You are here: Home Teaching Summer Term 2015 Seminar / Proseminar: Doing by …

Seminar / Proseminar: Doing by Thinking 2015

Within the proseminar and seminar, students will learn about many aspects of BCI systems, ranging from basic neurophysiology and simple experimental paradigms up to complex machine learning approaches for the decoding of brain signals.

Pro-seminar: Introduction to the Functional Decoding of Brain Signals
Seminar: Invasive and Non-Invasive Methods to Decode Brain Signals in Realtime

About the course

Brain-Computer Interfaces (BCI) provide the possibility to decode brain signals in real-time and use them to control a hardware device or software application. As the signals are noisy, high-dimensional and show non-stationary characteristics, the decoding is challenging. Typically, the informative features and the decoding algorithm must be calibrated for each individual user, sometimes even for each novel session of the same user. If non-stationarities are involved, then adaptive approaches are needed, which can compensate temporal effects within a single session. As BCI systems run as closed-loop systems, they provide various challenges for computer scientists and machine learners. BCI neurotechnology, however, also has the potential to provide novel treatment possibilities in a clinical context (e.g. in stroke rehabilitation), can be used as a tool to watch the acting brain for a neuroscientist, and provide means to analyze and improve human-machine interfaces (e.g. for consumer products). Within the proseminar and seminar, students will learn about many aspects of BCI systems, ranging from basic neurophysiology and simple experimental paradigms up to complex machine learning approaches for the decoding of brain signals.

Organizational Information - Update

  • A summary of the presentation (max 2 pages), should be handed in to your supervisor by May 29th.
  • Below you can find an updated list with the assign topics and the corresponding supervisors.

Organizational Information

  • Session Dates (update)

    • The talks session will be split into two sessions:
      • Tuesday, June 16th
      • Wednesday. July 15th
  • Reports (update)

    • English is the preferred language for the pro-seminar and mandatory for the seminar topics.
    • The reports should be about 10 pages long (latex, a4wide, 11pt) and are due one week after the second talk session, i.e., 22nd July.
    • The format of the report should be that of a scientific paper. Here are a couple of external resources about some general guidelines for writing the report:
  • Please register via the HISinOne ( click here ). The code for the seminar is 7300-S1-19 and 11LE13S-ID128356 for the pro-seminar.
  • The first meeting will be on April 29th, 1 p.m (s.t.), building 78, seminar room 14.
  • Seminar and Proseminar will be held as a compact course on two full days around end of June and first half of July.
  • Topics will be assigned in the first meeting. Below you can find the list of topics and the corresponding basic literature. Please contact your supervisor for more details on literature.
  • The list of topics is designed to be built upon each other which implies that the presentation quality of the basic topics is also fundamental.
  • Students are requested to prepare a talk of approx. 30 minutes and to write a report.
  • Grading: 60% presentation, 30% report, 10% contribution in discussions.
  • Prof. Dr. Wolfram Burgard will give a talk on general guidelines for the seminar presentations.
  • Prof. Burgard's talk on presentation guidelines will take place Wed., June 3rd, 1pm, Room 00.006 in Albertstr.23 (Institutsviertel). All the participants of the seminar are expected to attend to this talk.

Materials

Organizers / Supervisors

Co-Organizers

Materials

Topics to be assigned

Proseminar

  1. [TB] Introduction to EEG / ECoG
  2. [TB] Visual ERP: physiology + BCI application:
  3. [TB] Auditory ERP:
  4. [TB] Haptic ERP
  • [TB] P300
  • [TB] Error potentials
  • [MT] Language Processing
  • [WB] Basic analysis techniques for ERP signals
  • Seminar

    1. [MT] Classification of target-/non-target ERPs with shrinkage LDA:
    2. [MT] Subclass shrinkage for ERP classficiation:
    3. [MT] Spatial filtering for ERP signals (Laplace and XDAWN):
    4. [MT] Unsupervised classification of ERP signals:
    5. [WB] Error potentials and applications in robotics:
    6. [WB] Tactile ERP BCI to control a robot:
    7. [WB] Visual ERP BCI to control a robot:
    8. [WB][MT] Dynamic stopping for ERP paradigms:

    Topics assigned

    StudentTopicSupervisor
    Christoph Schulthess 1+4 Martin Glasstetter (TB), Laura Kahle(TB)  
    Marta Timón 2+5 Martin Glasstetter (TB), Joos Behnke (TB)  
    Johannes Meyer 3 Lukas Fiederer (TB)
    Michaela Freitag 6 Martin Völker (TB)
    Nils Schwabe 7 Michael Tangermann (MT)  
    Anatolii Domashnev 8+9 Michael Tangermann (MT), Daniel Kuhner (WB)  
    Anton Böhm 10 Andreas Meinel (MT)
    Torsten Koller 11 Atieh Bamdadian (MT)
    Branka Mirchevska 12 Sebastian Castaño (MT)
    Achim Leydecker 13 Camilo Gordillo (WB)
    Muneeb Shahid 14 Barbara Frank (WB)  
    Florian Krämer 15 Daniel Kuhner (WB), Felix Burget (WB)
    Toshika Srivastava 16 Michael Tangermann (MT), Barbara Frank (WB)

     

    Available Slides

    Filed under: