Computational Neuroscience

I am studying the Master's degree Computational Neuroscience in Berlin for a half year now. Previously, I was doing my Bachelor’s degree in Cognitive Computer Science in Bielefeld and was writing each semester about my experiences (in German). A post about Cognitive science in another blog inspired me to do so and I want to keep up with the tradition. Thus, here is my report about the first semester Computational Neuroscience! It is in English because the course of studies is also and it aims on attracting foreign students.

Overview

  1. What is Computational Neuroscience? Should I study it?
    1. You should study CNS if …
    2. You should not study CNS if …
  2. Where do I study it?
  3. I have to get a Bachelor's degree first! In what should I do it?
  4. How do I get into the Computational Neuroscience program?
  5. CNS at the BCCN
    1. Math Prep Course
    2. Individual Studies
    3. Programming Course and Project
    4. Models of Neural Systems
    5. Acquisition and Analysis of Neuronal Data
    6. Machine Intelligence
    7. GRK Lecture Series
    8. Ethical Issues
  6. Social Life
  7. Where to Live
  8. Where to Eat
  9. Where to Work if You Need Money
  10. My Personal Experience and Conclusion

What is Computational Neuroscience? Should I study it?

First of all, you might ask what the heck is Computational Neuroscience? (I will abbreviate it with CNS in the following.) I will try to give you some kind of a definition, but don't be surprised if other people have other definitions.

Probably, the neuroscience part of CNS is rather clear. It is the science about neurons and neuronal systems (most prominently the brain): How do neurons work? What kinds of neurons exist? How do they connect? How is the nervous system affected by diseases? These are some examples of questions asked by this branch of research. The ultimate goal is of course to understand the brain in its entirety.

The more mysterious part of CNS is the “computational”. It can be interpreted in at least two different ways. On the one hand, it specifies what neuroscientfic topic is examined: the way neurons do computations and process information. On the other hand, it specifies the most important method used in this science: computer simulations of neurons and networks of neurons.

This makes CNS a highly interdisciplinary enquiry. Methods and knowledge from biology, medicine, psychology, mathematics, physics, computer science and maybe even more fields have to be combined. I think, this makes it highly interesting, but can also be daunting as you will certainly touch scientific fields in which you are not overly comfortable.

Overall CNS is very theoretical: The modelling of neurons and networks is itself theoretic because models are usually mathematical models. But also the analysis of fMRI (functional magnetic resonance imaging, that is what produces these nice pictures of brain activation) involves quite a lot of math. Therefore you should be comfortable with math and computer science.

Nevertheless, don't be fooled into thinking there is not so much biology in there. For me as computer scientist it's quite a lot. At least, our models should be biological plausible and this requires us to know biology and also physics (maybe even a bit chemistry).

Beneath bridging between different scientific fields an important aim of CNS is to bridge scales. Neurobiology examines single neurons, psychology examines “the behavior of whole brains”. But how do we cross the scales from single neurons to whole brains? CNS has the potential to play in important role in answering that question.

I hope this gave you some impression. If not, take a look at Wikipedia or the website of our institute.

Now, that you know what CNS is you might wonder: “Should I study it?”

You should study CNS if …

  • you think, the traditional artificial intelligence stuff in computer science is biologically to implausible to ever get really intelligent.
  • you want to understand how one gets from single neurons to the behavior of a complete brain.
  • you want to do (academic) research. (The course of studies is heavily research oriented.)
  • you like many natural sciences and math.
  • you are looking for a challenging course of studies.
  • you want to study in a small group with nice and international colleagues.

You should not study CNS if …

  • you do not like math.
  • you cannot program or even worse cannot use a computer (although basic programming skills might suffice).
  • you are not interested in neurobiology or basic psychology.
  • you run away screaming when hearing electrical engineering terms like “capacitance”.
  • you are too easily frustrated by exercise sheets you cannot solve within an instant.
  • you want a well payed job (actually, I have no idea, how well payed jobs for CNS graduates are; but probably you could earn more money with a computer science degree in the economy).
  • you are not able to learn missing background knowledge by yourself.
  • you cannot speak and understand English (should be obvious ;)).

Where do I study it?

