Serum Cytokines and Adipokines in Periodontitis

Objective: To investigate whether there was an association between serum levels of pro-inflammatory cytokines, adipokines and periodontitis in 60-70 year-old men in Northern Ireland. Methods: Test subjects (n=61) with established periodontitis, categorised by loss of periodontal attachment (PAL) and pocketing of at least 5mm, were matched for age, smoking and BMI with controls (n=60) who were periodontitis resistant (no PAL or pocketing over 3mm). Serum levels of IL 1 alpha, TNF alpha, IL 6, adiponectin, resistin, leptin and C-reactive protein (CRP) were measured from serum using a multiplex array (Randox, UK). Mann Whitney analysis was used with the level of significance set at p<0.05. Results: There was considerable inter-individual variability in all the analytes measured. There was a significantly higher level of IL 6 in men with periodontitis (3.25 pg/ml, SD 2.29) than in those with no periodontitis (2.38 pg/ml, SD1.62), p=0.0041 (after Bonferroni correction for multiple testing p=0.029). Serum IL 6 was strongly correlated with CRP: r=0.52 (95% CI 0.38-0.64), p<0.0001. There was a lower serum level of adiponectin in men with periodontitis (3662.1 ng/ml, SD 2383.4) compared with those who were periodontitis resistant (4265.9 ng/ml, SD 2266.0), which just escaped statistical significance at p=0.068. There were no significant differences in the serum levels of any of the other analytes measured in relation to periodontal status. Conclusions: There was evidence in the 60-70 year-old male population investigated that periodontitis was associated with an increased serum level of IL 6.

Responsiveness of dark-adaptation threshold to vitamin A and beta-carotene supplementation in pregnant and lactating women in Nepal.

BACKGROUND: Impaired dark adaptation occurs commonly in vitamin A deficiency. OBJECTIVE: We sought to examine the responsiveness of dark-adaptation threshold to vitamin A and beta-carotene supplementation in Nepali women. DESIGN: The dark-adapted pupillary response was tested in 298 pregnant women aged 15-45 y in a placebo-controlled trial of vitamin A and beta-carotene; 131 of these women were also tested at 3 mo postpartum. Results were compared with those for 100 nonpregnant US women of similar age. The amount of light required for pupillary constriction was recorded after bleaching and dark adaptation. RESULTS: Pregnant women receiving vitamin A had better dark-adaptation thresholds (-1.24 log cd/m(2)) than did those receiving placebo (-1.11 log cd/m(2); P: = 0. 03) or beta-carotene (-1.13 log cd/m(2); P: = 0.05) (t tests with Bonferroni correction). Dark-adaptation threshold was associated with serum retinol concentration in pregnant women receiving placebo (P: = 0.001) and in those receiving beta-carotene (P: = 0.003) but not in those receiving vitamin A. Among women receiving placebo, mean dark-adaptation thresholds were better during the first trimester (-1.23 log cd/m(2)) than during the second and third trimesters (-1.03 log cd/m(2); P: = 0.02, t test). The mean threshold of nonpregnant US women (-1.35 log cd/m(2)) was better than that of all 3 Nepali groups (P: < 0.001, t test, for all 3 groups). CONCLUSIONS: During pregnancy, pupillary dark adaptation was strongly associated with serum retinol concentration and improved significantly in response to vitamin A supplementation. This noninvasive testing technique is a valid indicator of population vitamin A status in women of reproductive age.

