Assessing the depth of language processing in patients with disorders of consciousness

Abstract

Assessing residual consciousness and cognitive abilities in unresponsive patients is a major clinical concern and a challenge for cognitive neuroscience. Although neuroimaging studies have demonstrated a potential for informing diagnosis and prognosis in unresponsive patients, these methods involve sophisticated brain imaging technologies, which limit their clinical application. In this study, we adopted a new language paradigm that elicited rhythmic brain responses tracking the single-word, phrase and sentence rhythms in speech, to examine whether bedside electroencephalography (EEG) recordings can help inform diagnosis and prognosis. EEG-derived neural signals, including both speech-tracking responses and temporal dynamics of global brain states, were associated with behavioral diagnosis of consciousness. Crucially, multiple EEG measures in the language paradigm were robust to predict future outcomes in individual patients. Thus, EEG-based language assessment provides a new and reliable approach to objectively characterize and predict states of consciousness and to longitudinally track individual patients’ language processing abilities at the bedside.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Fig. 1: Paradigm and neural tracking of hierarchical linguistic structures.
Fig. 2: Procedure and auditory-evoked brain activity in the clinical study.
Fig. 3: Global patterns of brain states.
Fig. 4: Duration and occurrence of brain state maps.
Fig. 5: Diagnosis and outcome prediction.

Data availability

The data that support the findings of this study are available from the corresponding authors upon reasonable request.

Code availability

EEG data analyses were performed in BrainVision Analyzer, the freely available toolbox EEGLAB and MicrostateAnalysis in combination with custom MATLAB scripts. The software code that support the findings of this study is available from the corresponding authors upon reasonable request.

References

  1. 1.

    Giacino, J. T., Fins, J. J., Laureys, S. & Schiff, N. D. Disorders of consciousness after acquired brain injury: the state of the science. Nat. Rev. Neurol. 10, 99–114 (2014).

    PubMed  Article  Google Scholar 

  2. 2.

    Bruno, M. A., Gosseries, O., Ledoux, D., Hustinx, R. & Laureys, S. Assessment of consciousness with electrophysiological and neurological imaging techniques. Curr. Opin. Crit. Care 17, 146–151 (2011).

    PubMed  Article  Google Scholar 

  3. 3.

    Giacino, J. T. et al. The minimally conscious state: definition and diagnostic criteria. Neurology 58, 349–353 (2002).

    PubMed  Article  Google Scholar 

  4. 4.

    Owen, A. M. The search for consciousness. Neuron 102, 526–528 (2019).

    CAS  PubMed  Article  Google Scholar 

  5. 5.

    Strauss, M. et al. Disruption of hierarchical predictive coding during sleep. Proc. Natl Acad. Sci. USA 112, E1353–E1362 (2015).

    CAS  PubMed  Article  Google Scholar 

  6. 6.

    Goldfine, A. M., Victor, J. D., Conte, M. M., Bardin, J. C. & Schiff, N. D. Determination of awareness in patients with severe brain injury using EEG power spectral analysis. Clin. Neurophysiol. 122, 2157–2168 (2011).

    PubMed  PubMed Central  Article  Google Scholar 

  7. 7.

    Sitt, J. D. et al. Large scale screening of neural signatures of consciousness in patients in a vegetative or minimally conscious state. Brain 137, 2258–2270 (2014).

    PubMed  PubMed Central  Article  Google Scholar 

  8. 8.

    Dehaene, S. & Changeux, J. P. Experimental and theoretical approaches to conscious processing. Neuron 70, 200–227 (2011).

    CAS  PubMed  Article  Google Scholar 

  9. 9.

    Barttfeld, P. et al. Signature of consciousness in the dynamics of resting-state brain activity. Proc. Natl Acad. Sci. USA 112, 887–892 (2015).

    CAS  PubMed  Article  Google Scholar 

  10. 10.

    Lehmann, D., Ozaki, H. & Pal, I. EEG alpha map series: brain micro-states by space-oriented adaptive segmentation. Electroencephalogr. Clin. Neurophysiol. 67, 271–288 (1987).

    CAS  PubMed  Article  Google Scholar 

  11. 11.

    Michel, C. M. & Koenig, T. EEG microstates as a tool for studying the temporal dynamics of whole-brain neuronal networks: a review. Neuroimage 180, 577–593 (2018).

    PubMed  Article  Google Scholar 

  12. 12.

    Laureys, S. & Schiff, N. D. Coma and consciousness: paradigms (re)framed by neuroimaging. Neuroimage 61, 478–491 (2012).

    PubMed  Article  Google Scholar 

  13. 13.

    Schnakers, C. et al. Voluntary brain processing in disorders of consciousness. Neurology 71, 1614–1620 (2008).

    CAS  PubMed  Article  Google Scholar 

  14. 14.

    Owen, A. M. et al. Detecting awareness in the vegetative state. Science 313, 1402 (2006).

    CAS  PubMed  Article  Google Scholar 

  15. 15.

    Braiman, C. et al. Cortical response to the natural speech envelope correlates with neuroimaging evidence of cognition in severe brain injury. Curr. Biol. 28, 3833–3839 (2018).

  16. 16.

    Rohaut, B. et al. Probing ERP correlates of verbal semantic processing in patients with impaired consciousness. Neuropsychologia 66, 279–292 (2015).

    PubMed  Article  Google Scholar 

  17. 17.

    Coleman, M. R. et al. Do vegetative patients retain aspects of language comprehension? Evidence from fMRI. Brain 130, 2494–2507 (2007).

    PubMed  Article  Google Scholar 

  18. 18.

    Faugeras, F. et al. Event related potentials elicited by violations of auditory regularities in patients with impaired consciousness. Neuropsychologia 50, 403–418 (2012).

    PubMed  Article  Google Scholar 

  19. 19.

    Engemann, D. A. et al. Robust EEG-based cross-site and cross-protocol classification of states of consciousness. Brain 141, 3179–3192 (2018).

    PubMed  Article  Google Scholar 

  20. 20.

    Ding, N., Melloni, L., Zhang, H., Tian, X. & Poeppel, D. Cortical tracking of hierarchical linguistic structures in connected speech. Nat. Neurosci. 19, 158–164 (2016).

    CAS  PubMed  Article  Google Scholar 

  21. 21.

    Kiefer, M. The N400 is modulated by unconsciously perceived masked words: further evidence for an automatic spreading activation account of N400 priming effects. Brain Res. Cogn. Brain Res. 13, 27–39 (2002).

    PubMed  Article  Google Scholar 

  22. 22.

    Batterink, L. & Neville, H. J. The human brain processes syntax in the absence of conscious awareness. J. Neurosci. 33, 8528–8533 (2013).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  23. 23.

    Demertzi, A. et al. Human consciousness is supported by dynamic complex patterns of brain signal coordination. Sci. Adv. 5, eaat7603 (2019).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  24. 24.

    Khanna, A., Pascual-Leone, A., Michel, C. M. & Farzan, F. Microstates in resting-state EEG: current status and future directions. Neurosci. Biobehav/ Rev. 49, 105–113 (2015).

    Article  Google Scholar 

  25. 25.

    Koenig, T. et al. Millisecond by millisecond, year by year: normative EEG microstates and developmental stages. Neuroimage 16, 41–48 (2002).

    PubMed  Article  Google Scholar 

  26. 26.

    Brodbeck, V. et al. EEG microstates of wakefulness and NREM sleep. Neuroimage 62, 2129–2139 (2012).

    PubMed  Article  Google Scholar 

  27. 27.

    Britz, J., Van De Ville, D. & Michel, C. M. BOLD correlates of EEG topography reveal rapid resting-state network dynamics. Neuroimage 52, 1162–1170 (2010).

    PubMed  Article  Google Scholar 

  28. 28.

    Stender, J. et al. Diagnostic precision of PET imaging and functional MRI in disorders of consciousness: a clinical validation study. Lancet 384, 514–522 (2014).

    PubMed  Article  Google Scholar 

  29. 29.

    Scarpino, M. et al. Prognostic value of post-acute EEG in severe disorders of consciousness, using American Clinical Neurophysiology Society terminology. Neurophysiol. Clin. 49, 317–327 (2019).

    PubMed  Article  Google Scholar 

  30. 30.

    Wannez, S. et al. The repetition of behavioral assessments in diagnosis of disorders of consciousness. Ann. Neurol. 81, 883–889 (2017).

    PubMed  Article  Google Scholar 

  31. 31.

    Wu, X. et al. Intrinsic functional connectivity patterns predict consciousness level and recovery outcome in acquired brain injury. J. Neurosci. 35, 12932–12946 (2015).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  32. 32.

    Monti, M. M. et al. Willful modulation of brain activity in disorders of consciousness. N. Engl. J. Med. 362, 579–589 (2010).

    CAS  PubMed  Article  Google Scholar 

  33. 33.

