Repository logo
 
Publication

An unsupervised generative strategy for detection and characterization of rare behavioural events in mice in open fiel to assess effect of optogenetic activation of serotonergic neurons in the dorsal raphe nuclei

datacite.subject.fosCiências Sociais::Psicologiapt_PT
datacite.subject.fosCiências Médicaspt_PT
dc.contributor.advisorMainen, Zachary F
dc.contributor.advisorCorreia, Luís
dc.contributor.authorSoares, Lucas B. Nicolosi
dc.date.accessioned2020-01-16T14:12:15Z
dc.date.available2020-01-16T14:12:15Z
dc.date.issued2019-10-23
dc.date.submitted2019-09-05
dc.description.abstractThe purpose of our work is to provide an unsupervised deep learning tool that uses predictability of behavior as a meaningful metric to quantify the di erences between normal and abnormal behavior in the context of an experiment where mice receive optogenetic stimulation in their serotonergic neurons located in the dorsal raphe nuclei. We use generative adversarial networks to learn, on a training subset of the videos, a baseline behavioral repertoire by predicting future frames from subsequent frames in the past. By de ning a predictability index as dissimilarity between the quality of the generated prediction and the ground truth frame, we are able to determine in which frames a behavior not observed by the model during training is performed and therefore, we can detect the presence of stimulation by only analysing the uctuations of this index that indicate when the mouse is performing behaviors that are not present in the learnt baseline.pt_PT
dc.description.abstractO objetivo do presente trabalho é fornecer uma ferramenta de aprendizado profundo que utiliza a previsibilidade do comportamento animal como uma métrica para quanti car as diferenças entre o comportamento normal e anormal no contexto de um experimento em que os ratos recebem estimulação optogenética em seus neurónios serotoninérgicos localizados na região cerebral do núcleo dorsal da rafe. Usamos redes neurais generativas para aprender, treinando em segmentos de vídeo, um repertório comportamental básico, prevendo cenas futuras a partir de cenas subsequentes do passado. Ao defi nir um índice de previsibilidade como o grau de dissimilaridade entre a previsão gerada pelo modelo e a cena real no momento correspondente, conseguimos determinar em quais cenas o comportamento não observado pelo modelo durante o treinamento é realizado e, portanto, pudemos detectar a presença de estimulação apenas analisando as flutuações desse índice que indicam quando o rato está executando comportamentos que não estão presentes na base aprendida.pt_PT
dc.identifier.tid202329895pt_PT
dc.identifier.urihttp://hdl.handle.net/10451/40873
dc.language.isoengpt_PT
dc.subjectRedes neurais (Fisiologia)
dc.subjectAprendizagem automática
dc.subjectControlo electrónico do comportamento
dc.subjectSerotonina
dc.subjectInteligência artificial
dc.subjectTeses de mestrado - 2019
dc.titleAn unsupervised generative strategy for detection and characterization of rare behavioural events in mice in open fiel to assess effect of optogenetic activation of serotonergic neurons in the dorsal raphe nucleipt_PT
dc.typemaster thesis
dspace.entity.typePublication
rcaap.rightsopenAccesspt_PT
rcaap.typemasterThesispt_PT
thesis.degree.nameMestrado em Ciência Cognitivapt_PT

Files

Original bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
ulfl274966_tm.pdf
Size:
6.67 MB
Format:
Adobe Portable Document Format
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.2 KB
Format:
Item-specific license agreed upon to submission
Description: