**Preperation:**

I started about 1 1/2 months before the examination. Got the "Machine Learning and Pattern Recognition" by Bishop and the "Dynamical Systems in Neuroscience" by Izhikevich. Both were a very good help. For the rest I used the lecture slides and the script.

**Topics:**

50% Neuroinformatics

50% Neurodynamics

**Neuroinformatics:**

Derive Bayes theorem and explain each individual term.

What do you choose as an initial prior if you have no information about the underlying distribution?

How does the Maximum Likelihood approach in principle work?

What is an estimator?

What is the Fischer information? And does is it has to be positive(2nd derivative times -1)?

If you've done a regression how can your model be evaluated?

Explain Likelihood ratio, Deviance and AIC. Which one would you chose and why?

What effects appear if you do a linear polynomial regression? And why is a polynomial basis function prone to produce these two effects?

**Neurodynamics:**

What was the approach in Neurodynamics compared to Neuroinformatics? What is a dynamical system in general?

How did we derive the models? From Hodgkin Huxley -> 1D -> 2D -> ... -> Izhikevich.

What's the major advantage of the Izhikevich model?

What is periodic firing? And how does the system has to look like?

What is an isochrone and what is it good for?

**Examination:**

You have a piece of paper in front of you. The hardest task was to guess what Prof Pipa was aiming at. You should explain directly, fast and precise to use the very limited time in order to show how much you've understood. If you get stuck at a question, Pipa will postpone it to the end or you ask him to postpone it for later. He is very calm and supporting. If you don't know what he's aiming at you just ask to rephrase. Most of the time he does not state any advanced questions.

My second assessor did not ask any questions, he merely wrote the protocol. Overall I can highly recommend an oral exam in Neuroinformatics. Passing should be no problem anyway.

**Grade:**

1.0