The Multimodal Brain TumorImage Segmentation Benchmark (BRATS)

In this paper we report the set-up and results of the Multimodal Brain Tumor Image Segmentation Benchmark (BRATS) organized in conjunction with the MICCAI 2012 and 2013 conferences. Twenty state-of-the-art tumor segmentation algorithms were applied to a set of 65 multi-contrast MR scans of low- and high-grade glioma patients - manually annotated by up to four raters - and to 65 comparable scans generated using tumor image simulation software. Quantitative evaluations revealed considerable disagreement between the human raters in segmenting various tumor sub-regions (Dice scores in the range 74-85%), illustrating the difficulty of this task. We found that different algorithms worked best for different sub-regions (reaching performance comparable to human inter-rater variability), but that no single algorithm ranked in the top for all subregions simultaneously. Fusing several good algorithms using a hierarchical majority vote yielded segmentations that consistently ranked above all individual algorithms, indicating remaining opportunities for further methodological improvements. The BRATS image data and manual annotations continue to be publicly available through an online evaluation system as an ongoing benchmarking resource.

Clinical and pharmacokinetic evaluation of S-ketamine for intravenous general anaesthesia in horses undergoing field castration.

BACKGROUND Intravenous anaesthetic drugs are the primary means for producing general anaesthesia in equine practice. The ideal drug for intravenous anaesthesia has high reliability and pharmacokinetic properties indicating short elimination and lack of accumulation when administered for prolonged periods. Induction of general anaesthesia with racemic ketamine preceded by profound sedation has already an established place in the equine field anaesthesia. Due to potential advantages over racemic ketamine, S-ketamine has been employed in horses to induce general anaesthesia, but its optimal dose remains under investigation. The objective of this study was to evaluate whether 2.5 mg/kg S-ketamine could be used as a single intravenous bolus to provide short-term surgical anaesthesia in colts undergoing surgical castration, and to report its pharmacokinetic profile. RESULTS After premedication with romifidine and L-methadone, the combination of S-ketamine and diazepam allowed reaching surgical anaesthesia in the 28 colts. Induction of anaesthesia as well as recovery were good to excellent in the majority (n = 22 and 24, respectively) of the colts. Seven horses required additional administration of S-ketamine to prolong the duration of surgical anaesthesia. Redosing did not compromise recovery quality. Plasma concentration of S-ketamine decreased rapidly after administration, following a two-compartmental model, leading to the hypothesis of a consistent unchanged elimination of the parent compound into the urine beside its conversion to S-norketamine. The observed plasma concentrations of S-ketamine at the time of first movement were various and did not support the definition of a clear cut-off value to predict the termination of the drug effect. CONCLUSIONS The administration of 2.5 mg/kg IV S-ketamine after adequate premedication provided good quality of induction and recovery and a duration of action similar to what has been reported for racemic ketamine at the dose of 2.2 mg/kg. Until further investigations will be provided, close monitoring to adapt drug delivery is mandatory, particularly once the first 10 minutes after injection are elapsed. Taking into account rapid elimination of S-ketamine, significant inter-individual variability and rapid loss of effect over a narrow range of concentrations a sudden return of consciousness has to be foreseen.

Description and repeatability of a newly developed spinal cord injury scale for dogs.

The objectives of this study were to describe a new spinal cord injury scale for dogs, evaluate repeatability through determining inter-rater variability of scores, compare these scores to another established system (a modified Frankel scale), and determine if the modified Frankel scale and the newly developed scale were useful as prognostic indicators for return to ambulation. A group of client-owned dogs with spinal cord injury were examined by 2 independent observers who applied the new Texas Spinal Cord Injury Score (TSCIS) and a modified Frankel scale that has been used previously. The newly developed scale was designed to describe gait, postural reactions and nociception in each limb. Weighted kappa statistics were utilized to determine inter-rater variability for the modified Frankel scale and individual components of the TSCIS. Comparisons were made between raters for the overall TSCIS score and between scales using Spearman's rho. An additional group of dogs with surgically treated thoracolumbar disk herniation was enrolled to look at correlation of both scores with spinal cord signal characteristics on magnetic resonance imaging (MRI) and ambulatory outcome at discharge. The actual agreement between raters for the modified Frankel scale was 88%, with a weighted kappa value of 0.93. The TSCIS had weighted kappa scores for gait, proprioceptive positioning and nociception components that ranged from 0.72 to 0.94. Correlation between raters for the overall TSCIS score was Spearman's rho=0.99 (P<0.001). Comparison of the overall TSCIS score to the modified Frankel score resulted in a Spearman's rho value of 0.90 (P<0.001). The modified Frankel score was weakly correlated with the length of hyperintensity of the spinal cord: L2 vertebral body length ratio on mid-sagittal T2-weighted MRI (Spearman's rho=-0.45, P=0.042) as was the overall TSCIS score (Spearman's rho=-0.47, P=0.037). There was also a significant difference in admitting modified Frankel scores (P=0.029) and admitting overall TSCIS scores (P=0.02) between dogs that were ambulatory at discharge and those that were not. Results from this study suggest that the TSCIS is an easy to administer scale for evaluating canine spinal cord injury based on the standard neurological exam and correlates well with a previously described modified Frankel scale.

