Related work_BCI
 Share
The version of the browser you are using is no longer supported. Please upgrade to a supported browser.Dismiss

 
%
123
 
 
 
 
 
 
 
 
 
ABCDEFGHIJKLMNOPQRSTUVWXYZ
1
2
CategoryStudyDecoding
Problem
DatasetInput
Representation
Network
Architecture
Training
Strategies
BaselineAnalysis
/Visualization
FindingsPerformance
3
BCIDeep Learning with Convolutional Neural Networks
for EEG Decoding and Visualization,
Schirrmeister et al., HBM, 2017
Motor Imagery,
Motor Execution
BCIC-IV 2a,
In-house data (HGD)
EEG as an Image1. Deep ConvNet: Temporal Conv (T.C.)
-> Spatial Conv (S.C.)
-> three T.C.
-> Linear Mapping (L.M.)
2. Shallow ConvNet:
T.C. -> S.C. -> L.M.
Sliding windowFBCSP + rLDACorrelation between input-feature and unit-output,
and input-perturbation and network-prediction
Topography of each EEG class~70% (BCIC-IV 2a)
4
BCILearning Temporal Informatin for Brain-Computer
Interface using Convolutional Neural Networks,
Sakhavi et al., IEEE-TNNLS, 2018
Motor ImageryBCIC-IV 2aEEG as an ImageFBCSP -> three conv. -> L.M.Trial wise,
Temporal feature extraction using FBCSP,
Mutual information based feature selection,
CNN
FBCSP + rLDA,
FBCSP + SVM,
BO, TSSM
Kernal morphology,
Layer-wise Relavance Propagation (LRP)
Importance of kernel shape~75% (BCIC-IV 2a)
5
BCISignal Processing Approaches to Minimize or Suppress Calibration Time in Oscillatory Activity-based Brain-Computer Interfaces, Lotte et al., Proceedings of IEEE, 2015Motor Imagery, Mental task, Mental ImageryBCIC-IV 2a, In-house dataEEG as an ImageCalibration reduction:
1. Regularization approaches
2. User-to-user transfer
3. Semi-supervised learning
4. A priori physiological information
5. Artificial data generation

Calibration suppression:
1. Pooled design
2. Ensemble design
CSP + LDACSP feature visualization using w/ A.D. or w/o A.D.A.D. works.
6
BCICascade and Parallel Convolutional Recurrent Neural Networks on EEG-based Intention Recognition for Brain-Computer Interfaces, Zhang et al., AAAI, 2018Motor ImageryBCI 2000 (Schalk, 2004), 4 class1D EEG sequence to 2D EEG meshesCascade Convolutional Recurrent NN,
Parallel Convolutional Recurrent NN
Sliding windowBasic CNN, Basic RNN98% (BCI2000)
7
BCIEEG Source Imaging Enhances the Decoding of Complex Right-Hand Motor Imagery Tasks, Edelman et al., IEEE-TBE, 2016Motor ImageryIn-house data, 4 classTime-FrequencyICA -> ROI selection -> Feature extraction -> Gaussian classifierTrial wiseROI mapping, Visualization on brain cortext
8
BCILearning Representation from EEG with Deep Recurrent-Convolutional Neural Networks, Bashivan et al., ICLR, 2016Mental TaskIn-house dataBand-pass filtering -> Spectral Topography mapsCNN feature learning -> Temporal feature aggregation (LSTM) -> classificationTrial wiseSVM, Random Froest, Logistic Regression, DBNActivation Pattern
9
BCIOptimizing Spatial Filters for Robust EEG Single-Trial Analysis, Blankertz et al., IEEE-SPM, 2007Motor ImageryEEG as an ImageCSPTrial wiseSource projection
10
BCIRegularizing Common Spatial Patterns to Improve BCI Designs: Unified Theory and New Algorithms, Lotte et al., IEEE-TBE, 2011Motor ImageryBCIC-III 4a, BCIC-III 3a, BCIC-IV 2aEEG as an ImageRCSP and its variants + LDATrial wiseBasic CSP + LDAFilter projection~80% (avg. of all dataset)
11
BCIClassification of EEG with Recurrent Neural Networks, Greaves et al., Stanford, 2014ERPIn-houseElman RNN, Time-dependent Elman RNNMLP
12
BCIDeepSleepNet: A model for Automatic Sleep Stage Scoring Based on Raw Single-Channel EEG, Supratak et al., IEEE-TNNLS, 2017EEG - sleepMASS, Sleep-EDFEEG as an Image2 CNNs with different filter sizes -> bidirectional-LSTMs
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
Loading...