Description:
Connectomics is the subfield of neuroscience concerned with acquiring and analyzing the structure and function of neural circuits. This generation of computer scientists will be the first in history to have access to brain data in sufficient quantity and quality for large-scale structural and functional connectomics. Students in CS379C will have the opportunity to tackle the challenging machine-learning and signal-processing problems that arise in tracing neural circuits and inferring their function.
We'll be working with two teams of scientists and engineers who are building the tools to acquire this data. We have several relatively-small (10TB) EM datasets (including ground truth) that students interested in circuit tracing (structural connectomics) can use in projects. Scientists at AIBS have volunteered to help students in understanding the data and technologies used to collect it. In addition, engineers from my team at Google will supply examples of algorithms that have worked well in our experiments.
Inferring function from CI data is more challenging since until recently there haven't been suitable datasets to work with. We now have several such datasets provided by our collaborators that can be used in student projects. In addition, we'll be generating synthetic datasets for cortical circuits of 5-50K neurons using Hodgkin-Huxley models1
developed at AIBS and EPFL. These models and their associated simulators provide a controlled environment in which to experiment with and evaluate machine-learning technologies for functional connectomics.
The prerequisites are basic high-school biology, good math skills, and familiarity with machine learning. Some background in computer vision and signal processing will be important for projects in structural connectomics. Familiarity with modern artificial neural network technologies is a plus for projects in functional connectomics. As an added bonus, an extraordinary group of scientists and engineers have agreed to discuss their research on topics of particular relevance to connectomics.
Location and Time:
MW, 4:15-5:30pm, Building 100, Room 101K
Staff:
Instructor: Thomas Dean
Email: tld [at] google [dot] com
Office hours: by appointment
Course Assistant: Aditya Srinivas Timmaraju
Email: adityast [at] stanford [dot] edu
Textbooks:
There are no required textbooks for this course but
you are expected to do a lot of reading on your own
and these three texts are good to have around for
reference. I’ve yet to meet anyone who has
read them cover to cover but over the years,
I’ve probably read most of the chapters in one
edition or the other and found them consistently
useful. A copy of each book will be put on the
reference desk should you want to read a selection,
and, in the case of the latter two, you can also
often find preprint versions of individual chapters
on the web pages of the contributing authors:
-
- Neuroscience: Exploring the Brain (Third Edition), Bear, Connors and Paradiso.
-
- The Cognitive Neurosciences (Third Edition), Gazzaniga.
-
- Principles of Neural Science (Fourth Edition), Kandel, Schwartz and Jessell.
Grading:
- Class participation including presentation (30%)
- Project proposal due around midterm (20%)
- Project report due around finals week (50%)
1 These models were developed by Costas Anastassiou and his team at AIBS and by Sean Hill at EPFL. They consist of networks of reconstructed, multi-compartmental, virtually-instrumented and spiking pyramidal neurons and basket cells, plus ion- and voltage-dependent currents and local field potentials that allow us to generate the same sort of rasters we expect to collect during calcium imaging.