Prognostication of chronic disorders of consciousness using brain functional networks and clinical characteristics

  1. Ming Song
  2. Yi Yang
  3. Jianghong He
  4. Zhengyi Yang
  5. Shan Yu
  6. Qiuyou Xie
  7. Xiaoyu Xia
  8. Yuanyuan Dang
  9. Qiang Zhang
  10. Xinhuai Wu
  11. Yue Cui
  12. Bing Hou
  13. Ronghao Yu
  14. Ruxiang Xu  Is a corresponding author
  15. Tianzi Jiang  Is a corresponding author
  1. Chinese Academy of Sciences, China
  2. PLA Army General Hospital, China
  3. Guangzhou General Hospital of Guangzhou Military Command, China
  4. University of Electronic Science and Technology of China, China
  5. University of Queensland, Australia
33 figures, 9 tables and 3 additional files


Conceptual paradigm of the study.

CRS-R: Coma Recovery Scale Revised scale; GOS: Glasgow Outcome Scale.
Data analysis pipeline.

All datasets involved in this study included resting state fMRI and clinical data. For the fMRI data in the training dataset, data analysis first encompassed preprocessing and imaging feature selection and extraction. Partial least square regression was then used to generate the regression model using the selected imaging features and clinical features in the training dataset. In this way, a prediction score that depicts the possibility of consciousness recovery was computed for each patient. The optimal cut-off value for classifying an individual patient as responsive or non-responsive was then calculated, and the prognostic classification model was obtained. The two testing datasets were only used to validate externally the regression and classification model.
Imaging features involved in the prognostic regression model.

DMN.aMPFC, anterior medial prefrontal cortex in the default mode network; DMN.PCC, posterior cingulate cortex/precuneus in the default mode network; ExecuContr.DMPFC, dorsal medial prefrontal cortex in the executive control network; Auditory.MCC, middle cingulate cortex in the auditory network; Visual.R.V1, right lateral primary visual cortex in the visual network. DMN.aMPFC—ExecuContr.DMPFC: the functional connectivity between DMN.aMPFC and ExecuContr.DMPFC; Auditory.MCC—Visual.R.V1: the functional connectivity between Auditory.MCC and Visual.R.V1.
Prognostic regression model.

In the three subplots, each color denotes a particular predictor. (A) Regression formula. (B) Predictor importance for each predictor in prognostic regression model. The vertical axis represents the sMC F-test value. The larger the sMC F-value, the more informative the predictor with respect to the regression model. (C) The imaging features in the model are rendered on a 3D surface plot template in medial view.
The performance of the prediction model on the training dataset.

(A) Individual predicted scores for each DOC patient in the training dataset. The CRS-R score at the T0 time point is shown on the x axis and the predicted score on the y axis. The patients diagnosed as VS/UWS at the T0 time point are shown to the left of the vertical red solid line, whereas the patients diagnosed as MCS at this time point are shown to the right. The purplish red pentagram, imperial purple triangle and blank circle mark the patients with a GOS score ≥4,=3 and≤2, respectively, at the T1 time point. (B) Agreement between the CRS-R scores at the T1 time point and the predicted scores. The left panel shows the correlation between the CRS-R scores at the T1 time point and the predicted scores, and the right panel shows the differences between them using the Bland-Altman plot. (C) Bar chart showing the numbers or proportions of DOC patients in each band of predicted scores. In these two panels, the y axis shows the predicted score. (D) The area under the receiver-operating characteristic (ROC) curve. The star on the curve represents the point with the maximal sum of true positive and false negative rates on the ROC curve, which were chosen as the cut-off threshold for classification. Here, the corresponding predicted score = 13.9.
The performance of the prediction model on the two testing datasets.

(A) The individual predicted score (top panel) and agreement between the CRS-R scores at the T1 time point and the predicted scores (bottom panel) for the testing dataset ‘Beijing HDxt’. (B) The individual predicted score for each DOC patient in the testing dataset ‘Guangzhou HDxt’. The legend description is the same as for Figure 5.
The sensitivity and specificity in the ‘subacute’ patients (i.e. duration of unconsciousness T0 ≤3 months) and those in the chronic phase (i.e. duration of unconsciousness T0 >3 months), respectively.
Appendix 3—figure 1
The six brain functional network templates in this study.
Appendix 4—figure 1
Cumulative distribution of head motion per volume (framewise displacement) for normal controls and DOC patients separately in the training dataset ‘Beijing 750’ (A1), the testing dataset ‘Beijing HDxt’ (A2), and the testing dataset ‘Guangzhou HDxt’ (A3).

