Abraham et. al., 2016 |
fMRI |
Autism |
ROI signal correlations |
- |
Linear regression, SVM |
L1, L2 |
Children / Adults (6-64) |
871 (ABIDE) |
Learned, ICA, Atlas, k-means, Ward |
Varying |
ACC=68 |
10 folds, Leave one site out |
Alnæ s et. al., 2015 |
fMRI |
rest / low / high functional task loads |
ROI signal partial + full correlations from GLASSO |
- |
rLDA |
Covariance matrix shrinkage |
Adults (29+-10) |
101 (37 per class) |
Learned (dual-regression) |
46 |
ACC=100 / 70 / 70 |
LOSGO |
Anderson et. al., 2011 |
fMRI |
Autism |
ROI signal correlations |
t-test |
Age-based per-class linear model fit test |
- |
Adolecents (8-42) |
80 (40+, 40-) |
Voxel selection (Min. 5 mm radius) |
7266 |
ACC=79, SN=83, SP=75 |
LOO |
Anderson et. al., 2014 |
fMRI, Cortical Thickness |
ADHD |
Local network measures + cortical measurements + metadata |
NMF |
Decision tree |
- |
Adolescents ~(8-19) |
730 (276+, 472-) |
Atlas |
90 |
ACC=67, SN=76, SP=61 |
LOO |
Arbabshirani et. al., 2013 |
fMRI |
Schizophrenia |
ROI signal correlations |
Remove features effected by medication, various kernels |
Many (KNN, decision trees, RBF neural network, kernel SVM: all tied) |
- |
Adults (25-50) |
56 (28+, 28-) |
Learned (ICA) |
9 |
ACC=96, SN=100, SP=92, PPV=92, NPV=100 |
LOO |
Arslan et. al., 2015 |
fMRI |
Healthy Adults |
Locally connected cortical ROI signal correlations |
- |
Multi-graph spectral decomposition |
- |
Adults (22-35) |
40 |
Learned |
50, 100, 150, 200 |
DC=0.72 |
leave 20 (50%) out, 20 rounds |
Ashikh et. al., 2015 |
fMRI |
NC / eMCI / MCI / AD, |
Temporal variance of ROI signal correlations |
t-test |
OPTICS |
- |
Seniors (ADNI) |
123 (35, 34, 34, 29) |
Not specified |
200 |
ACC=100 |
- |
Ball et. al., 2016 |
fMRI |
Preterm / term, age |
ROI signal covariances |
Boruta feature selection |
SVM + RBF, Linear regression |
- |
Preterm (36-49 wks) |
131 (105+,26-) |
Learned (ICA) |
71 |
BAC=80, AUC=92, $r^2$=0.57, MSE=8.9 |
5 folds |
Bassett et. al., 2012 |
fMRI |
Schizophrenia |
Largest connected component size curve over thresholds |
- |
SVM |
L2 |
Adults (30-52) |
58 (29+, 29-) |
AAL |
90 |
ACC=75, SN=85, SP=64 |
1000 rounds, leave 28 (14+, 14-) out |
Brown and Kawahara et. al., 2016 |
dMRI |
Motor, Cognitive Function, Age |
Edge tract counts |
- |
BrainNetCNN |
L2 |
Preterm |
168 |
Atlas |
90 |
r=0.31, 0.19, 0.86, MAE=10.6, 10.5, 2.3 |
3 folds |
Brown et. al., 2015 |
dMRI |
Low Motor Function |
Network Measures |
PCA (After net. measures) |
SVM |
L2 |
Preterm |
168 (22+, 146-) |
Atlas |
90 |
ACC=72, SN=77, SP=69 |
1000 rounds, LOSGO |
Brown et. al., 2016 |
dMRI |
Low Motor, Cognitive Function |
Edge tract counts |
- |
Non-negative linear regression |
L1, backbone prior, connectivity prior |
Preterm |
168 (29+, 139-), 168 (13+, 155-) |
Atlas |
90 |
ACC=71, 60, AOC=14.3, 17.4, r=0.44, 0.34 |
1000 rounds, LOSGO |
Caeyenberghs et. al., 2013 |
dMRI + fMRI |
TBI |
structural and functional nodal degrees |
- |
