Persistent debate surrounds whether the frontal lobe supports the emergence or reporting of consciousness, raising the hypothesis that distinct frontal subregions may support these processes. We addressed this by combining electroencephalography (EEG) with eye-tracking in Report and No-Report paradigms. Eye-movement features distinguished conscious and unconscious trials in the no-report task. Event-related potential analyses showed that the Visual Awareness Negativity (VAN) was independent of reporting, whereas P3b occurred only with explicit reports. Importantly, the frontal Dorsal Attention Network (DAN) supported the emergence of consciousness, independent of post-perceptual reporting, as shown by multivoxel pattern analysis showing that a classifier's ability to decode visual consciousness generalized bidirectionally between report and no-report tasks. In contrast, frontal components of the Default Mode Network (DMN) and Frontoparietal Control Network (FPN) encoded visual consciousness only when explicit reports were required, indicating roles in reporting. These findings demonstrate a functional dissociation within the frontal lobe and refine the anatomical framework for the neural basis of visual consciousness.
Research efforts spanning more than seven decades have used functional neuroimaging to investigate whether putative extra‑sensory perception (ESP/receptive psi) has identifiable neural correlates. However, the field lacks a coherent and critical synthesis of its methodological approaches and reported effects. We conducted a systematic review consolidating 143 reports and qualitatively evaluated the methods of 129 individual studies. We organized studies by paradigm using two broad categories: (1) explicit psi paradigms (including forced-choice and free-response design subcategories), in which psi is assessed via overt responses; and (2) implicit psi paradigms (including distant stimulation, distant intentionality, and predictive anticipatory activity subcategories), in which psi is assessed solely via neurophysiology. Most work relied on EEG (91%), followed by fMRI (5%). We identified recurrent methodological limitations, in particular, small sample sizes, inadequate multiple‑comparison control, and analytical flexibility. Explicit paradigms rarely showed above‑chance behavior, yet investigators frequently proceeded to neural analyses, implying unacknowledged shifts in the operational definition of psi. Implicit paradigms often reported psi‑consistent effects, but findings were heterogeneous and seldom replicated. Overall, definitive conclusions about the neural correlates of ESP remain premature. Nevertheless, we identify potential leads-such as alpha‑band power in forced-choice designs, and a target‑related negative slow wave in event‑related designs-as testable candidates for future research. We provide a set of 13 methodological recommendations to promote cumulative progress, as well as 6 recommendations for future research directions.
Dynamic PET imaging with 11C-UCB-J enables in vivo quantification of synaptic vesicle glycoprotein 2A (SV2A), with prior reports of lower synaptic density in areas such as the brainstem nuclei and substantia nigra (SN) in Parkinson's disease (PD). Lowering PET dose reduces radiation exposure but increases noise and compromises quantification. This study evaluated a self-supervised two-step deep image prior (TS-DIP) denoising method for SV2A PET using 1/10 of the standard dose. Thirty healthy controls (HCs) and 30 PD patients underwent 60-minute PET scans to acquire full-count list-mode data, later down-sampled into ten independent 1/10-count dynamic datasets. TS-DIP was applied to denoise reduced-count frame images, and binding potential (BPND) maps were estimated. Performance was assessed by comparing group differences and correlations with motor severity against full-count results. Full-count data showed significant lower BPND in SN (-39%, p = 0.003) and red nucleus (RN; -27%, p = 0.009) in PDs versus HCs. With 1/10-count data, SN differences remained significant, but RN differences were inconsistent. TS-DIP introduced minimal bias, restored statistical significance across all noise realizations, and improved recovery of correlations with motor scores (SN: r = -0.43 ± 0.02; RN: r = -0.42 ± 0.04) compared with those from unprocessed 1/10-count data. Dynamic SV2A PET imaging at substantially reduced doses is feasible when combined with advanced DL-based denoising techniques such as TS-DIP, supporting its potential for broader clinical application.
Non-ordinary states of consciousness (NOC) offer a way to examine how large-scale brain dynamics reorganize as experience changes. We studied a participant able to reliably enter a self-induced NOC state characterized by vivid imagery, altered bodily perception, and a sense of unity. Across 20 fMRI sessions, we measured functional connectivity in four conditions (Baseline, Transition, NOC, and Residual) and compared the results with a matched control group. During the Transition phase, connectivity became more variable, indicating a temporary destabilization of network organization. In the NOC state, inter-network connectivity decreased broadly, with visual cortex showing reduced coupling to auditory, sensorimotor, orbitofrontal, thalamic, and cerebellar regions, and the somatomotor-dorsal network disengaging from auditory and language cortices, paralleling the reported visual phenomena and changes in bodily experience. In contrast, frontoparietal and salience networks showed increased coupling with precuneus/posterior cingulate, multimodal temporal cortex, and cerebellar hubs, in agreement with subjective reports of sustained inward-directed attention and stable absorption. Entropy and complexity analyses revealed systematic shifts that tracked the experiential sequence and returned to baseline in the Residual condition. This single-case study brings together something uncommon: controlled experimentation, voluntary induction of NOC states, and rich phenomenological data. Taken together, these elements offer a strong foundation for neurophenomenological research and illustrate why pairing structured paradigms with lived experience is useful for understanding non-ordinary states of consciousness.
Parkinson's disease (PD) is a complex neurodegenerative disorder characterized by both motor and non-motor symptoms, primarily attributed to dopaminergic dysfunction in the substantia nigra pars compacta. However, growing evidence indicates that serotonergic and noradrenergic alterations also contribute significantly to PD pathophysiology and progression. This growing understanding has driven the development of advanced neuroimaging techniques aimed at visualizing not only dopaminergic deficits but also serotonergic and noradrenergic alterations, providing deeper insights into PD pathophysiology and progression. Positron emission tomography (PET) and single-photon emission computed tomography (SPECT) have been instrumental in visualizing dopaminergic deficits, particularly dopamine transporter loss, aiding in diagnosis and disease progression tracking. While PET and SPECT remain crucial in assessing dopaminergic dysfunction, novel MRI techniques, including neuromelanin-sensitive MRI, iron-sensitive MRI, diffusion-weighted imaging, and pharmacological MRI, have emerged as promising non-invasive alternatives for evaluating monoaminergic dysfunction in PD. In this narrative review, we have discussed the recent neuroimaging advancements in assessing monoaminergic dysfunction in PD, highlighting how these advances not only improve our understanding of the distinct contributions of dopaminergic, noradrenergic, and serotonergic systems to motor and non-motor symptoms, but also offer promising biomarkers for disease progression and treatment response.
Our goal was to develop and validate machine learning models that are capable of fully automatic identification and segmentation of frontal, temporal, and posterior horns, the body of the lateral ventricle, the third and fourth ventricle, as well as the atrium on either side. Patients shunted for hydrocephalus were included. Data from two centers was used for development/external validation, respectively. Manual labelling of ventricular subregions on computed tomography (CT) was performed. First, an object detection algorithm (YOLOv5) was trained. This allowed for precise cropping of the subregions that could then be used as input for a 2D U-Net. For comparison, a nnU-Net was also trained. Precision, recall, mean average precision 50 and 50-95 (mAP50; mAP50-95) were used as performance metrics for the YOLO algorithm. Dice score, Jaccard score, and 95th percentile Hausdorff distance assessed performance for the U-Net. 80 CTs from patients at our center were included, as well as 50 from a second center. The mean age was 68.59 ± 15.89 and 75.94 ± 4.17 for the first and second centers, and 43 (52.5%) and 30 (60%) were male. MAP 50, mAP50-95 was 0.728, 0.453 for internal and 0.274, 0.124 for external validation across all classes. Best mean Dice scores were 0.92 ± 0.1 and 0.90 ± 0.05 for the body of the left lateral ventricle. Automatic segmentation and volumetry of ventricles including their subregions was feasible with high precision on computed tomography, potentially helping the clinical evaluation of even subtle changes in ventricular volume.