The normal controls were shown in left column, whereas the DOC patients were shown in right column. No healthy control data were available for the Guangzhou centre. In both patients and controls, head position was stable to within 1.5 mm for the vast majority (>95%) of brain volumes.
Appendix 4—figure 2
Correlations between motion artifact and neuroanatomical distance between the ROIs in this study.

Prior studies have shown that motion artifacts tend to vary with neuroanatomical distance between brain nodes. Here, we conducted quality control analyses as described in the previous study (Power et al., 2015). Specifically, we computed correlations between head motion (mean FD) and each resting state functional connectivity (RSFC) feature and plotted them as a function of neuroanatomical distance (mm) for subjects in the training dataset ‘Beijing 750’ (B1), the testing dataset ‘Beijing HDxt’ (B2), and the testing dataset ‘Guangzhou HDxt’ (B3). Smoothing curves (in red) were plotted using a moving average filter.
Appendix 4—figure 3
Histogram of the remaining number of fMRI volumes after scrubbing for each population, specifically ‘Beijing 750’ datatset (C1), ‘Beijing HDxt’ dataset (C2), and ‘Guangzhou HDxt’ dataset (C3).
Appendix 6—figure 1
The brain area connection features sorted by their Pearson's correlations to the CRS-R scores at the T1 time point in the training dataset ‘Beijing 750’.
Appendix 6—figure 2
The functional connectivity features sorted by their Pearson's correlations to the CRS-R scores at the T1 time point across the DOC patients in the training dataset ‘Beijing 750’.
Appendix 6—figure 3
The Circos map for the functional connectivity features that were significantly correlated to the CRS-R scores at the T1 time point across the DOC patients in the training dataset ‘Beijing 750’.
Appendix 7—figure 1
Histogram depicting the imaging features included in CARS-PLSR models.
Appendix 8—figure 1
The imaging subscores for all of the subjects in the three datasets.
Appendix 9—figure 1
The distribution of the predicted imaging subscores of the healthy controls at different sites.
Appendix 9—figure 2
The correlations between the fMRI signal-to-noise ratio (SNR) and the predicted imaging subscores in the healthy controls.


Table 1
Demographic and clinical characteristics of the patients in the three datasets.
(n = 63)
(n = 25)
(n = 24)
Gender, M/F36/2718/714/10
Age at the T0 (years)
 Mean (SD)42.8 (13.8)40.7 (15.2)39.3 (16.9)
 Range18.0 ~ 71.018.0 ~ 68.015.0 ~ 78.0
Time to MRI (months)
 Range1.0 ~ 77.01.0 ~ 44.01.0 ~ 10.0
 Mean (SD)7.4 (12.8)5.4 (8.4)2.3 (2.4)
Follow-up time (months)
 Range12.0 ~ 51.014.0 ~ 53.027.0 ~ 78.0
 Mean (SD)21.0 (9.8)41.7 (8.4)52.2 (14.5)
Diagnosis at T0
CRS-R total score
 Mean (SD)7.3 (2.9)6.5 (2.3)7.1 (4.1)
 Range3.0 ~ 18.03.0 ~ 14.03.0 ~ 17.0
Outcome at T1
CRS-R total score
 Mean (SD)9.9 (5.1)12.7 (6.4)N/A
 Range3.0 ~ 22.05.0 ~ 23.0N/A
GOS score
 GOS = 5000
 GOS = 4551
 GOS = 3875
 GOS <= 2501318
  1. Abbreviations: CRS-R, Coma Recovery Scale–Revised; GOS, Glasgow Outcome Scale; MCS, minimally conscious state; N/A, not available; SD, standard deviation; VS, vegetative state/unresponsive wakefulness syndrome.

