搜索结果：NeuroImage. Clinical

Borderline Personality Disorder (BPD) is a severe mental illness, although its neurobiological underpinnings remain largely unknown. Structural Magnetic Resonance Imaging (sMRI) in adolescents can offer insights into potential biomarkers to help advance early detection and targeted intervention. However, previous findings have been mixed, possibly due to clinical heterogeneity that may be better captured using a dimensional approach to personality functioning (PF). The current study explored grey matter volume (GMV) in youth with varying degrees of BPD pathology, and associations with dimensional PF. N = 93 females (14-21 years) comprising three groups (full-threshold BPD, sub-threshold BPD, and healthy controls) underwent sMRI and were assessed with the Semi-Structured Interview for Personality Functioning DSM-5 (STiP-5.1). Groups were combined to reflect dimensional personality pathology. Multiple linear regression analyses were conducted to determine associations between the STiP-5.1 total score, and each of its four elements with: (i) total GMV, (ii) GMV in individual brain regions defined by the Desikan-Killiany-Tourville atlas, (iii) selected regions of interest (ROIs). All analyses were statistically non-significant: STiP-5.1 total and total GMV (p = 0.61); STiP-5.1 total and individual brain regions (all corrected p values ≥0.82); STiP-5.1 total and ROIs (all corrected p values ≥0.91). Results were non-significant for each element, and a validity check using BPD criteria confirmed STiP-5.1 findings. We found no evidence of an association of dimensionally assessed PF with GMV in young females. The pursuit of clinical research efforts on other potential biomarkers using dimensional conceptualisations of PF may represent worthy endeavours.

Alzheimer's disease (AD) is a progressive neurodegenerative disorder characterized by cognitive decline and substantial brain atrophy. Early and accurate prediction of disease progression and staging is crucial for timely intervention and effective treatment planning. Previous studies, including those based on artificial intelligence techniques, have employed neuroimaging, biomarkers and clinical data to model AD progression; however, many of these approaches rely on strong parametric assumptions or lack robust statistical guarantees regarding model validity. To bridge this gap, this study proposes a novel framework for validating predictive and staging models of disease using a statistically agnostic methodology. The objective is to take the advantages of an unconventional method for robust validation of ML models related to AD. Validation is performed using the Statistical Agnostic Regression (SAR) methodology applied to the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset. The method tests for a linear relationship by resampling and estimating an upper bound on the expected risk (R) via a Bayesian bound under the worst-case scenario. The SAR power assesses the likelihood of detecting a true linear relationship using the test statistic R, via Monte Carlo simulations under the null distribution. Three predictive models related to structural neuroimaging are assessed: one for the Mini Mental State Examination (MMSE) score, another for the concentration of amyloid beta 1-42 protein in the cerebrospinal fluid, and a third for age. In addition, a model for staging based on Alzheimer's-related clinical groups is explored through the joint analysis of segmented gray matter and white matter images. The findings indicate that the SAR methodology not only facilitates robust validation of predictive ML models related to neuroimaging and AD but also enables an effective staging of the AD continuum. This SAR-proposed framework opens new perspectives for the validation of ML models for early diagnosis and provides a solid foundation for future research in computational neuroscience.

Status epilepticus (SE) is a critical neurological emergency with mortality rates reaching 15-25%. While several clinical scores exist for risk stratification of SE, they tend to perform poorly in a real-world scenario, where the prognostic role of electroencephalography (EEG) remains secondary. This study aimed to evaluate the clinical utility of existing prognostic scales and to determine if a simplified, structured EEG reporting framework provides incremental value in predicting mortality and length of stay (LOS). We retrospectively analyzed 182 adult SE patients (mean age 67 ± 15.9 years) between 2018 and 2024. We collected clinical scores and structured EEG features (based on Salzburg criteria and nomenclature from the American Clinical Neurophysiology Society position paper on critical care EEG). Outcomes included 30-day in-hospital mortality and LOS, analyzed using logistic regression, ROC curves, and Cox proportional-hazards models. In-hospital mortality was 23.1%, and mean LOS was 26.1 ± 29.8 days. Clinical scores alone demonstrated acceptable predictive value for mortality (AUC 0.60-0.76), but suffered from low specificity. Adding structured EEG features significantly improved mortality prediction (AUC = 0.88). Notably, plasma markers of systemic involvement primarily drove mortality prediction, whereas EEG patterns-specifically Salzburg criteria and spatio-temporal evolution-were strong predictors of LOS. Current clinical scores demonstrate limited accuracy in the risk stratification of SE due to high etiological variability. In contrast, structured EEG interpretation significantly enhances prognostic specificity and provides distinct predictive value for LOS. Future efforts should focus on integrating EEG parameters with existing clinical scores to harmonize and improve risk stratification models.

