Sleep-like cortical dynamics during wakefulness and their network effects following brain injury

This Perspective addresses how brain injuries lead to large-scale functional network disruption and behavioral impairments beyond local structural damage. Reviving and extending the concept of diaschisis, the authors propose that EEG slow waves observed after injury reflect sleep-like cortical bistability intruding during wakefulness. They frame slow waves as transient neuronal OFF-periods that disrupt effective connectivity, offering a parsimonious mechanism for widespread network alterations seen in neuroimaging. The study emphasizes the potential reversibility of these dynamics and suggests they are a tractable target for therapeutic modulation to improve recovery.

Historical observations (Broca, Brown-Séquard, von Monakow) established that deficits after focal lesions can produce remote dysfunction (diaschisis). Modern imaging shows that average stroke lesions (~10 ml) alter ~20% of brain connections, up to ~50% in severe cases, typically with ipsilesional hypersynchrony and reduced interhemispheric coupling. Early clinical EEG work (Walter, 1937) consistently linked focal lesions to delta slowing during wakefulness, but this was later overshadowed by CT/MRI. The neuronal basis of slow waves was clarified by Steriade et al. (1993), demonstrating cortical bistability with alternating ON (depolarized firing) and OFF (hyperpolarized silent) states. Mechanistic components include reverberation, activity-dependent adaptation (Na+- and Ca2+-dependent K+ currents), inhibition, and endogenous noise. Conditions promoting bistability include reduced neuromodulation (sleep, anesthesia), disfacilitation, increased inhibition, and deafferentation. Scalp EEG can infer OFF-periods via delta wave polarity and suppression of >20 Hz activity. Perturbational studies (TMS-EEG; intracortical stimulation) and PCI show breakdown of effective connectivity during sleep/anesthesia due to evoked OFF-periods, with converging evidence from animal models, cortical slices, and computational simulations. Recent clinical studies connect these mechanisms to brain injury: widespread bistability in UWS/VS, perilesional sleep-like dynamics after focal stroke, and mesoscale SEEG evidence of perilesional generation and long-range percolation of slow waves following controlled lesions. Additional literature explores relationships between slow waves, fMRI network alterations, and behavioral impairments, as well as possible protective roles of slow waves during sleep in recovery.

As a Perspective, the paper synthesizes evidence across methodologies rather than reporting a single empirical protocol. Methods reviewed include: (1) Scalp EEG and quantitative spectral analyses (delta power, delta/alpha ratio, spectral exponent) to detect slowing and infer OFF-periods via suppression of high-frequency (>20 Hz) activity. (2) Perturbational approaches using MRI-guided transcranial magnetic stimulation (TMS) combined with EEG to assess cortical reactivity, phase-locking, evoked slow waves, and OFF-periods; network complexity quantified with the perturbational complexity index (PCI). (3) Intracranial techniques: intracortical electrical stimulations with local field potential recordings in humans; stereo-EEG (SEEG) to examine perilesional bistability and propagation after radiofrequency thermocoagulation (RFTC), leveraging pre-lesion baselines. (4) Resting-state fMRI to quantify network alterations (hypersynchrony within ipsilesional hemisphere, decreased interhemispheric coupling) and relate them to lesion topology and slow wave dynamics. (5) Animal models employing chemogenetic inactivation to mimic deafferentation, tracking delta wave generation and propagation with electrophysiology and concurrent fMRI connectivity changes. (6) Computational modeling: mean-field models of coupled cortical modules demonstrating how increased adaptation produces OFF-periods and disrupts causality; whole-brain models based on human connectomes to study global effects. These methods collectively probe causal input-output alterations, bistability, network disruption, and potential reversibility across species and scales.

