Research ArticleCOGNITIVE NEUROSCIENCE

Temporal circuit of macroscale dynamic brain activity supports human consciousness

See allHide authors and affiliations

Science Advances  11 Mar 2020:
Vol. 6, no. 11, eaaz0087
DOI: 10.1126/sciadv.aaz0087
  • Fig. 1 Level of behavioral responsiveness across datasets and CAPs.

    (A) Dataset 1 (propofol-SHH) adopted Ramsay scale. Dataset 2 (propofol-WI) adopted observer’s assessment of alertness/sedation (OAAS). Dataset 3 (ketamine) adopted a button press task for every 30 s. Reaction time (RT) in milliseconds with respect to each instruction was recoded. By comparing the timing of verbal instruction and actual responsiveness during and after ketamine infusion, the periods during which a participant retained responsiveness (PreLOR), loss of responsiveness (LOR), and recovery of responsiveness were determined. Dataset 4 (neuropathological patients) adopted Coma Recovery Scale–Revised (CRS-R). Level of responsiveness is shown by the total score of six subscales (auditory, visual, motor, verbal, communication, and arousal). MCS, minimally conscious state. Error bars indicate ±SD. (B) Spatial maps of eight CAPs. The CAPs consist of DMN+, DAT+, FPN+, SMN+, VIS+, VAT+, GN+, and GN−. (C) The eight CAPs are composed of four pairs of mirror motifs with a strong negative spatial similarity, including DMN+ versus DAT+, VIS+ versus VAT+, FPN+ versus SMN+, and GN− versus GN+.

  • Fig. 2 Occurrence rates of CAPs.

    (A) Spearman rank correlations between the occurrence rates of joint mirror motifs or individual CAPs and the level of responsiveness. (B to E) The CAP occurrence rates in different conditions (conscious, intermediate, unresponsive, and recovered) and in different datasets. Intermediate conditions refer to propofol light sedation, PreLOR of ketamine induction, and patients with MCS; unresponsive conditions refer to propofol general anesthesia and deep sedation, LOR due to ketamine, and patients with UWS. Red squares in (A) and lines in (B) to (E) indicate significance at FDR-corrected α < 0.05. See fig. S4 and table S1 for full statistics. Error bars in (A) indicate 95% confidence interval, and error bars in (B) to (E) indicate ±SD.

  • Fig. 3 Transition probabilities among CAPs.

    (A) Full transition probability matrix for conscious condition (conscious), propofol-induced unresponsiveness (propofol), ketamine-induced unresponsiveness (ketamine), and patients with UWS. The on-diagonal entries are referred to as the persistence probabilities. (B) Diagonal-free transition probability matrix, where the off-diagonal entries are referred to as transition probabilities by controlling for autocorrelation due to the CAP’s persistence. (C) Schematic illustration of the significant preferred paths (>null) and nonpreferred paths (<null) for conscious versus null, propofol versus null, ketamine versus null, and patients with UWS versus null (all gray arrows). Red (higher than conscious condition) and green (lower than conscious condition) arrows indicate significant differences of persistence probabilities and transition probabilities comparing to baseline consciousness. The null model for each condition and the differences between conditions were generated by 1000 permutations across the entire dataset (see more details in Materials and Methods). Significance level was determined at P < 0.001 by considering multiple comparison corrections (99.9th and 0.1th percentile of the null distributions; two-sided).

  • Fig. 4 Descriptive complexity of trajectories among CAPs and their in-degree accessibility.

    (A) Descriptive complexity of trajectories (in bits) between each pair of CAPs in the conscious condition (conscious), propofol-induced unresponsiveness (propofol), ketamine-induced unresponsiveness (ketamine), and in patients with UWS. (B) Significant differences of the descriptive complexity of trajectories for propofol versus conscious, ketamine versus conscious, and patients with UWS versus conscious. The null models were generated by 1000 permutations across the entire dataset. Significance level was determined at P < 0.001. (C) Schematic illustration for (A). The accessibility of each CAP is defined as the inverse of descriptive complexity. The node size is proportional to in-degree accessibility. The Gephi Force Atlas layout algorithm (https://gephi.org) was used.

