Neural Signal Processing & Machine Learning
ELEC/BIOE 548 | ELEC 483
Tuesday/Thursday 10:50 - 12:05
Fall 2022 | BRC 286
〞
Do not take questions on the first day! It shows weakness!
- Barney Stinson
Do not take questions on the first day! It shows weakness!
- Barney Stinson
Am I in the right room?
Do not take questions on the first day! It shows weakness!
- Barney Stinson
Are you in the right room?
Instructor
Office Hrs: What works?
Shay(ok)
Meet the Team
TA|Grader
Office Hrs: TBD
Della
TA|Grader
Office Hrs: TBD
Kayla
Meet the Team
Instructor Emotional Support/Troublemaker/Pretty Girl/Bad Bun Bun
Truffle
Genuine Instructor Emotional Support/ Good Boy/Big Bunny
Thumper
Why you're here...
Instructor-splaining
- Appropriate amount of excitement about brains!
- How does the brain represent, process, encode, store and recall information?
- How can we (experimenters) process and understand signals from we record from the brain?
- Using "basic" machine learning topics to aid us
- If you only want machine learning (statistical, neural networks, etc.), this is not the class for you!
Why are you really here???
Inspiration!
Why do "Neural Signal Processing"
- Story time!
- Someone at the BRC's story
- My story
- WE can actually help!
Cleveland Clinic
Why do "Neural Signal Processing"
Assistive/reparative tech!
Why do "Neural Signal Processing"
Interventional/treatment
Why do "Neural Signal Processing"
More Treatment!
Why do "Neural Signal Processing"
Pharmacology development
Why do "Neural Signal Processing"
Neuromorphic systems
Taken from Dr. Kemere without shame of changing anything.
Brains
"Systems level" & "Cellular/Molecular Level"
Brains
Neuron/Cellular Level
But we must pivot to Logistics.
1
Introduction. Class & brains
2
Fundamental neurobiology. How do neurons fire? How/what do we record?
3
Modeling spike trains. First bit of analysis work and understanding firing properties of neurons.
5
Classification. Making machines learn. Which direction is a monkey trying to reach? Bayesian decoding.
4
Point processes. Continued modeling work of neurons.
6
Clustering/Mixture models. Making machines learn some more. Spike sorting.
Bi-weekly Schedule
7
Continuous decoding. Kalman filters. Machines continue to learn.
8
Spectral analysis? LFP interpretation in spectral domain. But also kinda in clustering.
Objectives & Outcomes
- Students are comfortable with neural data in many different forms, including "spikes" measured intracellularly, extracellularly, optically and LFP/EEG
- Students are comfortable building generative models that describe neural activity either from first principles or using experimental data
- Students are comfortable using generative models to optimally decode underlying information from neural activity
Assignments & Final Project
- 6 -- 8 homework assignment problem sets
- 70%
- elec548.github.io (previous course website has all these)
- CHANGES!
- Collab but don't copy! Ask us for help!
- Final project
- 30%
- Details will be provided later with slide deck
- ELEC 483 students can work in groups. All 548 students are expected to work individually
Grading Philosophy
How can I help you learn best?
The ability to observe without evaluating is the highest form of human intelligence
-- Jiddu Krishnamurti
Each homework problem will have expected outcomes and objectives listed. Your objective is to explain in a sentence or two how your answers or plots meet and demonstrate that outcome. If we are in agreement, credit shall be given fully, otherwise, we will provide feedback for reworking it and partial credit can be given. I will make this clearer next week!
Final project will work similarly except overall course objectives must be demonstrated.
Logistics
Canvas for announcements
Piazza for homework or class questions
Office hours for help
AND WE'RE DONE!
Questions? Comments? Concerns?
