Bridging the gap between statistics, machine learning and clinical practice: an interdisciplinary collaborative study on human brain mapping

• Processo: 10/51473-8
• Modalidade de apoio: Auxílio à Pesquisa - Regular
• Data de Início da vigência: 01 de setembro de 2010
• Data de Término da vigência: 30 de abril de 2012
• Área do conhecimento: Engenharias - Engenharia Biomédica - Engenharia Médica
• Acordo de Cooperação: King's College London
• Pesquisador responsável: João Ricardo Sato
• Beneficiário: João Ricardo Sato
• Pesquisador Responsável no exterior: Michael John Brammer
• Instituição Parceira no exterior: King's College London, Inglaterra
• Instituição Sede: Centro de Matemática, Computação e Cognição (CMCC). Universidade Federal do ABC (UFABC). Ministério da Educação (Brasil). Santo André , SP, Brasil
• Vinculado ao auxílio: 10/01394-4 - Métodos estatísticos e computacionais para a discriminação de alterações anatômicas, estados mentais e identificação da conectividade cerebral: uma abordagem integrativa utilizando ressonância magnética estrutural, funcional e EEG, AP.R
• Assunto(s): Ressonância magnética Epilepsia Reconhecimento de padrões Aprendizado computacional
• Palavra(s)-Chave do Pesquisador: Connectivity | Epilepsy | Machine Learning | Magnetic Resonance | Neuroimaging | Pattern Recognition

Resumo

A major recent technical innovation for mapping the structure and function of the human brain has been the ability to acquire magnetic resonance images (fMRI) with high spatial resolution. MRI images represent a very large amount of information and effective use of these data has become critical for comprehension of cognitive processes and diseases and for clinical application of MRI. A central process in clinical application of MRI is the ability to differentiate between patterns of response between populations (patients/controls) or mental states and to investigate connections between brain regions. This characterization can be made efficiently using Pattern Recognition and brain connectivity methods explored in FAPESP project number 2010/1394-4, with which the current project is associated. The aim of these methods is the prediction of subject group (e.g. patient/control) and mental states. This approach, sometimes called brain decoding, is proving useful in many clinical applications. The fMRI and MRI datasets to be analyzed in this project are from patients with epilepsy (FAPESP-CINAPCE Project). The proposed research project has a strong multidisciplinary profile and involves groups in Kings College London and Sao Paulo. Each of the research groups involved in this proposal is multidisciplinary and has done pioneering work in the area of human brain mapping and Neuroimaging. In addition, Prof. Michael John Brammer, the team leader from King's College London, is an associated researcher of the FAPESP project 2010/1394-4. The cooperation, interaction and team work between researchers from both São Paulo state and King's College that is crucial to guarantee the quality and high impact of this research has been shown to work well in past research projects. (AU)