Epidemiologic and methodological issues in serial measurements of blood pressure: The blood pressure study in Mexican children

The Blood Pressure Study in Mexican Children (BPSMC) is a short term longitudinal study of serial blood pressure collected in three observation periods by standardized examinations of 233 female children, 10 to 12 years of age, enrolled in public and private primary schools in Tlalpan, Mexico. Study objectives were: (1) to describe from baseline information the distribution and relationship of blood pressure to age and selected anthropometric factors, as well as to compare the BPSMC results with other blood pressure studies, (2) to examine the sources and amount of variation present in serial blood pressure of 123 children, and (3) to evaluate observer performance by means of intra- and inter-observer variability.^ Stepwise regression results from baseline revealed that of all anthropometric factors and age, weight was the best predictor for blood pressure.^ The results of serial blood pressure measurements show that, besides the known sources of blood pressure variability (subject, day, reading), the physiologic event of menarche has an important bearing upon the variability and characterization of blood pressure in young girls. The assessment of the effects of blood pressure variability and reliability upon the design and analysis of epidemiologic studies, became apparent among post-menarcheal girls; where blood pressure measurements taken from them have low reliability. Research is needed to propose alternatives for assessing blood pressure during puberty.^ Finally, observer performance of blood pressure and anthropometry were evaluated. Anthropometric measurements had reliabilities in excess of R = 0.96. Acceptable reliabilities (R = 0.88 to 0.95) were obtained for systolic and diastolic (phase 4 and 5) blood pressures. The BPSMC showed a 50 percent decrease in measurement error from the first to the third observation periods. ^

THE IMPACT OF FAMILY HISTORY ON MEDULLARY THYROID CANCER IN MEN2A PATIENTS

The American Thyroid Association recently classified all MEN2A-associated codons into increasing risk levels A-C and stated that some patients may delay prophylactic thyroidectomy if certain criteria are met. One criterion is a less aggressive family history of MTC but whether families with the same mutated codon have variable MTC aggressiveness is not well described. We developed several novel measures of MTC aggressiveness and compared families with the same mutated codon to determine if there is significant inter-familial variability. Pedigrees of families with MEN2A were reviewed for codon mutated and proportion of RET mutation carriers with MTC. Individuals with MTC were classified as having local or distant MTC and whether they had progressive MTC. MTC status and age were assessed at diagnosis and most advanced MTC stage. For those without MTC, age was recorded at prophylactic thyroidectomy or last follow-up if the patient did not have a thyroidectomy. For each pedigree, the mean age of members without MTC, with MTC, and the proportion of RET mutation carriers with local or distant and progressive MTC were calculated. We assessed differences in these variables using ANOVA and the Fisher’s exact test. Sufficient data for analysis were available for families with mutated codons 609 (92 patients from 13 families), 618 (41 patients from 7 families), and 634 (152 patients from 13 families). The only significant differences found were the mean age of patients without MTC between families with codon 609 and 618 mutations even after accounting for prophylactic thyroidectomy (p=0.006 and 0.001, respectively), and in the mean age of MTC diagnosis between families with codon 618 and 634 mutations even after accounting for symptomatic presentation (p=0.023 and 0.014, respectively). However, these differences may be explained by generational differences in ascertainment of RET carriers and the availability of genetic testing when the proband initially presented.

