Advanced geostatistical and machine-learning models for spatial data analysis of radioactively contaminated regions

Radioactive soil-contamination mapping and risk assessment is a vital issue for decision makers. Traditional approaches for mapping the spatial concentration of radionuclides employ various regression-based models, which usually provide a single-value prediction realization accompanied (in some cases) by estimation error. Such approaches do not provide the capability for rigorous uncertainty quantification or probabilistic mapping. Machine learning is a recent and fast-developing approach based on learning patterns and information from data. Artificial neural networks for prediction mapping have been especially powerful in combination with spatial statistics. A data-driven approach provides the opportunity to integrate additional relevant information about spatial phenomena into a prediction model for more accurate spatial estimates and associated uncertainty. Machine-learning algorithms can also be used for a wider spectrum of problems than before: classification, probability density estimation, and so forth. Stochastic simulations are used to model spatial variability and uncertainty. Unlike regression models, they provide multiple realizations of a particular spatial pattern that allow uncertainty and risk quantification. This paper reviews the most recent methods of spatial data analysis, prediction, and risk mapping, based on machine learning and stochastic simulations in comparison with more traditional regression models. The radioactive fallout from the Chernobyl Nuclear Power Plant accident is used to illustrate the application of the models for prediction and classification problems. This fallout is a unique case study that provides the challenging task of analyzing huge amounts of data ('hard' direct measurements, as well as supplementary information and expert estimates) and solving particular decision-oriented problems.

Expression of MYC, IgM, As Well As Non-Germinal Centre B-Cell Like Immunophenotype and Positive Immunofish Index Predict a Worse Progression Free Survival and Overall Survival in a Series of 670 De Novo Diffuse Large B-Cell Lymphomas Included in Clinical Trials: A GELA Study of the 2003 Program

Introduction: Diffuse large B-cell lymphomas (DLBCL) represent a heterogeneous disease with variable clinical outcome. Identifying phenotypic biomarkers of tumor cells on paraffin sections that predict different clinical outcome remain an important goal that may also help to better understand the biology of this lymphoma. Differentiating non-germinal centre B-cell-like (non-GCB) from Germinal Centre B-cell-like (GCB) DLBCL according to Hans algorithm has been considered as an important immunohistochemical biomarker with prognostic value among patients treated with R-CHOP although not reproducibly found by all groups. Gene expression studies have also shown that IgM expression might be used as a surrogate for the GCB and ABC subtypes with a strong preferential expression of IgM in ABC DLBCL subtype. ImmunoFISH index based on the differential expression of MUM-1, FOXP1 by immunohistochemistry and on the BCL6 rearrangement by FISH has been previously reported (C Copie-Bergman, J Clin Oncol. 2009;27:5573-9) as prognostic in an homogeneous series of DLBCL treated with R-CHOP. In addition, oncogenic MYC protein overexpression by immunohistochemistry may represent an easy tool to identify the consequences of MYC deregulation in DLBCL. Our aim was to analyse by immunohistochemistry the prognostic relevance of MYC, IgM, GCB/nonGCB subtype and ImmunoFISH index in a large series of de novo DLBCL treated with Rituximab (R)-chemotherapy (anthracyclin based) included in the 2003 program of the Groupe d'Etude des Lymphomes de l'Adulte (GELA) trials. Methods: The 2003 program included patients with de novo CD20+ DLBCL enrolled in 6 different LNH-03 GELA trials (LNH-03-1B, -B, -3B, 39B, -6B, 7B) stratifying patients according to age and age-adjusted IPI. Tumor samples were analyzed by immunohistochemistry using CD10, BCL6, MUM1, FOXP1 (according to Barrans threshold), MYC, IgM antibodies on tissue microarrays and by FISH using BCL6 split signal DNA probes. Considering evaluable Hans score, 670 patients were included in the study with 237 (35.4%) receiving intensive R-ACVBP regimen and 433 (64.6%) R-CHOP/R-mini-CHOP. Results: 304 (45.4%) DLBCL were classified as GCB and 366 (54.6%) as non-GCB according to Hans algorithm. 337/567 cases (59.4%) were positive for the ImmunoFISH index (i.e. two out of the three markers positive: MUM1 protein positive, FOXP1 protein Variable or Strong, BCL6 rearrangement). Immunofish index was preferentially positive in the non-GCB subtype (81.3%) compared to the GCB subtype (31.2%), (p<0.001). IgM was recorded as positive in tumor cells in 351/637 (52.4%) DLBCL cases with a preferential expression in non-GCB 195 (53.3%) vs GCB subtype 100(32.9%), p<0.001). MYC was positive in 170/577 (29.5%) cases with a 40% cut-off and in 44/577 (14.2%) cases with a cut-off of 70%. There was no preferential expression of MYC among GCB or non-GCB subtype (p>0.4) for both cut-offs. Progression-free Survival (PFS) was significantly worse among patients with high IPI score (p<0.0001), IgM positive tumor (p<0.0001), MYC positive tumor with a 40% threshold (p<0.001), ImmunoFISH positive index (p<0.002), non-GCB DLBCL subtype (p<0.0001). Overall Survival (OS) was also significantly worse among patients with high IPI score (p<0.0001), IgM positive tumor (p=0.02), MYC positive tumor with a 40% threshold (p<0.01), ImmunoFISH positive index (p=0.02), non-GCB DLBCL subtype (p<0.0001). All significant parameters were included in a multivariate analysis using Cox Model and in addition to IPI, only the GCB/non-GCB subtype according to Hans algorithm predicted significantly a worse PFS among non-GCB subgroup (HR 1.9 [1.3-2.8] p=0.002) as well as a worse OS (HR 2.0 [1.3-3.2], p=0.003). This strong prognostic value of non-GCB subtyping was confirmed considering only patients treated with R- CHOP for PFS (HR 2.1 [1.4-3.3], p=0.001) and for OS (HR 2.3 [1.3-3.8], p=0.002). Conclusion: Our study on a large series of patients included in trials confirmed the relevance of immunohistochemistry as a useful tool to identify significant prognostic biomarkers for clinical use. We show here that IgM and MYC might be useful prognostic biomarkers. In addition, we confirmed in this series the prognostic value of the ImmunoFISH index. Above all, we fully validated the strong and independent prognostic value of the Hans algorithm, daily used by the pathologists to subtype DLBCL.