    Chennu, S. et al. Brain networks predict metabolism, diagnosis and prognosis at the bedside in disorders of consciousness. Brain 140, 2120–2132 (2017).

    PubMed  Article  Google Scholar 

  34. 34.

    Di Perri, C. et al. Neural correlates of consciousness in patients who have emerged from a minimally conscious state: a cross-sectional multimodal imaging study. Lancet Neurol. 15, 830–842 (2016).

    PubMed  Article  Google Scholar 

  35. 35.

    Zalesky, A., Fornito, A., Cocchi, L., Gollo, L. L. & Breakspear, M. Time-resolved resting-state brain networks. Proc. Natl Acad. Sci. USA 111, 10341–10346 (2014).

    CAS  PubMed  Article  Google Scholar 

  36. 36.

    Hutchison, R. M. et al. Dynamic functional connectivity: promise, issues, and interpretations. Neuroimage 80, 360–378 (2013).

    PubMed  Article  Google Scholar 

  37. 37.

    Tagliazucchi, E. et al. Large-scale signatures of unconsciousness are consistent with a departure from critical dynamics. J. R. Soc. Interface 13, 20151027 (2016).

    PubMed  PubMed Central  Article  Google Scholar 

  38. 38.

    Cojan, Y. et al. The brain under self-control: modulation of inhibitory and monitoring cortical networks during hypnotic paralysis. Neuron 62, 862–875 (2009).

    CAS  PubMed  Article  Google Scholar 

  39. 39.

    Barttfeld, P. et al. Distinct patterns of functional brain connectivity correlate with objective performance and subjective beliefs. Proc. Natl Acad. Sci. USA 110, 11577–11582 (2013).

    CAS  PubMed  Article  Google Scholar 

  40. 40.

    Claassen, J. et al. Detection of brain activation in unresponsive patients with acute brain injury. N. Engl. J. Med. 380, 2497–2505 (2019).

    PubMed  Article  Google Scholar 

  41. 41.

    Davis, M. H. et al. Dissociating speech perception and comprehension at reduced levels of awareness. Proc. Natl Acad. Sci. USA 104, 16032–16037 (2007).

    CAS  PubMed  Article  Google Scholar 

  42. 42.

    Makov, S. et al. Sleep disrupts high-level speech parsing despite significant basic auditory processing. J. Neurosci. 37, 7772–7781 (2017).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  43. 43.

    Jin, P. et al. Low-frequency neural activity reflects rule-based chunking during speech listening. eLife 9, e55613 (2020).

    PubMed  PubMed Central  Article  Google Scholar 

  44. 44.

    Frank, S. L. & Yang, J. Lexical representation explains cortical entrainment during speech comprehension. PLoS ONE 13, e0197304 (2018).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  45. 45.

    Steppacher, I. et al. N400 predicts recovery from disorders of consciousness. Ann. Neurol. 73, 594–602 (2013).

    PubMed  Article  Google Scholar 

  46. 46.

    Giacino, J. T. et al. Practice guideline update recommendations summary: disorders of consciousness: report of the Guideline Development, Dissemination, and Implementation Subcommittee of the American Academy of Neurology; the American Congress of Rehabilitation Medicine; and the National Institute on Disability, Independent Living, and Rehabilitation Research. Neurology 91, 450–460 (2018).

    PubMed  PubMed Central  Article  Google Scholar 

  47. 47.

    Giacino, J. T. et al. Comprehensive systematic review update summary: disorders of consciousness: report of the Guideline Development, Dissemination, and Implementation Subcommittee of the American Academy of Neurology; the American Congress of Rehabilitation Medicine; and the National Institute on Disability, Independent Living, and Rehabilitation Research. Neurology 91, 461–470 (2018).

    PubMed  PubMed Central  Article  Google Scholar 

  48. 48.

    Lehmann, D. et al. EEG microstate duration and syntax in acute, medication-naive, first-episode schizophrenia: a multi-center study. Psychiatry Res. 138, 141–156 (2005).

    PubMed  Article  Google Scholar 

  49. 49.

    Perrin, F., Castro, M., Tillmann, B. & Luaute, J. Promoting the use of personally relevant stimuli for investigating patients with disorders of consciousness. Front. Psychol. 6, 1102 (2015).

    PubMed  PubMed Central  Article  Google Scholar 

  50. 50.

    Cortese, M. D. et al. Coma recovery scale-r: variability in the disorder of consciousness. BMC Neurol. 15, 186 (2015).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  51. 51.

