Systems Neuroscience | NX-435

Teacher: Mackenzie Mathis
TAs: Celia Benquet & Myriam Hamon

Language: English

Summary

Systems neuroscience is the study of the nervous system at the level of neural circuits and networks. It seeks to understand how groups of neurons work together to process information and generate behavior. This field of neuroscience combines techniques from multiple disciplines, including physiology, anatomy, genetics, and computer science, to investigate the complex interactions between brain cells and how they give rise to behavior. The course will use a variety of teaching methods, including lectures, discussions, primary literature reading, and hands-on coding activities.

Course Meeting Timing & Topic Schedule

We will meet for:

• Lectures on Thursdays at 1615-1800 in AAC137
• Sections on Fridays 1315-1500 in AAC137

Weekly Schedule: 

W1: Introduction to Systems Neuroscience & Memory
W2: Reward & Reinforcement Learning
W3: Sensorimotor modulation & control
W4:  Visual Systems
W5: Neural data analysis
W6: Behavioral data analysis 
W7: Encoding of Space (quiz 1 in section)
W8: Problem Set #1 work week
W9: NeuroAI
W10: spring break 🥳 - no assignments (April 17 - 28th)
W11: Problem Set # 2 work week 
W12: Brain Machine Interfaces for Systems Neuroscience
W13: Cognition & Skill Learning
W14: Emerging topics in systems neuroscience & FINAL EXAM May 23rd

Each week you will have:

• An in-person Lecture that requires no pre-reading or homework. Please just attend with an open mind and be ready to take notes. Slides will be provided on Moodle for reference after the lecture.
•  An in-person Section that will typically have 1 paper for assigned pre-reading. We will have discussions on that paper in depth, thus please come prepared to discuss the work. The papers are posted below within each week. Some Sections will have hands-on learning computational materials that aid in learning (neural or behavioral analysis).  Some may also have live demos. Please check the week schedule below to know more!

Problem Sets:

There will be two assignments:

• “Problem Set” 1 (25% of the course grade, DUE April 17th): Write a "News & Views Nature" style article (~1,000 words) on one of the assigned papers. News & Views are generally written by specialists for broader scientific audiences. The goal is for you to write about one of the assigned readings (from any time point in the course). The learning objective is to read primary research articles and learn to summarize them in the broader context of the field. For example, here is one from Prof. Mathis. Another example set here: News & Views Nature Neuroscience. See W8 below for more details!! 
   
• "Problem Set 2" (25% of the course grade, DUE May 17th): Now that you are about to venture off to do thesis projects in labs, let's build your skills at formulating science! The goal is to write a small research proposal to help you craft your critical thinking. Please write a 2-3 page (11pt font, 1.5 line spacing) proposal that includes:  (1) Background & Introduction; (2) Scientific Aim(s), Hypothesis, Methods; (3) Anticipated Results; (4) References. See W11 below for more details!!

• Please note our policy on the use of large language models: You are not allowed to use LLMs (ChatGPT) to generate or write this assignment. If you want to use ChatGPT (or equivalent) for “fine-tuning” for grammar and clarity, you are allowed to do so, but you must declare this at the end of your document. 🤖 You are fully responsible for the contents of this work. Please remember that plagiarism is a serious offense at EPFL.

Quiz:

There will be 1 short quiz during the term that account for 25% overall of your grade.

You are allowed to bring one A4 size hand-written page of notes (single-sided).

• Quiz 1: During Section W7: covering material from W1-7.

Final Exam:

• During the semester on May 23rd (Friday), we will have a final exam covering the material presented in Lectures and Sections throughout the course. This will be a written exam. It will count for 25% of your course grade.
• You are allowed to bring one A4 size hand-written page of notes (double-sided). This can be "hand-written" on an iPad too then printed!
  

Learning Outcomes:

By the end of the course, the student should be able to:

• Learn to effectively read primary literature.
• Understand the principles of neural coding and information processing.
• Understand the current state-of-the-art techniques for studying the brain, including electrophysiology, optogenetics, and functional imaging.
• Apply computational approaches to study specific brain functions such as sensory processing, motor control, and cognition.
• Understand the relationship between brain circuits and behavior.
• Develop computational skills for analyzing neural data.

