{"id":{"repo_id":"msu","oai_identifier":"oai:d.lib.msu.edu:etd_53360"},"canonical_url":"https://search.ndltd.org/etd/msu/oai:d.lib.msu.edu:etd_53360","repository":{"repo_id":"msu","name":"Michigan State University","base_url":"https://d.lib.msu.edu/oai"},"display":{"title":"INVESTIGATING INSECT OLFACTORY PROCESSING FOR NEURAL DYNAMICS UNDERLYING STIMULUS CLASSIFICATION AND ASSOCIATIVE LEARNING","abstract":"Thesis (Ph.D.)--Michigan State University. Biomedical Engineering - Doctor of Philosophy, 2026","abstract_html":"Thesis (Ph.D.)--Michigan State University. Biomedical Engineering - Doctor of Philosophy, 2026","abstract_has_math":false,"creators":["McLane-Svoboda, Autumn"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Saha, Debajit","Alessio, Adam","Li, Wen","Robison, Alfred"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026","date_published":"2026","updated_at":"2026-11-18T18:59:47Z","subjects":["Biomedical engineering"],"languages":["eng"],"rights":["In Copyright"],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["etd:53360","local:McLaneSvoboda_grad.msu_0128D_22416"],"render_values":[{"text":"etd:53360","href":null,"code":true},{"text":"local:McLaneSvoboda_grad.msu_0128D_22416","href":null,"code":true}]}],"person_identifier_entries":[]},"links":{"outbound_url":"https://doi.org/doi:10.25335/gv22-aw97","outbound_label":"DOI","outbound_source":"dc:identifier","record_url":"https://doi.org/doi:10.25335/gv22-aw97","record_label":"Repository record","record_kind":"doi","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://d.lib.msu.edu/oai?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai%3Ad.lib.msu.edu%3Aetd_53360","prefix":"oai_dc"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Saha, Debajit","Alessio, Adam","Li, Wen","Robison, Alfred"]},{"key":"dc:creator","label":"Author","values":["McLane-Svoboda, Autumn"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2026"]},{"key":"dc:relation","label":"Dc Relation","values":["Electronic Theses & Dissertations"]},{"key":"dc:type","label":"Dc Type","values":["text","Electronic dissertations"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Biomedical engineering"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["In Copyright"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["etd:53360","local:McLaneSvoboda_grad.msu_0128D_22416","https://doi.org/doi:10.25335/gv22-aw97"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Thesis (Ph.D.)--Michigan State University. Biomedical Engineering - Doctor of Philosophy, 2026","Remarkable progress has been made in replicating sensory modalities such as vision and hearing in engineered sensory systems, yet olfaction remains a challenge due to the complexity of the chemical dimensionality of odors. A possible solution arises through biological olfaction, specifically insect olfaction, as it can discriminate an extensive array of odors with remarkable precision, learn novel associations from only a few encounters, and generalize those associations to novel but related stimuli. Despite its demonstrated potential for applied chemical detection, the neural circuit mechanisms underlying these capabilities remain poorly understood, thereby limiting our ability to fully harness or replicate biological olfaction in engineered systems. This dissertation addressed that gap by investigating insect olfactory processing across applied, computational, and biological levels. First, we test the capacity of naïve antennal lobe (AL) neural responses to detect complex disease profiles. Then, we use naïve in vivo recordings to create a data-constrained computational model of the insect AL. Finally, we investigate the neural basis of associative learning and generalization through in vivo recordings before and after conditioning. Together, these approaches advance our mechanistic understanding of insect olfactory processing and lay the groundwork for the development of biologically informed, engineered detection systems.Biological olfaction has been successfully harnessed for disease detection, particularly through behavior-based methods; however, relying solely on behavior yields binary results without discrimination among disease subtypes. Here, we explore a different approach, harnessing the insect brain’s natural capacity for odor discrimination by using naïve neural responses to classify lung cancer. Extracellular insect AL recordings were used to classify seven different human cell lines, including non-small cell (NSC) and small cell (SC) lung cancer lines. VOCs emitted by lung cancer cell lines elicited distinct neural signatures that distinguished diseased from healthy cells. We were able to classify all cell lines with 82% accuracy, with most achieving individual classification above 80%. Additionally, we were able to distinguish between the two different cell