La forte prévalence des troubles des conduites alimentaires (TCA) chez les jeunes femmes et les faibles taux de rémission suite à un traitement ont encouragé les chercheurs à mieux comprendre les facteurs impliqués dans ce trouble mental. L’un des premiers modèles à mettre l’emphase sur des traits de personnalité associés au développement d’un TCA a été proposé par Hilde Bruch (1962, 1973, 1978) et a toujours une grande influence dans la recherche actuelle. Le modèle de Bruch inclue trois facteurs, soit l’insatisfaction corporelle, l’inefficacité et la conscience intéroceptive. Le but de cette thèse est d’apporter un support empirique au modèle de Bruch. En se basant sur une revue extensive des écrits scientifiques, cette thèse vise aussi à déterminer si deux facteurs reliés, soit l’alexithymie et le sentiment d’efficacité personnelle face à l’adoption de conduites alimentaires saines, améliorent la précision du modèle dans la prédiction de symptômes de TCA. Pour répondre empiriquement à cette question, il était d’abord nécessaire de disposer d’un questionnaire évaluant le sentiment d’efficacité personnelle en lien avec les conduites alimentaires qui peut être utilisé dans tout le spectre de présentation des TCA. Ainsi, le Eating Disorder Self-Efficacy Questionnaire (EDRSQ) a été adapté en français et ses propriétés psychométriques ont été évaluées. Une analyse factorielle confirmatoire a révélé une structure bi-factorielle, soit le sentiment d’efficacité personnelle en lien avec l’insatisfaction corporelle et avec l’adoption d’une alimentation normative. Chaque échelle a démontré une bonne fiabilité ainsi qu’une validité de construit cohérente avec la théorie. Par la suite, la capacité des facteurs proposés par Bruch à prédire les symptômes de TCA a été évaluée et comparée à des adaptations du modèle découlant des écrits. Au total, 203 étudiantes de premier cycle universitaire ont complété les versions validées en français du Eating Disorder Inventory 2, du Eating Attitudes Test, et du Toronto Alexithymia Scale en plus du EDRSQ. Les résultats montrent que le modèle de Bruch explique 46% de la variance des symptômes de TCA. Alors que l’insatisfaction corporelle et la conscience intéroceptive démontrent chacun une contribution importante dans la prédiction des symptômes de TCA, il a été démontré que l’inefficacité présente seulement une contribution spécifique négligeable. Le modèle de Bruch est amélioré par la substitution de l’inefficacité par le sentiment d’efficacité personnelle tel que mesuré par le EDRSQ; le modèle explique alors 64% de la variance des symptômes de TCA. Finalement, cette étude démontre que l’alexithymie n’a pas de contribution spécifique dans la prédiction des symptômes de TCA. Ainsi, la combinaison d’une faible conscience intéroceptive, de l’insatisfaction corporelle et d’un faible sentiment d’efficacité personnelle en lien avec les conduites alimentaires est fortement associée aux symptômes de TCA dans un échantillon non-clinique de jeunes femmes. Finalement, les implications conceptuelles et cliniques de ces résultats sont discutées. / High prevalence of Eating Disorders (EDs) amongst young women and poor treatment outcome rates have urged researchers to better understand premorbid factors involved in the pathology. One of the first models to emphasize premorbid personality factors in order to explain eating disorders was proposed by Hilde Bruch (1962, 1973, 1978) and is still very influential in today’s literature. Bruch’s model included three factors, namely body dissatisfaction, ineffectiveness and interoceptive awareness. The purpose of this thesis was to provide an empirical validation of Bruch’s theoretical model. Based on the literature, this thesis also aimed to determine if two related concepts, namely alexithymia and eating self-efficacy, improved the accuracy of the model in predicting ED symptoms. To empirically answer this question, it was first deemed necessary to have an eating self-efficacy questionnaire that could be used within all the spectrum of ED pathology. Therefore, the Eating Disorder Self-Efficacy Questionnaire (EDRSQ) was adapted to French and its’ psychometric properties were assessed. A confirmatory factor analysis revealed a bi-factorial structure, which were Body Image Self-Efficacy and Normative Eating Self-Efficacy. Both scales demonstrated evidence of reliability and theoretically consistent evidence of construct validity. Afterwards, Bruch’s factors ability to predict ED symptoms was assessed and then compared to empirically driven adaptations of the model. A total of 203 undergraduate females completed the Eating Disorder Inventory 2, Eating Attitudes Test, and Toronto Alexithymia Scale, as well as the EDRSQ. The results indicated that Bruch’s model accounted for 46% of the variance in eating symptomatology. While body dissatisfaction and interoceptive awareness both demonstrated an important contribution in predicting ED symptomatology, ineffectiveness was found to have a negligible specific contribution. Bruch’s model was improved by substituting ineffectiveness with eating self-efficacy as measured by the EDRSQ; it then accounted for 64% of the variance in eating disorder symptoms. Finally, this study demonstrated that alexithymia was not a specific predictor of eating disorder symptoms. Thus, a combination of lack of interoceptive awareness, body dissatisfaction and low eating self-efficacy is strongly associated with symptoms of eating disorders in a non-clinical sample of women. Finally, the conceptual and clinical implications of these findings were discussed.

