Machine Learning for Analysis of Brain Signals / Event-Related Potential (ERP) measures derived from the electroencephalogram (EEG) have been widely used in outcome prediction of brain disorders. Recently, the ERPs that are transient (EEG) responses to auditory, visual, or tactile stimuli, have been introduced as useful predictors of a positive coma outcome (i.e. emergence from coma). In this study, machine learning techniques were applied for detecting the Mismatch Negativity (MMN) component, which is a transient EEG response to auditory stimuli, and its existence has a high correlation with coma awakening, through analyzing ERPs signals recorded from healthy control brain signals. To this end, two different dimensionality reduction methods, Localized Feature Selection (LFS) and minimum-redundancy maximum-relevance (mRMR) were employed, where a localized classifier and the support vector machine (SVM) with radial basis function (RBF) kernel are used as classifiers. We trained both LFS and mRMR algorithms using signals of healthy brains and evaluated their performance for MMN detection on both healthy subjects and coma patients. The evaluation on healthy subjects, using leave-one-subject-out cross-validation technique, shows the detection accuracy performance of 86.6% (using LFS) and 86.5% (using mRMR). In addition to analyzing brain signals for MMN detection, we also implemented a machine learning algorithm for discriminating healthy subjects from those who have experienced TBI. The EEG signals used in the TBI study were recorded using an ERP paradigm. However, we treated the recorded signals as resting state signals. To this end, we used the mRMR feature selection method and fed the selected features into the SVM classifier that outputs the estimated class labels. This method gives us a poor performance compared to the methods that directly used ERP components (without considering them as resting signals.). We conclude that our hypothesis of treating ERP data as resting data is not valid for TBI detection. / Thesis / Master of Applied Science (MASc)

Assessing cognitive abilities in disorders of consciousness (DOC) relies on assessments of overt behaviour, such as the ability to follow commands. Neuroimaging has shown that absence of overt behavior does not necessarily indicate absence of covert cognition, raising questions about behaviour-only assessment. Several electroencephalographic (EEG) markers of higher cognitive functions (event-related potentials; ERPs) have shown the potential to differentiate between DOC states, as well as predict awakening and condition upon emergence. However, no one ERP has emerged with sensitivity and specificity high enough to be widely accepted, showing that further investigation is needed. More recently, evidence has emerged for fluctuations of ERP detectability in DOC over the course of several hours, and for prognostic power of changes in ERP presentation between testing sessions. This investigation builds on such findings towards improving evaluation of cognition in DOC. A testing battery combining several well-known auditory ERPs was administered to a comatose patient over a 24-hour period during two recording sessions one week apart, as well as to a sample of healthy young adults. The patient scored 3 and 6 on the Glasgow Coma Scale (GCS) during the first and second session, respectively. The results show that changes in GCS score were accompanied by changes in ERP detectability. The results also suggest detectability fluctuations over the course of 24 hours, which in turn suggests that repeated testing is necessary for complete evaluation. Future work should validate these findings with a larger sample; additionally, establishing population norms for single-subject prevalence, latency, and amplitude of ERPs would improve confidence in interpreting patient results. With the current understanding of both healthy and DOC ERPs, detecting ERP presence may contribute to a positive DOC prognosis with a degree of confidence, but caution must be exercised in making negative prognoses or high-stakes care decisions based on ERP absence. / Thesis / Master of Science (MSc) / Assessing cognitive abilities in disorders of consciousness such as coma currently relies on assessments of overt behaviour, such as the ability to follow commands or react to a stimulus. Neuroimaging has shown that absence of overt behavior does not necessarily indicate absence of covert cognition, raising questions about behaviour-only assessment. This study describes a neuroimaging testing battery aimed at evaluating a hierarchy of cognitive functions without the need for a behavioural response by measuring brain activity driven by auditory stimulation. This battery was administered to a comatose patient over a 24-hour period during two recording sessions one week apart, as well as to a sample of healthy young adults. The results show that changes in the patient’s condition between testing sessions was accompanied by detectable and quantifiable change in their stimulus-driven brain activity. The results also suggest fluctuations in the patient’s ability to produce detectable responses over the course of 24 hours, which in turn suggests that repeated testing is necessary for a complete evaluation. Overall, neuroimaging provides a promising avenue for non-behavioral assessment of cognition, which will greatly benefit a population whose physical faculties may be compromised.

