We aim to augment recovery in spinal cord (SC) injured patients. Electrical stimulation of the SC can facilitate recovery, but the mechanisms are not yet understood. One knowledge gap lies in the exact pathways that are recruited by stimulation. To close this gap, we have tested the effects of SC stimulation in people undergoing clinically indicated surgery. By testing the distribution and size of muscle responses to SC stimulation, we can infer which circuits are activated. We are also examining how SC injury changes those responses. We propose to use Bayesian methods to understand the interaction between muscle responses to stimulation and the MRI indicated pattern of damage. The project will involve construction of models linking multiple data modalities that predict muscle activity, followed by the modification of these models to account for patterns of damage. Construction of such models would enable a deeper understanding of SC stimulation leading to more effective stimulation paradigms.

Continue reading

Our goal is to use deep learning networks to understand which neurons in the brain encode fine motor movements in mice. We collected large datasets entailing calcium imaging data of active neurons and high-resolution videos when mice perform motor tasks. We want to use recent advances in deep learning to (1) estimate the poses of mouse body parts at a high spatiotemporal resolution (2) extract behaviorally-relevant information and (3) align them with neural activity data. Behavioral video analysis is made possible by transfer learning, the ability to take a network that was trained on a task with a large supervised dataset and utilize it on a small supervised dataset. This has been used e.g. in a human pose–estimation algorithm called DeeperCut. Recently, such algorithms were tailored for use in the laboratory in a Python-based toolbox known as DeepLabCut, providing a tool for high-throughput behavioral video analysis.

Continue reading

This project will be focused on creating a deep learning framework for tracking individual molecules and proteins as they move within a cell under various conditions. Using total internal reflection (TIRF) microscopy, we have accumulated more than 10 million trajectories over dozens of experimental preparations with differences in both the imaging approaches as well as the biological context. In our experiments we have captured particles under a wide variety of conditions including increased protein expression level, and a range of drug concentrations. Our biggest challenge is being able to stably track the movement of a particle as it passes by other particles or groups of particles, and to do this in a way that generalizes over novel conditions. The Data Science Institute Scholar chosen for this project would work with scientists in the Javitch laboratory and others across the Columbia campus to conceive of an approach for efficiently and effectively tracking particles. The resulting work would be of great interest to an increasing number of scientists working in this field who currently rely on methods based on feature engineering that are often inaccurate or inflexible compared to modern deep learning methods.

Continue reading

We are constantly exposed to inputs from the outside world, but we do not perceive everything we are exposed to. Some inputs are rather weak: we might perceive them at one point in time, but not at another. The state of our brains right before we receive such sensory inputs influences whether or not we perceive them. Brain oscillations are proposed to play a key role in setting these brain states; however, how exactly these brain rhythms influence our perception remains a topic of active research.

Continue reading

We are interested in investigating how deaths and hospitalizations resulting from opioid overdoses cluster across space and time in the US. This analysis will be conducted with the aid of two comprehensive databases: 1) detailed mortality data across the US; and 2) a stratified sample of all hospitalizations in the US, which can be subset to select for opioid overdoses. Analyses will be extended to drug type (prescription drugs, fentanyl etc.) and subject demographics (age, race, etc.). We have previously conducted similar cluster analysis for other health phenomena.

Continue reading

Effective representations and analyses of symbolic data, such as lexical data (words) and networks (graphs), have become of great interest in recent years, due both to advancements in data collection in Natural Language Processing (NLP), and the ubiquity of social networks. Such data often has no natural numerical representation, and is typically described in terms relational expressions or as pairwise similarities. It turns out that finding numerical representations of such data in “Hyperbolic” spaces—rather than into the more familiar Euclidean spaces—is a more effective way to preserve valuable relational information.

Continue reading

Given calcium imaging data of active neurons, can we detect groups of co-firing neurons, called neuronal ensembles? We have a number of datasets consisting of hundreds of neurons imaged for thousands of time steps, and seek to extend an existing CRF model to consider temporal relationships. The goal is to be able to detect neuronal ensembles that span multiple time steps, and that are not conditioned on external stimuli.

Continue reading