Discriminant function analysis |
- |
Adults (25+/-6) |
33 (17+, 16-) |
Motor Regions |
22 |
ACC=61, SN=43, SP=77 |
Not stated |
Chen et. al., 2011 |
fMRI |
AD |
ROI signal correlations |
Wilcoxon rank-sum test |
LDA |
- |
Seniors (69-82) |
55 (20NC, 15MCI, 20AD) |
AAL |
116 |
AUC=0.87, SN=85, SP=80 |
LOO |
Chen et. al., 2015 |
dMRI + fMRI |
TBI |
Edge FA, ROI signal correlations |
- |
Multi-view Clustering |
- |
Not specified |
40 (16+, 24-) |
DICCCOL |
358 |
- |
- |
Chen et. al., 2014 |
dMRI |
Healthy Adults |
Edge tract counts (binary) |
- |
Training group-wise consistent connections |
- |
Not specified |
120 |
DICCCOL, learned |
358, 272 |
- |
- |
Cheng et. al., 2012 |
fMRI |
ADHD |
ROI signal partial + full correlations |
t-test |
SVM |
L2 |
Adolescents ~(8-19) |
730 (276+, 472-) |
Atlas |
90 |
ACC=76, SN=63, SP=85 |
LOO |
Chung et. al., 2016 |
dMRI |
Low Motor Function |
Heat flow features |
Flow averaged between 6 partitions |
Gaussian naive Bayes |
- |
Infants (24 months) |
290 (55+, 235-) |
AAL |
116 |
ACC=82, SN=75, SP=83, FS=79 |
10 folds, 5 reps |
Cradock et. al., 2009 |
fMRI |
MDD |
ROI signal correlations + r-to-z |
Reliability filter |
SVM |
L2 |
Adults (22-54) |
40 (20+, 20-) |
AFNI + Relevance to MDD |
15 |
ACC=95, SN=92, SP=96, AUC=96 |
LOO |
Cuingnet et. al., 2011 |
dMRI |
Stroke |
Nodal features |
- |
SVM |
Exp. of Laplacian (over nodes) |
Adults (24 - 81) |
72 |
Per-voxel |
96 x 64 x 24 |
ACC=76 |
Not stated |
Dodero et. al., 2015 |
fMRI, dMRI, fMRI |
ASD, ASD, Schizophrenia |
Graph Laplacians |
- |
Kernel SVM |
L2 |
Adolescents, Adults (9-18, 9-18, 33+/-9) |
79 (42+, 37-), 94 (51+, 43-), 27 (12+, 15-) |
Atlas |
264, 264, 74 |
ACC=61, 74, 68 |
LOO |
Dosenbach et. al., 2011 |
fMRI |
Age groups |
ROI signal correlations |
t-test |
SVM |
L2 |
Children / Adults (7-30) |
122 (61+, 61-) |
Learned |
210 |
ACC=91, SN=90, SP=92 |
LOO |
Fei et. al., 2014 |
fMRI |
MCI |
gSpan frequent subnetworks from different thresholds |
DSM + WL kernel |
Graph-kernel-based |
- |
Seniors (74 +/- 8) |
37 (12+, 12-) |
AAL |
116 |
ACC=97, AUC=96 |
LOO |
Finn et. al., 2015 |
fMRI |
Healthy Adults |
ROI signal correlations |
None, various anatomical subnetworks |
Max correlation between feature vectors |
- |
Adults |
504 (126 restA, 126 restB, 126 taskA, 126 taskB) |
Atlas |
268 |
ACC=99 |
- |
Galvis et. al., 2016 |
dMRI |
Parkinson's Disease |
ROI signal correlations (no EPI correct, 3DMI, PDEC) |
ICC + t-test |
SVM |
L2 |
Seniors (51-71) |
189 (131+, 58-) |
FreeSurfer |
68 |
BAC=60 |
10 folds, 10 reps |
Gao et. al., 2015 |
dMRI |
NC / SMC / MCI / AD |
Edge tract counts normalized by ROI volumes |
- |
Multi-graph normalized-cut |
Graph Laplacian (over edges) |
Seniors (ADNI) |
154 (30 NC, 34 SMC, 62 MCI, 28 AD) |
FreeSurfer |
129 |
- |
- |
Gellerup et. al., 2016 |
fMRI |
Parkinson's Disease |
ROI signal correlations + network measures across 5 frequency bands |
mRMR |
Proximal SVM ensemble |
L2 |
Seniors (60+/-10) |
