{"id":{"repo_id":"penn","oai_identifier":"oai:repository.upenn.edu:20.500.14332/62317"},"canonical_url":"https://search.dev.ndltd.org/etd/penn/oai:repository.upenn.edu:20.500.14332/62317","repository":{"repo_id":"penn","name":"University of Pennsylvania","base_url":"https://repository.upenn.edu/server/oai/request"},"display":{"title":"From thoughts to actions: cracking the neural code across scales and modalities","abstract":"Understanding how neural activity encodes cognitive and behavioral statescould lead to breakthroughs in the development of brain-computer interfaces and transform the lives of patients suffering from neurological and neuromotor disorders. Despite progress, establishing robust mappings between neural activity and behavior remains an open challenge. To address this gap, this dissertation introduces methods that can be used to decode behavior from the macro- to the micro-scale using distinct neural recording modalities: noninvasive scalp EEG, invasive stereotactic EEG, and intracortical microelectrode arrays. At the macroscopic level, I first demonstrate how low-frequency phase dynamics in scalp EEG relate to fluctuations in attention. Using information-theoretic and signal processing approaches, I show that EEG phase encodes cognitive state, highlighting the role of phase as a viable signal for passive cognitive monitoring. Next, at the mesoscale, I introduce a flexible framework for multi-session, multi-subject neural decoding based on sEEG data. This approach accounts for variability in the number and placement of electrodes across individuals by learning neural representations that generalize across-subjects while maintaining subject-specific output mappings. This framework achieves generalization across subjects and sessions, enabling scalable decoding in real-world clinical sEEG cohorts. Finally, at the microscale, I introduce a causal, scalable, and energy-efficient neural decoding framework for intracortical recordings that enables cross-session, cross-subject, and cross-task transfer. This framework combines the versatility of transformers with the energy-benefits of spiking neural networks to enable robust, actionable decoding for closed-loop BCI systems that run on edge compute. Together, these works advance cognitive neuroscience and brain-computer interfaces by introducing tools that can be used for neural decoding across scales; from noninvasive, observational signals of cognition to invasive, high-fidelity representations of intention and action. By addressing decoding at each level of resolution and invasiveness, this dissertation lays the groundwork for flexible and generalizable neural interfaces that can adapt across populations, modalities, and use cases.","abstract_html":"Understanding how neural activity encodes cognitive and behavioral statescould lead to breakthroughs in the development of brain-computer interfaces and transform the lives of patients suffering from neurological and neuromotor disorders. Despite progress, establishing robust mappings between neural activity and behavior remains an open challenge. To address this gap, this dissertation introduces methods that can be used to decode behavior from the macro- to the micro-scale using distinct neural recording modalities: noninvasive scalp EEG, invasive stereotactic EEG, and intracortical microelectrode arrays. At the macroscopic level, I first demonstrate how low-frequency phase dynamics in scalp EEG relate to fluctuations in attention. Using information-theoretic and signal processing approaches, I show that EEG phase encodes cognitive state, highlighting the role of phase as a viable signal for passive cognitive monitoring. Next, at the mesoscale, I introduce a flexible framework for multi-session, multi-subject neural decoding based on sEEG data. This approach accounts for variability in the number and placement of electrodes across individuals by learning neural representations that generalize across-subjects while maintaining subject-specific output mappings. This framework achieves generalization across subjects and sessions, enabling scalable decoding in real-world clinical sEEG cohorts. Finally, at the microscale, I introduce a causal, scalable, and energy-efficient neural decoding framework for intracortical recordings that enables cross-session, cross-subject, and cross-task transfer. This framework combines the versatility of transformers with the energy-benefits of spiking neural networks to enable robust, actionable decoding for closed-loop BCI systems that run on edge compute. Together, these works advance cognitive neuroscience and brain-computer interfaces by introducing tools that can be used for neural decoding across scales; from noninvasive, observational signals of cognition to invasive, high-fidelity representations of intention and action. By addressing decoding at each level of resolution and invasiveness, this dissertation lays the groundwork for flexible and generalizable neural interfaces that can adapt across populations, modalities, and use cases.","abstract_has_math":false,"creators":["Mentzelopoulos, Georgios"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["Vitale, Flavia","Medaglia, John, D"],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025","date_published":"2025","updated_at":"2026-07-24T03:45:26Z","subjects":["Engineering"],"languages":["en"],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://repository.upenn.edu/handle/20.500.14332/62317","outbound_label":"Repository record","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Vitale, Flavia","Medaglia, John, D"]},{"key":"dc:creator","label":"Author","values":["Mentzelopoulos, Georgios"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2026-01-29T17:19:55Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2026-01-29T17:19:55Z"]},{"key":"dc:date.issued","label":"Date","values":["2025"]},{"key":"dc:type","label":"Dc