  • Fig. 5 Antiphasic coactivation accounts for anticorrelation.

    (A) Conventional static functional connectivity between DMN and DAT and within-network connectivity of DMN and DAT in different conditions (conscious, intermediate, unresponsive, and recovered) and in different datasets. Red lines indicate significance at FDR-corrected α < 0.05. Error bars indicate ±SD. (B) Pearson correlations between the joint occurrence rates of DMN+ and DAT+ and the strength of DMN-DAT functional connectivity (FC) (left), as well as within-network functional connectivity of DMN (middle) and DAT (right) across all participants.

  • Fig. 6 Stimulus modulations of CAPs and control analysis in psychiatric patients.

    (A) Stimulus-induced CAP occurrence rate changes (against stimulus onset, t = 0) in baseline conscious condition, light sedation, and general anesthesia (n = 15). Student’s t tests (against zero) for the CAP occurrence rate changes were performed during the peak period of stimulus-evoked fMRI signal activity (4 to 6 s). Asterisks indicate significance at α < 0.05 after FDR correction. (B) Stimulus-induced CAP occurrence rate changes in healthy controls (n = 12), patients with MCS (n = 4), and patients with UWS (n = 6). (C) Spatial similarity of the eight CAPs between the main cohort and psychiatric cohort data. (D) Comparisons of the CAP occurrence rates for healthy control participants (CONTROL) versus schizophrenic (SCHZ), bipolar disorder (BIPOLAR), and attention deficit/hyperactive disorder (ADHD) patients by Student’s t tests. Red solid lines indicate significant group differences at α < 0.05 after FDR correction, and red dash lines indicate uncorrected significance at P < 0.05. Error bars indicate ±SD.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/6/11/eaaz0087/DC1

    Fig. S1. k-means clustering approach and evaluations of k.

    Fig. S2. CAPs identified from k = 2 to k = 16 with and without GSR.

    Fig. S3. Spatial characteristics of the CAPs.

    Fig. S4. Scatterplots and statistics for the Spearman rank correlations between the occurrence rate of joint mirror motifs or individual CAPs and the level of responsiveness.

    Fig. S5. CAP occurrence rates of different conditions from k = 4 to k = 12 with and without GSR.

    Fig. S6. In-degree CAP accessibility of different conditions from k = 4 to k = 12 with and without GSR.

    Fig. S7. Classifying conscious versus unresponsive states using SVM.

    Fig. S8. Conventional static functional connectivity.

    Fig. S9. The association between CAP occurrence rates and anticorrelations from k = 4 to k = 12 with and without GSR.

    Table S1. A summary of statistics for Fig. 2 (B to E).

    Movie S1. CAP temporal dynamics.

  • Supplementary Materials

    The PDF file includes:

    • Fig. S1. k-means clustering approach and evaluations of k.
    • Fig. S2. CAPs identified from k = 2 to k = 16 with and without GSR.
    • Fig. S3. Spatial characteristics of the CAPs.
    • Fig. S4. Scatterplots and statistics for the Spearman rank correlations between the occurrence rate of joint mirror motifs or individual CAPs and the level of responsiveness.
    • Fig. S5. CAP occurrence rates of different conditions from k = 4 to k = 12 with and without GSR.
    • Fig. S6. In-degree CAP accessibility of different conditions from k = 4 to k = 12 with and without GSR.
    • Fig. S7. Classifying conscious versus unresponsive states using SVM.
    • Fig. S8. Conventional static functional connectivity.
    • Fig. S9. The association between CAP occurrence rates and anticorrelations from k = 4 to k = 12 with and without GSR.
    • Table S1. A summary of statistics for Fig. 2 (B to E).

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    • Movie S1 (.mp4 format). CAP temporal dynamics.

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article