Sediment volume variations inferred from DGPS measurements within Bossons glacier proglacial area (Mont-Blanc massif, France)

This dataset presents Differential Global Positioning System data (DGPS) acquired within the Bossons glacier proglacial area. Bossons glacier is a rapidly retreating glacier and its proglacial area is deglaciated for ~30 years. Bossons stream is one of the outlets of the subglacial drainage system. It starts as a 800 m steep cascade reach, then flows through an area with gentler slope : the Plan des Eaux (PdE). PdE is a 300 m long, 50 m wide proglacial alluvial plain with an increasing channel mobility in the downstream direction but decreasing slope gradient and incision. As it may act a sediment trap, studying periglacial and proglacial erosion processes in the Bossons catchment requires to quantify PdE sediment volume evolution. A several meter-sized block located within Bossons proglacial area was set up as GPS base : its location was measured by one antenna (Topcon Hyper Pro) by performing 600 consecutive measurements throughout one day. A second antenna (Topcon Hyper Pro) was then used to measure XYZ location of points in the proglacial area with a ~2 m grid. Radio communication between the two antennas allowed differential calculations to be automatically carried out on field using the Topcon FC-250 hand controller. This methodology yields 3 cm XY and 1.5 cm Z uncertainties. DGPS data have been acquired through 10 campaigns from 2004 to 2014; campaigns from 2004 to 2008 cover a smaller area than those from 2010 to 2014. Digital Elevation Model (DEM) have been interpolated from DGPS data and difference between two DEMs yields deposited and eroded volume within PdE. Maps of PdE volume variation between two campaigns show that incision mainly occurs in the upper and lower sections where as deposition dominates in the middle section. Deposition, denudation and net rate (deposition rate - denudation rate) are calculated by normalizing volumes by DEM areas. Deposition dominates results with a mean net rate of 29 mm/yr. However, strong inter-annual variability exists and some years are dominated by denudation : -36 mm/yr and -100 mm/yr for 2006 and 2011, respectively. Nonetheless, oldest campaigns (2004 to 2008) were carried out on the lower part part of the alluvial plain and ruling them out to keep only complete DEM (2010 to 2014) yields a mean net rate of ~15 mm/yr. This results is coherent with field observations of both strong deposition (e.g. flood deposits) and strong erosion (e.g. 30 cm incision) evidences. Bossons glacier proglacial area is thus dynamic with year-to-year geormorphological changes but may leans toward increasing its mean elevation through a deposition dominated system.

Zooplankton biovolume time series at station STARESO (Bay of Calvi, Corsica)

A long-term time series of subsurface zooplankton is performed in the oligotrophic Bay of Calvi (Corsica, Ligurian Sea, NW Mediterranean) from 2003 onwards. It is carried out from the marine station STARESO and is based on weekly measurement of zooplankton biovolumes. The main objectives are (i) to determine the seasonal cycle and inter-annual variability of the organisms, (ii) to study the dynamics of the populations, and (iii) to understand zooplankton interactions with other co-sampled hydrographic, meteorological and phytoplankton variables at the site.

(Table 1) Characteristics of South Pole personnel according to forced expiratory flow (FEF25-75) response to altitude

The impact of acute altitude exposure on pulmonary function is variable. A large inter-individual variability in the changes in forced expiratory flows (FEFs) is reported with acute exposure to altitude, which is suggested to represent an interaction between several factors influencing bronchial tone such as changes in gas density, catecholamine stimulation, and mild interstitial edema. This study examined the association between FEF variability, acute mountain sickness (AMS) and various blood markers affecting bronchial tone (endothelin-1, vascular endothelial growth factor (VEGF), catecholamines, angiotensin II) in 102 individuals rapidly transported to the South Pole (2835 m). The mean FEF between 25 and 75% (FEF25-75) and blood markers were recorded at sea level and after the second night at altitude. AMS was assessed using Lake Louise questionnaires. FEF25-75 increased by an average of 12% with changes ranging from -26 to +59% from sea level to altitude. On the second day, AMS incidence was 36% and was higher in individuals with increases in FEF25-75 (41 vs. 22%, P = 0.05). Ascent to altitude induced an increase in endothelin-1 levels, with greater levels observed in individuals with decreased FEF25-75. Epinephrine levels increased with ascent to altitude and the response was six times larger in individuals with decreased FEF25-75. Greater levels of endothelin-1 in individuals with decreased FEF25-75 suggest a response consistent with pulmonary hypertension and/or mild interstitial edema, while epinephrine may be upregulated in these individuals to clear lung fluid through stimulation of beta2-adrenergic receptors.