Uma perspectiva multidimensional para desenvolvimento de campanhas de marketing social: o caso do Grupo de Ajuda a Toxicodependentes (GATO)

O Grupo de Ajuda a Toxicodependentes (GATO) é uma instituição, sem fins lucrativos e de utilidade pública, que atua ao nível da prevenção, tratamento e (re)inserção de pessoas com problemas de consumo de drogas ilícitas, crianças e jovens em risco, população desfavorecida e inserida na medida de Rendimento Social de Inserção (RSI). No desenvolvimento da atividade do GATO são executadas campanhas de marketing de cariz social que, através da informação veiculada e meios utilizados, procuram sensibilizar a sua população-alvo, captar clientes, promover a sua aproximação à instituição e a apropriação de comportamentos preventivos, bem como difusão da imagem da instituição junto da restante comunidade. Neste pressuposto, o presente estudo permitirá, não só através de um inquérito por questionário aplicado à população-alvo constituída pelos clientes da Instituição GATO, como também de informação recolhida junto dos técnicos responsáveis, recolher dados que serão analisados mediante um esquema metodológico composto por duas etapas. Numa primeira fase aplicar-se-á uma Análise Não Linear de Correlações Canónicas (OVERALS) com o objetivo de detetar padrões entre os vários conjuntos de variáveis. Numa segunda etapa, com os resultados obtidos e com o recurso a uma Análise Não Hierárquica de Cluster, designadamente o procedimento Two-Step Clustering, será possível constituir grupos de indivíduos, os quais serão posteriormente validados mediante o Teste de Bonferroni de forma a constituírem-se Topologias e Tipologias de clientes que possibilitem orientações para o desenvolvimento de campanhas de marketing, consentâneas com o objeto social da organização.

Citotoxicidade do ácido peracético: avaliação metabólica, estrutural e de morte em fibroblastos L929

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Análise bilateral do timing de ativação dos músculos transverso abdominal/oblíquo interno durante o movimento rápido de flexão de cada membro superior – dominante e não-dominante

Introdução: Por mais de uma década assumiu-se que o transverso abdominal/oblíquo interno (TrA/OI) se comportava como um ventre muscular único, pré-ativando-se bilateralmente de forma simultânea para aumentar a stiffness da coluna lombar, conferindo-lhe estabilidade segmentar durante o movimento de um dos membros superiores. À luz dos estudos atuais sabe-se que tal mecanismo não ocorre, uma vez que o TrA/OI apresenta uma ativação por mecanismo de feedforward predominantemente contralateral ao membro superior movido. Apesar de morfologicamente o TrA/OI do lado não-dominante ser mais hipertrofiado do que o do lado dominante, nada se sabe acerca da influência da dominância nos timings ativação do TrA/OI. Objectivos: Confirmar se o TrA/OI tem um timing de ativação assimétrico durante o movimento rápido de flexão do membro superior (MRMS). Pretende-se ainda avaliar se o timing de ativação do TrA/OI é influenciado pela dominância de lateralidade manual. Métodos: Efectuou-se um estudo observacional descritivo, transversal e duplamente cego com 32 atletas de futebol voluntários com membro superior direito dominante, colocando-os apenas num grupo. Procedeu-se à recolha do sinal eletromiográfico de forma a avaliar os timings de ativação do TrA/OI bilateralmente aquando dos movimentos rápidos de flexão de ambos os membros superiores, à vez. Todos os dados foram tratados estatisticamente com o programa SPSS, versão 20.0 para Mac OS, com um grau de significância de 0,05. Resultados: Verificaram-se diferenças nos timings de ativação dos TrA/OI direito e esquerdo durante os MRMS direito e esquerdo (Teste ANOVA medidas repetidas: F=291,087; p<0,001). O timing de ativação do TrA/OI direito – 29,15(13,15)ms – foi superior ao esquerdo – 4,71(17,32)ms – durante MRMS direito (Teste Post Hoc Bonferroni: p<0,001). O timing de ativação do TrA/OI esquerdo – 31,98(12,50)ms – foi superior ao direito – 12,20(17,40)ms – durante MRMS esquerdo (p<0,001). O timing de ativação do TrA/IO direito aquando do MRMS direito foi superior ao observado durante MRMS esquerdo (p<0,001). O contrário sucedeu em relação ao timing ativação do TrA/IO esquerdo (p<0,001). O timing de ativação do TrA/IO esquerdo no MRMS direito foi inferior ao do TrA/IO direito aquando do movimento com o membro esquerdo (p<0,001). O TrA/IO direito possuiu um timing de ativação no MRMS direito inferior ao do TrA/IO esquerdo aquando do movimento com o membro esquerdo (p<0,001). Conclusão: Através deste estudo pôde-se concluir que o TrA/OI contralateral ao MRMS apresenta um timing de ativação inferior ao ipsilateral e ainda que durante o MRMS dominante o TrA/OI esquerdo e direito apresentam um timing de ativação inferior ao ocorrido durante o MRMS não-dominante.

