Frontiers in Human Neuroscience最新文献

Background: Upper limb motor impairment after stroke is a major cause of limitations in daily activities and reduced quality of life. Although traditional rehabilitation is effective, it is often insufficiently intensive and lacks focus on fine motor activation. Portable soft exoskeletons offer a promising approach to intensify recovery. This trial aimed to evaluate the effectiveness of the ReHand robotic system in subacute stroke rehabilitation.

Methods: This Randomized Controlled Trial (RCT) included 120 patients in the subacute period of stroke. Participants were stratified by age and motor deficit severity and randomized into the intervention group (robotic therapy + standard therapy) or control group (standard therapy only). The intervention was delivered 5 times per week over 8 weeks. The primary outcome was change in Fugl-Meyer Assessment for Upper Extremity (FMA-UE); secondary outcomes included Barthel Index (BI), Functional Independence Measure (FIM), National Institutes of Health Stroke Scale (NIHSS), modified Wolf Motor Function Test (mWMFT), Frenchay Arm Test (FAT), Disabilities of the Arm, Shoulder and Hand (DASH), and Hospital Anxiety and Depression Scale (HADS). Statistical analysis was performed in Python (v3.11). Normality was assessed with the Shapiro-Wilk test; the Mann-Whitney test was used for intergroup comparisons; Pearson's χ² test with Yates' correction was used for categorical variables. Effect size was calculated using Cliff's delta; rank-ANCOVA was performed via the Quade method. Significance was set at p < 0.05.

Results: The intervention group demonstrated significantly greater improvements in upper limb motor impairment compared to the control group, as reflected by larger gains in FMA-UE scores. Clinically meaningful improvements (≥5.25 points) were observed in 91.7% of patients in the intervention group versus 43.3% in the control group. All secondary outcomes also showed significant improvements in the intervention group (p < 0.001). No adverse events were reported in either group.

Conclusion: When combined with standard therapy, the ReHand robotic system may enhance upper limb recovery after stroke and appears to be a safe and feasible adjunct to multidisciplinary rehabilitation programs.

Clinical trial registration: ClinicalTrials.gov, identifier (NCT06937346).

The concept of connectome gradients, which represents the continuous spatial variation of brain connectivity, offers a robust framework for exploring the hierarchical organization of the cortex and its relationship with cognitive function. We hypothesize that structural gradients in frontal and parietal regions play a significant role in shaping individual cognitive abilities during early childhood. To evaluate this hypothesis, we identified macroscale structural connectome gradients in children aged 1-6 years, where the principal gradient exhibited a left-to-right axis, and the secondary gradient exhibited an anterior-to-posterior axis. Next, we employed machine learning approaches to predict the future cognitive outcomes assessed at ages 4, 6, and 8, specifically intelligence quotient (IQ), based on the structural connectome gradients measured at age 1. We achieved consistent and robust prediction results (mean Spearman's correlation > 0.25). The regional relevance maps highlighted regions in control network, and associated sensory processing networks. Our findings indicate that the structural connectome, which undergoes maturation during early childhood, plays a crucial role in the individual variability of IQ observed in early and middle childhood. Our approach underscores the utility of structural gradients as compact and interpretable representations of the brain's complex network architecture, effectively capturing individual differences that contribute to cognitive development.

[This corrects the article DOI: 10.3389/fnhum.2025.1629499.].

Brain-Computer Interfaces (BCIs) have traditionally been studied in clinical and laboratory contexts, but the rise of consumer-grade devices now allows exploration of their use in daily activities. Virtual reality (VR) provides a particularly relevant domain, where existing input methods often force trade-offs between speed, accuracy, and physical effort. This study introduces NeuroGaze, a hybrid interface combining electroencephalography (EEG) with eye tracking to enable hands-free interaction in immersive VR. Twenty participants completed a 360° cube-selection task using three different input methods: VR controllers, gaze combined with a pinch gesture, and NeuroGaze. Performance was measured by task completion time and error rate, while workload was evaluated using the NASA Task Load Index (NASA-TLX). NeuroGaze successfully supported target selection with off-the-shelf hardware, producing fewer errors than the alternative methods but requiring longer completion times, reflecting a classic speed-accuracy tradeoff. Workload analysis indicated reduced physical demand for NeuroGaze compared to controllers, though overall ratings and user preferences were mixed. While the differing confirmation pipelines limit direct comparison of throughput metrics, NeuroGaze is positioned as a feasibility study illustrating trade-offs between speed, accuracy, and accessibility. It highlights the potential of consumer-grade BCIs for long-duration use and emphasizes the need for improved EEG signal processing and adaptive multimodal integration to enhance future performance.

This study examined the use of electroencephalography (EEG) to measure real-time cognitive engagement during English-speaking tasks among Thai undergraduate students. A total of 41 participants took part in a mock job interview while wearing a five-channel EEG headset, which captured brain activity in the beta and gamma frequency bands. The EEG data were analyzed to identify variations in neural activation associated with speaking performance. Descriptive analyses revealed that participants with higher speaking scores generally exhibited increased beta and gamma wave activity, indicating greater attentional focus and semantic processing. However, nonparametric tests showed no statistically significant differences in neural activation between the lowest- and highest-scoring groups, suggesting that cognitive engagement may not always correspond directly with observable speaking performance. These results emphasize that even lower-performing learners may exhibit strong mental effort during communication, and that EEG can serve as a valuable tool for understanding hidden cognitive processes and informing more personalized language instruction and assessment.

