{"id":{"repo_id":"calpoly","oai_identifier":"oai:digitalcommons.calpoly.edu:theses-4756"},"canonical_url":"https://search.ndltd.org/etd/calpoly/oai:digitalcommons.calpoly.edu:theses-4756","repository":{"repo_id":"calpoly","name":"Cal Poly","base_url":"https://digitalcommons.calpoly.edu/do/oai/"},"display":{"title":"TopoDINO: Self-Supervised Topological Representation Learning for Neuronal Morphologies","abstract":"Neuronal cell types are categorized by transcriptomic identity, yet their morphological heterogeneity defies this classification. In response, researchers have adopted unsupervised graph representation learning as a tool to reveal morphological variation within single-class transcriptomic types. However, the complex geometry of neuronal morphology—especially long axons and dense dendrites—challenges graph neural networks, which struggle with message propagation across extended structures. To mitigate this, current approaches enforce sub-sampling on neuronal graphs and omit axons entirely, sacrificing critical biological features for computational efficiency. To overcome this trade-off, this thesis introduces TopoDINO, a self-supervised, topology-aware representation learning model designed to preserve the full hierarchical organization of neuronal morphology. Instead of relying on subsampled graph approximations, TopoDINO employs a novel topology-lifting approach that transforms neuronal graphs into combinatorial complexes, capturing the inherent hierarchical features of neurons. In addition, we pre-train TopoDINO on the SEU-D15K dataset. As a result, TopoDINO not only provides a biologically grounded embedding space by preserving neuronal compartmentalization, but also mitigates labeling inconsistencies across laboratories by learning in a fully data-driven manner from 12,353 dendritic reconstructions derived from 204 mouse brains, all registered to the Allen Mouse Common Coordinate Framework (CCF).","abstract_html":"Neuronal cell types are categorized by transcriptomic identity, yet their morphological heterogeneity defies this classification. In response, researchers have adopted unsupervised graph representation learning as a tool to reveal morphological variation within single-class transcriptomic types. However, the complex geometry of neuronal morphology—especially long axons and dense dendrites—challenges graph neural networks, which struggle with message propagation across extended structures. To mitigate this, current approaches enforce sub-sampling on neuronal graphs and omit axons entirely, sacrificing critical biological features for computational efficiency. To overcome this trade-off, this thesis introduces TopoDINO, a self-supervised, topology-aware representation learning model designed to preserve the full hierarchical organization of neuronal morphology. Instead of relying on subsampled graph approximations, TopoDINO employs a novel topology-lifting approach that transforms neuronal graphs into combinatorial complexes, capturing the inherent hierarchical features of neurons. In addition, we pre-train TopoDINO on the SEU-D15K dataset. As a result, TopoDINO not only provides a biologically grounded embedding space by preserving neuronal compartmentalization, but also mitigates labeling inconsistencies across laboratories by learning in a fully data-driven manner from 12,353 dendritic reconstructions derived from 204 mouse brains, all registered to the Allen Mouse Common Coordinate Framework (CCF).","abstract_has_math":false,"creators":["Binbisher, Yasser"],"institution":null,"degree_name":"MS in Computer Science","degree_level":null,"degree_discipline":"Computer Science","degree_department":null,"school":null,"contributors":["Franz J. Kurfess","Computer Science","College of Engineering"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":null,"date_issued":"","date_published":null,"updated_at":"2026-11-19T07:12:41Z","subjects":["Neuronal Morphology | Morphological Heterogeneity | Graph Representation Learning | Self-Supervised Learning | Combinatorial Complexes | Topology","Artificial Intelligence and Robotics","Bioinformatics","Biomedical Informatics","Biostatistics","Cell Biology","Computational Biology","Computational Neuroscience","Data Science","Molecular and Cellular Neuroscience","Other Neuroscience and Neurobiology","Theory and Algorithms"],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[],"person_identifier_entries":[]},"links":{"outbound_url":"https://digitalcommons.calpoly.edu/theses/3127","outbound_label":"Repository record","outbound_source":"dc:identifier","record_url":"https://digitalcommons.calpoly.edu/theses/3127","record_label":"Repository record","record_kind":"record","record_source":"dc:identifier","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://digitalcommons.calpoly.edu/do/oai/?verb=GetRecord&metadataPrefix=oai_etdms&identifier=oai%3Adigitalcommons.calpoly.edu%3Atheses-4756","prefix":"oai_etdms"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Franz J. Kurfess","Computer Science","College of Engineering"]},{"key":"dc:creator","label":"Author","values":["Binbisher, Yasser"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.available","label":"Dc