Extracorporeal membrane oxygenation support after heart explantation: a proposal on how to deal with hyperacute rejection of heart transplant

Purpose: In extreme situations, such as hyperacute rejection of heart transplant or major bleeding per-operating complications, an urgent heart explantation might be the only means of survival. The aim of this experimental study was to improve the surgical technique and the hemodynamics of an Extracorporeal Membrane Oxygenation (ECMO) support through a peripheral vascular access in an acardia model. Methods: An ECMO support was established in 7 bovine experiments (59±6.1 kg) by the transjugular insertion to the caval axis of a self-expanded cannula, with return through a carotid artery. After baseline measurements of pump flow and arterial and central venous pressure, ventricular fibrillation was induced (B), the great arteries were clamped, the heart was excised and right and left atria remnants, containing the pulmonary veins, were sutured together leaving an atrial septal defect (ASD) over the cannula in the caval axis. Measurements were taken with the pulmonary artery (PA) clamped (C) and anastomosed with the caval axis (D). Regular arterial and central venous blood gases tests were performed. The ANOVA test for repeated measures was used to test the null hypothesis and a Bonferroni t method for assessing the significance in the between groups pairwise comparison of mean pump flow. Results: Initial pump flow (A) was 4.3±0.6 L/min dropping to 2.8±0.7 L/min (P B-A= 0.003) 10 minutes after induction of ventricular fibrillation (B). After cardiectomy, with the pulmonary artery clamped (C) it augmented not significantly to 3.5±0.8 L/min (P C-B= 0.33, P C-A= 0.029). Finally, PA anastomosis to the caval axis was followed by an almost to baseline pump flow augmentation (4.1±0.7 L/min, P D-B= 0.009, P D-C= 0.006, P D-A= 0.597), permitting a full ECMO support in acardia by a peripheral vascular access. Conclusions: ECMO support in acardia is feasible, providing new opportunities in situations where heart must urgently be explanted, as in hyperacute rejection of heart transplant. Adequate drainage of pulmonary circulation is pivotal in order to avoid pulmonary congestion and loss of volume from the normal right to left shunt of bronchial vessels. Furthermore, the PA anastomosis to the caval axis not only improves pump flow but it also permits an ECMO support by a peripheral vascular access and the closure of the chest.

Explicit versus implicit spelling strategy instruction and its effects on grade one children's invented spellings

The intent of this study was to investigate .the effectiveness of teaching thirty-five Grade One children a variety of effective spelling strategies in comparison to tradit~onal spelling instruction. Strategy instruction included training in phonology, imagery and analogy. In addition, the type of instruction pro~ided (implicit versus explicit) was also examined. Children were seen in small groups of four or five, for four, twenty-five minute sessions. All children were tested prior and immediately following the training sessions, as well as at 14-day follow-ups. Pretest and posttest measures included a dictated spelling test (based on words used in training), a developmental spelling test and a sample of each child's writing. In addition, children completed a metacognitive spelling test as a measure of their strategy awareness. Performance scores on the pretest and posttest measures were compared to determine if any differences existed between the three spelling instruction groups using the Dunn-Bonferroni and Dunnett procedures. Findings revealed that explicit strategy instruction was the most effective spelling program for improving Grade One children's invented spellings. Children who received this instruction were able to spell targeted words more accurately, even after a 14-day follow-up, and were able to recall more effective spelling strategies than children who received either implicit strategy instruction or traditional strategy instruction.

