Our work centers on understanding how large populations of neurons in the brain perform computations and represent intention. We use these insights to develop high-performance, robust, and practical assistive devices for people with disabilities and neurological disorders.
Notes to prospective graduate students:
- All students joining the lab would first need to be accepted by one of the relevant graduate programs at Emory or Georgia Tech (Biomedical Engineering, Neuroscience, Electrical Engineering, Computer Science, Bioengineering, etc). It is not possible to join the lab until a student is first accepted to a graduate program.
- Grad school is complex, and choosing a great mentor is critical for your career. Do your homework! A helpful starting point is Ben Barres’s article, How to Pick a Graduate Advisor.
- Here’s a page from the Shackman lab with great tips and resources on the grad school applications process.
[2018-12] Had a wonderful visit to the University of Georgia to present our work using deep learning to uncover dynamics from neural population activity. Thanks very much to the UGA Deep Learning Group for the invite!
[2018-11] Congratulations to Tony Corsten, first student to graduate from the lab! Tony just defended his Masters thesis entitled, “A novel device for precise training and perturbing of motor cortically driven forelimb behaviors in the rat“. Tony gave an excellent presentation of stellar work, and I’m truly honored to have contributed to his training and development over the past 2 years. Very proud of him, and excited to see him move on to do great work as part of Randy Trumbower’s lab at Spaulding Rehabilitation Hospital at Harvard.
[2018-11] Our lab presented three abstracts at the 2018 Society for Neuroscience conference in San Diego. Really pleased by the response and feedback. Congrats to the students and postdoc!
- Feng Zhu is using deep learning to probe population dynamics underlying cognitive processes. This is a new foray for us, in collab with Sabine Kastner’s lab at Princeton. Using deep learning to characterize cognitive population activity in the pulvinar
- Lahiru Wimalasena is testing if deep learning/nonlinear dynamics models of M1 can improve prediction of muscle activity, esp. during complex tasks. Great collaboration with Lee Miller’s lab at Northwestern. Modeling neural ensemble dynamics in motor cortex leads to improved EMG decoding of multiple muscles in goal-directed reaching tasks
- Reza Keshtkaran is improving performance and robustness of our recently-developed LFADS technique, by developing methods for distributed, automated hyperparameter optimization of deep neural networks. Likely crucial for applying LFADS to different brain areas. Large-scale automated deep neural network training framework for robust inference of neural ensemble dynamics
[2018-10] Thankful to Emory’s Neuroscience Ph.D. program for the invitation to speak. Gave a talk as part of the Frontiers in Neuroscience Seminar Series, entitled “Using deep learning to characterize neural population activity.”
[2018-10] We just published our first paper, a review article in the Journal of Neuroscience: “Latent Factors and Dynamics in Motor Cortex and Their Application to Brain–Machine Interfaces“. We hope this is an accessible overview of recent science on neural population dynamics in motor cortex, techniques for inferring latent factors from neural population activity, and implications for BMIs.
[2018-09] Grateful to the NSF for supporting our lab’s work in an upcoming project, Discovering dynamics in massive-scale neural datasets using machine learning! This work is funded through the Neural and Cognitive Systems program, joint with our wonderful collaborators Prof. Lee Miller (Northwestern) and Prof. Matthew Kaufman (U Chicago). [NSF] [Emory news] [GT/Emory BME]
[2018-08] Had a great time participating in the forum on Artificial Intelligence in Medicine and Health Care, as part of the 2018 US-Korea Conference at St. Johns University, Queens, New York. Gave an invited talk entitled “Using AI to improve brain-machine interfaces by uncovering neural population dynamics.”
[2018-06] Yahia Ali, an undergraduate in the lab, was selected for the Charles I. Hancock Endowed Research Award in Neuroengineering from the Emory / Georgia Tech Biomedical Engineering Department. Congrats Yahia!
[2018-06] Enjoyed giving an invited talk on our deep learning work to understand neural population dynamics at Machine Learning in Science and Engineering 2018 at Carnegie Mellon University.
[2018-05] Our paper, Inferring single-trial neural population dynamics using sequential auto-encoders, was accepted at Nature Methods! (Pre-print available at bioRxiv.)
[2018-05] Our minisymposium proposal, Latent Factors and Dynamics in Motor Cortex and Their Application to Brain-Machine Interfaces, was accepted for the 2018 Society for Neuroscience Annual Meeting. Dr. Pandarinath will chair the session. Other speakers: K. Cora Ames (Columbia), Eva Dyer (Georgia Tech), Jonathan Kao (UCLA), Ali Farshchian (Northwestern), and Abigail Russo (Columbia). Congrats to all the speakers!
