Machine learning methods enable researchers to discover statistical patterns in large datasets and investigate a wide variety of questions in neuroscience. The videos below include a one-hour introduction to the field as well as multiple five-minute lightning talks featuring neuroscientists describing their research applying computer vision to neuroscience problems.
These videos are a complement to the virtual conference "Machine Learning in Neuroscience: Fundamentals and Possibilities," which took place on June 26, 2019.
The Fundamentals of Machine Learning
In this Machine Learning virtual conference session, Sanjoy Dasgupta gives an introduction to machine learning. He goes through a detailed prediction problem and surveys the landscape of the field.
Pose Estimation for Lab Animals Using Deep Learning
In this video, Talmo Pereira describes the use of pose estimation instead of classical tracking to quantify animal behavior from videos.
Learning to Predict Fly Behavior
In this video, Eyrún Eyjólfsdóttir describes a method for predicting the behavior of fruit flies.
Signed Boundary Distance Prediction to Reconstruct Synapses in EM
In this video, Larissa Heinrich explains her work in connectomics, reconstructing synapses from large electron microscopy volumes (in this case, the entire drosophila brain) using convolutional neural networks.
Mining Morphological Features of Neurons with PRISM
In this video, Beth Cimini talks about the use of machine learning for morphological profiling, extracting many features of the cells to build up its “fingerprint.”