Applying machine learning methods to avalanche forecasting

Avalanche forecasting is a complex process involving the assimilation of multiple data sources to make predictions over varying spatial and temporal resolutions. Numerically assisted forecasting often uses nearest neighbour methods (NN), which are known to have limitations when dealing with high dimensional data. We apply Support Vector Machines to a dataset from Lochaber, Scotland to assess their applicability in avalanche forecasting. Support Vector Machines (SVMs) belong to a family of theoretically based techniques from machine learning and are designed to deal with high dimensional data. Initial experiments showed that SVMs gave results which were comparable with NN for categorical and probabilistic forecasts. Experiments utilising the ability of SVMs to deal with high dimensionality in producing a spatial forecast show promise, but require further work.

De la perception à l'apprentissage [From perception to learning]

Dans le domaine de la perception, l'apprentissage est contraint par la présence d'une architecture fonctionnelle constituée d'aires corticales distribuées et très spécialisées. Dans le domaine des troubles visuels d'origine cérébrale, l'apprentissage d'un patient hémi-anopsique ou agnosique sera limité par ses capacités perceptives résiduelles, mais un déficit de reconnaissance visuelle de nature apparemment perceptive, peut également être associé à une altération des représentations en mémoire à long terme. Des réseaux neuronaux distincts pour la reconnaissance - cortex temporal - et pour la localisation des sons - cortex pariétal - ont été décrits chez l'homme. L'étude de patients cérébro-lésés confirme le rôle des indices spatiaux dans un traitement auditif explicite du « where » et dans la discrimination implicite du « what ». Cette organisation, similaire à ce qui a été décrit dans la modalité visuelle, faciliterait les apprentissages perceptifs. Plus généralement, l'apprentissage implicite fonde une grande partie de nos connaissances sur le monde en nous rendant sensible, à notre insu, aux règles et régularités de notre environnement. Il serait impliqué dans le développement cognitif, la formation des réactions émotionnelles ou encore l'apprentissage par le jeune enfant de sa langue maternelle. Le caractère inconscient de cet apprentissage est confirmé par l'étude des temps de réaction sériels de patients amnésiques dans l'acquisition d'une grammaire artificielle. Son évaluation pourrait être déterminante dans la prise en charge ré-adaptative. [In the field of perception, learning is formed by a distributed functional architecture of very specialized cortical areas. For example, capacities of learning in patients with visual deficits - hemianopia or visual agnosia - from cerebral lesions are limited by perceptual abilities. Moreover a visual deficit in link with abnormal perception may be associated with an alteration of representations in long term (semantic) memory. Furthermore, perception and memory traces rely on parallel processing. This has been recently demonstrated for human audition. Activation studies in normal subjects and psychophysical investigations in patients with focal hemispheric lesions have shown that auditory information relevant to sound recognition and that relevant to sound localisation are processed in parallel, anatomically distinct cortical networks, often referred to as the "What" and "Where" processing streams. Parallel processing may appear counterintuitive from the point of view of a unified perception of the auditory world, but there are advantages, such as rapidity of processing within a single stream, its adaptability in perceptual learning or facility of multisensory interactions. More generally, implicit learning mechanisms are responsible for the non-conscious acquisition of a great part of our knowledge about the world, using our sensitivity to the rules and regularities structuring our environment. Implicit learning is involved in cognitive development, in the generation of emotional processing and in the acquisition of natural language. Preserved implicit learning abilities have been shown in amnesic patients with paradigms like serial reaction time and artificial grammar learning tasks, confirming that implicit learning mechanisms are not sustained by the cognitive processes and the brain structures that are damaged in amnesia. In a clinical perspective, the assessment of implicit learning abilities in amnesic patients could be critical for building adapted neuropsychological rehabilitation programs.]