Ocean acidification impacts on sperm mitochondrial membrane potential bring sperm swimming behaviour near its tipping point

Broadcast spawning marine invertebrates are susceptible to environmental stressors such as climate change, as their reproduction depends on the successful meeting and fertilization of gametes in the water column. Under near-future scenarios of ocean acidification, the swimming behaviour of marine invertebrate sperm is altered. We tested whether this was due to changes in sperm mitochondrial activity by investigating the effects of ocean acidification on sperm metabolism and swimming behaviour in the sea urchin Centrostephanus rodgersii. We used a fluorescent molecular probe (JC-1) and flow cytometry to visualize mitochondrial activity (measured as change in mitochondrial membrane potential, MMP). Sperm MMP was significantly reduced in delta pH -0.3 (35% reduction) and delta pH -0.5 (48% reduction) treatments, whereas sperm swimming behaviour was less sensitive with only slight changes (up to 11% decrease) observed overall. There was significant inter-individual variability in responses of sperm swimming behaviour and MMP to acidified seawater. We suggest it is likely that sperm exposed to these changes in pH are close to their tipping point in terms of physiological tolerance to acidity. Importantly, substantial inter-individual variation in responses of sperm swimming to ocean acidification may increase the scope for selection of resilient phenotypes, which, if heritable, could provide a basis for adaptation to future ocean acidification.

(Table 2) Carbon and nitrogen isotopic ratios in blood sampled from arctic foxes (Vulpes lagopus) on Bylot Island between 2003-2008

Inter-individual variation in diet within generalist animal populations is thought to be a widespread phenomenon but its potential causes are poorly known. Inter-individual variation can be amplified by the availability and use of allochthonous resources, i.e., resources coming from spatially distinct ecosystems. Using a wild population of arctic fox as a study model, we tested hypotheses that could explain variation in both population and individual isotopic niches, used here as proxy for the trophic niche. The arctic fox is an opportunistic forager, dwelling in terrestrial and marine environments characterized by strong spatial (arctic-nesting birds) and temporal (cyclic lemmings) fluctuations in resource abundance. First, we tested the hypothesis that generalist foraging habits, in association with temporal variation in prey accessibility, should induce temporal changes in isotopic niche width and diet. Second, we investigated whether within-population variation in the isotopic niche could be explained by individual characteristics (sex and breeding status) and environmental factors (spatiotemporal variation in prey availability). We addressed these questions using isotopic analysis and Bayesian mixing models in conjunction with linear mixed-effects models. We found that: i) arctic fox populations can simultaneously undergo short-term (i.e., within a few months) reduction in both isotopic niche width and inter-individual variability in isotopic ratios, ii) individual isotopic ratios were higher and more representative of a marine-based diet for non-breeding than breeding foxes early in spring, and iii) lemming population cycles did not appear to directly influence the diet of individual foxes after taking their breeding status into account. However, lemming abundance was correlated to proportion of breeding foxes, and could thus indirectly affect the diet at the population scale.

Calcium isotope fractionation in cultured, modern and fossil scleractinian coral skeleton

The present study investigates the influence of environmental (temperature, salinity) and biological (growth rate, inter-generic variations) parameters on calcium isotope fractionation (d44/40Ca) in scleractinian coral skeleton to better constrain this record. Previous studies focused on the d44/40Ca record in different marine organisms to reconstruct seawater composition or temperature, but only few studies investigated corals. This study presents measurements performed on modern corals from natural environments (from the Maldives for modern and from Tahiti for fossil corals) as well as from laboratory cultures (Centre Scientifique de Monaco). Measurements on Porites sp., Acropora sp., Montipora verrucosa and Stylophora pistillata allow constraining inter-generic variability. Our results show that the fractionation of d44/40Ca ranges from 0.6 to 0.1 per mil, independent of the genus or the environmental conditions. No significant relationship between the rate of calcification and d44/40Ca was found. The weak temperature dependence reported in earlier studies is most probably not the only parameter that is responsible for the fractionation. Indeed, sub-seasonal temperature variations reconstructed by d18O and Sr/Ca ratio using a multi-proxy approach, are not mirrored in the coral's d44/40Ca variations. The intergeneric variability and intrageneric variability among the studied samples are weak except for S. pistillata, which shows calcium isotopic values increasing with salinity. The variability between samples cultured at a salinity of 40 is higher than those cultured at a salinity of 36 for this species. The present study reveals a strong biological control of the skeletal calcium isotope composition by the polyp and a weak influence of environmental factors, specifically temperature and salinity (except for S. pistillata). Vital effects have to be investigated in situ to better constrain their influence on the calcium isotopic signal. If vital effects could be extracted from the isotopic signal, the calcium isotopic composition of coral skeletons could provide reliable information on the calcium composition and budget in ocean.

