Factors associated with adverse pregnancy outcome and psychological outcomes during re-pregnancy: a cross-sectional study.

Disorders of consciousness (DoC) refer to a state in which a person’s awareness of themselves and their environment is significantly diminished or absent following a severe brain injury (, ). The etiology of impaired consciousness is typically categorized as either traumatic brain injury (TBI) or non-traumatic brain injury (NTBI) (). TBI results from external mechanical forces, whereas NTBI is generally related to medical conditions such as intracranial hemorrhage, cardiac arrest, or ischemic or hemorrhagic stroke (). According to the degree of impairment, patients with prolonged DoC are classified as either unresponsive wakefulness syndrome (UWS) or minimally conscious state (MCS) (, ). Patients with UWS exhibit sleep–wake cycles without conscious awareness, whereas patients with MCS demonstrate reproducible signs of consciousness, such as following commands or exhibiting purposeful responses (, ). Accurate diagnosis and reliable prognostic predictions for these patients remain a persistent challenge in contemporary clinical neuroscience and rehabilitation medicine. Traditional behavioral assessment scales, such as the Coma Recovery Scale-Revised (CRS-R) (), have inherent limitations when it comes to detecting cognitive functions in patients, particularly those with sensory or motor impairments (, ). To address these limitations, recent studies have increasingly used brain imaging or electrophysiological techniques to detect residual brain activity in patients with DoC (). Among these techniques, electroencephalography (EEG) has gradually become the focus of research due to its portability, non-invasiveness, and feasibility in routine clinical settings (, ). There are typically three types of EEG-based experimental paradigms used to assess preserved consciousness in patients with DoC: active paradigms involving task execution; passive paradigms based on stimulus-evoked brain responses; and resting-state assessments that infer the conscious state from spontaneous neural activity (). Unlike active tasks, which require sustained attention or the ability to follow instructions, passive paradigms allow neural processing to be assessed without demanding high-level cognitive engagement or active participation. Moreover, passive paradigms provide more direct, stimulus-related evidence of preserved cognitive processing than resting-state assessments do (). Given the varying levels of brain injury among patients with DoC, passive paradigms appear to be a promising approach for assessing residual cognitive function. Statistical learning (SL), which is commonly assessed in a passive paradigm, shows unique clinical potential. SL refers to the ability to extract patterns and structures from sensory input without explicit instruction or feedback (, ). This ability was first demonstrated in studies of infants using auditory speech sequences containing repeated trisyllabic pseudowords (), which showed that infants could segment words by tracking transitional probabilities (TP) between syllables, thereby learning the underlying patterns. TP quantifies the likelihood that a given element (e.g., a syllable or tone) follows another specific element within a sequence (). Over the following decades, researchers have demonstrated that SL occurs across multiple sensory modalities, such as vision () and touch (), and across the lifespan, from infancy and childhood to adulthood (). It has also been observed in various populations such as healthy individuals and those with post-stroke aphasia (), confirming its role as a universal, cross-domain cognitive mechanism. Crucially, SL does not require overt guidance or high cognitive effort. It can occur through passive exposure to structured input, without demanding active participation or explicit instructions (). In contrast, explicit learning demands active participation, sustained attention, and the ability to follow commands (). These capacities are often severely impaired in patients with DoC. This makes SL technically well-suited for investigating residual cognitive processing in patients with DoC. Recent work has highlighted the potential of SL paradigms in revealing residual cognitive abilities in patients with DoC. For example, Xu et al. () used a frequency-tagging EEG approach involving real and reversed disyllabic word sequences, revealing that patients with MCS and patients with emergence from MCS exhibited neural tracking of transitional probabilities between syllables. Similarly, Benjamin et al. () showed that some patients with MCS could learn statistical regularities in trisyllabic pseudowords when presented with structured versus random auditory streams, whereas patients with UWS showed only a weak trend. Together, these findings highlight the potential of SL paradigms for probing residual cognitive processing in patients with DoC. However, the extent to which SL-related neural measures can be used as reliable quantitative tools for clinical diagnosis and prognosis in pa