Thèse numérisée par la Division de la gestion de documents et des archives de l'Université de Montréal.

Disordered eating, including bingeing, dieting, purging, and clinical and subclinical forms of anorexia nervosa, bulimia nervosa, and binge eating disorder, is prevalent among college-aged women. To date, few interventions have successfully reduced risk factors related to disordered eating. One promising intervention utilizes principles of cognitive dissonance to reduce thin-ideal internalization among women at risk for eating disorders. Additionally, the benefits of yoga, including increased awareness of bodily processes, offer hope that this practice might reduce disordered eating symptomatology. The current study compared cognitive dissonance and yoga interventions for disordered eating attitudes and behaviors. Hierarchical regression analyses revealed that there were no significant differences between the yoga and control groups. However, participants in the dissonance group had significantly lower scores than both other groups on measures of disordered eating symptoms as well as thin-ideal internalization, body dissatisfaction, alexithymia, and anxiety. These findings have important implications for potential interventions on college campuses.

Tendo em vista que os textos jornalísticos de vulgarização científica são predominantemente argumentativos, o estudo da argumentação presente neles permite elucidar algumas características de seu funcionamento textual. A análise proposta aqui privilegia o exame de dois de seus componentes centrais, considerados fundadores de sua configuração argumentativa típica: os lugares-comuns e os percursos argumentativos. Os primeiros referem-se a idéias ou modos de raciocínio socialmente aceitos e tidos como válidos independentemente de comprovação, constituindo objetos de acordo primários, com base nos quais a arguimentação se torna possível. Os segundos dizem respeito à forma como os lugares se encadeiam e sucedem no interior de um quadro argumentativo, dando forma à macroestrutura argumentativa dos textos de vulgarização científica. A análise procura demonstrar que a formulação da questão argumentativa e a escolha dos lugares-comuns determinam a forma assumida pelos percursos geradores da estrutura argumentativa e que, nos textos jornalísticos de vulgarização científica, essa estrutura tem configuração em forma de labirinto. Para tanto, o estudo se divide em três partes: na primeira, analisa-se a noção de vulgarização científica e se procura entender como pode os chamados textos de vulgarização podem se constituir como objetos analisáveis. Na segunda, busca-se um modelo analítico capaz de levar à compreensão do funcionamento dos textos de vulgarização. Na terceira, discute-se o papel dos lugares-comuns nos percursos argumentativos analisados. / The study of argumentation in popular scientific texts allows one to identify certain characteristics concerning their textual operation. An analysis of the argumentative structure of texts, in particular of those found in specialist scientific publications, shows two components essential to this type of text: the topoi (both commonplace and specialist) and the sequence of arguments. The topoi, be they true, plausible or generally accepted ideas, are at the heart of any argumentation, and it is on the basis of the topoi that the proponent constructs his persuasive reasoning. The sequence of arguments constitutes the manner in which the topoï are connected and follow one another within the argumentative structure. It brings together the discursive competence (which can be sometimes linguistic, rhetorical, logical or encyclopaedic) as well as the emotions of the advocate. In this way, the topoi give form to the micro/macrostructures arguments of the aforesaid published texts. The analysis is made starting from the components of enunciation theory, pragmatism and rhetoric and makes it possible to show that the choice of the topoi by the proponent or presenter determines the configuration or labyrintic of the arguments as presented in the text.

