Sign inSign up
Inverse Methods for EEG and MEG Source Reconstruction thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Inverse Methods for EEG and MEG Source Reconstruction

Published on Dec 03, 20129235 Views

Guido Nolte

,

Stefan Haufe

In this lecture we review the most popular inverse methods for EEG and MEG source reconstruction. Inverse methods can be divided into three different catagories: a) overdetermined models, b) underdete

BBCI Summer School

Related categories

Machine LearningNeuroscienceNeurotechnology

Chapter list

Tutorial on EEG/MEG inverse source reconstruction00:00
EEG and MEG (1)00:37
EEG and MEG (2)01:41
Volume conduction02:29
Interpretability03:35
Characteristics of the EEG and MEG generation (1)04:59
Characteristics of the EEG and MEG generation (2)05:51
Characteristics of the EEG and MEG generation (3)06:28
Generative model of the EEG06:54
Forward modeling08:24
The Inverse Problem (1)09:13
The Inverse Problem (2)09:22
Source Reconstruction Paradigms09:48
Inverse methods10:40
Dipole modeling (1)10:58
Dipole modeling (2)11:00
Dipole Modeling, high noise11:53
Dipole Modeling, brain noise12:17
What is minimized? (1)12:59
What is minimized? (2)13:52
The Problem of Local Minima (1)14:37
The Problem of Local Minima (2)15:11
Including time16:23
Example: Event-related Potentials (ERP)17:37
Distributed Inverse Imaging (1)18:30
Distributed Inverse Imaging (2)18:33
Cost Function19:40
Constraints21:16
Spatial smoothness23:07
Origin of blurring24:46
Spatial sparsity25:29
Origin of sparsity26:48
No sparsity using L2-norm28:20
Limitations of smooth (linear) and sparse inverses28:43
Alternative constraints (1)30:29
Alternative constraints (2)31:34
Real-world Example32:00
Depth compensation (1)33:14
More „physiological“ constraints33:27
Summary (1. part)34:57
Part II35:48
Beamformers36:23
Dipole amplitude? (1)36:25
Dipole amplitude? (2)41:43
We measure source + noise How do we know the noise level?43:30
Nulling beamformer47:55
SAM and LCMV beamformer51:32
DICS=LCMV beamformer in frequency domain54:10
EEG-simulation of ERD (1 source) -154:54
EEG-simulation of ERD (1 source) -256:00
EEG-simulation of ERD (1 source) -358:03
MUSIC (Multiple Signal Classification)58:45
1. Find important patterns in data: PCA of covariance matrix (1)01:00:45
1. Find important patterns in data: PCA of covariance matrix (2)01:02:04
1. Find important patterns in data: PCA of covariance matrix (3)01:02:19
2. Does a combination of eigenvectors look like a dipole at a some location? (1)01:02:49
2. Does a combination of eigenvectors look like a dipole at a some location? (2)01:04:16
Scan: one slice01:04:59
Scan: whole brain01:05:27
Illustration for P=201:05:55
Truth - Music01:07:15
RAP-MUSIC01:08:20
Source explains data01:10:22
The End01:12:21