Musical improvisation requires complex cognitive operations, including idea generation, retrieval, and evaluation, supported by continuous sensory feedback. The neural reorganization that accompanies sensory loss has been described, but its impact on improvisation remains poorly understood. We present a case study of Matthew Whitaker (MW), a blind jazz piano prodigy with retinopathy of prematurity, to examine how cortical plasticity supports musical expertise. Functional magnetic resonance imaging (fMRI) was used to probe neural activity during three paradigms: music perception, proprioceptive sound-to-space mapping, and improvisation. As no suitable control group exists for a blind musical prodigy, internal controls contrasted experimental conditions to a matched control task for each paradigm. Results revealed recruitment of occipital visual regions during music perception, engagement of fusiform face areas during proprioceptive keyboard navigation, and coordinated activation and deactivation of frontal and occipital regions during improvisation. These findings demonstrate extensive cross-modal plasticity. As a musician who is blind, MW has undergone functional neural reorganization, recruiting his unused visual areas to aid him in his musical pursuits, from listening to navigating the keyboard to improvising.
Language comprehension is a complex cognitive process that engages multiple brain networks across multiple timescales and frequencies. Coordination across networks requires dynamic shifts in neural frequency that allow for layered, hierarchical communication. Limitations in spatial and temporal resolutions across brain imaging modalities have historically limited characterization of the real-time frequency dynamics of widespread language comprehension networks. Our objective was to implement a novel fused fMRI-EEG and Continuous Wavelet Transform (CWT) analysis to identify frequency "fingerprints" for key language networks during naturalistic language comprehension (comparing connected passages to scrambled words). CWT was performed on the EEG components to analyze frequency power changes over the 1 s post-stimuli window. Joint independent component analysis revealed three components that showed significantly greater engagement during passages with spatial expression across canonical language regions, a left-lateralized default mode sub-network (DMN), and a bilateral dorsal angular gyrus DMN subnetwork. Frequency analysis revealed that the language component corresponded with prolonged theta with corresponding beta and gamma bursts; the first DMN component displayed beta-gamma bursts; and the second DMN component showed dominant alpha activity. Network frequency profiles also differentially predicted language comprehension outcomes: 1.) Subject-level frequency profile difference from the language component was correlated with recall performance, and 2.) Language network expression correlation with reading comprehension was found to be conditional on alpha-dominant DMN expression. Our findings provide evidence that canonical brain networks that support language comprehension exhibit distinct, time-dependent cross-frequency oscillation patterns which are predictive of language ability. This work operationalizes a new approach that traces multiscale neuronal oscillations to distinct spatial networks.
Deep learning algorithms optimize data by enhancing resolution and suppressing noise associated with biological knowledge. The root issue is that, for example, CNNs learning mathematical patterns from statistical correlations in the data without regard to biological cues whatsoever, and merely apply filters such as max pooling, never grasping what the biological cues they are supposed to investigate are. This blind procedure can indeed be in technical language; however, it does not help to identify meaningful insights into neuroimaging, where interpretability is essential, and such inadequacies pose a grave challenge. In our research, rather than depending on the CNNs and FCNs only for the feature extractions, we have integrated biologically motivated features into voxel-based morphometry as well as deep learning. Our goal is to analyze T1-weighted MRI scans and T2-Flair images to investigate the characteristics of gray matter, white matter, cerebrospinal fluid, and white matter Hyperintensity in patients with mild cognitive impairment (MCI) who lie on the spectrum between normal aging and Alzheimer's disease (AD). So we extracted critical structural features such as white matter Hyperintensity, gray matter volume, white matter volume, cerebrospinal fluid (CSF) volume, and cortical thickness. These are biologically meaningful biomarkers that reflect the neurodegenerative alterations directly. To validate our method, after the detection of biological features, we have converted them into 3-bit, 4-bit, 8-bit, and 16-bit images. These images were used as inputs for both FCN and CNN models to investigate the early symptoms of AD from classified intracranial features.