Appendix 1—table 1
Demographic and clinical characteristics of patients in the ‘Beijing_750’ dataset.
Patient aliasGenderAge (years)DiagnoseEtiologyStructural lesions on MRITime to MRI (months)Number of CRS-R assessmentsCRS-R score at T0CRS-R subscore at T0CRS-R score at T1CRS-R subscore at T1Follow-up(months)GOSPredicted score
001M36VS/UWSAnoxiaDiffuse pons damage1670221022244632315418.26
002M29MCSTraumaBilateral-temporo-parietal damage94183551132245622339415.31
003F33VS/UWSTraumaBilateral-frontal lobe damage, atrophy12571022022245532312422.88
004F28MCSTraumaL-frontal-temporal lobe damage14153351032245622319416.58
005M23MCSAnoxiaDiffuse cortical and subcortical atrophy34102321022145522313417.08
006M45MCSStrokeL-temporo-parietal damage9492221021733422312313.94
007M39MCSStrokeBrainstem damage14173451131944512312314.39
008F27MCSTraumaL-basal ganglia damage106123321031834512319316.09
009M23MCSTraumaDiffuse cortical and subcortical atrophy6491321021944422313310.94
010M42MCSStrokeL-basal ganglia damage3771031021941632312314.72
011M53MCSStrokeDiffuse cortical and basal ganglia (caudates) damage75113321021433212314311.55
012F40VS/UWSStrokeDiffuse cortical and basal ganglia damage5671121021433312212314.67
013M22VS/UWSTraumaL-frontal-temporo-parietal lobe damage3471121021533412227315.48
014F64VS/UWSStrokeL-thalamus, basal ganglia lesions147112102112331021728.28
015F42VS/UWSAnoxiaDiffuse anoxic cortical lesions14711210292221021429.02
016M45VS/UWSAnoxiaDiffuse anoxic cortical lesions95500210271121021528.65
017F60VS/UWSAnoxiaDiffuse anoxic cortical lesions44610210261021021327.71
018M42VS/UWSStrokeR-cerebral hemisphere lesions647112102711210214212.44
019M51VS/UWSAnoxiaDiffuse cortical and subcortical atrophy34711210271121022824.28
020F35VS/UWSAnoxiaBilateral-frontal lobe damage, atrophy24711210271121021325.87
021M71VS/UWSTraumaDiffuse cortical and subcortical atrophy66310110041011011324.46
022F30VS/UWSAnoxiaBilateral-basal ganglia damage24400200270221023826.92
023F58VS/UWSTraumaDiffuse cortical and subcortical atrophy24300210040021011425.09
024M23MCSTraumaR-basal ganglia (caudates) damage5571031021122320212214.57
025F66VS/UWSTraumaBilateral-temporo-parietal damage14610210281131023225.71
026F25VS/UWSAnoxiaDiffuse cortical and subcortical atrophy34510200261120023626.75
027M48VS/UWSAnoxiaDiffuse cortical and subcortical atrophy45711210281131022927.83
028F28MCSAnoxiaDiffuse cortical and subcortical atrophy5492221021123310232211.36
029M57VS/UWSAnoxiaDiffuse cortical and subcortical atrophy114610210261021023324.70
030M61MCSStrokeBilateral-temporo-parietal lobe damage24111341021122311212210.34
031M40VS/UWSAnoxiaDiffuse cortical and subcortical atrophy44400110250111022725.70
032M39VS/UWSStrokeR-basal ganglia damage, atrophy34711210271121021228.03
033M41VS/UWSAnoxiaDiffuse cortical and subcortical atrophy24500210250021021326.44
034M26VS/UWSStrokeDiffuse cortical and subcortical atrophy544711210271121023827.28
035F50VS/UWSAnoxiaDiffuse cortical and subcortical atrophy86610210291222021225.77
036F53VS/UWSStrokeBilateral brainstem, midbrain damage34511210071121022828.02
037M67VS/UWSStrokeR- brainstem, cerebellar damage14511210030020011222.04
038M45MCSStrokeDiffuse cortical and subcortical atrophy2591321021022211213210.91
039F35VS/UWSAnoxiaDiffuse cortical and subcortical atrophy346102102811220219210.24
040F46MCSTraumaDiffuse axonal injury777112222121333221251214.76