types of lung cancer with high sensitivity and specificity (NSC: 0.88 and 0.98, SC: 0.9 and 0.99). The high classification rates demonstrated that naïve insect AL neural responses can distinguish complex profiles emitted by lung cancer subtypes, thereby establishing the inherent detection potential of the insect olfactory system without conditioning. Computational modeling offers a powerful avenue for deeper, more controlled exploration of AL processing dynamics that cannot be fully captured through in vivo experimentation alone. Here, we designed a one-to-one, biologically constrained RNN of the insect AL, trained on naïve in vivo data, to investigate these circuit mechanisms. The model successfully replicated biological AL dynamics, including diverse temporal responses and unique odor-induced neural patterns consistent with in vivo observations, thereby demonstrating odor-discrimination capability. Motif-specific manipulations of circuit connections revealed functionally distinct roles for inhibitory and excitatory circuit pathways. These findings establish a reusable biologically constrained framework that can be extended toward learning, integrated with biological components, or incorporated into engineered detection systems. Fast learning and generalization are powerful tools that can be harnessed for improved bio-inspired systems, yet the relationship between these two innate abilities remains unclear at the AL olfactory level. Understanding the mechanisms underlying the recognition of a stimulus as distinct, while still generalizing based on learned experiences, could help inform computational models to advance engineered systems or enhance detection capabilities in disease applications. In vivo AL recordings from neurons before and after appetitive conditioning revealed that associative learning and generalization arise through two distinct mechanisms, single-unit gain changes and population-level geometric shifts toward the conditioned stimulus. Notably, generalization was not uniform but arose as a hierarchy with these mechanisms differentially engaged by two separable dimensions of odor similarity: chemical structure and ecological context. Furthermore, learning-induced changes evolved over time, with transient ON responses displaying broad generalization and steady state ON responses becoming selective. These findings provide a neural mechanistic account of how the insect AL encodes discrimination and generalization, offering insight for engineering adaptive olfactory systems capable of learning and generalization across complex chemical applications. Overall, these findings advance our understanding of insect olfactory processing across naïve, modeled, and learned states. The mechanistic insights gained here lay the groundwork for incorporating learned associations into bio-inspired systems for advanced disease detection, extending the RNN toward a learning and generalization-capable model, and deepening our understanding of discrimination and generalization in biological olfactory systems.","Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).","Includes bibliographical references."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf","1 online resource (230 p.) : ill."]},{"key":"dc:title","label":"Title","values":["INVESTIGATING INSECT OLFACTORY PROCESSING FOR NEURAL DYNAMICS UNDERLYING STIMULUS CLASSIFICATION AND ASSOCIATIVE LEARNING"]}]}],"canonical_facts":{"dc:contributor":["Saha, Debajit","Alessio, Adam","Li, Wen","Robison, Alfred"],"dc:creator":["McLane-Svoboda, Autumn"],"dc:date":["2026"],"dc:description":["Thesis (Ph.D.)--Michigan State University. Biomedical Engineering - Doctor of Philosophy, 2026","Remarkable progress has been made in replicating sensory modalities such as vision and hearing in engineered sensory systems, yet olfaction remains a challenge due to the complexity of the chemical dimensionality of odors. A possible solution arises through biological olfaction, specifically insect olfaction, as it can discriminate an extensive array of odors with remarkable precision, learn novel associations from only a few encounters, and generalize those associations to novel but related stimuli. Despite its demonstrated potential for applied chemical detection, the neural circuit mechanisms underlying these capabilities remain poorly understood, thereby limiting our ability to fully harness or replicate biological olfaction in engineered systems. This dissertation addressed that gap by investigating insect olfactory processing across applied, computational, and biological levels. First, we test the capacity of naïve antennal lobe (AL) neural responses to detect complex disease profiles. Then, we use naïve in vivo recordings to create a data-constrained computational model of the insect AL. Finally, we investigate the neural basis of