Skull-involving meningiomas remain ill-defined, resulting in heterogeneous classifications and terminology. Existing schemes mainly describe bone-dura relationships and often overlook the relative burden of the intracranial soft-tissue component. This study aimed to develop and apply a deterministic, atlas-normalized, MRI-based radiological framework for the standardized description of imaging-defined osteomeningiomas and to explore associations between compartmental tumor distribution, radiological phenotype, and clinical presentation. We retrospectively reviewed adults with skull-involving meningiomas at our tertiary neurosurgical center between 2000 and 2024. Tumors were segmented on contrast-enhanced T1-weighted MRI, normalized to MNI152 space, and classified using a deterministic voxel-based radiological framework across osseous, juxta-osseous/dural, and intradural compartments. Tumors were classified as primary osteomeningioma (POM; isolated osseous compartment involvement) or secondary osteomeningioma (SOM; osseous plus adjacent juxta-osseous/dural compartment involvement), with subtypes SOM-I (no intradural extension), SOM-IIA (all three compartments involved, with an osseous component equal to or greater than the intradural component), and SOM-IIB (all three compartments involved, with an intradural component greater than the osseous component). These imaging-defined categories were intended as radiological descriptors of compartmental tumor distribution rather than distinctions between microscopic osseous invasion, reactive hyperostosis, or osseous metaplasia. All analyses were performed at the tumor level, with a predefined sensitivity analysis restricted to one index tumor per patient. Exploratory multivariable logistic regression models were fitted for brain edema, epileptic seizure, raised intracranial pressure, and exophthalmos. A total of 168 tumors from 149 patients were analyzed. Distribution was POM in 6 cases (3.6%), SOM-I in 37 cases (22.0%), SOM-IIA in 57 cases (33.9%), and SOM-IIB in 68 cases (40.5%). Convexity predominated in POM but was less common in other subtypes. SOM-IIB had the largest intracranial soft tissue component (29.9 ± 30.2 cm3) and the highest rate of brain edema, whereas SOM-IIA had the largest osseous component (24.8 ± 25.9 cm3). Epileptic seizures and signs of raised intracranial pressure were most frequent in SOM-IIB, exophthalmos in SOM-I, and subcutaneous mass in POM. In exploratory adjusted analyses, SOM-IIB remained associated with brain edema, epileptic seizure, and raised intracranial pressure, whereas SOM-I remained associated with exophthalmos. This voxel-based, atlas-normalized MRI framework provides a radiological standardization for the description of skull-involving meningiomas. Rather than establishing histological proof of bone or dural invasion, it standardizes compartmental tumor burden across osseous, juxta-osseous/dural, and intradural spaces. In exploratory analyses, the proposed imaging-defined subtypes were associated with distinct clinicoradiological presentation patterns, which warrant further pathological, multimodal, and external validation.
Parkinson's disease (PD) is a neurodegenerative disease characterised by molecular and structural brain changes detectable through advanced imaging. Understanding alterations in neurotransmitter systems and synaptic density, and their clinical relevance, is critical for identifying disease-specific biomarkers and therapeutic targets. This study included 33 PD patients (27 idiopathic PD (iPD) and 6 LRRK2 mutation carriers) (5.2 ± 3.6 years from diagnosis, 2.1 ± 0.7 Hoehn & Yahr OFF state) and 25 healthy controls (HC). Longitudinal data were collected for 20 iPD and 22 HC (10-33 months post-baseline; 20.2 ± 7.3 months). Participants underwent clinical assessments, structural magnetic resonance imaging, 11C-UCB-J positron emission tomography (PET) to assess synaptic density, 11C-DASB PET to assess serotonin transporter density, and 123I-FP-CIT single-photon emission computed tomography to assess dopamine transporter density. Analyses included baseline group comparisons, clinical correlations, and longitudinal assessments. At baseline, lower 123I-FP-CIT uptake in caudate and putamen (p < 0.001) and reduced 11C-DASB binding in the insular cortex (p = 0.003), parietal lobe (p = 0.009), caudate (p < 0.001), and putamen (p = 0.002) were observed in PD compared to HC. Some baseline correlations emerged between imaging metrics and symptom scales in PD, though these were limited. Despite progression in motor impairment, autonomic dysfunction, and overall disability in PD, no significant longitudinal changes or group × time interactions were detected for molecular imaging measures. This study confirmed dopaminergic and serotonergic dysfunction in PD. Synaptic density did not differ between groups or change over time, suggesting synaptic loss may be minimal at mild-to-moderate disease stages. These findings highlight how different molecular imaging markers reflect distinct aspects and timescales of PD pathophysiology.
