Canal de comunicação não invasivo para transferência de informações entre grupos de Zebrafish

• Ano de defesa: 2021
• Autor(a) principal: BARREIROS, Marta de Oliveira
• Orientador(a): BARROS FILHO, Allan Kardec Duailibe
• Banca de defesa: BARROS FILHO, Allan Kardec Duailibe , RIBEIRO, Sidarta Tollendal Gomes , SILVA, Aristofanes Correa , SOUZA, Francisco das Chagas de , PAES, Antônio Marcus de Andrade
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Maranhão
• Programa de Pós-Graduação: PROGRAMA DE PÓS-GRADUAÇÃO EM ENGENHARIA DE ELETRICIDADE/CCET
• Departamento: DEPARTAMENTO DE ENGENHARIA DA ELETRICIDADE/CCET
• País: Brasil
• Palavras-chave em Português: Zebrafish; Comportamento; Redes neurais aritificiais; Transferência de informação
• Palavras-chave em Inglês: Zebrafish; Behavior; Artificial neural networks; Information transfer
• Área do conhecimento CNPq: Metodologia e Técnicas da Computação
• Link de acesso: https://tedebc.ufma.br/jspui/handle/tede/tede/3474

Communication between individuals has made tremendous progress with technological advances in recent years. And, today there is a possibility that a brain communicates directly to another brain, using a brain-brain interface. BBI is a technology capable of transferring direct information between brains, combining neuroimaging and neurostimulation methods, where neural signals are extracted from an “ emitting ” brain and recoded into a “ receiving ” brain. Although there are different approaches to using this technique, it is still not entirely clear about the effectiveness of BBI and how the transfer of information is received by the “receiving” brain. Therefore, this work proposes a new communication channel, totally non-invasive, to transfer information between groups of fish, using an approach analogous to BBI. The proposed communication channel uses the position of each fish and video images as an analysis. Thus, the YOLOv2 network and the Kalman filter served to obtain the position and tracking of each fish in the tank, obtaining the behavior pattern of a group of zebrafish during the experiments, this pattern being modulated for generate information transmitted from one group of fish to another, synchronizing communication between aquariums. In addition, a VGG-16 convolutional network architecture and LSTM recurrent network use a sequence of video images to assess the behavior of fish during activities. To test the communication channel, food appetite conditioning proposals were created, where the fish should move from side to side according to the connected stimulus. And, meanwhile, their behavior was assessed by a recurrent neural network that classified the behavior into three behavioral conditions and sent a feedback to the sending group. To assess the response of the zebrafish, three strategies were developed with total, partial and unrestricted restrictions. The results showed that it was possible to synchronize bilateral information, showing that the recipient fish were able to respond with similar behavior in the groups sending the information. The performance of the communication channel was quantified in terms of the amount of information transmitted, as well as the precision of the response of the sending and receiving fish, until reaching the objective of the conditioned task. The results show that the proposed communication channel is promising and extendable to new behavioral perspectives.