041M49VS/UWSStrokeBilateral-brainstem, cerebellar damage1047112102711210228210.87
042M45VS/UWSStrokeDiffuse cortical and basal ganglia damage34711210281221021927.59
043M18VS/UWSAnoxiaDiffuse cortical and subcortical atrophy856111102912310212210.85
044M53VS/UWSAnoxiaBilateral-occipital lobe damage, atrophy24300200171121023421.98
045M46VS/UWSTraumaR-temporo-parietal damage44610120261012021327.23
046F29VS/UWSAnoxiaDiffuse cortical and subcortical atrophy284711210291231021228.31
047F47MCSStrokeR-basal ganglia damage4758113102112222121229.66
048M58VS/UWSStrokeBilateral-temporo-parietal lobe damage64711210281131022727.05
049M66VS/UWSAnoxiaL-frontal lobe damage44400200261021023824.79
050M34VS/UWSTraumaDiffuse axonal injury346112101812210214210.28
051F31MCSTraumaL-frontal-temporo-parietal lobe damage3511133202811220215215.56
052M33VS/UWSStrokeL-temporo-parietal lobe damage174610210281131021327.67
053F31VS/UWSAnoxiaDiffuse cortical and basal ganglia (caudates) damage14610210261021022728.36
054F28VS/UWSAnoxiaDiffuse cortical and subcortical atrophy34610210281122023229.23
055F26VS/UWSStrokeL-basal ganglia damage446102102610210212210.96
056M45VS/UWSTraumaDiffuse axonal injury14610210261021022929.05
057F69VS/UWSStrokeDiffuse cortical and subcortical atrophy446102102710220233212.43
058F68VS/UWSTraumaDiffuse axonal injury66711210291321021729.74
059M50VS/UWSStrokeL-frontal-temporo-parietal lobe damage34611110282220022727.01
060M60MCSTraumaBilateral brainstem, midbrain damage74111341021122311230211.69
061M44VS/UWSAnoxiaDiffuse cortical and subcortical atrophy246102102400210113210.48
062F35VS/UWSAnoxiaBilateral-basal ganglia damage35721110292311022729.07
063F43VS/UWSAnoxiaDiffuse cortical and subcortical atrophy247112102820211229210.09
Appendix 1—table 2
Demographic and clinical characteristics of patients in the ‘Beijing_HDxt’ dataset.
Patient aliasGenderAge (years)DiagnoseEtiologyStructural lesions on MRITime to MRI (months)Number of CRS-R assessmentsCRS-R score at T0CRS-R subscore at T0CRS-R score at T1CRS-R subscore at T1follow-up(months)GOSPredicted score
001M19VS/UWSTraumaL-temporo-parietal lobe damage6471121022245622340420.37
002M26MCSTraumaR-thalamus, basal ganglia lesions36102321022345632347417.12
003F22VS/UWSTraumaL-temporal lobe damage4461021022245622347414.05
004M41VS/UWSStrokeBilateral brainstem, midbrain damage3461121012345632350420.23
005M36MCSStrokeBilateral brainstem damage446003102234563233947.75
006M34VS/UWSAnoxiaDiffuse cortical and subcortical atrophy1461111021432312331317.25
007F18VS/UWSTraumaDiffuse axonal injury3450120021433212341314.86
008M58MCSTraumaR-frontal lobe damage12481131021533312340315.62
009M41MCSTraumaR-frontal-temporo-parietal lobe damage15112330121834422342318.89
010M46VS/UWSStrokeL-brainstem, cerebellar damage7461021021433212353317.05
011M25VS/UWSAnoxiaDiffuse cortical and subcortical atrophy4661021021422412346318.07
012M58VS/UWSTraumaL-brainstem damage1771121021935512340310.75
013M36VS/UWSTraumaL-frontal-temporo-parietal lobe damage647112102102321024429.58
014M58VS/UWSTraumaR-frontal-temporo-parietal lobe damage44610210261021024526.69
015M65VS/UWSStrokeDiffuse cortical and subcortical atrophy34310000251011024324.01
016F24VS/UWSTraumaDiffuse axonal injury4466102102812210244214.03
017F46VS/UWSStrokeL-pons damage247112102711210240212.11
018M53VS/UWSAnoxiaDiffuse cortical and subcortical atrophy34410100261021024125.38