45 (24+, 21-) |
Atlas |
264 |
ACC=84, SN=73, SP=93 |
5 folds |
Ghanbari et. al., 2014 |
dMRI |
ASD, Age |
Edge probability |
Built-into NMF |
NMF + graph embeddings |
- |
Adolescents (8-18) |
83 (24+, 59-) |
Atlas (Desikan) |
79 |
- |
- |
Guo et. al., 2012 |
fMRI |
MDD |
Nodal degree, efficiency, betweeness centrality from partial correlation at multiple thresholds |
- |
Many (SVM, RBF-SVM, LDA, random forest, logistic regression) |
L2 |
Adults (17 - 54) |
76 (38+,28-) |
AAL |
90 |
ACC=79 |
100 rounds, leave 30% out |
Iidaka, 2015 |
fMRI |
ASD |
ROI signal correlations |
Min. correlation threshold |
Probabilistic Neural Networks |
- |
Children (6-19) |
640 (312+, 328-) |
AAL |
90 |
ACC=90, SN=92, SP=88, PPV=88, NPV=92, cPPV=8, cNPV=100 |
2, 10 and 50 folds, Leave one out |
Jie et. al., 2014 |
fMRI |
MCI |
Local CC from multiple thresholds |
t-test, RFE |
Multi-kernel SVM |
L2 |
Seniors (74+/-8) |
37 (12+, 25-) |
AAL |
90 |
ACC=92, BAC=94, SN=100, SP=88, AUC=94 |
LOO |
Jie et. al., 2013 |
fMRI |
MCI |
Local CC from thresholded ROI signal correlations |
t-test + RFE + Linear and WL Kernel |
Multi-kernel SVM |
L2 |
Seniors (65-83) |
37 (12+, 25-) |
AAL |
116 |
ACC=92 |
LOO |
Jie et. al., 2014 |
fMRI |
MCI |
CC on Hypergraph |
M2TFS + linear kernel |
Multi-kernel SVM |
L2 |
Seniors (65-83) |
37 (12+, 25-) |
AAL |
116 |
ACC=95, SN=92, SP=96, AUC=96 |
LOO |
Jin et. al., 2016 |
dMRI |
ASD |
Edge FA, MD and TC, Multiple Scales |
t-test, LASSO logistic regression |
Multi-kernel SVM |
L2 |
Infants |
80 (40+, 40-) |
Atlas |
90, 203, 403 |
ACC=76, SN=72, SP=79, AUC=0.8 |
5 folds, 10 reps |
Kamiya et. al., 2016 |
dMRI |
TLE |
Local network measures (4x83=332) |
Dantzig selector + RBF Kernel |
Kernel SVM |
L2 |
Adults (21-45) |
58 (44+, 14-) |
FreeSurfer |
83 |
ACC=90, AUC=97 |
LOO |
Kaufmann et. al., 2016 |
fMRI |
Sleep Deprevation |
ROI signal partial correlations from GLASSO |
- |
rLDA |
Covariance matrix shrinkage |
Adults (20-24) |
60 (41+, 19-) |
Learned (MELODIC ICA) |
27 |
ACC=73 / 54 / 85 |
LOO |
Khazaee et. al., 2015 |
fMRI |
AD |
Local + global network measures (454) |
Fischer score |
SVM |
L2 |
Seniors (ADNI) |
40 (20+, 20-) |
AAL |
90 |
ACC=100 |
LOO |
Khazaee et. al., 2015 |
fMRI |
NC / MCI / AD |
Local + global network (2909) on global cost efficiency max. network |
Fischer score + FSFS |
SVM |
L2 |
Seniors (ADNI ~ 64-83) |
168 (41NC, 89MCI, 34AD) |
Atlas |
264 |
ACC=88, per-class SN, SP, PPV |
Cross-validation with holdout |
Li et. al., 2012 |
dMRI |
ASD |
Edge connectivity |
RFE |
SVM |
Graph Laplacian (over edges) |
Children (7-14) |
20 (10+, 10-) |
Atlas |
68 |
ACC=100 |
LOO |
Li et. al., 2014 |
fMRI + dMRI for landmarks |
PTSD |
Temporal sliding window nodal degree |
Manual selection of quasi-static states |
K-means + FDDL, finite state machines |
- |
Adults |
95 (45+, 51-) |
DICCCOL |
358 |
Avg. diff. of subnet histograms |
10 folds |
Liem et. al., 2016 |
fMRI |
Age |
ROI signal correlations |
- |