[2018-05] Excited to give a talk at the University of Chicago, “Inferring precise estimates of single-trial neural population dynamics using deep learning techniques.”
[2018-05] Lahiru Wimalasena (graduate student in the lab) gave a poster presentation at the Society for Neural Control of Movement Meeting, entitled “Modeling neural population dynamics in motor cortex leads to improved kinetic decoding in an unstructured motor task” (one of two presentations from our group).
[2018-04] Dilara Soylu (undergraduate student in the lab) won 3rd place at GT’s Undergraduate Research Symposium for her poster presentation entitled “Optimizing Deep Learning Models for Limited Precision Architectures”. Congrats Dilara !
[2018-03] Gave an invited talk at the COSYNE 2018 workshop, “Closed-loop control of neural systems and circuits for scientific discovery”, entitled “LFADS: A deep learning method to infer latent states and dynamics from neural population activity”
[2018-03] Lahiru Wimalasena (graduate student in the lab) and Dr. Pandarinath gave an invited seminar to the GT Mathematics and Applications Portal Seminar Series, entitled “Unsupervised discovery of ensemble dynamics in the brain using deep learning techniques.”
[2018-03] Had a lot of fun participating in the Emory Neuroethics Seminar Series, “The Future Now: NEEDs (Neuroscience and Emerging Ethical Dilemmas).” Gave a seminar entitled “To be implanted and wireless” (click for excellent writeup by Nathan Ahlgrim).
[2018-02] Congrats to Feng Zhu (graduate student in the lab) for receiving a travel award from the IEEE Brain Initiative to attend the COSYNE 2018 conference!
[2017-12] Yahia Ali, an undergraduate in the lab, and Siva Manivasagam (Rozell Lab, GT ECE), were selected for a prestigious University Interdisciplinary Research Award for Spring 2018 from Georgia Tech. Congrats Yahia and Siva!
[2017-11] New papers: “A comparison of intention estimation methods for decoder calibration in intracortical brain-computer interfaces” (IEEE Transactions on Biomedical Engineering). “Rapid calibration of an intracortical brain computer interface for people with tetraplegia” (Journal of Neural Engineering). “Feasibility of automatic error detect-and-undo system in human intracortical brain-computer interfaces” (IEEE Transactions on Biomedical Engineering).
[2017-08] Dr. Pandarinath gave an invited talk at Carnegie Mellon University, entitled “Improving the performance of brain-machine interfaces for communication through innovations in control algorithms and systems design”, as well as an informal seminar entitled “Inferring single-trial neural population dynamics using sequential auto-encoders.”
[2017-08] Dr. Pandarinath gave an invited talk at the Georgia Tech Mathematics and Applications Portal workshop on Dynamics and Control, entitled “High performance brain-machine interfaces through innovations in control algorithm design.”
[2017-06] A preprint of our manuscript on applying deep learning techniques to uncover neural dynamics is now available on BioRxiv: “Inferring single-trial neural population dynamics using sequential auto-encoders”.
[2017-05] Thrilled to participated in the first New Directions in Motor Control workshop at Emory. Gave a talk entitled “Deep learning methods to precisely estimate motor cortical population state and its dynamics” #ATLmotorcontrol
[2017-04] Gave a talk at the GT Neuro Seminar Series, “Advancing brain-machine interfaces toward clinical viability.”
[2017-03] Yahia Ali, an undergraduate in the lab, was selected for a prestigious President’s Undergraduate Research Award from Georgia Tech. Congrats Yahia!
[2017-01] Our paper “High performance communication by people with tetraplegia using an intracortical brain-machine interface” was accepted at the journal eLife.
[2017-01] Our abstract “Precise estimates of single-trial neural population state in motor cortex via deep learning methods” was accepted at Computational and Systems Neuroscience (Cosyne) 2017.
[2017-01] Our paper “Signal-independent noise in intracortical brain-computer interfaces causes movement time properties inconsistent with Fitts’ law” was accepted at the Journal of Neural Engineering.
[2016-11] Our paper “Feedback control policies employed by people using intracortical brain-machine interfaces” was accepted at the Journal of Neural Engineering.
[2016-08] The preprint of our paper, “LFADS – Latent Factor Analysis via Dynamical Systems,” is now available.
[2016-06] Gave an invited talk at the International BCI Society, “Using dynamical models of motor cortical activity to improve BCIs.”
[2016-06] David Sussillo gave an invited talk at the Grossman Center for the Statistics of the Mind at Columbia University, on our joint work “Inferring latent dynamics from single trial neural population activity using variational temporal autoencoders.”