Made available in DSpace on 2016-04-04T18:29:32Z (GMT). No. of bitstreams: 1 Jena Hanay Doutorado.pdf: 721081 bytes, checksum: 65f8d839cf686774d38170e56199f048 (MD5) Previous issue date: 2006-12-18 / Evaluates depression, anxiety, psychopathologic symptoms, alexithymia and defensive style of grade III obese patients, before and after bariatric surgery, as well as estimates the level of association among such variables. Methodological study design was correlational crosssectional type. 65 patients took part in the study (Group 1: surgical candidates vs. Group 2: postoperative patients), mainly females (92,3%), who answered the Beck Depression Inventory (BDI), the Beck Anxiety Inventory (BAI), the Symptoms Assessment Scale (EAS- 40), and the version in Portuguese of The Toronto Alexithymia Scale (TAS-26) and of the Defensive Style Questionnaire (DSQ-40). Gr 1 showed mild anxiety and depression levels and Gr 2 minimum level (respectively, p< 0,001 e p< 0,01). The total average score for EAS- 40 and F2 and F3 (obesity-compulsion and somatization) were higher in Gr 1 when compared to Gr 2 (repetitively, p< 0,005, p< 0,005 e p< 0,001). On TAS, the total average score for F1 (ability to identify and describe feelings and distinguish them from bodily sensations) were higher in Gr 1 than in Gr 2 (respectively, p< 0,01 e p< 0,005). According to DSQ-40, there was a tendency for an immature defensive style in Gr 1 in comparison to Gr 2 (p< 0,02). In relation to the association between the variables and BMI, the correlations were negative for Gr 1 and F2 of TAS (daydreaming, p < 0,05) and for the mature factor of DSQ- 40 and Gr 2 (p < 0,05), which shows a trend towards a more mature and adaptive style in Gr 2. Correlations were positive and significant in relation to immature defenses - acting out (Gr 1: p<0,005 and Gr 2: p<0,05) and in the autistic fantasy in Gr 1 (p<0,05). The results show the psychological status of the participants, who are apt for bariatric surgery and point to a decrease in psychopathological symptoms proportional to weight loss after surgery. / Avalia depress?o, ansiedade, sintomas psicopatol?gicos, alexitimia e o estilo defensivo de pacientes obesos grau III, antes e depois de cirurgia bari?trica, e estima o grau de associa??o entre essas vari?veis. O delineamento metodol?gico foi correlacional de tipo cross-sectional. Participaram do estudo 65 pacientes (Gr 1: pr?-cir?rgico vs. Gr 2: p?scir?rgico), predominantemente do sexo feminino (92,3%), que responderam o Invent?rio de Depress?o de Beck (BDI), o Invent?rio de Ansiedade de Beck (BAI), a Escala de Avalia??o de Sintomas (EAS-40), a Vers?o em Portugu?s da Escala de Alexitimia de Toronto (TAS-26) e a Vers?o em Portugu?s da Defensive Style Questionnaire (DSQ-40). O Gr 1 apresentou n?vel de depress?o e ansiedade leve e o Gr 2 n?vel m?nimo (respectivamente, p< 0,001 e p< 0,01). O escore m?dio total da EAS-40 e de F2 e F3 (obsessividade-compulsividade e somatiza??o) foram mais elevados no Gr 1 quando comparados aos do Gr 2 (respectivamente, p< 0,005, p< 0,005 e p< 0,001). Na TAS, o escore m?dio total e de F1 (habilidade de identificar e descrever sentimentos e distingui-los de sensa??es corporais) foram maiores no Gr 1 do que no Gr 2 (respectivamente, p< 0,01 e p< 0,005). De acordo com o DSQ-40, houve tend?ncia a um estilo defensivo imaturo no Gr 1 quando comparado ao Gr 2 (p< 0,02). Em rela??o ? associa??o das vari?veis com o IMC, as correla??es foram negativas no Gr 1 e o F2 da TAS (sonhar acordado, p < 0,05) e no fator maduro do DSQ-40 e o Gr 2 (p < 0,05), apontando uma propens?o de estilo mais maduro e adaptativo no Gr 2. As correla??es foram positivas e significantes nas defesas imaturas acting out (Gr 1: p<0,005 e Gr 2: p<0,05) e na fantasia aut?stica no Gr 1 (p<0,05). Os resultados mostram o modo de funcionamento psicol?gico dos participantes aptos ? cirurgia bari?trica e assinalam uma diminui??o da sintomatologia psicopatol?gica na propor??o da perda de peso ap?s a cirurgia.

