The need for attention to enable statistical learning is debated. Testing individuals with impaired consciousness offers valuable insight, but very few studies have been conducted due to the difficulties inherent in such studies. Here, we examined the ability of patients with varying levels of disorders of consciousness (DOC) to extract statistical regularities from an artificial language composed of randomly concatenated pseudowords by measuring frequency tagging in EEG. The objectives were firstly, to assess the automaticity of the segmentation process and the correlations between the level of covert consciousness and statistical learning capacities; secondly, to identify potential new diagnostic indicators. We observed that segmentation abilities were preserved in some minimally conscious patients, suggesting that auditory statistical learning is an inherently automatic low-level process. Due to significant inter-individual variability, word segmentation might not be robust enough for clinical use. In contrast, temporal accuracy of auditory syllable responses correlates strongly with coma severity.

Our brains constantly generate predictions of sensory input that are compared with actual inputs, propagate the prediction-errors through a hierarchy of brain regions, and subsequently update the internal predictions of the world. However, the essential feature of predictive coding, the notion of hierarchical depth and its neural mechanisms, remains largely unexplored. Here, we investigated the hierarchical depth of predictive auditory processing by combining functional magnetic resonance imaging (fMRI) and high-density whole-brain electrocorticography (ECoG) in marmoset monkeys during an auditory local-global paradigm in which the temporal regularities of the stimuli were designed at two hierarchical levels. The prediction-errors and prediction updates were examined as neural responses to auditory mismatches and omissions. Using fMRI, we identified a hierarchical gradient along the auditory pathway: midbrain and sensory regions represented local, shorter-time-scale predictive processing followed by associative auditory regions, whereas anterior temporal and prefrontal areas represented global, longer-time-scale sequence processing. The complementary ECoG recordings confirmed the activations at cortical surface areas and further differentiated the signals of prediction-error and update, which were transmitted via putative bottom-up γ and top-down β oscillations, respectively. Furthermore, omission responses caused by absence of input, reflecting solely the two levels of prediction signals that are unique to the hierarchical predictive coding framework, demonstrated the hierarchical top-down process of predictions in the auditory, temporal, and prefrontal areas. Thus, our findings support the hierarchical predictive coding framework, and outline how neural networks and spatiotemporal dynamics are used to represent and arrange a hierarchical structure of auditory sequences in the marmoset brain.

One of the biggest challenges in cognitive neuroscience is developing diagnostic tools for Disorders of Consciousness (DoC). Detecting dynamical connectivity brain states seems promising, specifically those linked to transient moments of enhanced cognitive states in patients. A growing body of evidence indicates that fMRI brain states properties are strongly modulated by the level of consciousness , as theoretically predicted by whole brain modeling. fMRI-based brain states, however, have very limited practical application due to methodological constraints. In this work we defined EEG-based brain states and explored their potential as a bedside, real-time tool to detect transient windows of enhanced brain states. We analysed data from 237 individual patients with chronic and acute DoCs -100 Unresponsive Wakefulness Syndrome (UWS), 96 Minimally Conscious State (MCS) and 41 acute- and 101 healthy controls obtained in three independent research centers (Fudan hospital in Shanghai, Pitié Salpêtrière in Paris and Purpan hospital in Toulouse). We determined five EEG functional connectivity brain states, and show that their probability of occurrence is strongly related to the level of consciousness. Distinctively, high entropy brain states are exclusively found in healthy subjects, while low-entropy brain states increase their probability with DoC’s severity, spanning from acute unarousable comatose state, to more chronic DoC’s patients, who are awake but show fluctuating (MCS) or absent awareness (VS). Furthermore, the brain state probability distribution of each individual subject —and even the presence of certain key brain states— significantly vary with the patients’ outcome. We also tested whether our procedure has an actual potential for real-time, bedside brain state detection, and proved that we can reliably estimate the concurrent brain state of a patient in real time, paving the way for a broad application of this tool for DoC patients’ diagnosis, follow-up, and neuroprognostication.

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.

Previous studies have examined the neural correlates for crossmodal paired-associate (PA) memory and the temporal dynamics of its formation. However, the neural dynamics for feedback processing of crossmodal PA learning remain unclear. To examine this process, we recorded event-related scalp electrical potentials for PA learning of unimodal visual-visual pairs and crossmodal visual-tactile pairs when participants performed unimodal and crossmodal tasks. We examined event-related potentials (ERPs) after the onset of feedback in the tasks for three effects: feedback type (positive feedback vs. negative feedback), learning (as the learning progressed) and the task modality (crossmodal vs. unimodal). The results were as follows: (1) feedback type: the amplitude of P300 decreased with incorrect trials and the P400/N400 complex was only present in incorrect trials; (2) learning: progressive positive voltage shifts in frontal recording sites and negative voltage shifts in central and posterior recording sites were identified as learning proceeded; and (3) task modality: compared with the unimodal PA learning task, positive voltage shifts in frontal sites and negative voltage shifts in posterior sites were found in the crossmodal PA learning task. To sum up, these results shed light on cortical excitability related to feedback processing of crossmodal PA learning.

