IEEE Systems, Man and Cybernetics Magazine - July 2020 - 33

with weights W, using St i = W (W T W) -1 W T S ivec, where
St i denotes the geodesic filtered SPD matrices from
S i . These filtered matrices St i are then projected back
onto the manifold using

specific narrow frequency band. Since the alpha rhythm
(8-12 Hz) is known to vary according to both workload [4]
and emotions [28], we applied CSP after bandpass filtering
in 8-12 Hz.
FBCSPs
The FBCSP is an algorithm that optimizes both spatial and
spectral filters. To do so, it first filters EEG signals into
multiple frequency bands using a filter bank. Here, we
used nine bandpass filters in 4-Hz-wide bands (in 4-8 Hz,
8-12 Hz, ..., 36-40 Hz) as in [13]. Then, for each bandpassed signal, CSP optimizes two spatial filter pairs. From
the resulting 36 features (nine bands multiplied by four
CSP filters per band), the four most relevant were selected
using minimal redundancy maximal relevance (mRMR)
[36] and used as input to an LDA. The FBCSP algorithm
proved its efficiency when winning the Fifth International
BCI competition [13].

	

t i are the filtered SPD matrices projected onto
where C
the manifold and exp m ^M h denotes the exponential
of matrix M. Finally, this approach uses a minimum
distance to the mean classifier to classify testing geot i . To do so, during the traindesic filtered matrices C
ing step, the class centroids G k of each class k are
computed by averaging the geodesic filtered covarit ik from each class k. During testing,
ance matrices C
the Riemannian distances between the testing geodet j and each class centroid G k are
sic filtered matrix C
t j is assigned
first calculated by using (2). The matrix C
k
class label for which the centroid G k is the nearest.
In our study, FgMDM was applied on EEG bandpass filtered in 8-12 Hz, as for the CSP.
◆◆ TSC: TSC first projects all training SPD matrices C i
onto the tangent space at point G (the mean of all
training matrices). Then, it uses any classifier such as
LDA, SVM, or LR on the vectorized upper-triangular
elements of the projected matrices [39]. We used LR
with L2 regularization (with the default C = 1.0 in
scikit-learn [40]). For FgMDM, TSC used data filtered
in 8-12 Hz.

Riemannian Geometry
Riemannian approaches represent EEG trials as covariance matrices, which are symmetric positive definite
(SPD) matrices, and manipulate them with an appropriate
geometry, the Riemannian geometry [32], [37]. Classifiers
based on such geometry are called RGCs.
First, in a Riemannian manifold we can estimate intrinsic non-Euclidean distances between two SPD matrices,
i.e., two points (here, C 1, and C 2 ), using the Riemannian distance
	

d 2 ^C 1, C 2 h = | log 2 m n ^C -1 1 C 2 h, 	(2)
n

where m n (M) is the nth eigenvalue of matrix M. The set of
tangent vectors to point G on the manifold defines the
manifold tangent space at G. More generally, any SPD
matrix C i can be projected onto the tangent space at
point G using
	

S i = Log G ^C i h = G 1/2 log m ^G -1/2 C i G -1/2 h G 1/2, (3)

where S i is the projection of C i on the tangent plane, and
log m ^ · h denotes the logarithm of a matrix.
In this article, we study two existing RGCs, the mean
classifier (FgMDM) and the TSC, and introduce two new
ones, the FBTSC and the FBFgMDM.
Existing Methods
◆◆ FGMDM: FgMDM [38] projects training matrices C i
onto the tangent space at point G (the mean of all
training data) using (3) to obtain matrices S i . Then,
a Fisher geodesic filter is obtained by optimizing an
LDA classifier on S ivec, the vectorized upper-triangular elements of S i, to discriminate classes using
such vectors. This results in a matrix of weights
W = LDA ^S ivec h . The projected SPD matrices S i from
both the training and the testing sets are then filtered
	