The contribution of skate blade properties to skating speed

The purpose of the study was to investigate the relative contribution of skate blade properties to on-ice skating speed. Thirty-two male ice hockey players (mean age = 19±2.65 yrs.) representing the Ontario Minor Hockey Association (OMHA; Midget AAA and Junior), Canadian Inter University Sport (CIS: Varsity), Ontario hockey league (OHL) and East Coast Hockey League (ECHL), and the playing positions of forwards (n=18) and defense (n=14) were recruited to participate. Skate related equipment worn by the players for the purpose of the research was documented and revealed that 80% of the players wore Bauer skates, Tuuk blade holders and LS2 skate blades. Subjects completed a battery of eight on-ice skating drills used to measure and compare two aspects of skating speed; acceleration [T1(s)] and total time to complete each drill [TT(s)] while skating on three skate blade conditions. The drills represented skills used in the game of hockey, both in isolation (e.g., forward skating, backward skating, stops and starts, and cornering) and in sequence to simulate the combination of skills used in a shift of game play. The three blade conditions consisted of (i) baseline, represented by the blades worn by the player throughout their current season of play; (ii) experimental blades (EB), represented by brand name experimental blades with manufacturers radius of contour and a standardized radius of hollow; and (iii) customized experimental blades (CEB), represented by the same brand name experimental blades sharpened to the players’ preference as identified in the baseline condition. No significant differences were found in acceleration time [T1(s)] or total time to complete [TT(s)] the isolated drills across blade conditions; however significant differences were revealed in both T1(s) and TT(s) measured during the execution of the sequenced drill across blade conditions. A iii Bonferroni post hoc test revealed that players skated significantly faster when skating on the CEB condition compared to the baseline condition (p≤.05). A questionnaire assessing subjects perceived comfort, confidence and effort expended while skating on the experimental blades revealed that players were significantly more comfortable when skating on the CEB versus the EB condition (p≤.05). Outcomes of the study provide evidence to suggest that the experimental skate blades customized with the players preferred blade sharpening characteristics results in faster skating speed in a combination drill representing skills performed in gameplay.

Exact Tests for Contemporaneous Correlation of Disturbances in Seemingly Unrelated Regressions.

This paper proposes finite-sample procedures for testing the SURE specification in multi-equation regression models, i.e. whether the disturbances in different equations are contemporaneously uncorrelated or not. We apply the technique of Monte Carlo (MC) tests [Dwass (1957), Barnard (1963)] to obtain exact tests based on standard LR and LM zero correlation tests. We also suggest a MC quasi-LR (QLR) test based on feasible generalized least squares (FGLS). We show that the latter statistics are pivotal under the null, which provides the justification for applying MC tests. Furthermore, we extend the exact independence test proposed by Harvey and Phillips (1982) to the multi-equation framework. Specifically, we introduce several induced tests based on a set of simultaneous Harvey/Phillips-type tests and suggest a simulation-based solution to the associated combination problem. The properties of the proposed tests are studied in a Monte Carlo experiment which shows that standard asymptotic tests exhibit important size distortions, while MC tests achieve complete size control and display good power. Moreover, MC-QLR tests performed best in terms of power, a result of interest from the point of view of simulation-based tests. The power of the MC induced tests improves appreciably in comparison to standard Bonferroni tests and, in certain cases, outperforms the likelihood-based MC tests. The tests are applied to data used by Fischer (1993) to analyze the macroeconomic determinants of growth.