    Edlow, B. L. et al. Early detection of consciousness in patients with acute severe traumatic brain injury. Brain 140, 2399–2414 (2017).

    PubMed  PubMed Central  Article  Google Scholar 

  52. 52.

    Giacino, J. T., Kalmar, K. & Whyte, J. The JFK Coma Recovery Scale-Revised: measurement characteristics and diagnostic utility. Arch. Phys. Med. Rehabil. 85, 2020–2029 (2004).

    PubMed  Article  Google Scholar 

  53. 53.

    Ding, N. et al. Characterizing neural entrainment to hierarchical linguistic units using electroencephalography (EEG). Front. Hum. Neurosci. 11, 481 (2017).

    PubMed  PubMed Central  Article  Google Scholar 

  54. 54.

    Teasdale, G. & Jennett, B. Assessment of coma and impaired consciousness. A practical scale. Lancet 2, 81–84 (1974).

    CAS  PubMed  Article  Google Scholar 

  55. 55.

    Bruno, M. A., Vanhaudenhuyse, A., Thibaut, A., Moonen, G. & Laureys, S. From unresponsive wakefulness to minimally conscious PLUS and functional locked-in syndromes: recent advances in our understanding of disorders of consciousness. J. Neurol. 258, 1373–1384 (2011).

    PubMed  Article  Google Scholar 

  56. 56.

    Wannez, S. et al. Prevalence of coma-recovery scale-revised signs of consciousness in patients in minimally conscious state. Neuropsychol. Rehabil. 28, 1350–1359 (2018).

    PubMed  Article  Google Scholar 

  57. 57.

    Büyüköztürk, Ş. & Çokluk Bökeoğlu, Ö. Discriminant function analysis: concept and application. Eurasian J. Educ. Res. 8, 73–92 (2008).

    Google Scholar 

  58. 58.

    Dixon, S. J. et al. Application of classification methods when group sizes are unequal by incorporation of prior probabilities to three common approaches: application to simulations and mouse urinary chemosignals. Chemometrics Intell. Lab. Syst. 99, 111–120 (2009).

    CAS  Article  Google Scholar 

  59. 59.

    Geurts, B. H., Andriessen, T. M., Goraj, B. M. & Vos, P. E. The reliability of magnetic resonance imaging in traumatic brain injury lesion detection. Brain Inj. 26, 1439–1450 (2012).

    PubMed  Article  Google Scholar 

  60. 60.

    Pialat, J. B. et al. Evolution of lesion volume in acute stroke treated by intravenous t-PA. J. Magn. Reson. Imaging 22, 23–28 (2005).

    PubMed  Article  Google Scholar 

  61. 61.

    Rosso, I. M. et al. Reduced amygdala volumes in first-episode bipolar disorder and correlation with cerebral white matter. Biol. Psychiatry 61, 743–749 (2007).

    PubMed  Article  Google Scholar 

Download references

Acknowledgements

We thank L. Melloni, J. Sitt, P. Barttfeld and J. Wu for their suggestions on the study and helpful comments on the manuscript. We also thank R. Zheng and W. Peng from the Institute of Neuroscience, the Institute of Neuroscience Primate Physiology Research Platform Core Facility and the nursing staff from Shanghai Huashan Hospital for their assistance in data acquisition. We are grateful to all the patients and their families and the healthy individuals for their participation.

This work was supported by the Key Research Program of Frontier Sciences (QYZDY-SSW-SMC001), the Strategic Priority Research Program (XDB32070200), the Pioneer Hundreds of Talents Program from the Chinese Academy of Sciences and the Shanghai Municipal Science and Technology Major Project (2018SHZDZX05) to L.W.; the National Natural Science Foundation of China (81571025) and the International Scientific and Technological Collaboration Foundation (18410711300) from the Shanghai Science and Technology Committee (STCSM) to X.W.; Shanghai Municipal Science and Technology Major Project (2018SHZDZX03) and ZJLab to Y.M.; and the Shanghai Sailing Program (17YF1426600) from the STCSM to Z.Q.

Author information

Affiliations

Authors

Contributions

L.W., X.W., S.D., J.Z. M.P. and N.D. conceptualized the study. L.W. and P.G. designed the experiments. D.Z., Z.Q., J.T., H.T., J.J., Y.W. and L.X. collected the data. D.Z., Z.Q., J.T., H.T., X.W. and Y.M. performed clinical ratings on patients. P.G. and Y.J. analyzed the data. L.W., X.W., S.D., M.P., P.G., Y.J. and N.D. wrote the manuscript.

