Friday, April 3, 2015

EEG electrode placement for motor imagery, Y02-M08-D03

Fr, 03.04.2014
Year 2, Month 8, Day 3
(1 tahun, 7 bulan dan 3 hari telah berlalu)

======= = =======

EEG electrode placement for motor imagery.

Emotiv EPOC:
- left:  AF3, F3, F7, FC5, P3 (CMS), P7, T7, O1
- right: AF4, F4, F8, FC6, P4 (DRL), P8, T8, O2
reference & ground: CMS, DRL
Sorted Channels (1-14): AF3;F7;F3;FC5;T7;P7;O1;O2;P8;T8;FC6;F4;F8;AF4

Easy Cap, in lab.
- left:  F3, C3, P3, O1, TP9
- right: F4, C4, P4, O2, TP10
- middle: FPz, AFz, Fz, FCz, Cz, Pz
The configuration of 16 electrodes is from Markus Kroesen's Bachelor Thesis, 2013, University of Oldenburg
"Vergleich verschiedener EEG-Verstärker zur Steuerung von Brain-Computer Interfaces"

Question: Why are they different?
Answer: They are different because the Emotiv is modified into Easymotiv and Markus Kroesen was developing P300-based BCI speller.
Read below.

Question: Are there different effect in motor imagery using Emotiv system? Do experiment!
Answer: I am considering other hardware than Emotiv, i.e. mBrainTrain, because since August 2014, the prices are not so different.

Easymotiv = Easy Cap and modified Emotiv


Well, it is time consider mBrainTrain, instead of Easymotiv


Similarity between both caps:
- left:  F3, P3, O1
- right: F4, P4, O2

Question: Are the references still P3 and P4?
Answer: Not necessarily. The ground and reference can be put anywhere with Easymotiv (Easy Cap + Emotiv)

- AF3, AFz, AF4 are aligned
- F7, F3, Fz, F4, F8 are aligned
- FC5, FC1, FCz, FC2, FC6 are aligned
- T7, Cz, T8 are aligned
- P7, P3, Pz, P4, P8 are aligned
- from previous experience, AFz was ground (GND) and Cz was reference

======= = =======

Diana Valbuena 2010 motor imagery cap:
left:  FC3, C5, C3, C1, CP3
right: Fc4, C6, C4, C2, CP4
middle: AFz, Cz
reference: right ear lobe / or left?
ground (GND): AFz
Laplacian spatial filter, with C3 and C4 are central, others are surrounding

Question: Right ear or left? Diana Valbuena paper is not consistent.
Answer: inconsistency is solved by Laplacian spatial filter. This filter make the signals reference-independent.

Question: Is Cz is also ground (GND)?
Answer: Cz can associate with motoric movement. It is unlikely to set Cz as reference or ground.

Similarity between Diana and Easy Cap:
- left:  C3
- right: C4
- middle: AFz, Cz

1. Diana Valbuena, 2010, "Selection of Reactive Frequency Bands during Motor Imagery". BRAINROBOT, Ch. 9, pp 109. Shaker Verlag.
- inconsistent reference placement but Laplacian spatial filter solve it.
- good idea of signal processing of ERD/ERS classification
- read other papers from citation to get idea of motor imagery, signal processing and electrode placement

2. G. Pfurtscheller, C. Brunner, A. Schlögl, F.L. da Silva, 2006, "Mu rhythm (de)synchronization and EEG sigle-trial classification of different motor imagery tasks". NeuroImage, vol. 31, pp. 153-159.
- Diana Valbuena got idea of EEG electrode placement from this paper

3. B. Hjorth, 1975, "An online transformation of EEG scalp potentials into orthogonal source derivations". Electroencephalography and Clinical Neurophysiology, vol. 39, pp. 526-530.
- Diana Valbuena got idea of Laplacian spatial filter from this paper
- this paper is from the pre-internet era