associative learning and generalization through in vivo recordings before and after conditioning. Together, these approaches advance our mechanistic understanding of insect olfactory processing and lay the groundwork for the development of biologically informed, engineered detection systems.Biological olfaction has been successfully harnessed for disease detection, particularly through behavior-based methods; however, relying solely on behavior yields binary results without discrimination among disease subtypes. Here, we explore a different approach, harnessing the insect brain’s natural capacity for odor discrimination by using naïve neural responses to classify lung cancer. Extracellular insect AL recordings were used to classify seven different human cell lines, including non-small cell (NSC) and small cell (SC) lung cancer lines. VOCs emitted by lung cancer cell lines elicited distinct neural signatures that distinguished diseased from healthy cells. We were able to classify all cell lines with 82% accuracy, with most achieving individual classification above 80%. Additionally, we were able to distinguish between the two different cell types of lung cancer with high sensitivity and specificity (NSC: 0.88 and 0.98, SC: 0.9 and 0.99). The high classification rates demonstrated that naïve insect AL neural responses can distinguish complex profiles emitted by lung cancer subtypes, thereby establishing the inherent detection potential of the insect olfactory system without conditioning. Computational modeling offers a powerful avenue for deeper, more controlled exploration of AL processing dynamics that cannot be fully captured through in vivo experimentation alone. Here, we designed a one-to-one, biologically constrained RNN of the insect AL, trained on naïve in vivo data, to investigate these circuit mechanisms. The model successfully replicated biological AL dynamics, including diverse temporal responses and unique odor-induced neural patterns consistent with in vivo observations, thereby demonstrating odor-discrimination capability. Motif-specific manipulations of circuit connections revealed functionally distinct roles for inhibitory and excitatory circuit pathways. These findings establish a reusable biologically constrained framework that can be extended toward learning, integrated with biological components, or incorporated into engineered detection systems. Fast learning and generalization are powerful tools that can be harnessed for improved bio-inspired systems, yet the relationship between these two innate abilities remains unclear at the AL olfactory level. Understanding the mechanisms underlying the recognition of a stimulus as distinct, while still generalizing based on learned experiences, could help inform computational models to advance engineered systems or enhance detection capabilities in disease applications. In vivo AL recordings from neurons before and after appetitive conditioning revealed that associative learning and generalization arise through two distinct mechanisms, single-unit gain changes and population-level geometric shifts toward the conditioned stimulus. Notably, generalization was not uniform but arose as a hierarchy with these mechanisms differentially engaged by two separable dimensions of odor similarity: chemical structure and ecological context. Furthermore, learning-induced changes evolved over time, with transient ON responses displaying broad generalization and steady state ON responses becoming selective. These findings provide a neural mechanistic account of how the insect AL encodes discrimination and generalization, offering insight for engineering adaptive olfactory systems capable of learning and generalization across complex chemical applications. Overall, these findings advance our understanding of insect olfactory processing across naïve, modeled, and learned states. The mechanistic insights gained here lay the groundwork for incorporating learned associations into bio-inspired systems for advanced disease detection, extending the RNN toward a learning and generalization-capable model, and deepening our understanding of discrimination and generalization in biological olfactory systems.","Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).","Includes bibliographical references."],"dc:format":["application/pdf","1 online resource (230 p.) : ill."],"dc:identifier":["etd:53360","local:McLaneSvoboda_grad.msu_0128D_22416","https://doi.org/doi:10.25335/gv22-aw97"],"dc:language":["eng"],"dc:relation":["Electronic Theses & Dissertations"],"dc:rights":["In Copyright"],"dc:subject":["Biomedical engineering"],"dc:title":["INVESTIGATING INSECT OLFACTORY PROCESSING FOR NEURAL DYNAMICS UNDERLYING STIMULUS CLASSIFICATION AND ASSOCIATIVE LEARNING"],"dc:type":["text","Electronic dissertations"]},"updated_at":"2026-11-18T18:59:47Z"}