Abundance and biomass of in- and epibenthic invertebrate macrofauna in the German Bight from 1969 to 2000

In order to examine the long-term development of offshore macrozoobenthic soft-bottom communities of the German Bight, four representative permanent stations (MZB-SSd, -FSd, -Slt, -WB) have been sampled continuously since 1969. Inter-annual variability and possible long-term trends were analysed based on spring-time samples from 1969 until 2000. This is part of the ecological long-term series of the AWI and is supplemented by periodic large-scale mapping of the benthos. The main factors influencing the development of the benthic communities are biological interactions, climate, food supply (eutrophication) and the disturbance regime. The most frequent disturbances are sediment relocations during strong storms or by bottom trawling, while occasional oxygen deficiencies and extremely cold winters are important disturbance events working on a much larger scale. Benthic communities at the sampling stations show a large inter-annual variability combined with a variation on a roughly decadal scale. In accordance with large-scale system shifts reported for the North Sea, benthic community transitions occurred between roughly the 1970ies, 80ies and 90ies. The transitions between periods are not distinctly marked by strong changes but rather reflected in gradual changes of the species composition and dominance structure.

Daily pan-Arctic sea-ice lead maps for 2003-2015, with links to maps in netCDF format

The presence of sea-ice leads represents a key feature of the Arctic sea ice cover. Leads promote the flux of sensible and latent heat from the ocean to the cold winter atmosphere and are thereby crucial for air-sea-ice-ocean interactions. We here apply a binary segmentation procedure to identify leads from MODIS thermal infrared imagery on a daily time scale. The method separates identified leads into two uncertainty categories, with the high uncertainty being attributed to artifacts that arise from warm signatures of unrecognized clouds. Based on the obtained lead detections, we compute quasi-daily pan-Arctic lead maps for the months of January to April, 2003-2015. Our results highlight the marginal ice zone in the Fram Strait and Barents Sea as the primary region for lead activity. The spatial distribution of the average pan-Arctic lead frequencies reveals, moreover, distinct patterns of predominant fracture zones in the Beaufort Sea and along the shelf-breaks, mainly in the Siberian sector of the Arctic Ocean as well as the well-known polynya and fast-ice locations. Additionally, a substantial inter-annual variability of lead occurrences in the Arctic is indicated.

Utility of thallium and geochemical analysis of Mariana arc sediments from ODP Hole 129-801A and 129-801B

We provide the first exploration of thallium (Tl) abundances and stable isotope compositions as potential tracers during arc lava genesis. We present a case study of lavas from the Central Island Province (CIP) of the Mariana arc, supplemented by representative sedimentary and altered oceanic crust (AOC) inputs from ODP Leg 129 Hole 801 outboard of the Mariana trench. Given the large Tl concentration contrast between the mantle and subduction inputs coupled with previously published distinctive Tl isotope signatures of sediment and AOC, the Tl isotope system has great potential to distinguish different inputs to arc lavas. Furthermore, CIP lavas have well-established inter island variability, providing excellent context for the examination of Tl as a new stable isotope tracer. In contrast to previous work (Nielsen et al., 2006b), we do not observe Tl enrichment or light epsilon 205Tl (where epsilon 205Tl is the deviation in parts per 10,000 of a sample 205Tl/203Tl ratio compared to NIST SRM 997 Tl standard) in the Jurassic-aged altered mafic ocean crust subducting outboard of the Marianas (epsilon 205Tl = - 4.4 to 0). The lack of a distinctive epsilon 205Tl signature may be related to secular changes in ocean chemistry. Sediments representative of the major lithologies from ODP Hole Leg 129 801 have 1-2 orders of magnitude of Tl enrichment compared to the CIP lavas, but do not record heavy signatures (epsilon 205Tl = - 3.0 to + 0.4), as previously found in similar sediment types (epsilon 205Tl > + 2.5; Rehkämper et al., 2004). We find a restricted range of epsilon 205Tl = - 1.8 to - 0.4 in CIP lavas, which overlaps with MORB. One lava from Guguan falls outside this range with epsilon 205Tl = + 1.2. Coupled Cs, Tl and Pb systematics of Guguan lavas suggests that this heavy Tl isotope composition may be due to preferential degassing of isotopically light Tl. In general, the low Tl concentrations and limited isotopic range in the CIP lavas is likely due to the unexpectedly narrow range of epsilon 205Tl found in Mariana subduction inputs, coupled with volcaniclastic, rather than pelagic sediment as the dominant source of Tl. Much work remains to better understand the controls on Tl processing through a subduction zone. For example, Tl could be retained in residual phengite, offering the potential exploration of Cs/Tl ratios as a slab thermometer. However, data for Tl partitioning in phengite (and other micas) is required before developing this application further. Establishing a database of Tl concentrations and stable isotopes in subduction zone lavas with different thermal parameters and sedimentary inputs is required for the future use of Tl as a subduction zone tracer.