Like me, you can study Computational Neuroscience as a Master's degree at the Bernstein Center for Computational Neuroscience in Berlin (BCCN Berlin). That is mainly an association of different research groups of all major universities in Berlin. Most prominently the Technische Universität (TU, “technical university“), Humboldt-Universität (HU), Freie Universität (FU, “free university“) and the Charité. In the Master's degree you will be enrolled at the TU, but the degree will jointly be awarded by the TU and HU. The BCCN offers also a PhD graduate program in the area of Computational Neuroscience.

But there are also some alternatives of more or less similar Master programs in the German speaking area (I cannot list every program in the whole world):

  • Neural Information Processing in Tübingen seems to be focused more on the biological foundations in comparison to Berlin. The program in Tübingen can be completed in 3 semesters (opposed to 4 in Berlin), but then you have probably not a single week of holidays. After completing the program you can directly continue with a PhD. Actually, they prefer students who also want to pursue a PhD afterwards. The Neural Information Processing program is also related to the Bernstein Center for Computational Neuroscience in Tübingen. If you are interested in Neurosciene, but not so much in Computational Neuroscience, Tübingen offers also a couple of other Neuroscience programs.
  • Neural Systems & Computation in Zürich (Switzerland). Can't say much about this except that the living expenses in Zürich probably exceed the ones in Berlin.
  • Science of Intelligence in Berlin is not a Master program on his own, but a specialization which can be studied as part of different Master programs. At the moment this is at least possible as part of the Computer Science program at the TU Berlin and maybe also as part of the Computer Science program of the FU Berlin. This already shows that this program is much more computer science and artificial intelligence oriented. Apart from that mainly psychology contributes to this program. Nevertheless, the program allows to visit most of the Computational Neuroscience lectures as far as I know. What I do not know is whether it is possible to only do English lectures in this program. At least the webpage seems to be German only.
  • The Computer Science Master program of the TU Berlin offers also the focus Intelligent Systems which is probably even more Computer Science/artificial intelligence. Again I cannot tell you whether you have to attend German lectures.
  • In Bielefeld (were I obtained my Bachelor's degree) is another place offering Intelligent Systems. Unfortunately, only in German.

If you are not so much interested in theoretical aspects, there is quite a number of other Neuroscience programs with a different focus. Just search for them.

In the following I will only talk about the Computational Neuroscience program here in Berlin because that is the program I am in.

I have to get a Bachelor's degree first! In what should I do it?

Maybe you do not have started with a Bachelor's degree, yet, and are wondering which one would be the best preparation for Computational Neuroscience. You should chose a degree with a high amount of math and some programming. I think, Computer Science is a good choice because those programs teach a mostly sufficient amount of math and also you get the required programming skills. Maybe you will learn even a bit physics or electrical engineering which is also helpful. But you might lack some biological knowledge. Math, Physics and Electrical Engineering are presumably also very good choices. However, then you might not just lack the biological knowledge, but also the programming skills. However, I suppose most Math programs also teach a sufficient amount of programming by now. Biology might also look like a good choice in preparation for CNS. It might teach you all the relevant neurobiology and neuroanatomy. Nevertheless, you might lack the math and programming skills. I think, those stuff is much harder to acquire on your own than the biological foundations.

Apart from these general considerations there are two programs I can especially recommend. Probably the more well known is Cognitive Science in Osnabrück. I believe, it gives you a good overall view of cognitive science and neuroscience. Also, as far as I know, it gives you enough possibilities to focus on specific parts (math, computer science, psychology, biology, ...). In my class are two people who did their Bachelor degree in Cognitive Science and among all the people of the BCCN are many more. There is also the German article about the program by a former Cognitive Science student (I mentioned it already in the introduction).

Finally, one of the best preparations – in my opinion – is the program I studied: Kognitive Informatik in Bielefeld (Cognitve Computer Science). It is mainly computer science and gives you therefore nearly all of the relevant math knowledge and sufficient programming skills. In the first semester you also have a course about basic neurobiology covering some of the required biology knowledge. Moreover, there is the course about neural networks and learning. This gives you an easy start in the Models of Neural Systems lecture where you will discuss the perceptron and maybe Hopfield networks first before continuing with more biological plausible models. But more important many topics in the Machine Intelligence course will be a repetition (more about that below). Depending on which elective courses you chose you can get even more preparation for CNS. I suggest to take one Statistics course as elective (e.g. “Teilmodul Statistik/Infomartik” of the department of biology) to get the formal math requirements for the application (even if this is a somewhat soft requirement). It is also quite helpful and with all the math you will have had before not very difficult (but it is not a repetition of content of other math courses). Unfortunately, the program is only in German. If you want to know more about Kognitive Informatik, you can read about it in my German blog.