SVR |
L2 |
Adult (19-82) |
2354 |
Atlas (BASC) |
197, 444 |
MAE=4.29, R2=0.87 |
5 folds |
Mitra et. al., 2016 |
dMRI |
TBI |
Edge FA |
NBS Edge t-test + PCA |
Random Forest |
- |
Adults |
215 |
AAL |
116 |
ACC=68, SN=80, SP=46, PPV=68, NPV=69 |
10 folds |
Mokhtari et. al., 2012 |
fMRI |
Rest/Attention |
ROI signal correlations |
Recursive feature ranking + WL Kernel |
Kernel SVM |
L2 |
Adults (20-30) |
38 (19+, 19-) |
Atlas |
24 |
ACC=100 |
LOO |
Moyer et. al., 2015 |
dMRI |
AD, Bipolar Disorder |
Edge tract counts |
- |
MMSB + (naive Bayes, SVM) |
- |
ADNI-2, ? |
96 (46+, 50-), 92 |
FreeSurfer |
68 |
ACC=82, 61, REC=80, 53, PRS=82, 69 |
10 folds |
Munsell et. al., 2015 |
dMRI |
TLE, surgical outcome |
Edge tract counts |
ElasticNet + linear kernel, or SCCA or deep model |
SVM |
L2 |
Adults (18 - 70) |
118 (70+, 48-) |
Atlas (Lusanne) |
82 |
ACC=80, SN=74, SP=84, PPV=90, NPV=70 |
10 folds |
Munsell et. al., 2016 |
dMRI + fMRI |
Autism |
Nodal betweeness of ROI signal correlations + Edge tract counts |
Clusters connecting modalities on bipartite graph between fMRI and dMRI nodes |
SVM |
L2 |
Adolescents (13-18) |
70 (35+, 35-) |
Atlas |
264 |
ACC=88, SN=89, SP=86, PPV=87, NPV=89 |
10 folds |
Ng et. al., 2014 |
fMRI |
Before/After Memory Task |
Covariance matrices |
Matrix whitening transport |
SVM |
L2 |
Not specified |
102 (51+, 51-) |
Atlas |
78 |
ACC=98 |
10000 rounds, leave 34 out |
Ng et. al., 2012 |
dMRI + fMRI |
Viewing a face vs rest |
Per - voxel time series. |
- |
SLR |
L1 + Graph Laplacian |
Not specified |
36 (18+, 18-) |
Ward clustering |
500 |
ACC=86 |
9 folds |
Pariyadath et. al., 2014 |
fMRI |
Smoking Status |
Correlations within and between subnetwork regions |
RFE |
AdaBoost + SVM |
L2 |
Adults (28-50) |
42 (21+, 21-) |
ICA + Clustering |
56 |
ACC=79, PPV=83 |
LOO |
Park et. al., 2015 |
dMRI, fMRI |
BMI |
Edge fiber density + fMRI mean nodal degrees |
t-test + permutation testing, region prior, functional-structural correlation |
PLSR |
- |
Adults (~29) |
120 (60+, 60-) |
AAL |
116 |
MAE=15%, RMS=5.3 |
LOO |
Prasad et. al., 2014 |
dMRI |
AD |
Normalized tract counts |
Region grouping, PCA |
Simulated annealing + SVM |
- |
Adults (ADNI-2) |
87 (37+, 50-) |
Freesurfer |
68 |
ACC=85, SN=88, SP=81 |
LOO |
Prasad et. al., 2015 |
dMRI |
NC / eMCI / MCI / AD, |
Normalized edge TC, max flow between ROIs, network measures |
- |
SVM |
L2 |
Seniors (64-85) |
200 (50, 74, 38, 38) |
Freesurfer |
68 |
ACC=78, SN=84, SP=69 |
10 folds |
Pruett et. al., 2015 |
fMRI |
Age (6 v.12m) and clincal risk (low, high) |
ROI signal correlations |
t-test + linear kernel |
SVM |
L2 |
Infants (6-12 months) |
128 (32 per group) |
Atlas |
230 |
ACC=81, 75, SN=78, 81, SP=84, 69 |
LOO |
Qiu et. al., 2015 |
fMRI |
Age |
ROI signal partial correlations from GLASSO |
LLE-SPD |
Linear regression |
L2 |
Adults (22-79) |
178 |
Atlas |
80 |
r=0.59, RMS=12.9 |
LOO |
Richiardi et. al., 2011 |
fMRI |
Resting vs watching movie |