Finite-Sample Diagnostics for Multivariate Regressions with Applications to Linear Asset Pricing Models

In this paper, we propose several finite-sample specification tests for multivariate linear regressions (MLR) with applications to asset pricing models. We focus on departures from the assumption of i.i.d. errors assumption, at univariate and multivariate levels, with Gaussian and non-Gaussian (including Student t) errors. The univariate tests studied extend existing exact procedures by allowing for unspecified parameters in the error distributions (e.g., the degrees of freedom in the case of the Student t distribution). The multivariate tests are based on properly standardized multivariate residuals to ensure invariance to MLR coefficients and error covariances. We consider tests for serial correlation, tests for multivariate GARCH and sign-type tests against general dependencies and asymmetries. The procedures proposed provide exact versions of those applied in Shanken (1990) which consist in combining univariate specification tests. Specifically, we combine tests across equations using the MC test procedure to avoid Bonferroni-type bounds. Since non-Gaussian based tests are not pivotal, we apply the “maximized MC” (MMC) test method [Dufour (2002)], where the MC p-value for the tested hypothesis (which depends on nuisance parameters) is maximized (with respect to these nuisance parameters) to control the test’s significance level. The tests proposed are applied to an asset pricing model with observable risk-free rates, using monthly returns on New York Stock Exchange (NYSE) portfolios over five-year subperiods from 1926-1995. Our empirical results reveal the following. Whereas univariate exact tests indicate significant serial correlation, asymmetries and GARCH in some equations, such effects are much less prevalent once error cross-equation covariances are accounted for. In addition, significant departures from the i.i.d. hypothesis are less evident once we allow for non-Gaussian errors.

Effet de la source du sélénium sur le statut du sélénium, de la GSH-Px et sur le système immunitaire des bovins de boucherie

Résumé L’objectif de cette étude était de déterminer les effets de la source de sélénium sur les concentrations de Se et de GSH-Px des vaches de boucherie (n =33) et leurs veaux et sur des paramètres immunitaires des veaux. Deux groupes de vaches ont reçu 3 mg/j/animal de Se organique ou inorganique dans le minéral. Le troisième groupe n'a pas été supplémenté en Se et leurs veaux ont été divisés en deux sous-groupes, l’un des deux a reçu une injection de sélénite de sodium (0,087 mg/Kg) à la naissance. Le Se et la GSH-Px ont été respectivement mesurés par HPLC-UV et par cinétique enzymatique. La phagocytose, la flambée respiratoire et le ratio CD4:CD8ont été évalués par des kits commerciaux et les IgG totales ont été mesurés par immunodiffusion radiale. La supplémentation de Se a augmenté significativement le Se sérique et colostral (P<0,02) et la GSH-Px(P≤0,04) pour les vaches et leurs veaux avec un effet significativement plus élevé pour le Se organique. Le Se du lait a augmenté de façon significative uniquement avec la source organique du Se (P≤0,0007). L’injection du Se chez les veaux a permis une augmentation significative mais temporaire (P<0,0001) du Se sérique. La supplémentation en Se n’a pas influencé les paramètres immunitaires mesurés (P>0,01, non significatif après correction de Bonferroni). Nous concluons que la supplémentation en Se améliore le niveau du Se colostral, lacté et sérique ainsi que la GSH-Px pour les vaches et leurs veaux sans effet sur les paramètres immunitaires mesurés des veaux. Mots clés: Sélénium, veaux de boucherie, phagocytose, flambée respiratoire, anticorps, ratio CD4:CD8, GSH-Px.

Contribution de l'hypoxie à la cicatrisation cutanée anormale chez le cheval : méthodes physiques d'évaluation