019F32VS/UWSTraumaL-temporo-parietal lobe damage346102102811220223213.76
020M41VS/UWSAnoxiaDiffuse cortical and subcortical atrophy244101002811220240212.06
021F33VS/UWSAnoxiaDiffuse cortical and subcortical atrophy756211002112322024724.55
022M49VS/UWSAnoxiaDiffuse cortical and subcortical atrophy24610210261021021428.97
023F25MCSAnoxiaBilateral thalamus, brainstem damage47144500231024002250212.42
024M63VS/UWSStrokeL-basal ganglia lesions54400110251011024828.22
025M68VS/UWSTraumaL-frontal-temporo-parietal lobe damage24500210260121024729.72
Appendix 1—table 3
Demographic and clinical characteristics of patients in the ‘Guangzhou_HDxt’ dataset.
Patient aliasGenderAge (years)DiagnoseEtiologyTime to MRI (months)CRS-R score at T0CRS-R subscore at T0Follow-up (months)GOSPredicted score
Appendix 1—table 4
Demographic of healthy controls in the ‘Beijing_750’ dataset.
Appendix 1—table 5
Demographic of healthy controls in the ‘Beijing_HDxt’ dataset.
Appendix 2—table 1
Brain networks and ROIs in this study.
Brain networkROI nameROI
Peak MNI coordinatesReferences
Default mode(Raichle, 2011; Demertzi et al., 2015)
Anterior medial prefrontal cortexaMPFC−1 54 27
Posterior cingulate cortex/precuneusPCC0 –52 27
Left lateral parietal cortexL.LatP−46 –66 30
Right lateral parietal cortexR.LatP49 –63 33
Executive control(Seeley et al., 2007; Raichle, 2011)
Dorsal medial PFCDMPFC0 27 46
Left anterior prefrontal cortexL.PFC−44 45 0
Right anterior prefrontal cortexR.PFC44 45 0
Left superior parietal cortexL. Parietal−50 –51 45
Right superior parietal cortexR. Parietal50 –51 45
Salience(Seeley et al., 2007; Raichle, 2011; Demertzi et al., 2015)
Left orbital frontoinsulaL.AIns−40 18 –12
Right orbital frontoinsulaR.AIns42 10 –12
Dorsal anterior cingulatedACC0 18 30
Sensorimotor(Raichle, 2011; Demertzi et al., 2015)
Left primary motor cortexL.M1−39 –26 51
Right primary motor cortexR.M138 –26 51
Supplementary motor areaSMA0 –21 51
Auditory(Raichle, 2011; Demertzi et al., 2015)
Left primary auditory cortexL.A1−62 –30 12
Right primary auditory cortexR.A159 –27 15
Middle cingulate cortexMCC0 –7 43
Visual(Demertzi et al., 2015)
Left primary visual cortexL.V1−13 –85 6
Right primary visual cortexR.V18 –82 6
Left associative visual cortexL.V4−30 –89 20
Right associative visual cortexR.V430 –89 20
Appendix 6—table 1
The brain area connection features and their Pearson's correlations to the CRS-R scores at the T1 time point across the DOC patients in the training dataset ‘Beijing 750’.
ROI namePearson's correlation coefficient and p value
**DMN.aMPFCr = 0.514, p=0.000
**ExecuContr.L.Parietalr = 0.429, p=0.000
**DMN.PCCr = 0.420, p=0.001
**DMN.R.LatPr = 0.407, p=0.001
**ExecuContr.DMPFCr = 0.405, p=0.001
*ExecuContr.R.Parietalr = 0.363, p=0.003
*Sensorimotor.SMAr = −0.332, p=0.008
*ExecuContr.R.PFCr = 0.320, p=0.011
*Auditory.R.A1r = 0.315, p=0.012
*DMN.L.LatPr = 0.298, p=0.018
*ExecuContr.L.PFCr = 0.291, p=0.021
*Sensorimotor.L.M1r = 0.267, p=0.035
Auditory.L.A1r = 0.206, p=0.105
Salience.R.AInsr = −0.187, p=0.142
Sensorimotor.R.M1r = 0.167, p=0.191
Visual.L.V4r = −0.151, p=0.236
Salience.dACCr = −0.104, p=0.418
Salience.L.AInsr = 0.075, p=0.560
Visual.R.V1r = 0.065, p=0.611
Auditory.MCCr = 0.053, p=0.682
Visual.R.V4r = −0.031, p=0.809
Visual.L.V1r = −0.028, p=0.830
  1. **: p<0.05, FDR corrected; *: p<0.05, uncorrected.