Date Available","values":["2025-06-13T07:00:00Z"]},{"key":"dc:date.created","label":"Dc Date Created","values":["2025-06-01T07:00:00Z"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Computer Science"]},{"key":"thesis:degree_name","label":"Degree Name","values":["MS in Computer Science"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Neuronal Morphology | Morphological Heterogeneity | Graph Representation Learning | Self-Supervised Learning | Combinatorial Complexes | Topology","Artificial Intelligence and Robotics","Bioinformatics","Biomedical Informatics","Biostatistics","Cell Biology","Computational Biology","Computational Neuroscience","Data Science","Molecular and Cellular Neuroscience","Other Neuroscience and Neurobiology","Theory and Algorithms"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://digitalcommons.calpoly.edu/theses/3127"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Neuronal cell types are categorized by transcriptomic identity, yet their morphological heterogeneity defies this classification. In response, researchers have adopted unsupervised graph representation learning as a tool to reveal morphological variation within single-class transcriptomic types. However, the complex geometry of neuronal morphology—especially long axons and dense dendrites—challenges graph neural networks, which struggle with message propagation across extended structures. To mitigate this, current approaches enforce sub-sampling on neuronal graphs and omit axons entirely, sacrificing critical biological features for computational efficiency. To overcome this trade-off, this thesis introduces TopoDINO, a self-supervised, topology-aware representation learning model designed to preserve the full hierarchical organization of neuronal morphology. Instead of relying on subsampled graph approximations, TopoDINO employs a novel topology-lifting approach that transforms neuronal graphs into combinatorial complexes, capturing the inherent hierarchical features of neurons. In addition, we pre-train TopoDINO on the SEU-D15K dataset. As a result, TopoDINO not only provides a biologically grounded embedding space by preserving neuronal compartmentalization, but also mitigates labeling inconsistencies across laboratories by learning in a fully data-driven manner from 12,353 dendritic reconstructions derived from 204 mouse brains, all registered to the Allen Mouse Common Coordinate Framework (CCF)."]},{"key":"dc:title","label":"Title","values":["TopoDINO: Self-Supervised Topological Representation Learning for Neuronal Morphologies"]}]}],"canonical_facts":{"dc:contributor":["Franz J. Kurfess","Computer Science","College of Engineering"],"dc:creator":["Binbisher, Yasser"],"dc:date.available":["2025-06-13T07:00:00Z"],"dc:date.created":["2025-06-01T07:00:00Z"],"dc:description.abstract":["Neuronal cell types are categorized by transcriptomic identity, yet their morphological heterogeneity defies this classification. In response, researchers have adopted unsupervised graph representation learning as a tool to reveal morphological variation within single-class transcriptomic types. However, the complex geometry of neuronal morphology—especially long axons and dense dendrites—challenges graph neural networks, which struggle with message propagation across extended structures. To mitigate this, current approaches enforce sub-sampling on neuronal graphs and omit axons entirely, sacrificing critical biological features for computational efficiency. To overcome this trade-off, this thesis introduces TopoDINO, a self-supervised, topology-aware representation learning model designed to preserve the full hierarchical organization of neuronal morphology. Instead of relying on subsampled graph approximations, TopoDINO employs a novel topology-lifting approach that transforms neuronal graphs into combinatorial complexes, capturing the inherent hierarchical features of neurons. In addition, we pre-train TopoDINO on the SEU-D15K dataset. As a result, TopoDINO not only provides a biologically grounded embedding space by preserving neuronal compartmentalization, but also mitigates labeling inconsistencies across laboratories by learning in a fully data-driven manner from 12,353 dendritic reconstructions derived from 204 mouse brains, all registered to the Allen Mouse Common Coordinate Framework (CCF)."],"dc:identifier":["https://digitalcommons.calpoly.edu/theses/3127"],"dc:subject":["Neuronal Morphology | Morphological Heterogeneity | Graph Representation Learning | Self-Supervised Learning | Combinatorial Complexes | Topology","Artificial Intelligence and Robotics","Bioinformatics","Biomedical Informatics","Biostatistics","Cell Biology","Computational Biology","Computational Neuroscience","Data Science","Molecular and Cellular Neuroscience","Other Neuroscience and Neurobiology","Theory and Algorithms"],"dc:title":["TopoDINO: Self-Supervised Topological Representation Learning for Neuronal Morphologies"],"thesis:degree_discipline":["Computer Science"],"thesis:degree_name":["MS in Computer Science"]},"updated_at":"2026-11-19T07:12:41Z"}