Corresponding authors

Correspondence to Xuehai Wu or Liping Wang.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature Neuroscience thanks Damian Cruse and Ziv Williams for their contribution to the peer review of this work.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1

The flowchart showing patients selection in data analysis.

Extended Data Fig. 2 Individual ITPC responses to hierarchical linguistic structures in Healthy controls (n = 47), MCS (n = 42) and UWS (n = 36) patients at three task levels.

In each inset, the dots in the left represent the ITPC values from individual subject at target frequencies (1/2/4 Hz), while the dots in the right represent the individual mean value at its respective neighbours. Solid black dots represent grand mean values. n.s., P > 0.1; ~, P < 0.1; *, P < 0.001; one-sided paired-sample t-test: see legend of Fig. 2b for precise statistical values.

Extended Data Fig. 3 ITPC responses to hierarchical linguistic structures in individual patients at three task levels.

Each bar denotes the response from one subject. Red dots indicate the significance (exact P < 0.05; one-sided exact test, the statistical significance (exact P) of the ITPC response at a target frequency is the probability that the target frequency response differs from the null distribution, which consisted of responses at all non-target frequencies, see Methods).

Extended Data Fig. 4 Brain state maps of healthy controls and patients in all four task conditions.

Number of subjects: nHealthy-Resting = 34, nHealthy-Task =47, nMCS-Resting = 41, nMCS-Task = 42, nUWS-Resting = 30, nUWS-Task =36. Top row: Template Anterior-Posterior (A-P) and Left-Right (L-R) maps obtained from healthy controls. Bottom three rows: Original four maps (Maps A, B, C, and D) of each group in each condition.

Extended Data Fig. 5 Duration and occurrence of brain state maps.

a, The duration of the A-P Map for the healthy control (nResting = 34, nTask = 47), MCS (nResting = 41, nTask = 42), and UWS (nUWS = 30, nTask =36) groups in all four task conditions. b, The occurrence of the L-R Map for healthy control, MCS, and UWS groups. Note that there were no differences between the three groups. Boxes represent IQR, central dots indicate the median, and whiskers indicate 1.5 × IQR. Colored dots indicate outliers. c, The duration of the L-R map for healthy controls (gray; nResting = 34, nTask = 47), recovery patients (+ve, green; nResting = 11, nTask = 19), and non-recovery MCS (-ve MCS, blue; nResting = 30, nTask = 23) and non-recovered UWS (-ve UWS, red; nResting = 30, nTask = 36) patients in all four conditions. One-way ANOVA, Bonferroni corrected: Healthy vs. +ve, PResting = 0.001, PWord = 0.018, PPhrase = 0.013, PSentence = 0.008; +ve vs. -ve MCS, PWord = 0.012, PPhrase = 0.009, PSentence = 0.002. d, The occurrence of the A-P map for healthy controls, recovery patients, non-recovery MCS patients, and non-recovery UWS patients. One-way ANOVA, Bonferroni corrected: Healthy vs. +ve, PResting = 6.8×10-5, PWord = 8.1×10-7, PPhrase = 5.7×10-7, PSentence = 1.8×10-7; +ve vs. -ve MCS, PWord = 0.046, PPhrase = 0.048, PSentence = 0.026. Panel c and d: colored dots represent individual subjects. Black dots represent mean values. Error bars represent S.E.M. All panels: *, P < 0.05; **, P < 0.01; ***, p < 0.001.

Extended Data Fig. 6 Correlation between the volumes of brain injury and the ∆Cρ.

a, The comparison of volumes of brain injury between MCS (n = 17) and UWS (n = 10) patients. The black dots and error bars denote the mean value and S.E.M. t25 = 0.64, P = 0.53, two-tailed two-sample t-test. b, The correlation between ∆Cρ and the volumes of brain injury in three task conditions. Pearson’s correlation test (two-tailed), nMCS = 17, nUWS = 10. c, An example patient (Patient 7): the MRI data and maps of the stroke patient without brain damage. d, An example patient (Patient 17): the MRI data and maps of the TBI patient with large brain damage. e, The comparison of ∆probability between the stroke patient without brain damage and the TBI patient with brain damage, as shown in c and d. P1: Patient 7, P2: Patient 17. f, ∆Cρ, the same format as e. g, The MRI data and brain states of a stroke patient with brain damage (an example patient, Patient 2). The orange box indicates the first EEG recording in unrecovered state. The green box indicates the last EEG recording in recovery state. h, The comparison of ∆probability in Patient 2 between the first EEG recording in unrecovered state and the last EEG recording in recovery state. i, ∆Cρ, the same format as h. W: Word, P: Phrase, S: Sentence. F: First recording, L: Last recording. The percentage under each spatial map indicates the probability of each map.