Os objetivos desse trabalho foram: (1) verificar a influência da música e mensagem oral sobre os Sinais vitais e Expressão facial dos pacientes em coma fisiológico ou induzido; (2) relacionar a existência de responsividade do paciente com a Escala de Coma de Glasgow ou com a Escala de Sedação de Ramsay, no que se refere ao estímulo musical e ao estímulo verbal. Foi realizado um Ensaio Clínico Controlado e Randomizado, a amostra consistiu-se de 30 pacientes de uma UTI, que foram divididos em 2 grupos: Grupo controle (sem estímulos auditivos) e Grupo experimental (com estímulos auditivos). Os familiares elaboraram uma mensagem gravada com a sua própria voz e escolheram uma música de acordo com a preferência do paciente, que foram gravadas no mesmo CD. Foram feitos dois CD’s idênticos em tempo de gravação; um com e outro sem estímulos para, então, ser feita a divisão aleatória dos grupos. Os pacientes foram avaliados segundo uma das Escalas, usaram fones de ouvido e foram submetidos a três sessões, em dias consecutivos. Durante as sessões, os dados referentes aos Sinais vitais e Expressão facial foram anotados em um instrumento de coleta de dados. Encontramos alterações estatisticamente significativas nos Sinais vitais (Saturação de O2 – sessão 1; Saturação de O2 – sessão 3; Frequência respiratória – sessão 3) durante a mensagem e na Expressão facial, sessão 1, durante a música e a mensagem. Aparentemente a mensagem foi um estímulo mais forte do que a música em relação à capacidade de produzir respostas fisiológicas sugestivas de audição. Em relação ao Pulso, já comentado em estudos semelhantes, não encontramos dados estatisticamente significantes, nem indicativos de tendências. Observarmos responsividade aos estímulos auditivos em todas as pontuações da Glasgow do Grupo experimental, no momento em que obtivemos diferenças estatisticamente significantes; portanto, podemos afirmar que essa variável pouco influenciou as respostas neste estudo. Quanto à Escala de Sedação de Ramsay, apenas 1 desses pacientes foi avaliado com essa Escala, o que dificulta qualquer análise. Este estudo sugere a necessidade e a importância de outras pesquisas nesta linha. / The objectives of this work was: (1) to check the music and verbal message influence on the vital signs and face expression of the patients in physiological or induced coma; (2) connect the existence of patient’s responsiveness with Glasgow Coma Scale or with Ramsay Sedation Scale, considering the musical and verbal stimuli. It is a randomized controlled clinical trial. The sample consisted of 30 patients of intensive care unit, who were divided in 2 groups: control group (without auditory stimulus) and experimental group (with auditory stimulus). The relatives elaborated a voice recorded message and choose a music according to the patient preference, which were recorded in the same Compact Disc. Two identical CD\'s were made at the same recorded time; one with and another one without stimulus in order to make the random division of the groups. The patients were evaluated according to one of the Scales; they used earphones and were submitted to three sessions in consecutive days. During the sessions, the data relating to the vital signs and the face expression were recorded in a data collection instrument. We found out significant statistics alterations on the vital signs (O2 saturation - session 1; O2 saturation - session 3; respiratory frequency - session 3) during the message and at the face expression, session 1, during the music and the message. Apparently, the message was a stimulus stronger than the music in respect to the capacity to produce suggestive physiological auditory answers. In respect to the pulse, already mentioned in similar studies, we did not find significant statistic data, or tendency marks. We observed responsiveness to the auditory stimulus in all the Glasgow scores in the experimental group, at the moment when we got significant differences; therefore, we may assert that this variable poorly influenced the answers in this study. About the Ramsay Sedation Scale, just 1 of these patients was evaluated by that, which makes any analysis difficult. This study suggests the necessity and the importance of other researches in this line.

