Magnetenzephalogramm, MEG

1
Magnetenzephalogramm, MEG
2
Equivalent dipole
3
(No Transcript)
4
MEG
5
(No Transcript)
6
(No Transcript)
7
(No Transcript)
8
EEG vs. MEG
9
Differences between EEG and MEG

• MEG is only sensitive to the tangential
  component of the dipoles, but insensitive to the
  radial component. EEG measures both. This implies
  that MEG recordings are mainly based on activity
  in the sulci, but not in the gyri (1/3 of the
  cortex).

• In contrast to EEG, MEG is insensitive to the
  inhomogeneities of skull and scalp which result
  in field spreading. As a consequence, the ERF is
  often more focal than the ERP.

10
Recording of ERPs and ERFs

• Strength
• - high temporal resolution
• - direct measure of neural activity
• Weakness
• - measures only a part of the neural
• activity (open fields)
• - poor spatial resolution (inverse problem)

11
Mismatch negativity (MMN)entdeckt durch Näätänen
1978
12
Winkler et al. 1999EEG Mismatch Negativity
Erwerb eines finnischen Vokalkontrastes durch
Ungarn
13
Phillips et al. 2000 J Cog Neuroscience
14
Phillips et al. 2000 J Cog Neuroscience
15
Phillips et al. 2000 J Cog Neuroscience
16
Phillips et al. 2000 J Cog Neuroscience
17
The inverse problem
Aim Finding the source distribution underlying a
given scalp potential map. Problem The inverse
problem in EEG and MEG has no unique solution.
For any given potential (or magnetic field)
distribution over the scalp surface, a variety of
possible neural source distributions exists that
can produce the same surface map. The number of
possible current source distributions that
matches a given set of surface data may be large.
18
Dipole analysis

• A head model is assumed, e.g. a three-shell
  model (brain, skull, scalp), or a more realistic
  head model.

• This model allows the calculation of the scalp
  electrical potential generated at a particular
  location on the scalp by an intracerebral source
  with a particular location, orientation and
  strength.

• In a number of iterative steps the source
  parameters can be changed until the difference
  between the modelled and the recorded waveforms
  is minimized.

19
Constraining the inverse problem
Actual solutions often involve information about
neurophysiology and anatomy to reduce the
solution space.

• Sources (dipoles) may change strength, but not
  location or orientation during a specified time
  interval (spatial-temporal constraint).

• Sources are all located at the same depth (e.g.
  in the neocortex). This approach is referred to
  as spatial deconvolution, de-blurring, or
  cortical imaging. It is based on the unique
  relation between surface potentials and sources
  at a fixed depth.

20
Dipole analysis
A source dipole is defined by its location,
orientation, and strength
Sources for the Bereitschaftspotential fit with
1 and 2 stationary dipoles (extension of the left
middle finger)
21
Die Kombination von hämodynamischen und
elektromagnetischen Daten kann
Informationen über räumliche UND zeitliche
Eigenschaften von Hirnaktivierungen bieten.
22
Geschätzte Zeitfenster der Wortproduktionsprozesse
Konzeptuelle Vorbereitung von der
Bildpräsentation bis zum lexikalischen
Konzept 175 ms (Thorpe et al.,1996 Schmitt et
al., 2000) Lemmazugriff 115 ms (Levelt et
al., 1992 Roelofs, 1992 Schmitt et al.,
2001) Formenkodierung Wortformzugriff
40 ms (van Turennout et al., 1998)
Syllabifizierung 125 ms (van Turennout et
al., 1997 Wheeldon Levelt, 1995)
Phonetische Enkodierung bis Beginn der
Aussprache 145 ms


Gesamt 600 ms (Jescheniak Levelt, 1994
Levelt et al., 1998 Damian et al., 2002)
23
Picture naming MEG
Salmelin, Hari, Lounasmaa, Sams (1994), Fig. 1
24
Gemessene (links) und erwartete (rechts)
Zeitfenster beiBildbenennung
MEG data from Salmelin et al., 1994 Levelt et
al., 1998 Maess et al., 2002
Indefrey, P. and Levelt, W.J.M. (2004) Cognition