The cerebellum is increasingly recognized as a key component of large-scale brain networks implicated in epilepsy, yet its electrophysiological characterization remains limited in non-invasive recordings. This limitation arises from the cerebellum's depth, complex folding, and unfavorable source orientations, which challenge conventional magnetoencephalography (MEG) and electroencephalography (EEG). Here, we quantitatively characterize cerebellar signal detectability across modalities and sensor configurations using anatomically informed source modeling at the population level. We analyzed clinical MEG and EEG recordings from a large cohort of patients with epilepsy undergoing presurgical evaluation (n = 54), selected from a larger consecutive clinical population. Cerebellar and cerebral source spaces were constructed using subject-specific anatomical models derived from routine clinical MRI, enabling consistent forward modeling across individuals. The signal-to-noise ratio (SNR) was estimated at individual source locations and summarized at the regional level. In addition to clinical superconducting quantum interference device (SQUID)-MEG and EEG, multiple on-scalp optically pumped magnetometer (OPM) configurations were evaluated through simulations, including layouts matched to clinical sensor geometries and layouts optimized for posterior fossa coverage. The effects of source orientation, source-to-sensor distance, and head size on SNR were systematically investigated. In routine clinical recordings, cerebellar SNR was consistently lower than superficial cortical reference levels, confirming challenges in detecting cerebellar activity using standard SQUID-MEG and EEG. Reducing source-to-sensor distance by placing OPMs at SQUID-equivalent locations, i.e., projecting SQUID sensor locations to the scalp, did not improve cerebellar SNR, indicating that proximity alone is insufficient for better detectability of deeper sources. In contrast, cerebellar-optimized OPM layouts produced substantial SNR gains in posterior cerebellar regions. The effects of source orientation influenced SNR differences between OPM and EEG (under identical sensor/electrode coverage) but were secondary to depth- and geometry-related constraints. Mediation analysis further demonstrated that relative sensor distance significantly mediated OPM-related advantages in posterior cerebellar regions, particularly in individuals with smaller head sizes. These findings demonstrate that cerebellar signal detectability is governed primarily by anatomical depth and geometry rather than sensor proximity alone. By combining anatomically informed source modeling with flexible, region-specific sensor layouts, this work provides a principled framework for evaluating and improving MEG and EEG sensitivity to cerebellar activity, with implications extending beyond epilepsy to non-invasive mapping of deep and highly folded brain structures.

Quantification using the Centiloid (CL) scale has become a valuable information to consider when interpreting amyloid-PET images and is now implemented in several software packages. This work aims to assess the comparability of CL from [18F]flutemetamol scans derived using several research and commercial quantification pipelines. This analysis relies on three datasets: a test-retest cohort, a group of clinically relevant patients with amnestic mild cognitive impairment (aMCI) and a subgroup from the BioFINDER-1 cohort enriched with scans with amyloid loads around potential clinical decision thresholds (0-50CL). Images from the Test-Retest and aMCI cohorts were processed across seven quantification pipelines: three commercial software platforms and four research tools, including the standard SPM8 workflow. The statistical analysis was based on three steps: 1) a repeatability analysis using the test-retest data; 2) a reproducibility analysis across all pipelines using the aMCI cohort; 3) an inter-software reliability analysis around three clinically relevant thresholds: 11, 25 and 37 CL using the aMCI and the BioFINDER-1 data. In the Test-Retest dataset composed of 10 Alzheimer's Disease (AD) patients, high test-retest repeatability and reliability were observed with an absolute bias of less than 5 CL. Within-individual coefficients of variation ranged from 2.6 to 4.4% and repeatability coefficients from ∼8 to ∼16 CL. CL quantification was generally reproducible across pipelines in a dataset of 80 aMCI individuals (R2 in [0.94-0.99], slope in [0.98-1.03], intercept in [-4, 4], but the 95% limits of agreement (LoAs) ranged between ∼±12 and ∼±21 CL. Agreement between software around the three clinically relevant thresholds was 92-100% (kappa 0.83-1) in the aMCI data (N = 80) and 75-99% (kappa 0.48-0.96) in the BioFINDER-1 subgroup (N = 110). In this study, CL quantification was shown to be robust across a range of currently available software platforms. Uncertainty estimates should always be considered when interpreting results. In clinical practice, the choice of quantification software should not impact patient management decisions.

