Seminar Computational Intelligence E (708.115)

Applications of Machine Learning in Neurobiology

SS 2007

Institut für Grundlagen der Informationsverarbeitung (708)

Lecturer: O.Univ.-Prof. Dr. Wolfgang Maass

Office hours: by appointment (via e-mail)


Location: IGI-seminar room, Inffeldgasse 16b/I, 8010 Graz
Date: Wednesday, 16:15-18:15 p.m.
starting on March 7, 2007 (organization meeting; you can also reserve a talk by sending email to )

Content of the seminar:

We will discuss in this seminar the application of advanced methods from machine learning and data mining  (pattern classification, feature extraction, clustering)
in order to detect patterns (e.g. "neural codes") in high-dimensional time series as they emerge from various types of multi-unit recordings in neurobiology (spike trains, local field potential, EEG).
We will focus on a few concrete study cases where our institute is involved in the analysis of data from neurobiology, and will examine the performance of various methods.
The available data and algorithms will be brought into a common format, so that it will be easy to compare the performance of different algorithms.
In addition we will have talks on some new approaches for pattern recognition and feature extraction in machine learning, that promise to provide improved performance.
The talks in this seminar can also serve as preparation for a project or master thesis on applied machine learning.
Each talk will have a lengt of 40 minutes (in order to serve also as an opportunity to extract and descibe the
most salient aspects of a paper).

Those among the papers listed below that are not publicly available, can be found in our pdf-archive (students who are not working at our institute can get a copy from Angelika Zehetner

Tentative topics for talks

At our organization meeting we came up with the following plan:

Later in March (to be announced): Michael Pfeiffer and Martin Ebner
will talk about their analysis of the Cricket data from the KFU-team;
and Dr Hartbauer from the KFU will present details of these data.

21.05.2007  Michael Pfeiffer and Gregor Hoerzer:
Clustering (spectral clustering, and message-passing clustering)
Presentation: PDF

23.05.2007  Martin Ebner
Methods developed for BCIs in Berlin, which may be relevant
for our data (in particular analysis of LFPs).
Presentation: PDF

06.06.2007  (16.15 p.m.) Klaus Schuch
Sparse Bayesian Learning and Relevance Vector Machine
Presentation: PDF

28.06.2007  (16.15 p.m.) Dejan Pecevski
Nonlinear feature selection with PSVM
Presentation: PDF
28.06.2007 Stefan Klampfl
Methods for estimating mutual information developed by Stefano Panzeri et al
Presentation: PDF

Remains to be scheduled:

Talk by Stefan Häusler on the data which he currently analyses

Talk by Robert Legenstein on his analysis of the data of Gregor Rainer

(some of the papers are quite long, and have to be split into several talks):

-- Information contained in local field potential (LFP) and spike trains of monkeys

-- Analysis of information contained in spike trains from the Omega-neuron  (= first stage of  the auditory system) in crickets

-- Separation of  memory-induced and  visually induced information in the visual cortex of monkeys

-- Encoding of the percept of a visual stimulus
(for the case of ambigous images, where the visual input does not change, but the internal percept in the brain changes)

-- Benjamin Blankertz, Guido Dornhege, Matthias Krauledat, Volker Kunzmann, Florian Losch, Gabriel Curio, and Klaus-Robert Müller. The berlin brain-computer interface: Machine-learning based detection of user  specific brain states. In Guido Dornhege, Jose del R. Millán, Thilo Hinterberger, Dennis McFarland, and Klaus-Robert Müller, editors, Towards Brain-Computer  Interfacing. MIT press, 2007. in press.

-- Classifying Single Trial EEG: Towards Brain Computer Interfacing , NIPS 2002
Benjamin Blankertz, Gabriel Curio, Klaus-Robert Müller

-- Hochreiter J. and Obermayer K. Support vector machines for dyadic data. Neural Comput., 2006.

-- S. Hochreiter and K. Obermayer. Nonlinear Feature Selection with the Potential Support Vector Machine. In: Feature extraction, Foundations and Applications, Springer, 2005.

-- Arabzadeh E, Panzeri S, Diamond ME. Deciphering the spike train of a sensory neuron: counts and temporal patterns in the rat whisker pathway.
J Neurosci. 2006 Sep 6;26(36):9216-26. Erratum in: J Neurosci. 2006 Sep 20;26(38):9835. PDF

-- M.A. Montemurro, R. Senatore, S. Panzeri. Tight data-robust bounds to mutual information combining shuffling and model selection techniques,
Neural Computation, in press

-- T. Natschläger and W. Maass. Dynamics of information and emergent computation in generic neural microcircuit models. Neural Networks, 18(10):1301-1308, 2005

                                                                                                                                                                                                                        2007-05-30, daniela