Visual stimuli are known to activate the auditory cortex of deaf people, presenting evidence of cross-modal plasticity. However, the mechanisms underlying such plasticity are poorly understood. In this functional MRI study, we presented two types of visual stimuli, language stimuli (words, sign language, and lip-reading) and a general stimulus (checkerboard) to investigate neural reorganization in the superior temporal cortex (STC) of deaf subjects and hearing controls. We found that only in the deaf subjects, all visual stimuli activated the STC. The cross-modal activation induced by the checkerboard was mainly due to a sensory component via a feed-forward pathway from the thalamus and primary visual cortex, positively correlated with duration of deafness, indicating a consequence of pure sensory deprivation. In contrast, the STC activity evoked by language stimuli was functionally connected to both the visual cortex and the frontotemporal areas, which were highly correlated with the learning of sign language, suggesting a strong language component via a possible feedback modulation. While the sensory component exhibited specificity to features of a visual stimulus (e.g., selective to the form of words, bodies, or faces) and the language (semantic) component appeared to recruit a common frontotemporal neural network, the two components converged to the STC and caused plasticity with different multivoxel activity patterns. In summary, the present study showed plausible neural pathways for auditory reorganization and correlations of activations of the reorganized cortical areas with developmental factors and provided unique evidence towards the understanding of neural circuits involved in cross-modal plasticity.

Studies have indicated that a cortical sensory system is capable of processing information from different sensory modalities. However, it still remains unclear when and how a cortical system integrates and retains information across sensory modalities during learning. Here we investigated the neural dynamics underlying crossmodal associations and memory by recording event-related potentials (ERPs) when human participants performed visuo-tactile (crossmodal) and visuo-visual (unimodal) paired-associate (PA) learning tasks. In a trial of the tasks, the participants were required to explore and learn the relationship (paired or non-paired) between two successive stimuli. EEG recordings revealed dynamic ERP changes during participants' learning of paired-associations. Specifically, (1) the frontal N400 component showed learning-related changes in both unimodal and crossmodal tasks but did not show any significant difference between these two tasks, while the central P400 displayed both learning changes and task differences; (2) a late posterior negative slow wave (LPN) showed the learning effect only in the crossmodal task; (3) alpha-band oscillations appeared to be involved in crossmodal working memory. Additional behavioral experiments suggested that these ERP components were not relevant to the participants' familiarity with stimuli per se. Further, by shortening the delay length (from 1300ms to 400ms or 200 ms) between the first and second stimulus in the crossmodal task, declines in participants' task performance were observed accordingly. Taken together, these results provide insights into the cortical plasticity (induced by PA learning) of neural networks involved in crossmodal associations in working memory.

The neural processes underlying pain memory are not well understood. To explore these processes, contact heat-evoked potentials (CHEPs) were recorded in humans with electroencephalography (EEG) technique during a delayed matching-to-sample task, a working memory task involving presentations of two successive painful heat stimuli (S-1 and S-2) with different intensities separated by a 2-s interval (the memorization period). At the end of the task, the subject was required to discriminate the stimuli by indicating which (S-1 or S-2) induced more pain. A control task was used, in which no active discrimination was required between stimuli. All event-related potential (ERP) analysis was aligned to the onset of S-1. EEG activity exhibited two successive CHEPs: an N2-P2 complex (∼400 ms after onset of S-1) and an ultralate component (ULC, ∼900 ms). The amplitude of the N2-P2 at vertex, but not the ULC, was significantly correlated with stimulus intensity in these two tasks, suggesting that the N2-P2 represents neural coding of pain intensity. A late negative component (LNC) in the frontal recording region was observed only in the memory task during a 500-ms period before onset of S-2. LNC amplitude differed between stimulus intensities and exhibited significant correlations with the N2-P2 complex. These indicate that the frontal LNC is involved in maintenance of intensity of pain in working memory. Furthermore, alpha-band oscillations observed in parietal recording regions during the late delay displayed significant power differences between tasks. This study provides in the temporal domain previously unidentified neural evidence showing the neural processes involved in working memory of painful stimuli.