Evaluación de la contribución y el impacto de las tecnologías del habla en la detección automática del Síndrome de la Apnea Obstructiva del Sueño = Contributions and impact assessment of speech technologies on the automatic detection of severe Obstructive Sleep Apnea syndrome

La presente Tesis analiza las posibilidades que ofrecen en la actualidad las tecnologías del habla para la detección de patologías clínicas asociadas a la vía aérea superior. El estudio del habla que tradicionalmente cubre tanto la producción como el proceso de transformación del mensaje y las señales involucradas, desde el emisor hasta alcanzar al receptor, ofrece una vía de estudio alternativa para estas patologías. El hecho de que la señal emitida no solo contiene este mensaje, sino también información acerca del locutor, ha motivado el desarrollo de sistemas orientados a la identificación y verificación de la identidad de los locutores. Estos trabajos han recibido recientemente un nuevo impulso, orientándose tanto hacia la caracterización de rasgos que son comunes a varios locutores, como a las diferencias existentes entre grabaciones de un mismo locutor. Los primeros resultan especialmente relevantes para esta Tesis dado que estos rasgos podrían evidenciar la presencia de características relacionadas con una cierta condición común a varios locutores, independiente de su identidad. Tal es el caso que se enfrenta en esta Tesis, donde los rasgos identificados se relacionarían con una de la patología particular y directamente vinculada con el sistema de físico de conformación del habla. El caso del Síndrome de Apneas Hipopneas durante el Sueno (SAHS) resulta paradigmático. Se trata de una patología con una elevada prevalencia mundo, que aumenta con la edad. Los pacientes de esta patología experimentan episodios de cese involuntario de la respiración durante el sueño, que se prolongan durante varios segundos y que se reproducen a lo largo de la noche impidiendo el correcto descanso. En el caso de la apnea obstructiva, estos episodios se deben a la imposibilidad de mantener un camino abierto a través de la vía aérea, de forma que el flujo de aire se ve interrumpido. En la actualidad, el diagnostico de estos pacientes se realiza a través de un estudio polisomnográfico, que se centra en el análisis de los episodios de apnea durante el sueño, requiriendo que el paciente permanezca en el hospital durante una noche. La complejidad y el elevado coste de estos procedimientos, unidos a las crecientes listas de espera, han evidenciado la necesidad de contar con técnicas rápidas de detección, que si bien podrían no obtener tasas tan elevadas, permitirían reorganizar las listas de espera en función del grado de severidad de la patología en cada paciente. Entre otros, los sistemas de diagnostico por imagen, así como la caracterización antropométrica de los pacientes, han evidenciado la existencia de patrones anatómicos que tendrían influencia directa sobre el habla. Los trabajos dedicados al estudio del SAHS en lo relativo a como esta afecta al habla han sido escasos y algunos de ellos incluso contradictorios. Sin embargo, desde finales de la década de 1980 se conoce la existencia de patrones específicos relativos a la articulación, la fonación y la resonancia. Sin embargo, su descripción resultaba difícilmente aprovechable a través de un sistema de reconocimiento automático, pero apuntaba la existencia de un nexo entre voz y SAHS. En los últimos anos las técnicas de procesado automático han permitido el desarrollo de sistemas automáticos que ya son capaces de identificar diferencias significativas en el habla de los pacientes del SAHS, y que los distinguen de los locutores sanos. Por contra, poco se conoce acerca de la conexión entre estos nuevos resultados, los sé que habían obtenido en el pasado y la patogénesis del SAHS. Esta Tesis continua la labor desarrollada en este ámbito considerando específicamente: el estudio de la forma en que el SAHS afecta el habla de los pacientes, la mejora en las tasas de clasificación automática y la combinación de la información obtenida con los predictores utilizados por los especialistas clínicos en sus evaluaciones preliminares. Las dos primeras tareas plantean problemas simbióticos, pero diferentes. Mientras el estudio de la conexión entre el SAHS y el habla requiere de modelos acotados que puedan ser interpretados con facilidad, los sistemas de reconocimiento se sirven de un elevado número de dimensiones para la caracterización y posterior identificación de patrones. Así, la primera tarea debe permitirnos avanzar en la segunda, al igual que la incorporación de los predictores utilizados por los especialistas clínicos. La Tesis aborda el estudio tanto del habla continua como del habla sostenida, con el fin de aprovechar las sinergias y diferencias existentes entre ambas. En el análisis del habla continua se tomo como punto de partida un esquema que ya fue evaluado con anterioridad, y sobre el cual se ha tratado la evaluación y optimización de la representación del habla, así como la caracterización de los patrones específicos asociados al SAHS. Ello ha evidenciado la conexión entre el SAHS y los elementos fundamentales de la señal de voz: los formantes. Los resultados obtenidos demuestran que el éxito de estos sistemas se debe, fundamentalmente, a la capacidad de estas representaciones para describir dichas componentes, obviando las dimensiones ruidosas o con poca capacidad discriminativa. El esquema resultante