The risk factors for recurrent stroke in intracranial atherosclerosis (ICAS) may differ between the anterior and posterior circulations. This exploratory study aimed to investigate and compare these risk factors, leveraging plaque characteristics identified by high-resolution vessel wall imaging (HR-VWI). We retrospectively analyzed 166 patients with ICAS-related ischemic stroke, categorized into ACIS (n&#xa0;=&#xa0;123) and PCIS (n&#xa0;=&#xa0;43) groups. Each group was subdivided into recurrent (ACIS: 25; PCIS: 11) and non-recurrent (ACIS: 98; PCIS: 32) cohorts based on follow-up (mean: 16.69&#xa0;&#xb1;&#xa0;10.18 months). Clinical and imaging risk factors were compared, with multivariable Cox regression analysis employed to identify independent predictors of recurrence. Predictive performance was assessed using the area under the receiver operating characteristic (ROC) curve (AUC). For ACIS, older age, multiple infarcts, and significant plaque enhancement were associated with recurrence (all q&#xa0;<&#xa0;0.05). For PCIS, diabetes mellitus and intraplaque hemorrhage (IPH) were associated with recurrence (all q&#xa0;<&#xa0;0.05). Multivariable analysis confirmed older age (>55.5 years), multiple infarcts, and significant plaque enhancement as independent risk factors for ACIS recurrence, whereas diabetes mellitus and IPH were independent predictors for PCIS recurrence. A combined predictive model for ACIS achieved a fair-to-good predictive performance (AUC: 0.821, 95% CI: 0.72-0.92), while the model for PCIS yielded a more uncertain performance (AUC: 0.770, 95% CI: 0.61-0.93) with a wide confidence interval, reflecting the small sample size. The drivers of recurrent stroke in ICAS appear to differ between the anterior and posterior circulations. These exploratory findings suggest that integrating clinical data with HR-VWI plaque features may enhance risk stratification. However, due to significant limitations, these results should be interpreted with caution and require validation in larger, prospective cohorts before being applied to clinical practice. This study underscores the potential need for circulation-specific secondary prevention strategies.

WHO grade II-III insular glioma (InG) can displace the basal ganglia, resulting in distinct linear or curved boundary shapes. This study retrospectively explored this morphological distinction and its clinical relevance. MRI and clinical data were collected from Beijing Tiantan Hospital over the past 20&#xa0;years. The boundary shape on the FLAIR sequence was characterized using the fractal dimension (FD) and curved parameters. Statistical analyses included the &#x3c7;2 test, Cox regression, Kaplan-Meier analysis, and correlation analysis. Regression models were constructed using the stepwise Wald method. Model performance was evaluated through internal validation, including bootstrap resampling and 5-fold cross-validation. A total of 330 patients were included. Based on the FD, patients were classified into linear ('L'; 49%) and curved ('C'; 51%) subgroups. Compared with the L subgroup, the C subgroup exhibited more favorable biological features and progression-free survival (PFS) (p&#xa0;<&#xa0;0.001). Using routine pathological variables, regression models were constructed to identify crucial factors in different subgroups and perform disease stratification for Statistical explanation. Internal validation supported model stability. With regard to surgical responses, although the rate of gross total resection (GTR) was higher in the L subgroup, the PFS was not increased, whereas that of the C subgroup demonstrated an increase (p&#xa0;<&#xa0;0.001). The boundary shape between WHO grade II-III InG and the basal ganglia may reflect differences in biological features, survival outcomes, and surgical responses. Thus, the boundary shape should be considered in clinical practice.