Introdução: A lesão cerebral traumática decorrente do traumatismo crânio encefálico (TCE) é uma das principais causas de morte e incapacidade em todo o mundo. A estimativa do prognóstico é, por definição, um desafio multivariável. Objetivos: Estabelecer modelo prognóstico para avaliação das probabilidades de sobrevida após TCE com base nas características de admissão, incluindo lesões extracranianas, de doentes não sedados e estáveis quanto a hemodinâmica e respiração, apresentando tomografia de crânio com alterações. Doentes e Métodos: Um estudo de coorte avaliou 1275 doentes com TCE e tomografia do crânio (TC) anormais na admissão na unidade de emergência do HCFMUSP e analisou a mortalidade final. Realizou-se uma análise de regressão logística para determinar a relação de cada variável independente no desfecho final. Resultados: Quatro variáveis foram consideradas significativas no modelo: idade (anos), Escala de Coma de Glasgow (3-15), Escala de Marshall (estratificada em 2,3 ou 4,5,6, de acordo com o melhor grupo positivo (Valor preditivo) e anisocoria (sim / não). Obteve-se uma fórmula que está em um modelo logístico (USP índex do TCE) que estima a probabilidade de morte dos doentes de acordo com características da admissão. Conclusão: O modelo de probabilidade matemática (USP Index do TCE) é prático e acessível aos serviços de neurocirurgia. Foi criado para melhorar a qualidade da informação sobre a possibilidade de morte durante a hospitalização dos doentes acometidos pelo TCE, e pode ser aplicado pelos médicos para auxiliar à tomada de decisões / Introduction: Traumatic brain injury (TBI) is a major cause of death and disability worldwide. Estimation of prognosis is by definition a multivariable challenge. Objectives: The authors aimed to develop prognostic model for assessment of survival chances after TBI based on admission characteristics, including extracranial injuries of non-sedated and stable patients regarding hemodynamics and respiration, which would allow application of the model before in-hospital therapeutic interventions. Patients and Methods: 1275 patients were evaluated in a cohort study with TBI and abnormal CT scans upon admission to the emergency unit of Hospital das Clinicas of University of Sao Paulo. Mortality was analyzed as the final outcome. A logistic regression analysis was undertaken to determine the adjusted weigh of each independent variable in the outcome. Results: Four variables were found to be significant in the model: age (years), Glasgow Coma Scale (3-15), Marshall Scale (MS, stratified into 2,3 or 4,5,6; according to the best group positive predictive value) and anysochoria (yes/no). The formula in a logistic model (USP index to head injury) estimates the probability of death of patients according to characteristics that influence on mortality. Conclusion: The mathematical probability model (USP Index of TBI) is practical and accessible to neurosurgery services. It was created to improve the quality of information about the possibility of death during hospitalization of patients affected by the TBI and can be applied by doctors to aid in decision making

Os objetivos desse trabalho foram: (1) verificar a influência da música e mensagem oral sobre os Sinais vitais e Expressão facial dos pacientes em coma fisiológico ou induzido; (2) relacionar a existência de responsividade do paciente com a Escala de Coma de Glasgow ou com a Escala de Sedação de Ramsay, no que se refere ao estímulo musical e ao estímulo verbal. Foi realizado um Ensaio Clínico Controlado e Randomizado, a amostra consistiu-se de 30 pacientes de uma UTI, que foram divididos em 2 grupos: Grupo controle (sem estímulos auditivos) e Grupo experimental (com estímulos auditivos). Os familiares elaboraram uma mensagem gravada com a sua própria voz e escolheram uma música de acordo com a preferência do paciente, que foram gravadas no mesmo CD. Foram feitos dois CD’s idênticos em tempo de gravação; um com e outro sem estímulos para, então, ser feita a divisão aleatória dos grupos. Os pacientes foram avaliados segundo uma das Escalas, usaram fones de ouvido e foram submetidos a três sessões, em dias consecutivos. Durante as sessões, os dados referentes aos Sinais vitais e Expressão facial foram anotados em um instrumento de coleta de dados. Encontramos alterações estatisticamente significativas nos Sinais vitais (Saturação de O2 – sessão 1; Saturação de O2 – sessão 3; Frequência respiratória – sessão 3) durante a mensagem e na Expressão facial, sessão 1, durante a música e a mensagem. Aparentemente a mensagem foi um estímulo mais forte do que a música em relação à capacidade de produzir respostas fisiológicas sugestivas de audição. Em relação ao Pulso, já comentado em estudos semelhantes, não encontramos dados estatisticamente significantes, nem indicativos de tendências. Observarmos responsividade aos estímulos auditivos em todas as pontuações da Glasgow do Grupo experimental, no momento em que obtivemos diferenças estatisticamente significantes; portanto, podemos afirmar que essa variável pouco influenciou as respostas neste estudo. Quanto à Escala de Sedação de Ramsay, apenas 1 desses pacientes foi avaliado com essa Escala, o que dificulta qualquer análise. Este estudo sugere a necessidade e a importância de outras pesquisas nesta linha. / The objectives of this work was: (1) to check the music and verbal message influence on the vital signs and face expression of the