Subjects with temporal lobe epilepsy (TLE) often experience cognitive impairment in different domains. Currently, the mechanisms underlying neuropsychological dysfunction in TLE remain poorly understood. The main objective is to characterize the multivariate relationship between brain connectivity patterns and cognitive impairment detected by robotic testing in subjects with TLE. Kinarm robotic technology was used to evaluate motor, cognitive, and sensory domains of healthy controls and individuals with TLE. Structural connectivity (SC) and functional connectivity (FC) were obtained from multi-shell diffusion MRI and resting-state fMRI, respectively. After principal component analysis for dimension reduction of connectivity features, sparse canonical correlation analyses were used to identify the patterns of multivariate association between brain connectivity and cognitive dysfunctions. Patients with TLE demonstrated worse performance mainly in the domains of memory, executive function and attention, and to a lesser extent in the perceptual-motor domain. We found that memory and executive function alterations were associated with an intra-hemispheric SC pattern between somatomotor network and default, limbic and frontoparietal networks. We also found that an intra-hemispheric SC pattern of the posterior parietal cortex was related to perceptual-motor and attention skills with FC between this region and the precentral ventral region of DAN and frontal operculum insula of VAN also associated to impairment in these domains. This study identifies multivariate patterns of structural and functional connectivity that correlate with domain-specific cognitive impairment, as measured by robotic screening, in individuals with TLE. These findings support the conceptualization of TLE as a network disorder, contextualizing multidomain cognitive deficits within a network-level framework rather than interrogating specific functional circuits. This may in the future permit more personalized treatments or prediction of cognitive changes in response to planned treatment changes.
Functional MRI (fMRI) and structural MRI (sMRI) offer complementary insights into brain function and anatomy, but their integration for schizophrenia identification remains challenging due to modality heterogeneity. Many existing methods fall short of effective modeling of the interaction between two modalities. We propose CAMF, a Cross-Attentive Multi-modal Fusion framework that employs self-attention to capture intra-modal patterns and cross-attention to learn inter-modal relationships. In addition, we introduce the gradient-guided score-class activation map to enhance interpretability by highlighting salient features. Our approach significantly improves the accuracy in classifying schizophrenia, as demonstrated by the evaluation of multi-modal brain imaging datasets from four cohorts of schizophrenia studies. Furthermore, the model identifies functional networks and anatomical regions aligned with established biomarkers. CAMF provides an accurate and interpretable framework for multimodal brain imaging analysis, offering new insights into schizophrenia-related alterations.
Fatigue is a common and disabling symptom in multiple sclerosis (MS), quantifiable by patient reported outcome instruments such as the Fatigue Scale for Motor and Cognitive Functions (FSMC). The pathophysiology of fatigue remains poorly understood, and effective treatments are limited. Emerging evidence implicates disrupted excitation-inhibition balance in the premotor cortex as a potential culprit of fatigue in MS. Converging evidence now show that such network imbalance can be modulated with repetitive transcranial magnetic stimulation (TMS). The efficacy of premotor rTMS retuning excitation-inhibition balance, thus improving MS-related fatigue, has yet to be examined in a clinical trial. This randomized, double-blinded, sham-controlled, parallel-group trial investigates the efficacy of premotor TMS in treating fatigue in MS. Fifty-eight patients with MS will receive either active or sham TMS targeting the left dorsal premotor cortex (PMd). On five consecutive days, participants will undergo 30-min sessions using a novel low-frequency (0.72 Hz) paired-pulse repetitive TMS protocol with an interstimulus interval of 33 ms. The primary endpoint is the change in FSMC score 6 days post-intervention. Secondary outcomes include additional fatigue assessments and quantification of regional γ-aminobutyric acid (GABA) and glutamate concentrations of the targeted PMd, via ultra-high-field (7T) magnetic resonance spectroscopy. We hypothesize that active treatment will result in greater fatigue reduction than sham treatment and correlate positively with an increase in regional GABA in the stimulated premotor region. Exploratory endpoints include structural and functional connectivity changes assessed with 7T resonance imaging and motor cortical excitability changes measured with TMS. This study will assess the feasibility and efficacy of a novel low-frequency paired-pulse TMS protocol for fatigue in MS. Repeated neurophysiological measurements of cortical excitation-inhibition balance will yield mechanistic insights and guide future repetitive TMS trials targeting MS-related fatigue. http://www.clinicaltrials.gov, NCT06569550.