Background: The bidirectional communication between the brain and gut in Crohn's disease is increasingly acknowledged, highlighting how gut inflammation can influence brain function and psychological health, and vice versa, through the gut-brain axis. The insula is critical for processing pain, its emotional evaluation, and for regulating neurometabolites involved in these processes. The role of insular neurometabolites in gastrointestinal symptoms, particularly pain, in Crohn's disease patients, however, is not well-understood, highlighting the need for further investigation. Therefore, this study aims to enhance our understanding of the connection between Crohn's disease and brain function by investigating neurometabolic profiles in the insula of patients with Crohn's disease.

Methods: In this study, 7 Tesla proton magnetic resonance spectroscopy (¹H-MRS) was utilized to examine the left insular cortex in 14 individuals with Crohn's disease and 14 age- and gender-matched healthy controls during resting state. Participants also completed neuropsychological evaluations, including the Gastrointestinal Symptom Rating Scale (GSRS) and the Pain Catastrophizing Scale (PCS).

Results: No significant differences were found in the absolute concentrations of insular neurometabolites between Crohn's disease patients and healthy controls. However, in patients with Crohn's disease, GSRS scores were negatively correlated with the neurometabolites aspartate (Asp) and N-acetylaspartylglutamate (NAAG) in the insula. Furthermore, significant positive correlations were observed between scores on the PCS magnification subscale and concentrations of neurometabolites-namely glutamine (Gln) and the combined glutamate and glutamine signal (Glx) -as measured by (¹H-MRS).

Conclusion: The neurometabolic alterations observed in the insular cortex of Crohn's disease patients suggest increased insular activity, which may enhance interoceptive awareness and pain sensitivity, potentially contributing to heightened pain catastrophizing.

Background: Developmental dyslexia is a common neurodevelopmental disorder that impairs reading ability despite adequate intelligence and education, affecting up to 17% of children worldwide. Advances in neuroscience have revealed complex mechanisms involving phonological, visual, and temporal processing, with cross-linguistic variability. At the same time, technological innovation is driving a shift toward AI-powered diagnostics, immersive learning tools, and neurostimulation-based interventions.

Methods: This narrative review synthesizes evidence from recent research published between 2015 and 2025, focusing on four thematic areas: (1) neurobiological underpinnings of dyslexia, (2) diagnostic innovations using AI and eye- or handwriting-based deep learning, (3) neurostimulation and immersive VR/AR interventions, and (4) policy, equity, and ethical considerations. Studies were identified through major academic databases and thematically analyzed to highlight trends, strengths, and limitations.

Results: AI-based diagnostic tools using eye-tracking and handwriting features have achieved reported accuracies exceeding 80% in multiple pilot studies. VR/game-based programs and neurostimulation interventions (TMS, tDCS) have shown promising short-term effects on reading fluency and phonological processing, though evidence for long-term literacy transfer remains limited. Across studies, methodological heterogeneity and small sample sizes constrain generalizability. Significant disparities in access persist across socioeconomic, linguistic, and geographic contexts.

Conclusions: While these technologies offer promising avenues for more personalized and scalable dyslexia care, their integration must be accompanied by stronger evidence, ethical safeguards, and equity-focused policies. Technology should augment, not replace, human interaction in inclusive education. Future research should prioritize larger trials, cross-linguistic validation, and sustainable implementation strategies.

Few laboratory studies permitting granular analyses of alcohol use on neurobehavioral processes in older adults have been reported. This study, reporting baseline data from an on-going longitudinal project, seeks to address this gap. Toward that end, working memory (WM) processes were targeted using the continuous recognition version of the Mnemonic Similarity Task (MST). Healthy male and female drinkers aged 65-80 years completed self-report measures of substance use, negative affect and demographics prior to testing. Drinking patterns were quantified on the basis of typical standard drinks/day (D/D). Behavioral data were obtained in a two-button forced choice paradigm. Neurophysiological data were obtained for each stimulus presentation with analyses focusing on a mid-frontal negative shift occurring ∼ 300-500 ms post stimulus (FN400) and a posterior positive shift occurring ∼ 550-800 ms after stimulus presentation (LPC). To constrain the models, for the behavioral analyses correlations between D/D, measures of negative affect, stimulus condition ("new," "identical," or "similar") and performance were conducted. They indicated that only accuracy in labeling "new" items was related to D/D. Subsequent least squares regression revealed that D/D was inversely related to accuracy for new items. In a sensitivity analysis removing THC users, the D/D effect was retained. Correlations incorporating mean amplitudes for the FN400 and LPC failed to reveal identifiable patterns. Consequently, separate mixed models (e.g., stimulus condition) for FN400 and LPC were conducted. D/D was not predictive of the FN400 for any stimulus condition. It was negatively related to the LPC mean amplitude. In post-hoc analyses, the effect was most notable for "new" stimuli. After removing THC users, the magnitude and direction of the D/D effect was retained, although the p value fell short of significance. Primary models failed to reveal sex main or interaction effects. However, exploratory post-hoc analyses justify their continued study. These data lend preliminary support for the hypothesis that sustained drinking among older adults may negatively impact neurobehavioral processes. They are also consistent with expectations that alcohol effects may be modest and constrained by specific process. Importantly, these outcomes will be expanded through on-going longitudinal study, extending investigation to study of alcohol-related cognitive decline.