t i = Exp C ^St i h = G 1/2 exp m ^ G -1/2 St i G -1/2 h G 1/2, (4)
C

New Methods
◆◆ FBFgMDM: Contrary to FgMDM, which exploits EEG
signals in a single frequency band, FBFgMDM applies
FgMDM in multiple bands separately and combines the
resulting distances to exploit additional spectral information. This approach should possibly improve classification performances, as FBCSP did to improve CSP.
To do so, FBFgMDM first filters EEG signals in multiple bands using a filter bank, as for FBCSP. Here, we
used the same bands as the FBCSP, i.e., 4-8 Hz,
8-12 Hz, ..., 36-40 Hz. Then, for the EEG signals in
each frequency band j, this method first uses a regular
FgMDM, i.e., it computes the Riemannian distances
t ij i between a geodesic filtered SPD matrix
d 2 _ G kj, C
t ij and each class centroid G kj . Then, from all nine
C
bands j, the four that are most useful for classification
are selected with mRMR feature selection [36] on the
t ij i used as features, on
Riemannian distances d 2 _ G kj, C
the training set. For testing, we compute the squared
Riemannian distances for the four bands selected
using mRMR only and sum them as
t i i = |d 2 _ G kj, C
t ij i, (5)
c 2 _ G k, C

	

j!X

where Ω is the set of frequency bands selected with
t i i to
mRMR. We thus obtain k new distances c 2 _ G k, C
Ju ly 2020

IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE	

33



IEEE Systems, Man and Cybernetics Magazine - July 2020

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - July 2020

Contents
IEEE Systems, Man and Cybernetics Magazine - July 2020 - Cover1
IEEE Systems, Man and Cybernetics Magazine - July 2020 - Cover2
IEEE Systems, Man and Cybernetics Magazine - July 2020 - Contents
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 2
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 3
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 4
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 5
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 6
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 7
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 8
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 9
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 10
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 11
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 12
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 13
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 14
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 15
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 16
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 17
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 18
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 19
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 20
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 21
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 22
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 23
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 24
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 25
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 26
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 27
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 28
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 29
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 30
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 31
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 32
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 33
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 34
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 35
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 36
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 37
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 38
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 39
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 40
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 41
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 42
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 43
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 44
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 45
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 46
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 47
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 48
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 49
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 50
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 51
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 52
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 53
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 54
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 55
IEEE Systems, Man and Cybernetics Magazine - July 2020 - 56
IEEE Systems, Man and Cybernetics Magazine - July 2020 - Cover3
IEEE Systems, Man and Cybernetics Magazine - July 2020 - Cover4
https://www.nxtbook.com/nxtbooks/ieee/smc_202110
https://www.nxtbook.com/nxtbooks/ieee/smc_202107
https://www.nxtbook.com/nxtbooks/ieee/smc_202104
https://www.nxtbook.com/nxtbooks/ieee/smc_202101
https://www.nxtbook.com/nxtbooks/ieee/smc_202010
https://www.nxtbook.com/nxtbooks/ieee/smc_202007
https://www.nxtbook.com/nxtbooks/ieee/smc_202004
https://www.nxtbook.com/nxtbooks/ieee/smc_202001
https://www.nxtbook.com/nxtbooks/ieee/smc_201910
https://www.nxtbook.com/nxtbooks/ieee/smc_201907
https://www.nxtbook.com/nxtbooks/ieee/smc_201904
https://www.nxtbook.com/nxtbooks/ieee/smc_201901
https://www.nxtbook.com/nxtbooks/ieee/smc_201810
https://www.nxtbook.com/nxtbooks/ieee/smc_201807
https://www.nxtbook.com/nxtbooks/ieee/smc_201804
https://www.nxtbook.com/nxtbooks/ieee/smc_201801
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_1017
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0717
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0417
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0117
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_1016
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0716
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0416
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0116
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_1015
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0715
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0415
https://www.nxtbook.com/nxtbooks/ieee/systems_man_cybernetics_0115
https://www.nxtbookmedia.com