======= = =======

Irina Schierholz 2013 motor imagery cap:
- left:        FC1, C3, CP1, CP3, CP5, P3
- right:       FC2, C4, CP2, CP4, CP6, P4
- middle: AFz, FCz, Cz,                Pz
reference: FCz
ground (GND): AFz
Laplacian spatial filter, with C3 and C4 are central, others are surrounding

Similarity between Diana's and Irina's cap:
- left:        C3, CP3
- right:       C4, CP4
- middle: AFz, Cz,
- ground (GND): AFz

1. Irina Schierholz, 2013, "Feasibility of humanoid robot feedback for motor imagery training". Master Thesis, University of Oldenburg

2. Catharina Zich, 2013, "Wireless EEG with individual electrode positions facilitates motor imagery learning". Master Thesis, University of Oldenburg

3. S. Debener, F. Minow, R. Emkes, K. Gandras, M. de Vos, 2012, "How about taking a low-cost, small, and wireless EEG for a walk?". Psychophysiology, 49(11), 1449-1453
- about Easymotiv (Easy Cap + modified Emotiv)

4. Catharina Zich, Maarten de Vos, Cornelia Kranczioch, Stefan Debener, 2014, "Wireless EEG with individualized channel layout enables efficient motor imagery training", Clinical Neurophysiology,

======= = =======


mBrainTrain Smarting Easy cap: RBE-24
- left:  FP1, F7, FC1, C3, T7, CP1, CP5, TP9,  P3, O1
- right: FP2, F8, FC2, C4, T8, CP2, CP6, TP10, P4, O2
- middle: AFz (DRL/GND), Fz, FCz (CMS/REF), Cz, CPz, Pz
reference: FCz  (CMS)
ground: AFz  (DRL)
Sorted channels (1-24): FP1;FP2;Fz;F7;F8;FC1;FC2;Cz;C3;C4;T7;T8;CPz;CP1;CP2;CP5;CP6;TP9;TP10;Pz;P3;P4;O1;O2
RBE = Real-Behavior-EEG

- FC3, FC4, CP3, and CP4 are not there
- Laplacian structure (like Valbuena's) for C3 and C4 cannot be used. So, do think about Common Spatial Pattern (CSP) to automate spatial filter.
- Bipolar structure to think about: C3-FC1 and C4-FC2.


About Common Spatial Pattern (CSP):

CSP tutorial:


======= = =======

Bipolar electrodes for motor imagery, from Doyle, et al, 2005
- left:  FC3-C3  (from FC3 to C3)
- right: FC4-C4  (from FC4 to C4)

Louise Doyle, Kielan Yarrow, Peter Brown, 2005, "Lateralization of event-related beta desynchronization in the EEG during pre-cued reaction time tasks", Clinical Neurophysiology, Vol 116, Issue 8, pp 1879-1888.

======= = =======

EEG electrode placement summary, in August 2014


CMS and DRL for EEG electrodes:
- CMS = Common Mode Sense, active electrode, used as reference (REF)
- DRL = Driven Right Leg, passive electrode, used as ground (GND)


standard EEG 10-20 system:

My previous experiment on steady-state evoked potential (SSVEP):
Condro, 2010, "Improvement of Response Times in SSVEP-based Brain-Computer Interface". Master Thesis. University of Bremen.


SocMed notes:

======= = =======

EEG electrode placement for motor imagery.
without picture:
With picture:


See previous post about electrode position for mu wave

======= = =======

My other thought about electrode position for P300

======= = =======


======= = =======

Kembali kubaca catatanku mengenai posisi elektroda EEG. Ada sedikit hal yang bertambah, yaitu mengenai Common Spatial Pattern (CSP) untuk motor imagery. Aku juga teringat kalau catatan terbaru mengenai motor imagery, belum terangkum. Kini tulisanku tentang motor imagery kuperbaharui.

Bremen, 3 April 2014


No comments:

Post a Comment