{"id":{"repo_id":"radboud","oai_identifier":"oai:repository.ubn.ru.nl:2066/319687"},"canonical_url":"https://search.dev.ndltd.org/etd/radboud/oai:repository.ubn.ru.nl:2066/319687","repository":{"repo_id":"radboud","name":"Radboud University Nijmegen","base_url":"https://repository.ubn.ru.nl/oai/request"},"display":{"title":"Automated rodent behavior recognition: Machine learning strategies, behavioral challenges and practical solutions","abstract":"Contains fulltext : 319687.pdf (Publisher’s version ) (Open Access)","abstract_html":"Contains fulltext : 319687.pdf (Publisher’s version ) (Open Access)","abstract_has_math":false,"creators":["Dam, E.A. van"],"institution":"Nijmegen : Radboud University Press","degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Gerven, M.A.J. van","Noldus, L.P.J.J."],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025","date_published":"2025","updated_at":"2026-07-24T04:01:40Z","subjects":["Donders Series","Radboud Dissertation Series","Cognitive artificial intelligence"],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["10.54195/9789465150864","9789465150864","Donders Series ; Radboud Dissertation Series ; 728,"],"render_values":[{"text":"10.54195/9789465150864","href":"https://doi.org/10.54195/9789465150864","code":true},{"text":"9789465150864","href":null,"code":true},{"text":"Donders Series ; Radboud Dissertation Series ; 728,","href":null,"code":true}]}]},"links":{"outbound_url":"https://hdl.handle.net/2066/319687","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Gerven, M.A.J. van","Noldus, L.P.J.J."]},{"key":"dc:creator","label":"Author","values":["Dam, E.A. van"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2025"]},{"key":"dc:publisher","label":"Institution","values":["Nijmegen : Radboud University Press"]},{"key":"dc:type","label":"Dc Type","values":["Doctoral thesis"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Donders Series","Radboud Dissertation Series","Cognitive artificial intelligence"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://repository.ubn.ru.nl//bitstream/handle/2066/319687/319687.pdf","https://hdl.handle.net/2066/319687","10.54195/9789465150864","9789465150864","Donders Series ; Radboud Dissertation Series ; 728,"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Contains fulltext : 319687.pdf (Publisher’s version ) (Open Access)","Artificial intelligence (AI) is increasingly capable of understanding the world. This is eagerly used by behavioral researchers who want to automate their behavioral recordings. However, existing measurement systems either measure low-level behaviors such as posture and movement or are overly specialized for a dataset and not adaptable to ever-changing research needs. This thesis investigates the automated recognition of complex rodent behaviors using various machine learning strategies. It presents an automated behavior recognition (ABR) system that recognizes ten specific rodent behaviors and is currently used in over 300 academic and industrial laboratories worldwide. It investigates the use of deep learning for generic behavior classification and highlights three behavioral aspects that hamper recognition regardless the amount of training data. Practical solutions are described to annotate new behaviors, one of which is a hybrid active learning strategy that combines manual annotation with AI assistance for efficient and accurate annotation. Finally, the dissertation provides an outlook on new ways of behavior analysis with fully unsupervised detection of behavioral effects, and reflects on the ethical implications of automated behavior monitoring. This thesis shows that by combining the power of AI with human expertise and guidance, we can build ethical, usable, and beneficial applications for society.","Radboud University, 16 juni 2025","Promotores : Gerven, M.A.J. van, Noldus, L.P.J.J.","103 p."]},{"key":"dc:title","label":"Title","values":["Automated rodent behavior recognition: Machine learning strategies, behavioral challenges and practical solutions"]}]}],"canonical_facts":{"dc:contributor":["Gerven, M.A.J. van","Noldus, L.P.J.J."],"dc:creator":["Dam, E.A. van"],"dc:date":["2025"],"dc:description":["Contains fulltext : 319687.pdf (Publisher’s version ) (Open Access)","Artificial intelligence (AI) is increasingly capable of understanding the world. This is eagerly used by behavioral researchers who want to automate their behavioral recordings. However, existing measurement systems either measure low-level behaviors such as posture and movement or are overly specialized for a dataset and not adaptable to ever-changing research needs. This thesis investigates the automated recognition of complex rodent behaviors using various machine learning strategies. It presents an automated behavior recognition (ABR) system that recognizes ten specific rodent behaviors and is currently used in over 300 academic and industrial laboratories worldwide. It investigates the use of deep learning for generic behavior classification and highlights three behavioral aspects that hamper recognition regardless the amount of training data. Practical solutions are described to annotate new behaviors, one of which is a hybrid active learning strategy that combines manual annotation with AI assistance for efficient and accurate annotation. Finally, the dissertation provides an outlook on new ways of behavior analysis with fully unsupervised detection of behavioral effects, and reflects on the ethical implications of automated behavior monitoring. This thesis shows that by combining the power of AI with human expertise and guidance, we can build ethical, usable, and beneficial applications for society.","Radboud University, 16 juni 2025","Promotores : Gerven, M.A.J. van, Noldus, L.P.J.J.","103 p."],"dc:identifier":["https://repository.ubn.ru.nl//bitstream/handle/2066/319687/319687.pdf","https://hdl.handle.net/2066/319687","10.54195/9789465150864","9789465150864","Donders Series ; Radboud Dissertation Series ; 728,"],"dc:publisher":["Nijmegen : Radboud University Press"],"dc:subject":["Donders Series","Radboud Dissertation Series","Cognitive artificial intelligence"],"dc:title":["Automated rodent behavior recognition: Machine learning strategies, behavioral challenges and practical solutions"],"dc:type":["Doctoral thesis"]},"updated_at":"2026-07-24T04:01:40Z"}