee and the Quality Assurance and Outcomes Improvement Subcommittee of the American Association of Physicists in Medicine [AAPM] to study the localization accuracy with non-radiographic methods such as infrared systems in external beam radiation therapy). An important limitation of this solution is that the actual pose of applicator is superimposed on a patient’s preoperative image. An intraoperative image would provide updated anatomical information and would allow estimating the 3D dose distribution. The second specific study of this thesis evaluated the feasibility of acquiring intraoperative CT images with a CT simulator in real IOERT scenarios. There were no complications in the whole procedure related to the transport step using the subtable and its stretcher or the acquisition of intraoperative CT images in the CT simulator room. The acquired intraoperative studies were used to evaluate the improvement achieved in the dose distribution estimation when compared to that obtained from preoperative CT images, identifying the dominant factor in those estimations (air gap and the surface irregularities, not tissue heterogeneities). Finally, the last specific study focused on assessing several kilovoltage (kV) CT technologies other than CT simulators to acquire intraoperative images for estimating IOERT dose distribution. That would be necessary when a mobile electron LINAC was available in the operating room as transferring the patient to the CT simulator room could not be approved. Our results with an abdominal phantom revealed that a portable CT (BodyTom) and even a LINAC with on-board kV cone-beam CT (TrueBeam) would be suitable for this purpose. Programa Oficial de Doctorado en Multimedia y Comunicaciones Presidente: Joaquín López Herráiz.- Secretario: María Arrate Muñoz Barrutia.- Vocal: Óscar Acosta Tamayo

In the last decades, neuroimaging, and particularly functional magnetic resonance, have lead to a high number of findings on brain organization that have helped the scientific community to better understand brain function. Despite these advances, the foundations of neural activity and connectivity are not yet understood, and little is known about its role in the etiology of psychiatric disorders. A vast amount of literature on neuroimaging demonstrates the presence of abnormal and dysfunctional brain patterns of activity, revealed by a wide number of different analytical approaches. However, the establishment of consistent relations between abnormal connectivity patterns and symptoms or diagnoses is still not confirmed. For this reason, the present project studied these relations in young individuals suffering from First Episode Psychosis (FEP) and healthy control participants by implementing a model-dependent atlas-based approach using the publicly available Seven Network Atlas. The statistical analysis of preprocessed images revealed various tendencies consistent with previous results from literature as well as new findings. Youth in psychosis presented hyperconnectivity in most resting-state networks when compared to controls, being significantly enhanced in Fronto-Parietal network (FPN) and Default Mode network (DMN). Also, within patients hyperconnectivity was negatively related to general cognitive performance and positively linked to clinical symptom severity. We estimate that the results of this work will eventually contribute to a better understanding of the early stages of psychosis in children and adolescents, critical for optimizing current interventions and treatments. En los últimos años, la comunidad científica ha hecho grandes avances para alcanzar una mejor comprensión del funcionamiento del cerebro gracias al gran número de descubrimientos relativos a la organización del mismo, llevados a cabo mediante neuroimagen y en particular por medio de resonancia magnética funcional. A pesar de estos avances, todavía se desconocen las bases de la actividad y conectividad neuronal así como su papel en la etiología de los trastornos psicóticos. Una gran cantidad de literatura sobre análisis de neuroimagen mediante diversos métodos demuestra de manera consistente la presencia de patrones de actividad cerebral alterados y disfuncionales; sin embargo, es imposible establecer de manera sólida, a través de diferentes estudios, la relación entre patrones específicos de conectividad funcional alterada con los diferentes síntomas y condiciones de estos trastornos. Por este motivo, el presente proyecto ha estudiado estos vínculos en niños y adolescentes que padecen un primer episodio de psicosis así como en controles sanos mediante un enfoque basado en un atlas público de 7 redes cerebrales. El análisis estadístico de las imágenes preprocesadas reveló varias tendencias coherentes con resultados previos extraídos de la literatura así como nuevos hallazgos. Los jóvenes con psicosis mostraron tener hiperconectividad en la mayoría de las redes en reposo en comparación con los controles, siendo dicha diferencia especialmente significativa en las redes Fronto-Parietal y por Defecto. Además, se observaron patrones de correlación negativa entre la conectividad y los procesos cognitivos, mientas que la relación de la conectividad con la gravedad de los síntomas mostró ser positiva. Los resultados de este trabajo contribuyen a una mejor comprensión de las etapas iniciales de la psicosis en niños y adolescentes, fundamentales para la optimización de los actuales tratamientos e intervenciones. Ingeniería Biomédica

Although the majority of our upper-limb movements are performed with two arms, only a few neurophysiological studies have addressed neuronal mechanisms of bimanual motor control. In particular, little is known about directional tuning of cortical neurons during bimanual motor tasks. This contrasts with the vast literature on the same topic for unimanual movements. Georgopoulos and his colleagues pioneered the description of neuronal mechanisms of unimanual movements in terms of directional tuning curves, where firing rate of a neuron is expressed as a function of reach angle for center-out movements. According to Georgopoulos, individual neurons in the motor cortex are broadly tuned to movement direction, with the firing rate reaching a maximum value for the direction called preferred. This tuning pattern is well approximated by a cosine curve. No such description has been developed for bimanual movements. The main objective of the project is to address this gap in knowledge. For this purpose, an analysis will be presented of neuronal recordings performed in rhesus monkeys trained to perform bimanual tasks. Multiple linear regression fitting methods will be used to describe the dependency of neuronal rates on the movement directions of two arms, comparing the results obtained with each method. Neuronal tuning for unimanual movements (performed with either the left or right arm) will also be compared with the tuning for bimanual movements, testing whether bimanual tuning could be derived from the tuning for each arm tested separately, or if it is principally different from the dependencies for unimanual movements. The results from this work will eventually lead to improved neural decoding algorithms that could be utilized in brain-machine interfaces. A pesar de que la mayoría de los movimientos que realizamos con los miembros superiores sea con los dos brazos, pocos estudios neurofisiológicos han analizado los mecanismos neuronales del control motor bimanual. En particular, sabemos poco acerca del ajuste direccional de las neuronas motoras durante movimientos bimanuales. Esto contrasta con la extensa investigación que se ha realizado sobre el mismo tema para movimientos unimanuales. Georgopoulos y sus colaboradores lideraron la descripción de movimientos unimanuales en términos de curvas de ajuste direccional, donde la actividad neuronal se expresa en función del ángulo del movimiento del brazo. Según Georgopoulos las neuronas del cortex motor se ajustan de manera “amplia” a la dirección del movimiento, siendo máximo el ritmo de activación de la neurona para movimientos realizados en la “dirección preferida” de la neurona. Esta curva de ajuste se aproxima a un coseno. Una descripción de este tipo para movimientos bimanuales no ha sido realizada aún. El objetivo principal del proyecto es abordar esta laguna de conocimiento. Para ello se analizarán grabaciones neuronales de macacos Rhesus durante movimientos bimanuales. Se utilizarán varios modelos de regresión linear múltiple para analizar la relación entre la actividad neuronal y la dirección del movimiento de los brazos. Asimismo se contrastarán los resultados obtenidos con los diferentes métodos. El ajuste direccional de las neuronas durante movimientos unimanuales será comparado con el obtenido para movimientos bimanuales. Específicamente, se estudiará si el ajuste direccional en movimientos bimanuales puede derivarse del ajuste direccional de cada brazo durante movimientos aislados, o si es mayoritariamente independiente de la representación neuronal durante movimientos unimanuales. Los resultados de este trabajo podrán contribuir al desarrollo de mejores algoritmos para descodificar la actividad neuronal, los cuales son una parte esencial de las Interfaces Cerebro-Computadora. Ingeniería Biomédica