ofrece una tasa de error por debajo del 18%, sirviéndose de clasificadores notablemente menos complejos que los descritos en el estado del arte y de una única grabación de voz de corta duración. En relación a la conexión entre el SAHS y los patrones observados, fue necesario considerar las diferencias inter- e intra-grupo, centrándonos en la articulación característica del locutor, sustituyendo los complejos modelos de clasificación por el estudio de los promedios espectrales. El resultado apunta con claridad hacia ciertas regiones del eje de frecuencias, sugiriendo la existencia de un estrechamiento sistemático en la sección del tracto en la región de la orofaringe, ya prevista en la patogénesis de este síndrome. En cuanto al habla sostenida, se han reproducido los estudios realizados sobre el habla continua en grabaciones de la vocal /a/ sostenida. Los resultados son cualitativamente análogos a los anteriores, si bien en este caso las tasas de clasificación resultan ser más bajas. Con el objetivo de identificar el sentido de este resultado se reprodujo el estudio de los promedios espectrales y de la variabilidad inter e intra-grupo. Ambos estudios mostraron importantes diferencias con los anteriores que podrían explicar estos resultados. Sin embargo, el habla sostenida ofrece otras oportunidades al establecer un entorno controlado para el estudio de la fonación, que también había sido identificada como una fuente de información para la detección del SAHS. De su estudio se pudo observar que, en el conjunto de datos disponibles, no existen variaciones que pudieran asociarse fácilmente con la fonación. Únicamente aquellas dimensiones que describen la distribución de energía a lo largo del eje de frecuencia evidenciaron diferencias significativas, apuntando, una vez más, en la dirección de las resonancias espectrales. Analizados los resultados anteriores, la Tesis afronta la fusión de ambas fuentes de información en un único sistema de clasificación. Con ello es posible mejorar las tasas de clasificación, bajo la hipótesis de que la información presente en el habla continua y el habla sostenida es fundamentalmente distinta. Esta tarea se realizo a través de un sencillo esquema de fusión que obtuvo un 88.6% de aciertos en clasificación (tasa de error del 11.4%), lo que representa una mejora significativa respecto al estado del arte. Finalmente, la combinación de este clasificador con los predictores utilizados por los especialistas clínicos ofreció una tasa del 91.3% (tasa de error de 8.7%), que se encuentra dentro del margen ofrecido por esquemas más costosos e intrusivos, y que a diferencia del propuesto, no pueden ser utilizados en la evaluación previa de los pacientes. Con todo, la Tesis ofrece una visión clara sobre la relación entre el SAHS y el habla, evidenciando el grado de madurez alcanzado por la tecnología del habla en la caracterización y detección del SAHS, poniendo de manifiesto que su uso para la evaluación de los pacientes ya sería posible, y dejando la puerta abierta a futuras investigaciones que continúen el trabajo aquí iniciado. ABSTRACT This Thesis explores the potential of speech technologies for the detection of clinical disorders connected to the upper airway. The study of speech traditionally covers both the production process and post processing of the signals involved, from the speaker up to the listener, offering an alternative path to study these pathologies. The fact that utterances embed not just the encoded message but also information about the speaker, has motivated the development of automatic systems oriented to the identification and verificaton the speaker’s identity. These have recently been boosted and reoriented either towards the characterization of traits that are common to several speakers, or to the differences between records of the same speaker collected under different conditions. The first are particularly relevant to this Thesis as these patterns could reveal the presence of features that are related to a common condition shared among different speakers, regardless of their identity. Such is the case faced in this Thesis, where the traits identified would relate to a particular pathology, directly connected to the speech production system. The Obstructive Sleep Apnea syndrome (OSA) is a paradigmatic case for analysis. It is a disorder with high prevalence among adults and affecting a larger number of them as they grow older. Patients suffering from this disorder experience episodes of involuntary cessation of breath during sleep that may last a few seconds and reproduce throughout the night, preventing proper rest. In the case of obstructive apnea, these episodes are related to the collapse of the pharynx, which interrupts the air flow. Currently, OSA diagnosis is done through a polysomnographic study, which focuses on the analysis of apnea episodes during sleep, requiring the patient to stay at the hospital for the whole night. The complexity and high cost of the procedures involved, combined with the waiting lists, have evidenced the need for screening techniques, which perhaps would not achieve outstanding performance rates but would allow clinicians to reorganize these lists ranking patients according to the severity of their condition. Among others, imaging diagnosis and anthropometric characterization of patients have evidenced the existence of anatomical patterns related to OSA that have direct influence on speech. Contributions devoted to the study of