Whatever program you might chose for your Bachelor's degree you will probably miss some knowledge. But this is normal for such an interdisciplinary program. Actually, in the program is time devoted to individual studies to close those gaps. Thus, do not be afraid of having some gaps, but be prepared to close them by yourself.

How do I get into the Computational Neuroscience program?

You have decided to study Computational Neuroscience and you are wondering how to get into the program? You have to apply by March, the 15th, for the next winter term. To do so you have to fulfill some prerequisites (no warranty for correctness, also check the official sources):

  • You will need a Bachelor's degree. Actually, you can apply without having the Bachelor degree completed because the application deadline is probably before you receive your degree if you want to pursue the Master directly following the Bachelor. In this case you have to provide your certificate within a certain time frame (I believe, you have time until end of the first semester to hand in all certificates) and the application has to contain some kind of preliminary certificate with the current grades.
  • You have to be a native speaker of English or proof that you have a certain level of proficiency. There are different ways to do so. For example they accept the internet based TOEFL test with at least 88 points.
  • You need “sufficient” (24 credit points) mathematical knowledge including linear algebra, analysis, probability theory and statistics. See the official pages for the details.

Then you will need some documents:

  • A letter describing your motivation.
  • A tabular CV.
  • A transcript of records (and certificates) of previous studies.
  • Two letters of recommendation.
  • Certificates of the prerequisites mentioned above.
  • A couple of filled out application forms (see the official sources).

This is quite some stuff. Thus, better start early to get all of the documents together and take a look at the formal requirements. There are quite specific instructions how copies have to be certified. Once you have everything you apply by using the online platform uni-assist (you still have to mail the documents). The first reply that I have been selected for the Master program came by email at the beginning of May.

Make sure you get all the formalities right because uni-assist does a pre-selection based on these formalities and only those applications will be forwarded to the BCCN selection committee. In case you notice that something went wrong you should contact our teaching coordinator Vanessa Casagrande. She might still be able to retrieve your application from uni-assist and clarify any issues.

As far as I know there were about 70 applications this year (for the winter term 2012/2013) and each year 10 new students are allowed to join the program. If you want more details on how the students are selected, you can take a look at the Admission Regulations. The most important single part is the grade of the Bachelor's degree, but the mathematical courses and other qualifications are also quite important. This allows to counterbalance shortcomings in one area with other achievements and highlights the overall picture of the applicant. In the end each application is ranked by a score and the highest ranked students get selected.

CNS at the BCCN

The structure of the Computational Neuroscience program is as follows at the Bernstein Center in Berlin:

Module structure of Computational Neuroscience

In the first year you will have normal courses. In the second year you are mainly doing project work in different research groups (“Lab Rotations”) and writing your Master thesis. You can take a look at the official module page to get an impression of what the schedule in the first two semesters might look like. Note that there might be listed a few lectures which are not compulsory. The amount of time spent on exercises is of course not visible in the schedule and that is quite a significant amount of time.

You might have noted that each module is given with a number of LP. This stands for the German “Leistungspunkte” meaning credit points. Sometimes you will also find the abbreviation ECTS for European Credit Transfer System. The credit points are supposed to indicate the amount of work needed to complete the module whereby one credit point is supposed to be 30 hours of work.

It is quite nice that you can select from a variety of courses from all major Berlin Universities in the elective courses Individual Studies and Courses on Advanced Topics. The former module will be discussed a bit more below. The latter will (probably) discussed in a later blog post after I completed it. But take a look at the section about the GRK Lecture where I am also telling a bit about this module.