Cerebral small vessel disease (CSVD) is a primary contributor to vascular cognitive impairment. Although extensive research has examined white matter alterations in CSVD, cortical mechanisms underlying cognitive dysfunction remain incompletely characterized. To address this gap, we conducted a systematic review and meta-analysis of 26 studies investigating whether structure-cognition relationships in CSVD could be interpreted through biologically defined functional brain networks. By mapping structural features to the Yeo-7 functional atlas, we offer a network-based perspective on cognitive impairment in this population. Our integrated results demonstrate significant associations between structural alterations and all cognitive domains in CSVD patients. Notably, higher-order cognitive processes (e.g., executive function, attention and processing speed) involved more extensive functional networks than other domains. These findings help synthesize heterogeneous neuroanatomical literature on CSVD through contemporary network neuroscience frameworks, suggesting structure-cognition relationships may align with functional network architecture.
Motor asymmetry is a hallmark of Parkinson's disease (PD), but ~20% of patients present with symmetric motor signs, which are associated with faster disease progression and poorer dopaminergic response. The impact of motor symmetry on activities of daily living (ADL) outcomes following subthalamic deep brain stimulation (STN-DBS) remains unclear. We hypothesised that patients with symmetric PD experience less ADL improvement post-STN-DBS than asymmetric PD patients. This was a prospective, quasi-experimental, non-randomised, controlled, international multicentre study with a 6-month follow-up. The primary outcome was the Scales for Outcomes in Parkinson's Disease-Motor ADL scale. Secondary outcomes included Unified Parkinson's Disease Rating Scale motor examination and Parkinson's Disease Questionnaire-8 (PDQ-8). We defined symmetric PD as a right-to-left hemibody motor score equalling 1. We analysed within-group longitudinal changes, between-group outcome differences, effect size and correlations between PDQ-8 and motor changes. We confirmed results in a propensity-score matched subcohort with well-balanced demographic and clinical parameters. We included 200 patients with asymmetric and 54 with symmetric PD. In symmetric PD, ADL remained stable, which was not associated with the observed PDQ-8 improvement. In contrast, in asymmetric PD, ADL improved with a moderate effect size, which correlated moderately with PDQ-8 improvement. In symmetric PD, the absolute risk of experiencing no clinically relevant postoperative ADL improvement was 23.8% higher. This study provides class IIb evidence of worse ADL outcome of STN-DBS in patients with symmetric compared with asymmetric PD. Clinicians should counsel patients with symmetric PD on their elevated risk of ADL non-response when discussing STN-DBS as a treatment option.
Infantile hydrocephalus (IH) can lead to lasting brain structure changes despite early treatment. We investigated deep gray matter morphology in children with treated IH using volumetric analysis and 3D shape modeling. Twenty-one IH patients (diagnosed and treated within the first 2 years of life) and 21 age- and sex-matched controls underwent 3T MRI. We measured volumes of five bilateral subcortical structures (caudate, thalamus, putamen, pallidum, hippocampus) and performed shape analysis with spherical harmonic point distribution models (SPHARM-PDM) to map regional surface deformations. Group differences were assessed by t-tests for volume and vertex-wise general linear models for shape (false discovery rate q < 0.05). IH patients had significantly smaller volumes than controls in all examined structures (p < 0.001). Shape analysis revealed extensive localized differences in the caudate, thalamus, putamen, and pallidum. In IH, the surfaces adjacent to the enlarged ventricles bulged outward, while more distal parts showed inward compression (p < 0.05, corrected). The caudate head lateral surface was displaced outward in IH, whereas the caudate tail was medially compressed. The thalamus and pallidum similarly showed lateral expansion anteriorly and medial inward deformation posteriorly. The hippocampus exhibited a ∼25% volume reduction in IH but no significant regional shape differences. Early hydrocephalus results in persistent atrophy and shape distortion of deep gray matter structures. This is the first application of shape analysis in IH, revealing region-specific deformations consistent with mechanical stretching and compression from ventricular expansion. These findings underscore that measures of parenchymal integrity provide a more direct marker of hydrocephalus-related brain injury than ventricular size alone.