- EEG slow waves after brain injury reflect sleep-like cortical bistability during wakefulness, characterized by neuronal OFF-periods that suppress high-frequency activity and disrupt effective connectivity. - Perturbational evidence: During wakefulness, TMS triggers complex, long-range interactions; during NREM sleep/anesthesia, responses are local, stereotyped slow waves with OFF-periods and low PCI. In UWS patients, similar sleep-like OFF-periods occur across stimulated regions despite open eyes, with markedly reduced phase-locking and PCI, indicating widespread bistability in intact-appearing cortex. - Recovery linkage: Longitudinal TMS-EEG in UWS shows decreasing OFF-period duration/depth and increasing causal interactions and PCI paralleling recovery of consciousness. Intracranial stimulation studies after traumatic brain injury show comparable normalization with recovery. - Stroke: Awake patients show perilesional EEG slowing; TMS-EEG reveals local sleep-like responses with slow waves and >20 Hz suppression over perilesional cortex (OFF-periods), disrupting local cortico-cortical interactions and correlating with deficits. Reduction of perilesional sleep-like dynamics associates with recovery of interactions and clinical improvement. - Latent bistability: Even when spontaneous EEG slow waves are not evident, cortical stimulation can reveal evoked OFF-periods, demonstrating that bistability can be functionally significant yet latent. - Spatial extent: SEEG after RFTC shows perilesional slow waves during wakefulness within ~28–30 mm radius around lesion (about 10× lesion size) with OFF-periods, and propagation to distant connected sites up to ~60 mm, correlating with pre-lesion effective connectivity maps. - Network-level effects: Animal chemogenetic deafferentation induces delta waves that propagate with consistent time lags along anterior-posterior axes, increasing low-frequency inter-areal coherence; concurrent fMRI shows functional overconnectivity mirroring ipsilesional hypersynchrony seen in human stroke. - Epidemiological imaging: Resting-state fMRI indicates an average ~10 ml stroke lesion alters ~20% of brain connections, up to ~50% in severe injury; typical patterns include ipsilesional hypersynchrony and decreased interhemispheric coupling. - Mechanisms: Postlesional slow waves arise from decreased ascending neuromodulation (enhanced adaptation), disfacilitation due to loss of lateral excitation (shifted E/I balance), inflammatory and metabolic factors (IL-1, TNFα, microglial modulation of norepinephrine, ATP-sensitive K+ channel activation under hypoxia). - Behavioral relevance: In healthy sleep-deprived subjects/animals, local slow waves predict motor failures, sensory processing changes, and attentional lapses with topographical specificity, suggesting stronger clinical impacts in patients when OFF-periods encompass larger neuronal populations and propagate. - Potential benefits: Sleep slow waves support restorative functions (metabolic downscaling, glymphatic flow, synaptic remodeling). Enhanced NREM slow waves during sleep relate to better stroke outcomes; perilesional slow wave boosting during sleep in rodents improves recovery, though wake-state slow waves may differ functionally.

The proposed framework links structural brain injury to secondary, potentially reversible disruptions of neuronal dynamics via sleep-like cortical bistability during wakefulness. OFF-periods alter cortical input-output properties, causing neurons to transiently silence and lose causal linkage to inputs, which breaks down effective connectivity and network complexity. This mechanism parsimoniously explains remote network disruptions (diaschisis) observed with fMRI and EEG across injuries of differing etiologies. Evidence across TMS-EEG, SEEG, intracranial stimulation, animal models, and computational studies supports that postlesional slow waves can be locally generated and propagate to connected distant regions, with topologies matching known cortical connectivity. Clinically, widespread bistability correlates with loss of consciousness in UWS, while reduction of OFF-periods accompanies recovery, implying modifiability. In focal stroke, perilesional bistability associates with specific deficits and improves with rehabilitation-linked normalization, suggesting targeted interventions to reduce bistability and restore wake-like network dynamics. The framework underscores that while structural damage is fixed, functional dynamics and their network consequences can change, offering therapeutic windows to reawaken functionally offline circuits and optimize rehabilitation.

This Perspective advances an integrated, mechanistic account of post-injury network disruption: structural lesions reduce ascending neuromodulation and/or lateral excitation, engaging cortical bistability and slow waves during wakefulness; these dynamics can propagate along connectivity backbones, producing topologically specific network interference and behavioral impairments. Importantly, bistability and its consequences are potentially reversible. The work suggests actionable paths for therapy: measuring and mapping sleep-like dynamics with perturbational tools (e.g., PCI), targeting perilesional and network nodes to reduce OFF-periods, and employing neuromodulation (drugs, transcranial stimulation, temporal interference, closed-loop neurofeedback) guided by individualized in silico models. Future research should: (1) systematically link slow wave propagation topographies to fMRI network changes in patients; (2) predict behavioral deficits from spatial-temporal patterns of slow waves; (3) delineate critical periods when wake-state slow waves could be beneficial versus detrimental; (4) probe subcortical participation in propagation; and (5) develop precision strategies to allow local protective slow waves acutely while limiting distant spread and chronic persistence.

The Perspective synthesizes diverse studies without presenting a single new empirical dataset, so causal links in patients remain partly inferential. Direct, systematic demonstrations in clinical populations of how perilesional slow waves map onto specific fMRI network alterations and behavioral deficits are still lacking. The protective versus detrimental roles of slow waves during wakefulness post-injury are unresolved and may depend on timing, location, and brain state. Spatial resolution limits of scalp EEG and variability in lesion characteristics complicate mapping. Mechanistic contributions (neuromodulatory loss, disfacilitation, inflammation, metabolic factors) likely co-occur and are difficult to disentangle without longitudinal multimodal data. Generalizability across etiologies, severity, and chronicity requires further validation, and the feasibility, safety, and efficacy of targeted neuromodulation to reduce pathological slow waves in humans remain to be established.