Detection of HIV drug resistance during antiretroviral treatment and clinical progression in a large European cohort study.

OBJECTIVE(S): To investigate the relationship between detection of HIV drug resistance by 2 years from starting antiretroviral therapy and the subsequent risk of progression to AIDS and death. DESIGN: Virological failure was defined as experiencing two consecutive viral loads of more than 400 copies/ml in the time window between 0.5 and 2 years from starting antiretroviral therapy (baseline). Patients were grouped according to evidence of virological failure and whether there was detection of the International AIDS Society resistance mutations to one, two or three drug classes in the time window. METHODS: Standard survival analysis using Kaplan-Meier curves and Cox proportional hazards regression model with time-fixed covariates defined at baseline was employed. RESULTS: We studied 8229 patients in EuroSIDA who started antiretroviral therapy and who had at least 2 years of clinical follow-up. We observed 829 AIDS events and 571 deaths during 38,814 person-years of follow-up resulting in an overall incidence of new AIDS and death of 3.6 per 100 person-years of follow-up [95% confidence interval (CI):3.4-3.8]. By 96 months from baseline, the proportion of patients with a new AIDS diagnosis or death was 20.3% (95% CI:17.7-22.9) in patients with no evidence of virological failure and 53% (39.3-66.7) in those with virological failure and mutations to three drug classes (P = 0.0001). An almost two-fold difference in risk was confirmed in the multivariable analysis (adjusted relative hazard = 1.8, 95% CI:1.2-2.7, P = 0.005). CONCLUSION: Although this study shows an association between the detection of resistance at failure and risk of clinical progression, further research is needed to clarify whether resistance reflects poor adherence or directly increases the risk of clinical events via exhaustion of drug options.

Per aprendre anglès, quan abans millor? L'aprenentatge de l'anglès com a llengua estrangera. When Learning English, is it a case of 'The earlier the better?' Learning English as a foreign Language

En contra del que molts pares i mestres pensen, i en contra del que s'ha trobat en contextos d'adquisició natural de segones llengües i d'immersió, començar al més aviat possible l'aprenentatge d'idiomes, sobretot pel que fa a les tasques cognitivament més exigents, com l'escriptura, no sembla l'opció més eficaç. Els resultats del Grup de Recerca en Adquisició de Llengües (GRAL) estudià aprenents d'anglès que començaren als vuit i onze anys i trobaren, sistemàticament, que a llarg termini, després de les mateixes hores d'instrucció, eren els més grans, que havien començat als onze anys, els que obtenien millors resultats en totes les proves orals i escrites d'anglès excepte en alguna de prova de reconeixement fonètic. La maduresa cognitiva dels alumnes més grans i els diferents mecanismes d¿aprenentatge implícit i explícit entre els nens més petits i els adults ajuden a explicar per què només començar abans l'aprenentatge d'idiomes no sembla suficient per obtenir millors resultats.

Deep learning via semi-supervised embedding

We show how nonlinear embedding algorithms popular for use with shallow semi-supervised learning techniques such as kernel methods can be applied to deep multilayer architectures, either as a regularizer at the output layer, or on each layer of the architecture. This provides a simple alternative to existing approaches to deep learning whilst yielding competitive error rates compared to those methods, and existing shallow semi-supervised techniques.

Implementing statistical learning methods through Bayesian networks. Part 1: a guide to Bayesian parameter estimation using forensic science data

As a thorough aggregation of probability and graph theory, Bayesian networks currently enjoy widespread interest as a means for studying factors that affect the coherent evaluation of scientific evidence in forensic science. Paper I of this series of papers intends to contribute to the discussion of Bayesian networks as a framework that is helpful for both illustrating and implementing statistical procedures that are commonly employed for the study of uncertainties (e.g. the estimation of unknown quantities). While the respective statistical procedures are widely described in literature, the primary aim of this paper is to offer an essentially non-technical introduction on how interested readers may use these analytical approaches - with the help of Bayesian networks - for processing their own forensic science data. Attention is mainly drawn to the structure and underlying rationale of a series of basic and context-independent network fragments that users may incorporate as building blocs while constructing larger inference models. As an example of how this may be done, the proposed concepts will be used in a second paper (Part II) for specifying graphical probability networks whose purpose is to assist forensic scientists in the evaluation of scientific evidence encountered in the context of forensic document examination (i.e. results of the analysis of black toners present on printed or copied documents).