Intra-Operative Electron Radiation Therapy (IOERT) is a technique for cancer treatment that delivers a high dose of radiation directly to the post-resected tumour bed or to an unresected tumour during surgery. Recently, a tracking system (OptiTrack, NaturalPoint Inc., USA) was installed into the operating room that allows for navigation and thus assist oncologists in the treatment planning of this procedure. The system is an optical tracking system composed of eight independent cameras, and then a calibration process is required for a correct tracking performance. The current state consists on calibrating just before every IOERT procedure, increasing surgery preparation time by approximately one hour. In order to study the feasibility of defining new calibration protocols based on calibrating every certain period of time, this projects assess the temporal stability of tracking system calibrations and determines an optimum methodology for the detection of miscalibration within the tracking system. To achieve this, software applications were developed to control the Optitrack system by means of the Application Programming Interface provided by the vendor. Also, a group of experiments were performed using developed applications to acquire tracking data. Results show the temporal stability of the tracking system calibration during five days, and the feasibility of detecting tracking system miscalibrations by using the proposed methodology. This study will allow the hospital to save time and resources by stablishing protocols for calibration in which calibration is performed every 2-5 days. The calibration state of the system will be evaluated before every surgery in order to determine if the system needs to be calibrated or not following a proposed methodology (with an average duration of 5 minutes). Ingeniería Biomédica

As far as all the background information about atlases and gene expression databases has already been analysed, now we need to define further the project, its objectives and facts of interest. Gene expression databases, in most cases, do not provide any kind of integration with anatomical information of where those genes are expressed. The characterization of the whole transcriptome for structures like the brain is of limited utility if we have no anatomical information. Combining the databases with the anatomical information provided by an anatomical atlas, we can have lots of advantages. First of all, the most immediate advantage that such integration would introduce would be user-friendliness. Part of this problem is solved with the aGEM tool already developed, that integrates different databases into a single user interface.Visual representation of the gene locations would improve user experience if it is integrated with aGEM. As a second advantage, this integration could facilitate the connection between imaging and gene expression information when defining or analysing results from preclinical experiments. The definition of an imaging protocol in order to study the phenotype of a transgenic animal model could benefit from the results of this project, since the researcher could look for anatomical structures related to the genes that have been manipulated. The results of image quantification are usually an statistical parametric map, that presents the statistical significance of a certain analysis for every voxel. Significant areas from this image could be related to the underlying genes by means of the proposed integration tool. So, the main objective of the project is to connect all the information provided by aGEM and by the atlas. For the issue, it is needed to find how the information is stored and related in aGEM, in order to extract enough information of interest to program a first version of the tool. Also, it is required to study which atlases are available and which one is the most suitable for our purposes. Once all these steps have been done, the kind of program which is going to be developed needs to be analysed. There are several possibilities, like a program in Java, C++, Matlab or a plugin for its integration in ImageJ. Once all the necessary information is extracted, an integration step must be done for the program to be operative. Then, when the information mapping is ready, the interface of the program can be written. The first version of program should be able to perform certain query types: 1. Anatomy query: Given a list of anatomy structures, the user should be able to select any of them and the program would launch a query showing the genes and their information expressed in the structure, which would be shown in the atlas images. 2. Gene query: The user selects a gene and the program would detect in which structures is the gene expressed and show them in the atlas images. The purpose in a first stage is to integrate part of the information contained in aGEM in a beta trial version of the interface, in order to check its real utility. Ingeniería Biomédica

En estudios preclínicos neurológicos de imagen de resonancia magnética (MRI) en pequeños animales es común el uso de la segmentación cerebral para su posterior análisis volumétrico y/o registro con otras modalidades de imagen. Este proceso suele realizarse de forma manual, por lo que a menudo se emplea una gran cantidad de tiempo dependiendo del estudio. En este trabajo se propone un nuevo método de segmentación automática basado en registro para facilitar dicho proceso. La propuesta se ha comparado con dos métodos: segmentación manual, que se emplea como referencia, y una segmentación basada en PCNN (Pulse Couple Neural Network) propuesta específicamente para imágenes de rata. El método propuesto consigue buenos resultados en índice de solapamiento y volumen cerebral comparado con el manual, y ofrece además una reducción considerable en el tiempo de ejecución comparado con PCNN. Ingeniería Técnica en Sistemas de Telecomunicación