Type","values":["Dissertation/Thesis"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Engineering"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://repository.upenn.edu/handle/20.500.14332/62317"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["2025"]},{"key":"dc:description.abstract","label":"Abstract","values":["Understanding how neural activity encodes cognitive and behavioral statescould lead to breakthroughs in the development of brain-computer interfaces and transform the lives of patients suffering from neurological and neuromotor disorders. Despite progress, establishing robust mappings between neural activity and behavior remains an open challenge. To address this gap, this dissertation introduces methods that can be used to decode behavior from the macro- to the micro-scale using distinct neural recording modalities: noninvasive scalp EEG, invasive stereotactic EEG, and intracortical microelectrode arrays. At the macroscopic level, I first demonstrate how low-frequency phase dynamics in scalp EEG relate to fluctuations in attention. Using information-theoretic and signal processing approaches, I show that EEG phase encodes cognitive state, highlighting the role of phase as a viable signal for passive cognitive monitoring. Next, at the mesoscale, I introduce a flexible framework for multi-session, multi-subject neural decoding based on sEEG data. This approach accounts for variability in the number and placement of electrodes across individuals by learning neural representations that generalize across-subjects while maintaining subject-specific output mappings. This framework achieves generalization across subjects and sessions, enabling scalable decoding in real-world clinical sEEG cohorts. Finally, at the microscale, I introduce a causal, scalable, and energy-efficient neural decoding framework for intracortical recordings that enables cross-session, cross-subject, and cross-task transfer. This framework combines the versatility of transformers with the energy-benefits of spiking neural networks to enable robust, actionable decoding for closed-loop BCI systems that run on edge compute. Together, these works advance cognitive neuroscience and brain-computer interfaces by introducing tools that can be used for neural decoding across scales; from noninvasive, observational signals of cognition to invasive, high-fidelity representations of intention and action. By addressing decoding at each level of resolution and invasiveness, this dissertation lays the groundwork for flexible and generalizable neural interfaces that can adapt across populations, modalities, and use cases."]},{"key":"dc:description.degree","label":"Dc Description Degree","values":["Doctor of Philosophy (PhD)"]},{"key":"dc:title","label":"Title","values":["From thoughts to actions: cracking the neural code across scales and modalities"]}]}],"canonical_facts":{"dc:contributor.advisor":["Vitale, Flavia","Medaglia, John, D"],"dc:creator":["Mentzelopoulos, Georgios"],"dc:date.accessioned":["2026-01-29T17:19:55Z"],"dc:date.available":["2026-01-29T17:19:55Z"],"dc:date.issued":["2025"],"dc:description":["2025"],"dc:description.abstract":["Understanding how neural activity encodes cognitive and behavioral statescould lead to breakthroughs in the development of brain-computer interfaces and transform the lives of patients suffering from neurological and neuromotor disorders. Despite progress, establishing robust mappings between neural activity and behavior remains an open challenge. To address this gap, this dissertation introduces methods that can be used to decode behavior from the macro- to the micro-scale using distinct neural recording modalities: noninvasive scalp EEG, invasive stereotactic EEG, and intracortical microelectrode arrays. At the macroscopic level, I first demonstrate how low-frequency phase dynamics in scalp EEG relate to fluctuations in attention. Using information-theoretic and signal processing approaches, I show that EEG phase encodes cognitive state, highlighting the role of phase as a viable signal for passive cognitive monitoring. Next, at the mesoscale, I introduce a flexible framework for multi-session, multi-subject neural decoding based on sEEG data. This approach accounts for variability in the number and placement of electrodes across individuals by learning neural representations that generalize across-subjects while maintaining subject-specific output mappings. This framework achieves generalization across subjects and sessions, enabling scalable decoding in real-world clinical sEEG cohorts. Finally, at the microscale, I introduce a causal, scalable, and energy-efficient neural decoding framework for intracortical recordings that enables cross-session, cross-subject, and cross-task transfer. This framework combines the versatility of transformers with the energy-benefits of spiking neural networks to enable robust, actionable decoding for closed-loop BCI systems that run on edge compute. Together, these works advance cognitive neuroscience and brain-computer interfaces by introducing tools that can be used for neural decoding across scales; from noninvasive, observational signals of cognition to invasive, high-fidelity representations of intention and action. By addressing decoding at each level of resolution and invasiveness, this dissertation lays the groundwork for flexible and generalizable neural interfaces that can adapt across populations, modalities, and use cases."],"dc:description.degree":["Doctor of Philosophy (PhD)"],"dc:identifier.uri":["https://repository.upenn.edu/handle/20.500.14332/62317"],"dc:language.iso":["en"],"dc:subject":["Engineering"],"dc:title":["From thoughts to actions: cracking the neural code across scales and modalities"],"dc:type":["Dissertation/Thesis"]},"updated_at":"2026-07-24T03:45:26Z"}