Made available in DSpace on 2016-04-04T18:29:40Z (GMT). No. of bitstreams: 1 BERENICE VICTOR CARNEIRO.pdf: 3590625 bytes, checksum: c12256d2735ce90af0df41e15e96a213 (MD5) Previous issue date: 2008-02-15 / The aim of this study was to analyze psychometric properties of the Brazilian version of the Observer Alexithymia Scale (OAS) by estimating its internal consistency, test-retest reliability and inter-rater reliability; construct validity through factorial analysis, criteria validity and convergent validity in a clinical population with substance dependence or abuse. The OAS is a brief observer report designed to be used by clinicians as well as patient s relatives and acquaintances, to identify Alexithymia according to 5 dimensions: distant, uninsightful, somatizing, humorless, and rigid. It is composed of 33 items rated on a 4-point Likert scale, ranging from 0 (never, not at all like the person) to 3 (all of the time, completely like the person). Data were gathered at an outpatient public service for substance dependence and a non-profit religious inpatient institution also for substance dependence. The sample was comprised of three groups: G1 - 200 relatives or acquaintances of substance dependence and abusers (alcohol=46.5%; drugs=53.5 %), aged from 18 to 82 (M = 48, SD = 12), both sexes (F=88%; M=12%); G2 39 outpatients at the public service for substance dependence or abuse (alcohol=66,6%; drugs=33,3%), both sexes (F=15.3%; M=84.6%); G3 9 clinicians at the public service (clinical experience ranging from 1-15 years). Reliability studies suggested good internal consistency and temporal stability of the scale for both G1 (alpha = .83; rs = .79) and G3 (alpha= .85; rs= .80), but low inter-rater reliability when scored by relatives and clinicians (rs = .14). The OAS s exploratory factor analysis suggested a structure of 5 dimensions (48% total variance) as the original one, and 30 items with a minimum factor loadings of .40. When the degree of alexithymia was estimated as a criteria, the study indicated that relatives do not view those with substance dependence differently (p = 0.16) according to type of dependence (alcohol or drugs). For the convergent study, G2 participants completed the Toronto Alexithymia Scale (TAS) and the scores were compared against OAS-30, according to G1 and G3. Results suggested a negligible association (rs = .29; p >0.05) between TAS-22 and OAS, according to G1 and a lack of association (rs = -.18; p >0.20) between TAS-22 and OAS-30 scored by G3, suggesting that both scales would be measuring different constructs. The Portuguese version of the OAS-30 demonstrated good internal consistency and temporal stability. Its structure seems to be compatible to the original one, but a few items need revision, especially on factor 4. Some considerations will be made regarding the educational level of respondents, as well as the degree of acquaintance with the person who is substance dependent. / O estudo analisou propriedades psicom?tricas da vers?o brasileira da Observer Alexithymia Scale (OAS). Foram estimadas, a consist?ncia interna, precis?o por testereteste e precis?o entre avaliadores; validade de construto por meio de an?lise da estrutura fatorial, validade de crit?rio e validade convergente, em popula??o cl?nica por abuso ou depend?ncia de sust?ncias psicoativas. A OAS ? uma escala breve para uso do profissional cl?nico assim como pessoas que conhecem bem o paciente, para diagnosticar alexitimia segundo cinco dimens?es: distanciamento, sem insight, somatiza??o, sem gra?a e rigidez. ? composta de 33 itens que devem ser respondidos atrav?s de uma escala Likert de 4 pontos, sendo o grau de intensidade indicado pela escolha entre 0 (nunca, em nada parecido) e 3 (todo o tempo, totalmente parecido). Os dados foram coletados em servi?o p?blico ambulatorial para dependentes de subst?ncia e uma institui??o sem fins lucrativos, com v?nculo religioso, para o