If you have study related questions, there are at least three persons you can go to. One is mentor which gets assigned to you at the beginning of the Master course. It is usually a professor at the BCCN, but can sometimes also be a PhD student. The mentor is supposed to give you study related advice and discuss what to do as Individual Studies. Depending on the person assigned to you this can be more or less helpful. It really depends on how much the mentor wants to do this. Some really like it, whereas others not even answer any of your emails. But our teaching coordinators are trying to ensure an adequate mentoring by exclude those persons. In any case, you are free to change your mentor anytime.

The other two helpful persons are our teaching coordinators Vanessa Casagrande and Julia Schaeffer. If you have any questions or problems related to the study program you can always go to Vanessa. Also, she is providing us with useful information how to organize our courses, sometimes acts as bridge between the lecturers or the examination office and us. Probably it also due to her communication with the examination office that our courses from all the different universities get recognized without problems. It seems that this is usually really troublesome in other study programs.

To give you a better impression what we learn in the individual modules and lectures I will discuss the modules I attended during the winter term in the following. Apart from those there is nearly every week at least one neuroscientific talk you can attend in Berlin.

Math Prep Course

It all started with the Math Prep Course at the beginning of October. It is not compulsory, but I would suggest visiting it. It gives you already 4 of the 6 credit points for the Individual Studies. Most of the math you will need during the program (analysis, linear algebra, differential equations, probability theroy and some other stuff) will be discussed there. That way you can close gaps in your math knowledge and reactivate your rusty math skills.

The lectures were held by Prof. Schimansky-Geier. His English is not the best and his lectures are not very energetic, but all in all quite okay. The lectures before noon were followed by a tutorial. We had two different tutors, but I think I met only one (I missed out on some days of the course). Unfortunately, that tutor's English skills were really bad leading to many misunderstandings. Either she didn't understand our questions or we didn't understand her explanations. :( As far as I know the next Math Prep Course will be held by a new professor at the BCCN. So, things might improve.

About two weeks after the Math Prep Course of eight days we wrote an ungraded exam. It wasn't very hard and as far as I know nobody in our year failed the exam.

Individual Studies

You have to earn 6 ungraded credit points in the module Individual Studies. It's purpose is to give you the possibility to close gaps in your knowledge required to study the program. Basically, you can take any course you want, but it should be approved by your mentor. Many of us did the Math Prep Course because it is before the regular semester. Therefore, it is not adding to your high workload later. It already gives you two thirds of the points needed, too.

It is also possible to get credit points for reading a book or attending a summer school. You should talk to your mentor if you want to do this. I have read a couple of chapters in the Principles of Neural Science by E. Kandel but have still to meet with my mentor to finally get the credit points.

Programming Course and Project

The Programming Course is held by Robert Martin. It consists of lectures in which you learn Python and some other things like SVN, Test Driven Development or UML. If you never have done any programming before, it will probably be hard to follow. In contrast it can be quite boring if you are already familiar with programming.

Besides the lectures there is a tutorial and you have to do some exercises. These are quite easy and do not add much to the overall workload. Do not expect to get much feedback on them.

After the semester you have to do a two week programming project in a small group (2 to 4 people). You can freely chose a project. It is quite common to replicate the work of some neuroscientific modelling paper (in my group we replicated this paper). In the next winter semester we have to do a presentation of it to the new students to give them an idea what they can do.

Models of Neural Systems

The Models of Neural Systems course consisted out of four parts: a theoretical lecture, an “experimental” lecture, a computer practical and an analytic tutorial. The theoretical lecture was the main lecture. It covered topics like:

  • Basic artificial networks including the perceptron and memory networks.
  • Phenomenological model of single and complex cells in the visual cortex.
  • Nernst equation for estimating reversal potentials.
  • Goldman-Hodgkin-Katz current and voltage equations (also to estimate reversal potentials).
  • Leaky integrate and fire neuron model.
  • Hodgkin-Huxley model for action potential generation.
  • Modelling stochastic ion channels.
  • Phase space analysis of differential equations.
  • Cable equation to model extended neurons.
  • Networks of neurons modelled by differential equations.

As you can see the modelling of single neurons and their action potentials were in the focus. Differential equations are quite important for this. The lecture was mostly held by Prof. Kempter, but some parts like the cable equation were done by Prof. Lindner. In my opinion both lectures are good average.