La guérison des plaies cutanées appendiculaires chez le cheval, à la différence de celle des plaies corporelles, se complique régulièrement. Un retard de cicatrisation s’y observe et un tissu de granulation exubérant tend à s’y développer, le tout menant à une cicatrice pathologique hypertrophiée. La pathogénie exacte du tissu de granulation exubérant chez le cheval demeure inconnue à ce jour. Une hypoxie tissulaire pourrait favoriser son développement tout comme elle semble contribuer au développement de cicatrices cutanées pathologiques similaires observées chez l’Homme. L’objectif de cette étude était d’évaluer la perfusion vasculaire et la disponibilité locale en oxygène de plaies cutanées appendiculaires et corporelles en cours de cicatrisation normale et pathologique chez le cheval, à l’aide de la thermographie infrarouge et de la spectroscopie par réflectance dans le proche infrarouge. Six juments âgées de 3 à 4 ans ont été utilisées. Trois plaies cutanées ont été créées à l’aspect dorso-latéral du canon des membres thoraciques (plaies appendiculaires), et sur la paroi costale de l’un des hémithorax (plaies corporelles). Chez chaque jument, un canon a été aléatoirement bandé dans le but d’induire la formation de tissu de granulation exubérant dans les plaies s’y trouvant, tel que rapporté. La perfusion vasculaire et la disponibilité locale en oxygène ont été évaluées séquentiellement par thermographie infrarouge et spectroscopie par réflectance dans le proche infrarouge pour chaque plaie de chaque site (thorax; membre bandé; membre non bandé) au cours du processus de cicatrisation cutanée. Un modèle linéaire à doubles mesures répétées associé à une correction séquentielle de Bonferroni a révélé des différences significatives de perfusion vasculaire et de disponibilité locale en oxygène entre les plaies appendiculaires et corporelles. Ainsi la perfusion vasculaire et la disponibilité locale en oxygène étaient significativement plus élevées dans les plaies corporelles (P<0.05) et la perfusion vasculaire était significativement plus élevée dans les plaies appendiculaires non bandées que dans celles bandées (P<0.05). Nous avons récemment rapporté une plus grande occlusion de la micro-vascularisation au niveau des plaies appendiculaires chez le cheval. Nous rapportons maintenant que la perfusion vasculaire et la disponibilité locale en oxygène sont significativement inférieures dans les plaies appendiculaires, en particulier lorsqu’un tissu de granulation exubérant s’y développe. Compilés, ces résultats sous-tendent l’hypothèse que les plaies appendiculaires souffrent d’une altération de la perfusion vasculaire à l’origine possible d’une hypoxie tissulaire qui pourrait favoriser une cicatrisation cutanée anormale, telle la formation d’un tissu de granulation exubérant.

Évaluation de l'acuité visuelle chez la personne âgée atteinte de troubles de la cognition

Objectif : L'évaluation de l'acuité visuelle (AV) chez la personne âgée atteinte de troubles cognitifs peut être limitée par le manque de collaboration ou les difficultés de communication du patient. Très peu d'études ont examiné l'AV chez les patients atteints de déficits sévères de la cognition. L’objectif de cette étude était d’évaluer l'AV chez la personne âgée vulnérable ayant des troubles cognitifs à l'aide d'échelles variées afin de vérifier leur capacité à répondre à ces échelles. Méthodes: Trois groupes de 30 sujets chacun ont été recrutés. Le premier groupe était constitué de sujets jeunes (Moy.±ET: 24.9±3.5ans) et le second, de sujets âgés (70.0±4.5ans), ne présentant aucun trouble de la cognition ou de la communication. Le troisième groupe, composé de sujets atteints de démence faible à sévère (85.6±6.9ans), a été recruté au sein des unités de soins de longue durée de l’Institut Universitaire de Gériatrie de Montréal. Le test du Mini-Mental State Examination (MMSE) a été réalisé pour chaque sujet afin de déterminer leur niveau cognitif. L’AV de chaque participant a été mesurée à l’aide de six échelles validées (Snellen, cartes de Teller, ETDRS-lettres,-chiffres,-Patty Pics,-E directionnel) et présentées selon un ordre aléatoire. Des tests non paramétriques ont été utilisés afin de comparer les scores d’AV entre les différentes échelles, après une correction de Bonferroni-Holm pour comparaisons multiples. Résultats: Le score moyen au MMSE chez les sujets atteints de démence était de 9.8±7.5, alors qu’il était de 17.8±3.7 et 5.2±4.6 respectivement, chez les sujets atteints de démence faible à modérée (MMSE supérieur ou égal à 13; n=11) et sévère (MMSE inférieur à 13; n=19). Tous les sujets des groupes 1 et 2 ont répondu à chacune des échelles. Une grande majorité de sujets avec démence ont répondu à toutes les échelles (n=19) alors qu’un seul sujet n’a répondu à aucune échelle d’AV. Au sein du groupe 3, les échelles d’AV fournissant les scores les plus faibles ont été les cartes de Teller (20/65) et les Patty Pics (20/62), quelque soit le niveau cognitif du sujet, alors que les meilleurs scores d’AV ont été obtenus avec les échelles de Snellen (20/35) et les lettres ETDRS (20/36). Une grande proportion de sujets avec démence sévère ont répondu aux cartes de Teller (n=18) mais le score d’AV obtenu était le plus faible (20/73). Au sein des trois groupes, l’échelle de lettres-ETDRS était la seule dont les scores d’AV ne différaient pas de ceux obtenus avec l’échelle de Snellen traditionnelle. Conclusions: L’acuité visuelle peut être mesurée chez la personne âgée atteinte de troubles cognitifs ou de la communication. Nos résultats indiquent que les échelles les plus universelles, utilisant des lettres comme optotypes, peuvent être utilisées avec de bons résultats chez les personnes âgées atteintes de démence sévère. Nos résultats suggèrent de plus que la mesure d’acuité visuelle doit être tentée chez toutes les personnes, peu importe leur niveau cognitif.