patients in physiological or induced coma; (2) connect the existence of patient’s responsiveness with Glasgow Coma Scale or with Ramsay Sedation Scale, considering the musical and verbal stimuli. It is a randomized controlled clinical trial. The sample consisted of 30 patients of intensive care unit, who were divided in 2 groups: control group (without auditory stimulus) and experimental group (with auditory stimulus). The relatives elaborated a voice recorded message and choose a music according to the patient preference, which were recorded in the same Compact Disc. Two identical CD\'s were made at the same recorded time; one with and another one without stimulus in order to make the random division of the groups. The patients were evaluated according to one of the Scales; they used earphones and were submitted to three sessions in consecutive days. During the sessions, the data relating to the vital signs and the face expression were recorded in a data collection instrument. We found out significant statistics alterations on the vital signs (O2 saturation - session 1; O2 saturation - session 3; respiratory frequency - session 3) during the message and at the face expression, session 1, during the music and the message. Apparently, the message was a stimulus stronger than the music in respect to the capacity to produce suggestive physiological auditory answers. In respect to the pulse, already mentioned in similar studies, we did not find significant statistic data, or tendency marks. We observed responsiveness to the auditory stimulus in all the Glasgow scores in the experimental group, at the moment when we got significant differences; therefore, we may assert that this variable poorly influenced the answers in this study. About the Ramsay Sedation Scale, just 1 of these patients was evaluated by that, which makes any analysis difficult. This study suggests the necessity and the importance of other researches in this line.

Thèse d'exercice : Médecine. Neurologie : Paris 12 : 2006. / Titre provenant de l'écran-titre. Bibliogr. f. 194-205.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Partindo da premissa de que o paciente em coma é um sujeito, e que o seu processo comatoso é uma alteração do estado de vida, em contraposição ao imaginário social que lhe atribui estereótipos de sono próximo a morte, inércia, suspensão de sua vida mental e impotência. Esta dissertação tem como objetivo demonstrar que o coma é demarcado no limite entre o caos e o desejo na medida em que cada paciente vivencia o processo comatoso de forma singular e demarcado conforme sua experiência subjetiva, implementando movimentos em direção a processos de vida ou de morte. Para tanto foram utilizados análise de casos clínicos, entrevistas com médicos, familiares e pacientes que saíram do coma, além de pesquisa bibliográfica, utilizando a psicanálise como arcabouço teórico que fundamenta as inferências e conclusões propostas nesta dissertação. / Assuming that the comatose patient is a human been, and that his comatose process is an altered state of life, in contrast to the picture that gives you social stereotypes sleep near death, inertia, suspension of your mental life and impotence. This dissertation aims to demonstrate that the coma is demarcated the border between chaos and desire in that each patient experiences the process of natural Comatose shape and marked according to their subjective experience, moves toward implementing processes of life or death . For this analysis of clinical cases, interviews with doctors, patients and family members who came out of the coma were used in addition to literature, using psychoanalysis as a theoretical framework that underlies the inferences and conclusions proposed in this dissertation.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Partindo da premissa de que o paciente em coma é um sujeito, e que o seu processo comatoso é uma alteração do estado de vida, em contraposição ao imaginário social que lhe atribui estereótipos de sono próximo a morte, inércia, suspensão de sua vida mental e impotência. Esta dissertação tem como objetivo demonstrar que o coma é demarcado no limite entre o caos e o desejo na medida em que cada paciente vivencia o processo comatoso de forma singular e demarcado conforme sua experiência subjetiva, implementando movimentos em direção a processos de vida ou de morte. Para tanto foram utilizados análise de casos clínicos, entrevistas com médicos, familiares e pacientes que saíram do coma, além de pesquisa bibliográfica, utilizando a psicanálise como arcabouço teórico que fundamenta as inferências e conclusões propostas nesta dissertação. / Assuming that the comatose patient is a human been, and that his comatose process is an altered state of life, in contrast to the picture that gives you social stereotypes sleep near death, inertia, suspension of your mental life and impotence. This dissertation aims to demonstrate that the coma is demarcated the border between chaos and desire in that each patient experiences the process of natural Comatose shape and marked according to their subjective experience, moves toward implementing processes of life or death . For this analysis of clinical cases, interviews with doctors, patients and family members who came out of the coma were used in addition to literature, using psychoanalysis as a theoretical framework that underlies the inferences and conclusions proposed in this dissertation.

