BME 3636

#tCFspring2021 Professor Levy is definitely a very knowledgeable, experienced and intelligent researcher in the field of computational neuroscience. This course is well structured: he spends the first third of the lecture talking about--and a lot of the times correcting--homework assignments, the second part lecturing the new contents and the last third about specific homework assignment for next week.

The course is very loaded: we were introduced mainly to two hippocampal CA3 models and one synaptogenesis model. All of them MATLAB based. So if you don't have some basic foundation in programming in matlab of python and you don't spend time learning MATLAB and the models themselves, you won't benefit from being introduced to these models. Aside from examining and playing with these models, we mainly learned about synaptogenesis rules--we were asked to implement the categorized and uncategorized weight change rules--we also learned about information theory including concepts such as mutual entropy and statistical independence. I myself had only a little bit of matlab background prior to this course, so this course had mainly been like a (time consuming) math and programming course for me. If you truly want to do justice to this course, I would recommend allotting about enough time to read the lab lecture notes all of which he posts at the beginning of each course, doing the activities in there, and take enough time playing with the models yourself. And also, don't forget to go to office hours for homework assignments, he has one hour of OH each week for this semester but since not a lot of people go, he can really give your individualized help.

My biggest problem is that he speaks so slow, and my mind wanders off if I have to wait for him to do the steps of the calculations. One solution to this is to work alongside of him, don't try to just let him do the calculations and proofs, do them yourself, that way you'll use your lecture time way more efficiently and don't have to go back and play the recording double speed to learn (like I did).

We had zero class discussion and it's mostly Levy lecturing--alas, it's covid time in zoom uni.

Since this is a covid time course, the grading consists of just homework and a final project, with no quizzes or and exams. the homeworks are graded leniently, and you are allowed to submit revisions. You have 6 days to complete each HW assignment but never, never, never wait till the last day. The final project is for you to run simulations on a model of your choosing, and conduct an experiment to analyze a variable or a certain way of implementing the model. But it shouldn't be about programming a new/revised model, as this is still a neuroscience course...and he doenst want you to spend all of your time debugging.

If you're interested in computational neuroscience and mathematical modeling, this is a good course, as he mentioned, a background in neuroscience would be helpful--especially in terms of making sense of the biological significance of the implementation of the models--but not necessary. He does go through a little bit of neuroscience and information flow in the natural world but you can definitely pick those up on youtube. If you have a solid background in linear algebra and probability you'll have a lot of fun in this course.

Good luck. This can be a very rewarding (and time consuming class) if you put proper effort into it.

This class was extremely time-consuming and extremely difficult (and I say that as someone who did really well...). 2/3 of the class dropped it, and I'd advise not even attempting it unless you're very, very comfortable with MATLAB. However, I highly recommend it if you really want to *understand* neuroscience or neural networks. It will completely change how you see brains, as well as make sense of a lot of psychology.

I like that this class is one block of 3 hours. I also think the subject matter was very interesting, and Levy is certainly an expert on the subject. However, this class is poorly developed and grading expectations are incredibly arbitrary, and Levy does not always explain things to students as if this were the first time they ever saw this material. Beware to students who are not familiar with Matlab!