Embodying feminist research: learning from action research, political practices, diffractions and collective knowledge.

In the past three decades, feminists and critical theorists have discussed and argued the importance of deconstructing and problematizing social science research methodology in order to question normalized hierarchies concerning the production of knowledge and the status of truth claims. Nevertheless, often, these ideas have basically remained theoretical propositions not embodied in research practices. In fact there is very little published discussion about the difficulties and limits of their practical application. In this paper we introduce some interconnected reflections starting from two different but related experiences of embodying 'feminist activist research'. Our aim is to emphasise the importance of attending to process, making mistakes and learning during fieldwork, as well as experimenting with personalized forms of analysis, such as the construction of narratives and the story-telling process.

Embodying feminist research: learning from action research, political practices, diffractions and collective knowledge.

In the past three decades, feminists and critical theorists have discussed and argued the importance of deconstructing and problematizing social science research methodology in order to question normalized hierarchies concerning the production of knowledge and the status of truth claims. Nevertheless, often, these ideas have basically remained theoretical propositions not embodied in research practices. In fact there is very little published discussion about the difficulties and limits of their practical application. In this paper we introduce some interconnected reflections starting from two different but related experiences of embodying 'feminist activist research'. Our aim is to emphasise the importance of attending to process, making mistakes and learning during fieldwork, as well as experimenting with personalized forms of analysis, such as the construction of narratives and the story-telling process.

Anesthesia prevents auditory perceptual learning.

BACKGROUND: An auditory perceptual learning paradigm was used to investigate whether implicit memories are formed during general anesthesia. METHODS: Eighty-seven patients who had an American Society of Anesthesiologists physical status of I-III and were scheduled to undergo an elective surgery with general anesthesia were randomly assigned to one of two groups. One group received auditory stimulation during surgery, whereas the other did not. The auditory stimulation consisted of pure tones presented via headphones. The Bispectral Index level was maintained between 40 and 50 during surgery. To assess learning, patients performed an auditory frequency discrimination task after surgery, and comparisons were made between the groups. General anesthesia was induced with thiopental and maintained with a mixture of fentanyl and sevoflurane. RESULTS: There was no difference in the amount of learning between the two groups (mean +/- SD improvement: stimulated patients 9.2 +/- 11.3 Hz, controls 9.4 +/- 14.1 Hz). There was also no difference in initial thresholds (mean +/- SD initial thresholds: stimulated patients 31.1 +/- 33.4 Hz, controls 28.4 +/- 34.2 Hz). These results suggest that perceptual learning was not induced during anesthesia. No correlation between the bispectral index and the initial level of performance was found (Pearson r = -0.09, P = 0.59). CONCLUSION: Perceptual learning was not induced by repetitive auditory stimulation during anesthesia. This result may indicate that perceptual learning requires top-down processing, which is suppressed by the anesthetic.

Relationship between imaging biomarkers, age, progression and symptom severity in Alzheimer's disease.

The early diagnostic value of glucose hypometabolism and atrophy as potential neuroimaging biomarkers of mild cognitive impairment (MCI) and Alzheimer's disease (AD) have been extensively explored using [18F]fluorodeoxyglucose positron emission tomography (FDG-PET) and structural magnetic resonance imaging (MRI). The vast majority of previous imaging studies neglected the effects of single factors, such as age, symptom severity or time to conversion in MCI thus limiting generalisability of results across studies. Here, we investigated the impact of these factors on metabolic and structural differences. FDG-PET and MRI data from AD patients (n = 80), MCI converters (n = 65) and MCI non-converters (n = 64) were compared to data of healthy subjects (n = 79). All patient groups were split into subgroups by age, time to conversion (for MCI), or symptom severity and compared to the control group. AD patients showed a strongly age-dependent pattern, with younger patients showing significantly more extensive reductions in gray matter volume and glucose utilisation. In the MCI converter group, the amount of glucose utilisation reduction was linked to the time to conversion but not to atrophy. Our findings indicate that FDG-PET might be more closely linked to future cognitive decline whilst MRI being more closely related to the current cognitive state reflects potentially irreversible damage.