Appendix 6—table 2
Functional connectivity features and their Pearson's correlations to the CRS-R scores at the T1 time point across the DOC patients in the training dataset ‘Beijing 750’.
Functional connectivityPearson's correlation coefficient and p-value
DMN.aMPFC - ExecuContr.DMPFCr = −0.489, p=0.000
*DMN.L.LatP - Visual.L.V4r = −0.421, p=0.001
*Auditory.MCC - Visual.R.V1r = 0.375, p=0.002
*ExecuContr.R.PFC - ExecuContr.R.Parietalr = 0.361, p=0.004
*ExecuContr.DMPFC - Auditory.MCCr = −0.351, p=0.005
*ExecuContr.L.PFC - Salience.dACCr = −0.335, p=0.007
*Sensorimotor.R.M1 - Sensorimotor.SMAr = −0.330, p=0.008
*Sensorimotor.R.M1 - Auditory.L.A1r = 0.319, p=0.011
*Salience.dACC - Visual.R.V1r = 0.319, p=0.011
*ExecuContr.DMPFC - Sensorimotor.L.M1r = −0.310, p=0.013
*DMN.R.LatP - Visual.R.V4r = −0.306, p=0.015
*ExecuContr.L.Parietal - Sensorimotor.L.M1r = −0.302, p=0.016
*DMN.aMPFC - Salience.dACCr = −0.292, p=0.020
*DMN.aMPFC - Sensorimotor.L.M1r = −0.286, p=0.023
*DMN.aMPFC - DMN.PCCr = 0.275, p=0.029
*ExecuContr.R.Parietal - Visual.R.V4r = −0.270, p=0.033
*DMN.aMPFC - Sensorimotor.R.M1r = −0.268, p=0.034
*ExecuContr.R.Parietal - Sensorimotor.R.M1r = −0.263, p=0.037
*Sensorimotor.L.M1 - Sensorimotor.SMAr = −0.261, p=0.039
*DMN.R.LatP - Sensorimotor.R.M1r = −0.261, p=0.039
*ExecuContr.R.Parietal - Visual.L.V4r = −0.257, p=0.042
*Salience.dACC - Visual.L.V4r = 0.256, p=0.043
*ExecuContr.DMPFC - Sensorimotor.R.M1r = −0.255, p=0.043
*DMN.aMPFC - Visual.L.V1r = 0.251, p=0.047
*Salience.R.AIns - Sensorimotor.L.M1r = 0.250, p=0.049
*DMN.L.LatP - Sensorimotor.SMAr = 0.248, p=0.050
  1. Specifically, the functional connectivity features were the functional connectivity between any pair of ROIs. As there were more than 200 functional connectivity features, because of space limitations, only the functional connectivity features that were significantly correlated to the CRS-R scores at the T1 time point are shown. **: p<0.05, FDR corrected; *: p<0.05, uncorrected.

Additional files

Supplementary file 1

Some examples of the warped ROIs in the default mode network for one healthy control and three DOC patients with a GOS score of 2,3 and 4, respectively.
Supplementary file 2

Details about single-domain prognostic models and comparisons of the single-domain and combination models.
Transparent reporting form

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Ming Song
  2. Yi Yang
  3. Jianghong He
  4. Zhengyi Yang
  5. Shan Yu
  6. Qiuyou Xie
  7. Xiaoyu Xia
  8. Yuanyuan Dang
  9. Qiang Zhang
  10. Xinhuai Wu
  11. Yue Cui
  12. Bing Hou
  13. Ronghao Yu
  14. Ruxiang Xu
  15. Tianzi Jiang
Prognostication of chronic disorders of consciousness using brain functional networks and clinical characteristics
eLife 7:e36173.