Introdução: A lesão cerebral traumática decorrente do traumatismo crânio encefálico (TCE) é uma das principais causas de morte e incapacidade em todo o mundo. A estimativa do prognóstico é, por definição, um desafio multivariável. Objetivos: Estabelecer modelo prognóstico para avaliação das probabilidades de sobrevida após TCE com base nas características de admissão, incluindo lesões extracranianas, de doentes não sedados e estáveis quanto a hemodinâmica e respiração, apresentando tomografia de crânio com alterações. Doentes e Métodos: Um estudo de coorte avaliou 1275 doentes com TCE e tomografia do crânio (TC) anormais na admissão na unidade de emergência do HCFMUSP e analisou a mortalidade final. Realizou-se uma análise de regressão logística para determinar a relação de cada variável independente no desfecho final. Resultados: Quatro variáveis foram consideradas significativas no modelo: idade (anos), Escala de Coma de Glasgow (3-15), Escala de Marshall (estratificada em 2,3 ou 4,5,6, de acordo com o melhor grupo positivo (Valor preditivo) e anisocoria (sim / não). Obteve-se uma fórmula que está em um modelo logístico (USP índex do TCE) que estima a probabilidade de morte dos doentes de acordo com características da admissão. Conclusão: O modelo de probabilidade matemática (USP Index do TCE) é prático e acessível aos serviços de neurocirurgia. Foi criado para melhorar a qualidade da informação sobre a possibilidade de morte durante a hospitalização dos doentes acometidos pelo TCE, e pode ser aplicado pelos médicos para auxiliar à tomada de decisões / Introduction: Traumatic brain injury (TBI) is a major cause of death and disability worldwide. Estimation of prognosis is by definition a multivariable challenge. Objectives: The authors aimed to develop prognostic model for assessment of survival chances after TBI based on admission characteristics, including extracranial injuries of non-sedated and stable patients regarding hemodynamics and respiration, which would allow application of the model before in-hospital therapeutic interventions. Patients and Methods: 1275 patients were evaluated in a cohort study with TBI and abnormal CT scans upon admission to the emergency unit of Hospital das Clinicas of University of Sao Paulo. Mortality was analyzed as the final outcome. A logistic regression analysis was undertaken to determine the adjusted weigh of each independent variable in the outcome. Results: Four variables were found to be significant in the model: age (years), Glasgow Coma Scale (3-15), Marshall Scale (MS, stratified into 2,3 or 4,5,6; according to the best group positive predictive value) and anysochoria (yes/no). The formula in a logistic model (USP index to head injury) estimates the probability of death of patients according to characteristics that influence on mortality. Conclusion: The mathematical probability model (USP Index of TBI) is practical and accessible to neurosurgery services. It was created to improve the quality of information about the possibility of death during hospitalization of patients affected by the TBI and can be applied by doctors to aid in decision making

Comets are important for solar system studies because their interiors hold evidence of the conditions in which they formed in the outer solar system. However, the coma obscures the nucleus from view when observations are most easily performed, thus it is important to understand the nature of cometary comae. This study examines the spatial distribution of CN radicals in the coma of comet Encke and determines the likelihood that CN is a photodissociative daughter of HCN in the coma. Observations of CN were obtained from October 22-24, 2003, using the 2.7 m Cassegrain telescope at McDonald Observatory in Fort Davis, TX. The classical vectorial model was modified by introducing a fan-like feature in order to explain Encke’s aspherical coma. The results are consistent with HCN being the photodissociative parent of CN, based on the OH/CN ratios and the physical parameters used to match the model profiles with the observations.