Extended Data Fig. 7 Diagnosis and outcome prediction using SVM.

a, The confusion matrix of diagnosed consciousness classification generated by the cross-validated SVM. The feature combinations we used were [∆Cρ + DurationL-R+ OccurrenceA-P + ITPC1Hz + ITPC2Hz + ITPC4Hz] for Sentence task. b, The performance of outcome prediction on training data using SVM classifier with the best feature combinations. Left: Outcome prediction accuracies by EEG on 38 EEG recordings (15 outcome-positive patients). Right: Comparison of individual predictions and actual outcomes. The patients with UWS are shown to the left of dashed line, and the patients with MCS are shown to the right. The dots above the threshold (gray line, prediction score = 0.3) represent the patients with predicted positive outcomes, while the others represent those with predicted negative outcomes. The actual outcome-negative patients are marked by orange dots, and the actual outcome-positive patients are marked by green diamonds. Solid green diamonds represent the outcome in patients that regained wakefulness. The feature combinations we used were: [∆Probability + DurationL-R + TransitionA-P] for Word condition, [∆Probability + OccurrenceA-P + DurationL-R + TransitionL-R + ITPC4Hz] for Phrase condition, [OccurrenceA-P + DurationL-R + TransitionL-R + ITPC1Hz + ITPC2Hz] for Sentence condition.

Extended Data Fig. 8 Comparison of EEG-based and CRS-based classifiers for diagnosis and outcome prediction.

a, Performance of clinical diagnosis using the CRS-R total-score. The classification model (LDA) was trained on the first 38 patients with cross validation (Left) and then tested (without retraining) on a novel dataset of 25 patients (Right). b, Left: Performance of outcome prediction using the optimal CRS-R sub-score (Visual subscale). Right: The prognostic validity of this model. The classifier was trained with cross-validation on the first dataset of 38 patients, then tested for generalization on a new dataset of 25 patients. c, Comparison of the prediction performance for models with the same number of features, based either on CRS-R (7 features, as in b) or the EEG recording under the word condition (7 features: 1 ITPC and 6 microstates). The optimal feature combination was ∆Probability + OccurrenceA-P. d, To test whether the superior EEG generalization ability was due to the larger number of EEG features used in the model, we then ran another way of selecting features by merely using the features with the first two highest weights in the model, and compared their performance of generalization. For the model using EEG see c, and for CRS-R, the best two features were Visual and Arousal subscales. e, Comparison of the performance of outcome prediction, using a standard LDA without feature selection, using all 7 features under the word condition of EEG versus all 7 features of the CRS-R scores (1 total-score and 6 sub-scores: auditory, visual, motor, oromotor, communication and arousal). Generalizations using EEG recorded during the other two task conditions (phrase: AUC = 89%, χ2 = 13.1, P = 5.5×10-4, accuracy = 84%; sentence: AUC = 93%, χ2 = 13.1, P = 5.5×10-4, accuracy = 84%; chi-squared test) showed similar results as that during the word condition.

Extended Data Fig. 9 Comparisons of performance of outcome prediction using EEG versus EEG plus CRS-R scores.

Upper: The confusion matrix of outcome prediction by EEG scores. Lower: The confusion matrix of outcome prediction by the combination of EEG and CRS-R scores.

Extended Data Fig. 10 Multiple CRS-R ratings across time. Each inset indicates one patient.

a, Individual patients (n = 15). Within each inset, every blue dot indicates one CRS-R rating, and the gray line indicates the GLM fitting of all ratings in the entire period. Day 0 and red vertical dashed lines indicate the day of first EEG recording. b, The comparison of CRS-R scores between the EEG recording day and the day within a week (on average within 2.67 days). Colored lines indicate the ratings of individual patients. Black line indicates the mean. No significant difference was found between the two ratings (n = 15, t14 = 0.899, P = 0.384; two-sided paired-sample t-test).

Supplementary information

Supplementary Information

Supplementary Tables 1–4.

Reporting Summary

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Gui, P., Jiang, Y., Zang, D. et al. Assessing the depth of language processing in patients with disorders of consciousness. Nat Neurosci 23, 761–770 (2020). https://doi.org/10.1038/s41593-020-0639-1

Download citation