Every week we had to solve an analytical problem sheet and programming problem sheet with exercises related to the lecture. The time needed for solving was quite much (several hours). The programming exercises were done with a partner and in the last five weeks we got a project instead of exercises. We also had to do a short presentation about it. Officially this presentation does not influence the grade, but someone saw the lecturer taking notes about the quality of presentations. Decide for yourself what you think about this. ;)

The theoretical lecture was concluded by an oral exam of 30 minutes. This was quite manageable. I heard from no one who got a grade worse than 2.0. I personally got some nasty questions about a topic I overlooked a bit when learning for the exam, but I was able to deduce the answers from other knowledge. That way I also got a absolutely satisfying grade.

Until now I did not write anything about the experimental lecture. That's because it is not directly related to the other parts. The experimental lecture was held by another lecturer each week presenting their own research. It was nice to get such a variety of impressions. The content of these lectures was not relevant for the exam.

Acquisition and Analysis of Neuronal Data

This module lasts two semesters and will end with an oral exam. Of course I cannot tell you something about the second term, yet. The topics covered in the first term included cellular recordings (extracellular, patch clamp etc.), EEG and fMRI. Thus, it was more acquisition than analysis and the next semester it will be the other way around.

The lecturers were Prof. Brecht, Prof. Curio and Dr. Blankenburg. Depending on the person the lectures were average to excellent. You have to get a bit used to Prof. Brecht's lecturing style. He will usually ask a couple of questions during the lecture and wants you to make guesses if you do not know the answer. At least this makes it a bit more interactive and lets you think about material.

Besides the lectures we saw demonstrations of the cellular recording techniques in anaesthetised rats. This was quite interesting once you got used to the preparation of the animals. If you do not want to attend these animal experiments for whatever reason, probably no one will force you. But it is good to see for yourself how such experiments are performed. Moreover, much more measures to prevent any pain or discomfort to the animal were applied than I imagined.

Machine Intelligence

Another module lasting two semesters is Machine Intelligence. In the winter term we focused on supervised machine learning techniques, whereas we will focus on unsupervised learning in the summer term. More precisely, the topics of the first half were:

  • Artifical Neural Networks
    • Multilayer Perceptron
    • Radial Basis Function Networks
  • Statistical Learning Theory
  • Support Vector Machines
  • Probabilistic Methods
    • Bayesian Networks
    • Bayesian Inference

The importance of Machine Intelligence for Computational Neuroscience is twofold. On the one hand, it provides a theoretical background and the artificial intelligence view on learning (and similar topics). On the other hand, the methods of Machine Intelligence are often used for data analysis in neurosciences.

The lecturer of this course was Prof. Obermayer. He structured the topics in a very good way in my opinion (better than in a similar lecture I heard in Bielefeld). Unfortunately, his presentation is very “dry” and I am missing good examples or demos of the learning algorithms (this was definitely better in Bielefeld).

Each week an exercise sheet has to be solved and will be discussed in a tutorial. Usually these are analytical exercises, but sometimes also programming exercises. The programming exercises have to be solved in small groups. After the second part of the lecturer we have to take an oral exam covering both parts.

GRK Lecture Series

Every first and third Wednesday of a month the three hour GRK Lecture takes place. It is held each time by a different lecturer giving an introduction to things related to their research. Therefore there is quite a variety of topics which are varyingly interesting. Nevertheless, it is nice to get an introduction to such a variety of topics. Take a look at the official webpage to get an impression of the topics.

The GRK Lecture Series is mandatory for PhD students, but can also be taken by Master students. It can than be part of the Courses on Advanced topics and gives 6 credit points for two semesters. You can chose between ungraded or graded. Each way you have to solve some exercise sheet for each lecture and your grade will be calculated from your scores on those sheets. The problem with this is that you are almost never sure what exactly is expected of you in the exercises. Usually you know this after one or two exercise sheets because in most courses all sheets are rated by one person. But in the GRK lecture it is each time someone else.

If you intend to take this course, anticipate the additional workload. But you can manage it. I think, approximately half of our class is doing this lecture additionally and is able to cope with it. Also, it is nice to get already 6 of the 10 credit points for the Courses on Advanced topics in the first year. Another thing to keep in mind is that there are some lectures during the semester break (as the PhD students do not have semesters and semester breaks).