tratamento residencial da depend?ncia de subst?ncia. A amostra foi composta de tr?s grupos: G1 200 familiares ou amigos (F=88%; M=12%) de dependentes ou que abusam de subst?ncia (?lcool=46,5% e drogas=53,5%); G2 39 dependentes de subst?ncia (?lcool=66,6%; drogas=33,3%), ambos os sexos (F=15,3%; M=84,6%) em tratamento no servi?o ambulatorial e G3 nove terapeutas do servi?o ambulatorial com experi?ncia cl?nica entre 1 e 15 anos. Os estudos de consist?ncia interna e precis?o indicaram que a OAS possui boa consist?ncia interna, tanto entre os participantes do G1 (alfa = 0,83) quanto do G3 (alfa = 0,85) e boa estabilidade temporal (rs = 0,79; rs = 0,80), mas baixa precis?o entre avaliadores quando avaliada por familiares e profissionais (rs = 0,14). A an?lise fatorial explorat?ria indicou uma estrutura com 5 dimens?es (vari?ncia total = 48%) e 30 itens com cargas fatoriais m?nimas de 0,40, reproduzindo a estrutura da escala original. Na validade de crit?rio, o grau de alexitimia foi estudado em fun??o do tipo de depend?ncia (?lcool ou drogas qu?micas), tomado como medida de crit?rio. N?o se observou rela??o entre o tipo de depend?ncia e alexitimia avaliada pela OAS-30 (p = 0,16). Para o estudo de validade convergente, os participantes do G2 completaram a Escala de Alexitimia de Toronto (TAS-22). Os resultados sugerem muito baixa associa??o entre os escores da TAS-22 e a OAS-30 avaliada pelo G1 (rs = 0,29; p >0,05) e entre a TAS-22 e a OAS-30, completada pelo G3 (rs = -0,18; p >0,20), indicando que as escalas estariam medindo diferentes construtos. Os resultados apontam que a OAS-30 det?m boa precis?o e que a estrutura fatorial ? compat?vel com a vers?o original. No entanto, h? necessidade de revis?o de alguns itens, especialmente no Fator 4. Considera??es a respeito da influ?ncia do n?vel de escolaridade do respondente, assim como o n?vel de familiaridade do respondente com o dependente de subst?ncia psicoativa s?o realizadas.

L'objectif de cette thèse était d‟étudier les processus d'activation et d'inhibition lexicales sous-tendant la lecture de mots et de déterminer le rôle du système affectif sur ces processus chez l'adulte. Pour cela, nous avons testé les effets de fréquence du voisinage orthographique et de l'émotionalité de ce voisinage dans plusieurs tâches cognitives. Les résultats ont montré un effet de fréquence du voisinage orthographique qui était inhibiteur dans des tâches de reconnaissance visuelle de mots (Expériences 1-4) et facilitateur dans des tâches de catégorisation de couleur de mots (Expériences 6-8). L'inhibition lexicale ralentirait la reconnaissance du mot stimulus et diminuerait ainsi son effet d'interférence dans des tâches de type Stroop. De plus, la valence et le niveau d'arousal du voisin plus fréquent modifiaient également la vitesse de reconnaissance du mot stimulus (Etude préliminaire, Expérience 1-5) et la catégorisation de sa couleur (Expériences 6, 7 et 9). Le système affectif s'activerait lors de la lecture de mots avec un voisin émotionnel et modifierait la propagation d'activation et d'inhibition lexico-émotionnelles. De plus, les résultats indiquaient que ces effets de voisinage orthographique étaient sensibles aux caractéristiques des participants. Une diminution de l'effet de fréquence du voisinage selon l'âge a été montrée et interprétée en termes de déficits conjoints d'activation et d'inhibition lexicales (Expériences 4, 5, 8 et 9). Enfin, l'effet du voisinage émotionnel obtenu suggérait une préservation des processus lexico-émotionnels avec l'avancée en âge (Expériences 4, 5 et 9), mais cet effet était corrélé négativement avec le niveau d'alexithymie des individus (Expériences 2, 4 et 6). Dans l'ensemble, ces données soulignent l‟importance de la prise en compte du système affectif dans les modèles de reconnaissance visuelle des mots. / The aim of this thesis was to study lexical activation and inhibition processes underlying word reading