PI: Carlson, David E
Email: david.carlson@duke.edu
Institution: Duke University
Title: Uncovering Population-Level Cellular Relationships to Behavior via Mesoscale Networks.
Grant #: EB026937
Status: Project has just entered its third year. Several manuscripts and code bases are now available describing the mathematical and algorithmic advances in electrical connectome (electome) modeling, their use in designing neurostimulation protocols, and novel neuroscience results.
Relevant recent talks discribing the research:
NeurIPS Plenary Talk 2019: Mapping Emotions: Discovering Structure in Mesoscale Electrical Brain Recordings
Neuroscience Conference 2021: Learning Electrical Connectomes of Neuropsychiatric Disorders and Behaviors
Deliverables:
Published or in-press:
Mague, SD*, Talbot, A*, Blount, C, Duffney, LJ, Walder-Christensen, KK, Adamson, E, Bey, AL, Ndubuizu, N, Thomas, G, Hughes, DND, Sinha, S, Fink, AM, Gallagher, NM, Fisher, RL, Jiang, Y, Carlson, DE#, and Dzirasa, K#. Brain-wide electrical dynamics encode an appetitive socioemotional state. Neuron 2022
Yuan, S, Li, Y, Wang, D, Bai, K, Carin, L, Carlson, D. Learning to Weight Filter Groups for Robust Classification. IEEE/CVF Winter Conference on Applications of Computer Vision 2022.
Zhou T, Li Y, Wu Y, Carlson D. Estimating Uncertainty Intervals from Collaborating Networks. Journal of Machine Learning Research 2021.
Gallagher, NM, Dzirasa, K, Carlson, D. Directed Spectral Measures Improve Latent Network Models Of Neural Populations. Neural Information Processing Systems 2021.
Carson, WE, Talbot, A, Carlson, D. AugmentedPCA: A Python Package of Supervised and Adversarial Linear Factor Models. Neural Information Processing Systems Workshop on Learning Meaningful Representations of Life 2021.
Isaev, DY, Tchapyjnikov, D, Cotton, M, Tanaka, D, Martinez, N, Bertran, M, Saprio, G, and Carlson, D. Attention-Based Network for Weak Labels in Neonatal Seizure Detection. Machine Learning in Healthcare 2020.
Preprints:
Talbot A, Dunson DB, Dzirasa K, Carlson DE. Supervised autoencoders learn robust joint factor models of neural activity. arXiv 2004.05209
Block, C.L., Eroglu, O., Mague, S.D., Sriworarat, C., Blount, C., Malacon, K.E., Beben, K.A., Ndubuizu, N., Talbot, A., Gallagher, N., Jo, Y.C., Nyangacha, T. , Carlson, D., Dzirasa, K., Eroglu, C., Bilbo, S.D.. Prenatal Environmental Stressors Impair Postnatal Microglia Function and Adult Behavior in Males. bioRxiv. https://doi.org/10.1101/2020.10.15.336669
Code Repositories:
Directed Spectrum (Gallagher et al 2021): https://github.com/neil-gallagher/directed-spectrum
Electome Models (Talbot et al 2020): https://github.com/carlson-lab/encodedSupervision
Augmented PCA (Carson et al 2021): https://github.com/wecarsoniv/augmented-pca, https://augmented-pca.readthedocs.io/en/latest/
Slide Updates:
2021: https://carlson.pratt.duke.edu/sites/carlson.pratt.duke.edu/files/images/TMM_Carlson_Update.pdf
2020: https://carlson.pratt.duke.edu/sites/carlson.pratt.duke.edu/files/images/carlson_tools.pdf
Link to Data/Model Reuse abstract, [Link]
2021 Brain PI Meeting
Update: Several manuscripts and code bases are now available describing the mathematical and algorithmic advances in electrical connectome (electome) modeling, their use in designing neurostimulation protocols, and novel neuroscience results.
Link to Poster: TMM Section
Poster Number 4012
Identifying Autism-Specific Neural Signatures with Adversarial Machine Learning
Poster Number 4029
Directed Communication Measures Uncover Latent Networks Of Neural Populations
Poster Number 4084
Estimating a Brain Network Predictive of Stress and Genotype with Supervised Autoencoders