Ethical Issues

The module on Ethical Issues is a one week block course which is intended to be taken after the third term. However, some of this year's Master students including me decided to do it already after the first semester. Actually, I would recommend it doing this way. It does not add to the workload during the semester and gives you more flexibility when you are doing you lab rotations. This could be especially important if you are planning to do a lab rotation abroad. Moreover, there were some things which you ideally should have heard or thought about before the lab rotations.

This year the module was for the first time the “Winter School Ethics and Neuroscience” and open for external students (not just of the Berlin universities). It consisted mainly of different lectures. The topics included:

  • The Neurobiology of Values – Scientific Paradigms
  • Introduction to Applied Ethics
  • Good Scientific Practice
  • Patient Data Security
  • Deep Brain Stimulation and Ethics
  • Ethical Issues of Animal Experiments
  • Neuroenhancement
  • Legal Aspects of Neuroimaging
  • Social and Ethical Implications of Brain-Hardware Interfaces
  • Ethic Committees

Depending on the lecturer the parts were more or less interesting to listen to. I found the style of presenting of two lecturers quite bad, but other lecturers were excellent (e.g. the Introduction to Applied Ethics by Prof. Pauen or Good Scientific Practice by Prof. Dirnagl). In fact, this ethics course was the module I liked most so far.

Apart from the lectures we had to do some group work and present it. Unfortunately, they assigned the groups at random by counting and you had only a vague idea who was part of your group. This made it hard to organize the group work.

Social Life

Maybe you are not just wondering what you will learn in the Computational Neuroscience program, but also about the kind of people studying it. With just 10 people each year it seems that you do not have many possibilities to chose your friends and to some degree this might be true. However, I can say only good things about the students from my year and those of previous years. The class of my year is meeting nearly every weekend for some fun activity. I heard similar things from previous years.

Also, there is not much competition among us students as one might expect. Rather, we help each other if someone has a question about an exercise or similar stuff. As discussed before in some exercises we are even “forced” to work together.

Apart from the students of your year you can meet other people (mostly older students and PhD students) at various occasions. Once a month the “BCCN Stammtisch” takes place where we meet in a bar, in December was a Christmas party, in the summer there might be barbecues and there is the BCCN retreat. The nice thing about these events is that you can easily meet the older students and PhD students. They can give you valuable information about their exams and lab rotations. Or you have just fun partying together with them. :)

If you want to get to know even more people, you could try the sports courses of one of the universities. They offer a large number to chose from.

Another thing I should mention, even though it has not necessarily to do with the social life, is that I am quite impressed by the connections between research groups. I was surprised quite a number of times that one person knows or works together with another person from a completely different research group. I think, such close networks exist not everywhere and being part of such a network is a really good thing.

Where to Live

Most courses take place at the BCCN which is located at the HU Nord campus. That again is located near the Charité Mitte (which in turn is near to the central station). However, one day in the week you will have courses at the TU campus which is at the west side of Tiergarten. This makes the district called Moabit quite popular among CNS students because it is more or less in-between both locations ensuring short routes and it is also quite central. Me included three persons of our year are living there and at least two PhDs, too.

I think, the rents in Moabit are low priced. But they can vary at least by a factor of two (a colleague is paying about the same as I for half the space). I am not sure whether it is the reason for the affordable rents or the other way around: Quite many immigrants and unemployed live Moabit. Moreover, the official website of Berlin says that the criminality is high. But so far personally I felt save in Moabit and in my opinion the district Wedding is worse. I visited some flats there and did not like the area there too much.

Another area is currently becoming popular among the students in our year: Three have moved near to Rosenthaler Platz in Berlin Mitte. It is a nice area near to the BCCN, but not as near to the TU.

Unfortunately, I am not able to say something about all the other districts except one thing you should keep in mind: Berlin is a large city and I would not suggest to take a flat to far away from the BCCN. Then it can easily take you an hour each to get to the university and that is really a lot of wasted time. You might need that time in the first (maybe also second term to do the exercise sheets if you still want to have some leisure time.