Approximation du calcul de la taille échantillonnale pour les tests à hypothèses multiples lorsque r parmis m hypothèses doivent être significatives

Généralement, dans les situations d’hypothèses multiples on cherche à rejeter toutes les hypothèses ou bien une seule d’entre d’elles. Depuis quelques temps on voit apparaître le besoin de répondre à la question : « Peut-on rejeter au moins r hypothèses ? ». Toutefois, les outils statisques pour répondre à cette question sont rares dans la littérature. Nous avons donc entrepris de développer les formules générales de puissance pour les procédures les plus utilisées, soit celles de Bonferroni, de Hochberg et de Holm. Nous avons développé un package R pour le calcul de la taille échantilonnalle pour les tests à hypothèses multiples (multiple endpoints), où l’on désire qu’au moins r des m hypothèses soient significatives. Nous nous limitons au cas où toutes les variables sont continues et nous présentons quatre situations différentes qui dépendent de la structure de la matrice de variance-covariance des données.

Inférence topologique

Les données provenant de l'échantillonnage fin d'un processus continu (champ aléatoire) peuvent être représentées sous forme d'images. Un test statistique permettant de détecter une différence entre deux images peut être vu comme un ensemble de tests où chaque pixel est comparé au pixel correspondant de l'autre image. On utilise alors une méthode de contrôle de l'erreur de type I au niveau de l'ensemble de tests, comme la correction de Bonferroni ou le contrôle du taux de faux-positifs (FDR). Des méthodes d'analyse de données ont été développées en imagerie médicale, principalement par Keith Worsley, utilisant la géométrie des champs aléatoires afin de construire un test statistique global sur une image entière. Il s'agit d'utiliser l'espérance de la caractéristique d'Euler de l'ensemble d'excursion du champ aléatoire sous-jacent à l'échantillon au-delà d'un seuil donné, pour déterminer la probabilité que le champ aléatoire dépasse ce même seuil sous l'hypothèse nulle (inférence topologique). Nous exposons quelques notions portant sur les champs aléatoires, en particulier l'isotropie (la fonction de covariance entre deux points du champ dépend seulement de la distance qui les sépare). Nous discutons de deux méthodes pour l'analyse des champs anisotropes. La première consiste à déformer le champ puis à utiliser les volumes intrinsèques et les compacités de la caractéristique d'Euler. La seconde utilise plutôt les courbures de Lipschitz-Killing. Nous faisons ensuite une étude de niveau et de puissance de l'inférence topologique en comparaison avec la correction de Bonferroni. Finalement, nous utilisons l'inférence topologique pour décrire l'évolution du changement climatique sur le territoire du Québec entre 1991 et 2100, en utilisant des données de température simulées et publiées par l'Équipe Simulations climatiques d'Ouranos selon le modèle régional canadien du climat.