The Default Mode Network (DMN) is a large-scale intrinsic brain network critically involved in internally oriented cognition, including autobiographical memory. Core DMN regions such as the hippocampus and medial prefrontal cortex are central to memory retrieval, schema construction and self-referential processing. Individuals with Highly Superior Autobiographical Memory (HSAM) provide a unique model to investigate the neural mechanisms underlying exceptional memory ability. However, the intrinsic functional connectivity and temporal dynamics of the DMN in HSAM remain largely unexplored. To provide new insight into the baseline network mechanisms that supports HSAM irrespective of memory retrieval, in this study we examined both static and dynamic features of DMN functional architecture in 12 HSAM individuals and 31 matched controls during resting-state fMRI. Using a multilevel analytical framework encompassing link-level, node-level, and whole-network level measures, we characterized connectivity strength, temporal variability, and co-activation dynamics within the DMN. HSAM individuals showed enhanced and more temporally stable functional connectivity among memory-related, schema-related, and self-referential DMN regions, including the hippocampus, temporal pole, and ventromedial prefrontal cortex. These findings suggest that HSAM is associated with a more integrated and stable DMN organization, potentially supporting continuous memory replay and the consolidation of autobiographical experiences. This enhanced DMN coherence may represent a neural signature of HSAM.
Central variant posterior reversible encephalopathy syndrome (PRES) is a rare subtype (4% of cases) affecting brainstem and deep structures, presenting with severe hypertension but minimal neurological deficits, creating diagnostic challenges. A 34-year-old man presented with a two-month history of severe, refractory headache and malignant hypertension (256/150 mmHg). Brain MRI revealed diffuse T2-weighted and FLAIR hyperintensity in the pons and right middle cerebellar peduncle (MCP). Using serial arterial spin labeling (ASL) and diffusion tensor imaging (DTI), we provide the first longitudinal evidence of acute-phase biphasic hemodynamics in central variant PRES: concurrent pontine CBF of 27.1 mL/100 g/min (contralateral reference: 25.2 ± 1.8 mL/100 g/min) and right MCP CBF of 29.3 mL/100 g/min (contralateral reference: 39.1 ± 2.1 mL/100 g/min). This regional perfusion imbalance is consistent with autoregulatory failure and blood-brain barrier (BBB) compromise, leading to vasogenic edema. Critically, the coexistence of these opposing patterns reveals complementary hemodynamic phenotypes of cerebrovascular dysregulation. Post-treatment, pontine CBF was 25.3 mL/100 g/min (contralateral reference: 25.3 ± 1.8 mL/100 g/min) and right MCP CBF was 38.6 mL/100 g/min (contralateral reference: 39.2 ± 2.2 mL/100 g/min), both within the reference range of the contralateral regions, accompanied by progressive increase of fractional anisotropy (FA) on DTI at both the 6-day and 90-day follow-up imaging time points. This case demonstrates biphasic hemodynamic changes in central variant PRES, supporting autoregulatory failure as the mechanism. Serial ASL/DTI provide valuable biomarkers for monitoring recovery in this rare phenotype.
Preadolescent irritability is a robust transdiagnostic neurodevelopmental predictor of later psychopathology, linked to altered reward processing-a common neurocognitive substrate across psychiatric disorders. However, its neurobiological mechanisms remain unclear. Deep learning (DL) excels in predicting neurodevelopmental vulnerabilities and detecting nonlinear brain-behavior relationships but often lacks explainability. Here, we integrate optimized prediction with explainability to characterize neural mechanisms of irritability using task-based fMRI from a large preadolescent sample (N = 1934; mean age = 9.95 years, 101 Persistently High Irritability [PHI], 1833 Persistently Low Irritability [PLI]). We trained three classifiers-artificial neural network (ANN), random forest (RF), XGBoost-to distinguish PHI from PLI using functional connectivity (FC) during reward anticipation. FC was assessed between four seeds (bilateral amygdala, ventral striatum) and 18 cortical/subcortical regions. Shapley additive explanations (SHAP) identified key connectivity predictors accounting for nonlinear effects. ANN (AUC = 0.73, p < .001) outperformed RF (AUC = 0.63, p = .03), XGBoost (AUC = 0.65, p = .01). SHAP revealed increased contralateral FC (e.g., right amygdala-left middle frontal gyrus) and decreased ipsilateral FC (e.g., left ventral striatum-left insula) generally predicted PHI, except amygdala connectivity, where higher ipsilateral FC predicted PHI. These findings highlight interplay between reward and emotion regulation circuits in persistent irritability, underscoring the potential of explainable DL to improve irritability prediction and enhance understanding of its neural mechanisms.