Where to Eat

Both near the HU Nord campus and the TU campus is a canteen (called “Mensa” in German) where you can get a meal for usually 2 to 3 euros (or a bit more for special meals and extras) as student. Even though the food there is not really bad, it is not very special and you can get sometimes the feeling that every dish tastes the same. In case you get bored by the food there, you should try Thai Tasty and an Oriental place in Luisenstraße. That is very close to the HU Nord campus and BCCN and you get there a good meal for about 5 euros.

Of course there is an enormous number of restaurants in Berlin with quite a number of good ones. Some of these are even quite cheap. One place you should definitely try is Dolores where you get really good burritos!

Where to Work if You Need Money

In case you need money, you can try to get a job as student assistant in one of the research groups. I regularly get mails about open positions on the mailing lists and it seems to me that you will always find a position if you want to. Usually you will work 40 hours a month and get paid with approximately 10 euros per hour.

Despite that, keep in mind that the CNS master program is intended to be a full time study program and it is indeed a lot of work. I was able to manage 40 hours a month in the first semester. But most of the time I had the feeling that the workload was too much. Therefore, I reduced the work time to a maximum of 30 hours by now. So far I'm quite happy with it this term.

If you apply for job, you should check where your workplace is. Spending each time one unpaid hour to get there or to get back might not be what you want to do. My workplace is at in the Charité Mitte next to the BCCN/HU Nord campus. That is very comfortable because I can walk over there in five minutes and use the free time between courses to work.

My Personal Experience and Conclusion

This whole article has been influenced by my views for sure and at some points I already mentioned some personal experiences. Nevertheless, I want to conclude with some more details on how I personally experienced the first term.

To be honest: I was a bit disappointed in several ways, yet I really like the program. Before I started I imagined the lectures to be really great in such a small and prestigious Master program. In fact, they were mostly good with some not so good and some excellent lectures in-between. Thinking about this it is clear that there is no real reason why the lectures in the program have to be better or worse than elsewhere. The lecturers got their professorship for scientific achievements and there is obviously no reason why this should correlate with giving especially good lectures. Anyways, the small lectures can be an advantage because the lecturers can really discuss questions. I also liked, that one lecture is not necessarily given by only one lecturer, but that each part is done by a lecturer who is focused with his research on that specific topic.

One of the reasons for me starting this Master was that I was missing the biological foundations and plausibility in pure computer science and artificial intelligence. Strangely, it turned out that I got the feeling of CNS being too much biology and too less computer science for me. But the biology stuff got more interesting over time as I got more familiar with it. Nonetheless, I am pretty sure that I do not want to do single cell modelling, but more higher level stuff (which will be covered in the second term).

And last but not least, it was just too much work in the first term. In my Bachelor's degree I was able to do several courses more than the standard curriculum proposed and to work in parallel. Because the amount of credit points in the curriculum per semester was the same in my Bachelor's degree as in the CNS program I expected the workload to be roughly the same. Unfortunately I was mistaken. Due to the workload I was quite unhappy and stressed out the last semester and this caused myself believing that I made the wrong choice with the master. So far the second term is much better. Reducing working hours was a good choice and I have the impression that I am spending a little less time on exercises. By the way, I think the amount of work in the Master program roughly matches the credit points (in my case). This means I worked less in my Bachelor than I should have accounting to the number of credit points. ;)

The last paragraphs might have given you the impression that I do not like the CNS master program. But at the moment (a few weeks into the second term), I think it is a really good and interesting program. All I want to say is that you should consider four things to have a good experience:

  1. The lectures are on average not necessarily better than anywhere else. But the small number of students in the lecture can be of advantage.
  2. Think about how much biology, computer science, math ... (depending on your background) you can endure. Computational Neuroscience is that broad that you probably have to learn some things which do not seem that interesting to you. If the single neuron models are not your thing you can hope for the second term in which you will do Models of Higher Brain Functions (which I really like so far).
  3. The master program will be a lot of work. Consider this if you plan to get a job.
  4. If you think about dropping out of the program think well about it. That might just be a hard phase as for me, but after some time you might again realize how cool that stuff is and why you are doing it.

In conclusion, Computational Neuroscience is a really awesome master program as long as you approach it with a realistic set of expectations and prepare yourself for the workload. You can learn much and get an insight in many neuroscience related topics and laboratories you might not get somewhere else. Last but not least, you can meet nice people there.

Update: Some words about the second semester.