Impaired Frontoparietal Connectivity in Traumatic Individuals with Disorders of Consciousness: A Dynamic Brain Network Analysis

Abstract

Recent advances in neuroimaging have demonstrated that patients with disorders of consciousness (DOC) may retain residual consciousness through activation of a complex functional brain network. However, an understanding of the hierarchy of residual consciousness and dynamic network connectivity in DOC patients is lacking. This study aimed to investigate residual consciousness and the dynamics of neural processing in DOC patients. We included 42 patients with DOC, categorized by aetiology. Event-related potentials combined with time-varying electroencephalography networks were used to probe affective consciousness in DOC and examine the related network mechanisms. The results showed an obvious frontal P3a component among patients in minimally conscious state (MCS), while a prominent N1 was observed in unresponsive wakefulness syndrome (UWS). No late positive potential (LPP) was detected in these patients. Next, we divided the results by aetiology. Patients with nontraumatic injury presented an obvious frontal P3a response compared to those with traumatic injury. With respect to the dynamic network mechanism, patients with UWS, both with and without trauma, exhibited impaired frontoparietal network connectivity during the middle to late emotion processing period (P3a and LPP). Surprisingly, unconscious post-traumatic patients had an evident deficit in top-down connectivity. This, it appears that early automatic sensory identification is preserved in UWS and that exogenous attention was preserved even in MCS. However, high-level cognitive abilities were severely attenuated in unconscious patients. We also speculate that reduced frontoparietal connectivity may be useful as a biomarker to distinguish patients in an MCS from those with UWS given the same aetiology.

Keywords: consciousness; event-related potentials; frontoparietal network; traumatic brain injury.

Figure 1.

The spatial locations of the selected electrodes. (A) The electrodes used for dynamic network analysis. (B-D) The spatial locations of the electrodes: (B) back, (C) upper, and (D) right.

Figure 2.

Grand average ERP to target stimuli at electrodes Fz, Cz, and Pz. (A) An obvious N1 and P3a evoked by emotional sound in MCS patients. (B) A significant emotion-evoked N1 in UWS subjects (P<0.05, FDR corrected).

Figure 3.

Grand average ERP waveforms at electrode Fz. (A) Both traumatic and nontraumatic MCS patients had increased P3a amplitudes at the frontal-central electrode. (B) The waveform difference between traumatic and nontraumatic groups with the same diagnosis. P3a amplitudes were enhanced in the nontraumatic participants compared with the traumatic participants. The grey bars indicate regions of significant difference between conditions (P <0.05, FDR corrected). (C) The scalp topography of P3a. Positive activation was detected between 280 ms and 320 ms in frontoparietal electrodes for nontraumatic MCS and UWS. NT=nontraumatic, T=traumatic.

Figure 4.

Significantly different network patterns in four conditions (rows) and for five time points (columns). Red lines illustrate increased connectivity in MCS compared to UWS, and blue lines illustrate decreased connectivity in MCS. The arrows indicate the direction of information flow. From 300 ms to 1000 ms, both traumatic MCS (A) and nontraumatic MCS (B) showed increased frontoparietal connectivity. (C) and (D) denote networks with stronger connectivity in UWS than MCS patients; the results from trauma patients are presented in the upper row (C), and those from nontraumatic UWS patients with occipital activation are in the bottom row (D).

Figure 5.

Time-varying network comparisons between traumatic and nontraumatic patients. Red lines illustrate increased connectivity in nontraumatic patients, and blue lines illustrate decreased connectivity. Increased connection from frontal regions to temporoparietal regions between 300 and 1000 ms was observed in nontraumatic MCS (A) and nontraumatic UWS (B). (C) Significant activation in the occipital area in traumatic MCS. (D) No significant increased network connectivity in traumatic UWS.

Figure 6.

Traumatic patients showed a strong correlation between brain network properties and total CRS-R scores.