Multi-band time series |
t-test + FDR |
Polythetic trees |
Tree pruning + ensembles |
Adults (18-36) |
15 |
Altas |
90 |
ACC=97 |
LOO |
Richiardi et. al., 2012 |
fMRI |
MS |
ROI signal correlations |
- |
Ensemble of functional trees |
- |
Adults (29-45) |
36 (22+, 14-) |
AAL |
90 |
ACC=83, SN=82, SP=86 |
LOO |
Robinson et. al., 2010 |
dMRI |
Age (young adults / seniors) |
Edge FA |
PCA |
MLDA |
- |
Adults (20-30, 59-90) |
96 |
Atlas |
83 |
ACC=87, SN=90, SP=88, Bayes err.=0.87 |
10 folds |
Rosa et. al., 2013 |
fMRI |
MDD |
ROI signal sparse inverse covariance |
- |
SVM |
L1 |
Adults (29-58) |
38 (19+, 19-) |
Atlas |
137 |
ACC=82, SN=74, SP=89 |
LOSGO |
Rosa et. al., 2015 |
fMRI |
MDD, Depression Spectrum |
ROI signal sparse inverse covariance |
- |
SVM |
L1 |
Adults (29-58), (27-49) |
38 (19+, 19-), 60 (30+, 30-) |
Atlas |
137 |
ACC=85, SN=83, SP=87, sparsity=0.6%, stability=57 |
LOSGO |
Sacchet et. al., 2015 |
dMRI |
MDD |
Global Network Measures (9) |
- |
SVM |
L2 |
Adults (18-55) |
32 (14+, 18-) |
Atlas |
68 |
ACC=72, SN=71, SP=72 |
LOO |
Shahnazian et. al., 2012 |
fMRI |
Rest / Attention |
Bi-variate Granger causality network |
Random walk kernel (tried other kernels) |
Kernel SVM |
L2 |
Adults (20-30) |
38 (19+, 19-) |
Atlas |
24 |
ACC=100 |
LOO |
Shen et. al., 2010 |
fMRI |
Schizophrenia |
ROI signal correlations |
KTRCC + LLE |
C-Means |
- |
Adults (19-30) |
52 (32+, 20-) |
AAL |
116 |
ACC=92, AUC=96 |
LOO |
Smyser et. al., 2016 |
fMRI |
Age |
ROI signal correlations + r-to-z |
t-test + linear kernel |
SVM |
L2 |
Preterm Infants (36 - 41 weeks) |
100 (50+, 50-) |
Spheres in Talairach atlas space |
214 |
ACC=84, SN=90, SP=78 |
LOO |
Sweet et. al., 2013 |
fMRI, EEG |
Epileptic Regions |
ROI signal correlations |
- |
Baysean model for abnormal regions given edges |
- |
Not specified |
44 (6+, 38-) |
Surface subdivision (50-100mm patches) |
1153 |
- |
- |
Takerkart et. al., 2012 |
fMRI |
Differing Auditory Stimuli |
ROI time series, ROI barycenters |
Functional, geometrical, structural kernels |
Kernel SVM |
L2 |
Not specified |
45 (9 per class) |
Clustering (Ward's) |
5-30 |
ACC=45 |
LOO |
Tunc et. al., 2016 |
dMRI |
Sex |
Mean anatomical subnetwork inter-connectivity, cognitive + motor scores |
- |
SVM |
L2 |
Adolescents (15+/-3.5) |
900 |
Atlas (Desikan) |
95 |
ACC=79, 64 |
10 folds |
Vanderweyen et. al., 2015 |
fMRI |
TBI and AD |
ROI signal partial correlations |
LASSO |
SVM |
L2 |
Seniors (~70) |
69 (40NC, 15AD, 14TBI) |
Atlas |
264 |
ACC=82, SN=40, SP=98, PPV=86, NPV=81 |
LOO |
Varoquaux et. al., 2010 |
fMRI |
Stroke |
ROI signal correlations |
- |
Gaussian model over manifold of (control) covariance matrices |
- |
Not specified |
30 (10+, 20-) |
Seeded Regions |
33 |
Class log likelihood |
LOO |
Vegas-Pons et. al., 2014 |
fMRI |
Differing Auditory Stimuli |
Thresholded ROI signal correlations |
WL kernel |
Kernel SVM |
- |
Not specified |
38 |
Clustering |
? |
ACC=74 |
LOO |
Wang et. al., 2014 |