and to determine the role of affective system on these processes in adults. For this, we investigated the effects of orthographic neighbourhood frequency and emotionality of this neighbourhood in several cognitive tasks. Results showed an orthographic neighbourhood frequency effect that was inhibitory in visual word recognition tasks (Experiments 1-4) and facilitatory in colour categorization tasks (Experiments 6-8). Lexical inhibition likely slows down the recognition of the stimulus word as well as diminishing its interference effect in Stroop-like tasks. Moreover, emotional valence and arousal level of the higher-frequency neighbour also modified the speed of stimulus word recognition (Preliminary study, Experiments 1-5) and its colour categorization (Experiments 6, 7 and 9). Thus, the affective system would be activated during reading of words with an emotional neighbour and would modify the spread of lexico-emotional activation and inhibition. Moreover, results indicated that these orthographic neighbourhood effects were sensitive to participants‟ characteristics. A decreaseof the orthographic neighbourhood effect depending on age was shown and interpreted in terms of deficits in both activation and inhibition processes (Experiments 4, 5, 8 and 9). Finally, the emotional neighbourhood effect that was obtained suggested a preservation of lexico-emotional processes with advance in age (Experiments 4, 5, and 9), but this effect was negatively correlated with individuals' level of alexithymia (Experiments 2, 4, and 6). Taken together, thes data underline the importance of taking the affective system into account in models of visual word recognition.

On a routine and daily basis, women are exposed to sexually objectifying experiences, which result in a number of harmful psychosocial outcomes (Fredrickson & Roberts, 1997). Five-hundred and forty-sex women attending a large, Southeastern university participated in this study that investigated a conceptual model of how childhood sexual abuse (CSA) contributes to sexual risk behaviors (SRBs) via self-objectification (S0). In order to assess the causal relationships among variables, measured variable path analyses were conducted in order to test two theoretical models. The following instruments were used in this investigation: the Sexual Abuse Subscale of the Childhood Trauma Questionnaire (a measure assessing experiences of childhood sexual abuse [Bernstein, Stein, Newcomb, Walker, Pogge, Ahluvia et al., 2003]); the Body Surveillance Subscale of the Objectified Body Consciousness Scale (a measure assessing self-objectification [McKinley & Hyde, 1996]); the Body Shame Subscale of the Objectified Body Consciousness Scale (a measure assessing body shame [McKinley & Hyde, 1996]); the Toronto Alexithymia Scale-20 (assesses alexithymic symptoms, or difficulty identifying, describing, and expressing one’s emotions [Bagby, Parker, & Taylor, 1994]), the Contraceptive Self-Efficacy Scale (assesses overall sexual self-efficacy, such as the ability to insist upon sexual protection [Levinson, 1986]), and the Sexual Risk Survey (assesses risky sexual practices [Turkchik & Garske, 2009]). Results revealed that the data fit the second model better than the first. Specifically, data revealed that CSA directly predicted SRBs and was not mediated via SO, but was partially mediated by alexithymia and body shame. That is, CSA predicted increased alexithymia and body shame. Increased alexithymia predicted SRBs, whereas body shame decreased SRBs. Results also revealed that alexithymia and body shame mediated the relationship between SO and SRBs. Specifically, self-objectification led to increased alexithymia and body shame, and alexithymia increased SRBs while body shame decreased SRBs. Last, results revealed that body shame fully mediated the relationship between both CSA and SO and sexual self-efficacy. Pathways were significant at the p < .05 level.