how this disorder affects scpeech are scarce and somehow contradictory. However, since the late 1980s the existence of specific patterns related to articulation, phonation and resonance is known. By that time these descriptions were virtually useless when coming to the development of an automatic system, but pointed out the existence of a link between speech and OSA. In recent years automatic processing techniques have evolved and are now able to identify significant differences in the speech of OSAS patients when compared to records from healthy subjects. Nevertheless, little is known about the connection between these new results with those published in the past and the pathogenesis of the OSA syndrome. This Thesis is aimed to progress beyond the previous research done in this area by addressing: the study of how OSA affects patients’ speech, the enhancement of automatic OSA classification based on speech analysis, and its integration with the information embedded in the predictors generally used by clinicians in preliminary patients’ examination. The first two tasks, though may appear symbiotic at first, are quite different. While studying the connection between speech and OSA requires simple narrow models that can be easily interpreted, classification requires larger models including a large number dimensions for the characterization and posterior identification of the observed patterns. Anyhow, it is clear that any progress made in the first task should allow us to improve our performance on the second one, and that the incorporation of the predictors used by clinicians shall contribute in this same direction. The Thesis considers both continuous and sustained speech analysis, to exploit the synergies and differences between them. On continuous speech analysis, a conventional speech processing scheme, designed and evaluated before this Thesis, was taken as a baseline. Over this initial system several alternative representations of the speech information were proposed, optimized and tested to select those more suitable for the characterization of OSA-specific patterns. Evidences were found on the existence of a connection between OSA and the fundamental constituents of the speech: the formants. Experimental results proved that the success of the proposed solution is well explained by the ability of speech representations to describe these specific OSA-related components, ignoring the noisy ones as well those presenting low discrimination capabilities. The resulting scheme obtained a 18% error rate, on a classification scheme significantly less complex than those described in the literature and operating on a single speech record. Regarding the connection between OSA and the observed patterns, it was necessary to consider inter-and intra-group differences for this analysis, and to focus on the articulation, replacing the complex classification models by the long-term average spectra. Results clearly point to certain regions on the frequency axis, suggesting the existence of a systematic narrowing in the vocal tract section at the oropharynx. This was already described in the pathogenesis of this syndrome. Regarding sustained speech, similar experiments as those conducted on continuous speech were reproduced on sustained phonations of vowel / a /. Results were qualitatively similar to the previous ones, though in this case perfomance rates were found to be noticeably lower. Trying to derive further knowledge from this result, experiments on the long-term average spectra and intraand inter-group variability ratios were also reproduced on sustained speech records. Results on both experiments showed significant differences from the previous ones obtained from continuous speech which could explain the differences observed on peformance. However, sustained speech also provided the opportunity to study phonation within the controlled framework it provides. This was also identified in the literature as a source of information for the detection of OSA. In this study it was found that, for the available dataset, no sistematic differences related to phonation could be found between the two groups of speakers. Only those dimensions which relate energy distribution along the frequency axis provided significant differences, pointing once again towards the direction of resonant components. Once classification schemes on both continuous and sustained speech were developed, the Thesis addressed their combination into a single classification system. Under the assumption that the information in continuous and sustained speech is fundamentally different, it should be possible to successfully merge the two of them. This was tested through a simple fusion scheme which obtained a 88.6% correct classification (11.4% error rate), which represents a significant improvement over the state of the art. Finally, the combination of this classifier with the variables used by clinicians obtained a 91.3% accuracy (8.7% error rate). This is within the range of alternative, but costly and intrusive schemes, which unlike the one proposed can not be used in the preliminary assessment of patients’ condition. In the end, this Thesis has shed new light on the underlying connection between OSA and speech, and evidenced the degree of maturity reached by speech technology on OSA characterization and detection, leaving the door open for future research which shall continue in the multiple directions that have been pointed out and left as future work.