fMRI |
MCI |
Local CC from thresholded ROI signal correlations |
t-test + RFE + gSpan + Linear, WL Kernels |
Multi-kernel SVM |
L2 |
Seniors (65-83) |
37 (12+, 25-) |
AAL |
116 |
ACC=97, AUC=92 |
LOO |
Watanabe et. al., 2016 |
dMRI |
TBI |
Edge probability |
1000 features with top rank-sum correlations with clinical scores |
NMF |
Class discriminative subnetworks |
Adults |
111 (79+, 32-) |
Atlas (Desikan) |
86 |
ACC=82, SN=82, SP=81, BAC=82 |
LOSGO |
Wee et. al., 2016 |
fMRI |
ASD |
ROI signal correlations from temporal clusters |
LASSO + various kernels |
Kernel SVM |
L2 |
Children, Young Adults (4 - 22) |
92 (45+, 47-) |
AAL |
116 |
ACC=71, SN=80, SP=61, PPV=79, NPV=65 |
10 folds |
Wee et. al., 2010 |
dMRI |
MCI |
Local CC (for 6 connectome types) |
Ranked correlation + RFE |
SVM |
L2 |
Seniors (74+/-8.6) |
27 (10+,17-) |
AAL |
90 |
ACC=89, AUC=93 |
LOO |
Wee et. al., 2013 |
fMRI |
MCI |
Temporal sliding window region activations |
Fused multiple graphical LASSO + t-test |
SVM |
L2 |
ADNI-2 (68-80) |
59 (29+, 30-) |
AAL |
116 |
ACC=90, BAC=79, SN=76, SP=83, AUC=79 |
LOO |
Wee et. al., 2013 |
fMRI |
AD, MCI |
Granger causality networks |
t-test on MAR model order distribution + RFE |
SVM |
L2 |
Seniors (65-83) |
37 (12+, 25-) |
AAL |
116 |
ACC=92, AUC=90 |
LOO |
Wee et. al., 2012 |
dMRI + fMRI |
MCI |
Structural and functional nodal CC |
t-test, Linear + Polynomial + RBF kernels |
Multi-kernel SVM |
L2 |
Seniors (64-83) |
27 (10+, 17-) |
AAL |
90 |
ACC=96, SN=100, SP=94, AUC=95, YDI=94, BAC=95, FS=97 |
LOO |
Yang et. al., 2016 |
fMRI |
MCI |
CC on fused randomly thresholded networks |
LASSO |
SVM |
L2 |
Seniors (68-80) |
59 (29+, 30-) |
AAL |
116 |
ACC=83, SN=86, SP=80, AUC=81 |
LOO |
Yoldemir et. al., 2015 |
dMRI + fMRI |
7 Functional Tasks |
fMRI time series, weighted nodal degree |
- |
CSORD + SVM |
Non-negative weights, fixed weight magnitude |
Adults |
40 |
Clustering |
200 |
ACC=79, ICC=0.43, DC=0.63 |
100 rounds, leave 5 (12.5%) out |
Yoldemir et. al., 2013 |
fMRI |
Healthy Adults |
fMRI time series |
- |
SORD |
Non-negative weights, fixed weight magnitude |
Adults |
40 |
Clustering |
200 |
ICC of net. measures, OmI=0.72 |
Test-retest |
Zhan et. al., 2015 |
dMRI + MRI |
NC / MCI / AD |
Edge tract counts |
High order SVD |
SLR |
L1 |
ADNI-2 |
202 |
Atlas |
113 |
ACC=71, SN=68, SP=72, AUC=76 |
20 rounds, leave 15% out |
Zhu et. al., 2014 |
fMRI, dMRI for landmarks |
MCI, Schizophrenia |
ROI signal correlations |
Edge t-test + CFS |
SVM |
L2 |
Adults |
28 (10+,18-), 20 (10+,10-) |
DICCCOL |
358 |
ACC=96, 100 |
LOO |
Zhu et. al., 2014 |
fMRI + dMRI for landmarks |
MCI |
ROI signal correlations |
CFS |
SVM |
L2 |
Seniors (55-84), (66-84) |
28 (10+, 18-), 24 (10+, 10-) |
DICCCOL |
358 |
ACC=100, 96 |
LOO |
Ziv et. al., 2013 |
dMRI |
Neonatal Encephalopathy |
Counts of binary subgraphs |
RFE, PCA |
SVM |
L2 |
Neonates |
24 |
Clustering (Recursive partitioning) |
100 |
ACC=79 |
Cross-validation |