La forte prévalence des troubles des conduites alimentaires (TCA) chez les jeunes femmes et les faibles taux de rémission suite à un traitement ont encouragé les chercheurs à mieux comprendre les facteurs impliqués dans ce trouble mental. L’un des premiers modèles à mettre l’emphase sur des traits de personnalité associés au développement d’un TCA a été proposé par Hilde Bruch (1962, 1973, 1978) et a toujours une grande influence dans la recherche actuelle. Le modèle de Bruch inclue trois facteurs, soit l’insatisfaction corporelle, l’inefficacité et la conscience intéroceptive. Le but de cette thèse est d’apporter un support empirique au modèle de Bruch. En se basant sur une revue extensive des écrits scientifiques, cette thèse vise aussi à déterminer si deux facteurs reliés, soit l’alexithymie et le sentiment d’efficacité personnelle face à l’adoption de conduites alimentaires saines, améliorent la précision du modèle dans la prédiction de symptômes de TCA. Pour répondre empiriquement à cette question, il était d’abord nécessaire de disposer d’un questionnaire évaluant le sentiment d’efficacité personnelle en lien avec les conduites alimentaires qui peut être utilisé dans tout le spectre de présentation des TCA. Ainsi, le Eating Disorder Self-Efficacy Questionnaire (EDRSQ) a été adapté en français et ses propriétés psychométriques ont été évaluées. Une analyse factorielle confirmatoire a révélé une structure bi-factorielle, soit le sentiment d’efficacité personnelle en lien avec l’insatisfaction corporelle et avec l’adoption d’une alimentation normative. Chaque échelle a démontré une bonne fiabilité ainsi qu’une validité de construit cohérente avec la théorie. Par la suite, la capacité des facteurs proposés par Bruch à prédire les symptômes de TCA a été évaluée et comparée à des adaptations du modèle découlant des écrits. Au total, 203 étudiantes de premier cycle universitaire ont complété les versions validées en français du Eating Disorder Inventory 2, du Eating Attitudes Test, et du Toronto Alexithymia Scale en plus du EDRSQ. Les résultats montrent que le modèle de Bruch explique 46% de la variance des symptômes de TCA. Alors que l’insatisfaction corporelle et la conscience intéroceptive démontrent chacun une contribution importante dans la prédiction des symptômes de TCA, il a été démontré que l’inefficacité présente seulement une contribution spécifique négligeable. Le modèle de Bruch est amélioré par la substitution de l’inefficacité par le sentiment d’efficacité personnelle tel que mesuré par le EDRSQ; le modèle explique alors 64% de la variance des symptômes de TCA. Finalement, cette étude démontre que l’alexithymie n’a pas de contribution spécifique dans la prédiction des symptômes de TCA. Ainsi, la combinaison d’une faible conscience intéroceptive, de l’insatisfaction corporelle et d’un faible sentiment d’efficacité personnelle en lien avec les conduites alimentaires est fortement associée aux symptômes de TCA dans un échantillon non-clinique de jeunes femmes. Finalement, les implications conceptuelles et cliniques de ces résultats sont discutées. / High prevalence of Eating Disorders (EDs) amongst young women and poor treatment outcome rates have urged researchers to better understand premorbid factors involved in the pathology. One of the first models to emphasize premorbid personality factors in order to explain eating disorders was proposed by Hilde Bruch (1962, 1973, 1978) and is still very influential in today’s literature. Bruch’s model included three factors, namely body dissatisfaction, ineffectiveness and interoceptive awareness. The purpose of this thesis was to provide an empirical validation of Bruch’s theoretical model. Based on the literature, this thesis also aimed to determine if two related concepts, namely alexithymia and eating self-efficacy, improved the accuracy of the model in predicting ED symptoms. To empirically answer this question, it was first deemed necessary to have an eating self-efficacy questionnaire that could be used within all the spectrum of ED pathology. Therefore, the Eating Disorder Self-Efficacy Questionnaire (EDRSQ) was adapted to French and its’ psychometric properties were assessed. A confirmatory factor analysis revealed a bi-factorial structure, which were Body Image Self-Efficacy and Normative Eating Self-Efficacy. Both scales demonstrated evidence of reliability and theoretically consistent evidence of construct validity. Afterwards, Bruch’s factors ability to predict ED symptoms was assessed and then compared to empirically driven adaptations of the model. A total of 203 undergraduate females completed the Eating Disorder Inventory 2, Eating Attitudes Test, and Toronto Alexithymia Scale, as well as the EDRSQ. The results indicated that Bruch’s model accounted for 46% of the variance in eating symptomatology. While body dissatisfaction and interoceptive awareness both demonstrated an important contribution in predicting ED symptomatology, ineffectiveness was found to have a negligible specific contribution. Bruch’s model was improved by substituting ineffectiveness with eating self-efficacy as measured by the EDRSQ; it then accounted for 64% of the variance in eating disorder symptoms. Finally, this study demonstrated that alexithymia was not a specific predictor of eating disorder symptoms. Thus, a combination of lack of interoceptive awareness, body dissatisfaction and low eating self-efficacy is strongly associated with symptoms of eating disorders in a non-clinical sample of women. Finally, the conceptual and clinical implications of these findings were discussed.