Expected student activities:

Attend lectures and sections (both are critical), read scientific articles assigned for each section, complete the two problem sets, participate in quizzes, and take the final exam!

Assessment methods:

The final mark is a combination of three evaluations: 2 problem sets (50%), 1 quiz (25%), and 1 final exam (25%).

Additional Reading Material:

• Principles of Neural Science 5th or 6th Edition

• The Future of the Brain

Office Hours w/Mackenzie

Thursdays, 3-4 PM SV 2811 by appointment

RECOMMENDED COURSES
CS-433, BIO-411, BIO-482 or BIO-311

IMPORTANT CONCEPTS TO START THE COURSE

Good intro to neuroscience background, programming in Python, good mathematical background

Introduction to Systems Neuroscience:

• Welcome to systems neuroscience! This lecture covers course logistics, a quick background on neuroscience, and dives into the engram (memories!). Lecture Slides are posted below after the lecture!
• For ref: Video recording of lecture 1 from 2024!

Section Paper:

•  Steve Ramirez et al, Creating a False Memory in the Hippocampus. Science 341,387-391(2013).

• Decision-making is hard, as we all know! We will explore how can we test how the brain might enable decision-making by formalizing this computationally with with reinforcement learning paired with neural recordings.
  

•  Cohen, J., Haesler, S., Vong, L. et al. Neuron-type-specific signals for reward and punishment in the ventral tegmental area. Nature 482, 85–88 (2012). https://doi.org/10.1038/nature10754

• Demo notebook on computing neural Peri-Stimulus Time Histogram (PSTH), PCA clustering, and ROC analysis: https://github.com/MMathisLab/Nx-435_EPFL/blob/main/Notebooks/Demo_PSTH.ipynb

Motor learning and neuro-modulation

• This week we will dive into how to study mechanisms of rapid motor learning (called adaptation) and "slower" across day learning. Critically, we will discuss how studies in mice relate to human translational approaches. 
• For ref: Week 3 video lecture from 2024.

Section Paper:

•  Bowles S, Hickman J, Peng X, Williamson WR, Huang R, Washington K, Donegan D, Welle CG. Vagus nerve stimulation drives selective circuit modulation through cholinergic reinforcement. Neuron. 2022 Sep 7;110(17):2867-2885.e7. 
• *The First author, Dr. Bowles, will be there to discuss with you!

Visual Systems Neuroscience

• This week we will take a deeper dive into how to study neural circuits of ethological visual behaviors. Specifically, we will cover historical and modern approaches to studying vision in zebrafish, a critical model organism in modern science.
• For Ref: Video Recording of the Lecture from 2024

•  Timothy W. Dunn, Christoph Gebhardt, Eva A. Naumann, Clemens Riegler, Misha B. Ahrens, Florian Engert, Filippo Del Bene. Neural Circuits Underlying Visually Evoked Escapes in Larval Zebrafish. Neuron, Volume 89, Issue 3, 2016, Pages 613-628, ISSN 0896-6273, https://doi.org/10.1016/j.neuron.2015.12.0

Section Hands-On:

• Developing an experiment and testing visual behaviors in zebrafish

Lecture: Neural & Behavioral Analysis:

• This week we will do a deeper dive into methods for neural analysis. Moving beyond the PSTH, this lecture will cover cutting-edge machine learning methods for neural data analysis. Here is a Video for part 1 (GLMs).

• 45 min: Fitting neural data with GLMs: https://github.com/MMathisLab/Nx-435_EPFL
  
• 45 min: Population analysis of neural data with CEBRA: https://github.com/MMathisLab/Nx-435_EPFL
• Extra notebook on CEBRA good coding practices: Best Practices for training CEBRA models.
  
• If you missed section, you can watch a video that covers the population analysis material here. You can then work through the Notebook we provide on GitHub! The can be covered on the final exam!

The supporting papers can be found below; you are welcome to use them for your Problem Set 1 "News & Views" assignment, but we will not specifically have a paper discussion about them.

• Pillow, J., Shlens, J., Paninski, L. et al. Spatio-temporal correlations and visual signalling in a complete neuronal population. Nature 454, 995–999 (2008). 
• Schneider, S., Lee, J.H. & Mathis, M.W. Learnable latent embeddings for joint behavioural and neural analysis. Nature 617, 360–368 (2023).

