Yusi Chen
A guided roadmap for learning computational neuroscience, including recommended topics, resources and labs.
Getting Started
LLM experiences
In the era of LLMs, using them correctly and prompting them effectively can significantly boost efficiency. My strongest impression, however, is that they tend to hallucinate, make things up, and often agree with you uncritically. It’s essential to read through and understand everything they generate. You’re welcome to share your own experiences with me via email.
- Searching: I find the DeepThought mode in the GPT family especially helpful for literature search, as it provides references to actual papers and helps reduce hallucinations.
- Coding: In my view, Cursor is very useful since it can search across the entire project repo, test and launch functions, write documentation, and comment code. However, hallucinations still occur, and the tool doesn’t always accomplish what it claims. Always make sure you understand the code it produces, or ask the agent to provide a step-by-step walkthrough of the project code.
- Writing: As a non-native English speaker, I often rely on GPTs to polish my writing (as I am right now). That said, their wording tends to be general rather than precise. For academic writing, precision and conciseness are critical, so I always double-check and refine the output.
Prerequisites
- Mathematics: Linear algebra, calculus, probability and statistics
- Programming: Python or MATLAB proficiency
- Neuroscience: Basic understanding of neurons, synapses, neural circuits and neural anatomy.
This roadmap is a living document. Suggestions and contributions are welcome!
Contact: Feel free to reach out if you have questions about computational neuroscience or need guidance on specific topics.