Le premier objectif de cette étude était d’évaluer la relation entre l’alexithymie et différents troubles du sommeil chez des patients diagnostiqués (N= 580) selon la polysomnographie et la classification de l’American Academy of Sleep Medicine (AASM) et chez des sujets contrôle (N= 145) en utilisant l’Échelle d’Alexithymie de Toronto à 20 items (TAS-20). Le deuxième objectif était d’estimer le lien entre l’alexithymie et des caractéristiques de rêves suivant un Questionnaire sur les Rêves de 14 items. Les résultats confirment un lien entre l’alexithymie et les troubles du sommeil. Sa prévalence était supérieure dans le groupe clinique comparativement au groupe contrôle, et était différente selon les troubles. Les hommes cotaient plus haut que les femmes à l’Échelle d’Alexithymie de Toronto à 20 items (TAS-20) et sur ses sous-échelles DDF (difficulty describing feeling) et EOT (externally oriented thinking). L’EOT pourrait être impliquée dans les troubles de sommeil en étant l’unique sous-échelle, où un effet principal des diagnostics était significatif dans le groupe clinique. Pour les rêves, le score du TAS-20 corrélait positivement avec le facteur « détresse des cauchemars »; et négativement avec « rappel de rêves » et « signification des rêves ». Les sous-échelles du TAS-20 avaient des corrélations différentes: positive entre DIF et « détresse des cauchemars », négative entre DDF et « rappel de rêves » et EOT avec « signification des rêves ». À part quelques exceptions, ces modèles sont obtenus pour les groupes cliniques et non-cliniques, et pour les hommes et les femmes dans ces deux groupes. Ces résultats suggèrent un modèle consistant, et reproductible, de relations entre l’alexithymie et les composantes des rêves. / Using a large clinical group of patients suffering from sleep disorders (N= 580) and non-clinical comparison subjects (N= 145), the first aim of this study was to evaluate the relationship between alexithymia and several different sleep disorders diagnosed from polysomnography following the American Academy of Sleep Medicine classification system and using a Canadian French translation of the 20-item Toronto Alexithymia scale (TAS-20). The second objective was to evaluate the link between alexithymia and dream characteristics using a 14-item Dreaming Questionnaire. Results confirm a relationship between alexithymia and sleep disorders. Its prevalence was higher in the clinical than the non-clinical group and differed according to sleep disorder diagnoses. Men scored higher than women on the TAS-20 total score and on the DDF (difficulty describing feeling) and EOT (externally oriented thinking) subscales. EOT could be implicated in sleep disorders pathology as it was the only subscale for which the sleep diagnosis effect was significant in the clinical group. Concerning the second objective, TAS-20 total score correlated positively with nightmare distress and negatively with dream recall; and correlated negatively with dream meaning. TAS-20 subscales were differentially correlated with the 3 dream factors of the Dreaming Questionnaire: DIF with increased nightmare distress, DDF with decreased dream recall and EOT with decreased dream meaning. With some exceptions, these patterns were obtained for clinical and non-clinical groups and for men and women. These results suggest a consistent and replicable pattern of relationships between alexithymia and dreaming components.