Lecture: Neural & Behavioral Analysis:

• This week we will hone in on behavioral analysis tools for neuroscience - a critical aspect of understanding how neural activity relates to behaviors.

Section Paper(s): 

• “Deep learning tools for the measurement of animal behavior in neuroscience.” Current Opinion in Neurobiology (2020).
• "A Primer on Motion Capture" Neuron (2020).
• "Keypoint-MoSeq" Nature Methods (2024) <-- please read for Journal Club

• Demo Notebook - using DeepLabCut w/your awesome TAs!
• (Optional) Extra notebook: label your own data! 👩‍💻

Encoding of Space in the Brain

How do we know where we are? Do we have an internal map in our brain? In this lecture, we will discuss how the brain encodes space. 

No in-person lecture: please see this video: Video of the Lecture 

Section Paper:

•  Hafting, T., Fyhn, M., Molden, S. et al. Microstructure of a spatial map in the entorhinal cortex. Nature 436, 801–806 (2005). https://doi.org/10.1038/nature03721.

Problem Set #1 Week: please use the time to work on your assignment! See Guideline below. 

There is no in-person lecture.

Section: come get feedback on your problem set from your TAs and peers! 

Lecture: NeuroAI

What is neuroAI, and how can it help us understand biological intelligence? In this lecture we will cover the background and key examples in sensory systems for how we use task-driven approaches in neuroscience to build better models of intelligent systems.

VIDEO RECORDING: http://mediaspace.epfl.ch/media/t/0_iu2d9ajq

Lecture Paper reading: 

• Required: Decoding the brain: From neural representations to mechanistic models. Mackenzie Weygandt Mathis, Adriana Perez Rotondo, Edward F. Chang, Andreas S. Tolias, Alexander Mathis. Cell 2024
• BONUS (not required reading, but you can use it for your Problem Set): Walker, E.Y., Sinz, F.H., Cobos, E. et al. Inception loops discover what excites neurons most using deep predictive models. Nat Neurosci 22, 2060–2065 (2019). https://doi.org/10.1038/s41593-019-0517-x

Spring Break Starts/EPFL Holiday: No Section!

NO CLASSES -- Spring Break 🎉

Problem Set #2: How to write a scientific proposal

There is no lecture  this week; please use the time to write your proposal.

• This assignment teaches you a useful skill that will help with master projects, master thesis planning & writing, PhD candidacy exams, and even onwards to faculty-level project grant writing! HERE is a LaTeX template if you would like to use it for writing your assignment.
• In section you can get feedback on your proposal!

NOTE: Problem Set #2 Assignment DUE MAY 17th

Brain Machine Interfaces for Systems Neuroscience

How can BCIs be used to study the brain, and how can a better understanding of neural circuits help create better BCIs? 

Please watch before class: Part 1 is a video lecture by Prof Mathis, which covers the critical background and cutting-edge approaches to all-optical BCIs, and studies in language, vision, and motor systems. 

• Lecture Video Part 1. 

In class:  Dr. Spencer Bowles, gives us a deeper dive into electrical arrays, neural dynamics, and neural subspaces, covering this paper in depth: Vivek R. Athalye, Preeya Khanna, Suraj Gowda, Amy L. Orsborn, Rui M. Costa, Jose M. Carmena. Invariant neural dynamics drive commands to control different movements. Current Biology 2023. 

• For Ref: Lecture part 2 from 2024.

Section Hands-On:

Building "BCIs" for muscles! You can record EMGs and build a decoder for the muscle-computer-interface! Please join! 💪 

Cognition & Skill Learning

This lecture will bring together sensory, motor, and decision-making to study skill learning. Specifically, the lecture will use principles of sensory & motor systems to engineer models of skill learning. Then, how to map this back to biological systems. 

Section:

• Paper discussion on Acquiring musculoskeletal skills with curriculum-based reinforcement learning. Silvio Chiappa, Pablo Tano, Nisheet Patel, Abigaïl Ingster, Alexandre Pouget, Alexander Mathis. Neuron 2024

Emerging topics in systems neuroscience & final exam review

• During the first ~hour we will cover some "breaking news" systems neuroscience papers that were published during the semester. 
• In the second half, join us for pizza and Q&A for the final exam review and a little celebration to wrap up the course! 🍕🥳

Section: 

• FINAL EXAM! covering W1-13