A neuropathological staging system for Parkinson's disease (PD) proposes that PD may disseminate in a sequential regional pattern in the brain during six disease stages. Diffusion tensor imaging (DTI) can identify PD-associated patterns of brain alterations at the group level. (1) To develop a framework for a hypothesis-guided DTI-based approach targeted to automatically analyze in vivo the fiber tracts that are typically involved at each neuropathological stage of PD, and (2) to apply this framework to a large sample of MRI datasets of patients with PD in comparison with controls. A tract of interest (TOI)-based fiber tracking approach was used to analyze tracts that become involved during the course of PD, representative of the six neuropathological stages. The TOI-based technique of tractwise fractional anisotropy statistics (TFAS) was applied to calculate fractional anisotropy (FA) alterations for the investigated tracts. 206 DTI datasets were analysed (i.e., 102 DTI scans from patients with PD at different clinical and cognitive stages and 104 DTI scans from age- and sex-matched healthy controls). The TOI-related analyses showed differences between PD patients and controls for the tracts corresponding to each neuropathological stage, i.e., in the olfactory tract, the initiation states in the medulla-oblongata, the catecholaminergic tract, the corticopontine tract, the inferior longitudinal fascicle, and the superior longitudinal fascicle. Based on these differences, a sequential pattern classification could be performed into stages with sequentially increased tract alterations, indicating a more advanced neurodegenerative process. The herein used TOI-based framework could enable tracking disease stages in PD in vivo, as an approach to the propagation concept of neuropathological stages in PD. In future applications, this framework may be used not only for individual clinical work-up purposes, but also may enlarge the spectrum of potential non-invasive surrogate markers as a neuroimaging-based read-out for PD studies at the group level within a clinical context.

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 &#xb1; 3.6 years from diagnosis, 2.1 &#xb1; 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 &#xb1; 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&#xa0;<&#xa0;0.001) and reduced 11C-DASB binding in the insular cortex (p&#xa0;=&#xa0;0.003), parietal lobe (p&#xa0;=&#xa0;0.009), caudate (p&#xa0;<&#xa0;0.001), and putamen (p&#xa0;=&#xa0;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&#xa0;&#xd7;&#xa0;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.

Tourette syndrome (TS) is a neurodevelopmental disorder characterized by motor and phonic tics, affecting up to 1% of the adult population. While its etiology is unknown, increasing evidence highlights the key role of brain connectivity changes. Given its involvement in motor and speech inhibition, the Frontal Aslant Tract (FAT), a white matter tract connecting the posterior Superior and Inferior Frontal Gyri, may present structural alterations in individuals with TS. To investigate this hypothesis, a prospective case-control study was conducted from 2020 to 2023, enrolling 11 consecutive right-handed adults diagnosed with TS at a single referral center, alongside 12 age- and sex-matched healthy controls. We examined microstructural alterations of the FAT using metrics derived from two diffusion MRI techniques: Neurite Orientation Dispersion and Density Imaging (NODDI) and Constrained Spherical Deconvolution (CSD). We also examined correlations between diffusion metrics and clinical scores, as well as relationships among the diffusion metrics within both the TS and control groups. After False Discovery Rate corrections, patients' bilateral FAT showed reduced Fractional Anisotropy (FA) and Neurite Density (ND), and higher Mean Diffusivity (MD), indicating a more isotropic water diffusion and a decreased density in axons and dendrites. No significant correlations between diffusion metrics and clinical measures were found. FA and ND were positively correlated, and they also showed negative correlations with MD and Orientation Dispersion Index (ODI). These results strongly indicate that patients with TS may have reduced integrity of the bilateral FAT, which could be considered a potential target of stimulation techniques in TS treatment.

Secondary cortical degeneration caused by the remote effects of subcortical infarcts has been implicated in long-term outcomes after acute ischemic stroke. However, this process remains insufficiently studied in recent small subcortical infarcts (RSSI). We aimed to verify RSSI-induced cortical damage, determine whether it can be captured by neuroimaging markers, and explore its association with clinical outcomes. RSSI patients with longitudinal Magnetic Resonance Imaging (MRI) were included. Cortical degeneration was assessed using linear mixed-effects models, incorporating a direct approach based on individual diffusion weighted imaging and an indirect approach using the normative connectome from the Human Connectome Project (HCP). Principal component analysis (PCA) was employed to extract features of cortical alterations. The resulting component scores were used in general linear models to assess associations with neuroimaging markers and clinical outcomes. A total of 76 RSSI patients were analyzed. RSSI was found to induce progressive cortical thinning and volume loss in structurally connected regions. PCA identified a component reflecting parenchymal atrophy associated with diffusion-based markers of white matter integrity, as well as the presence of track/cap signs. Moreover, faster cortical degeneration in lesion-connected regions was significantly associated with a greater increase in Hamilton Anxiety Rating Scale (HAMA) scores (&#x3b2;&#xa0;=&#xa0;-2.38, 95% CI&#xa0;=&#xa0;-4.30 - -0.47, p&#xa0;=&#xa0;0.017). RSSI induces secondary cortical damage through structurally connected fiber tracts, which is detectable by neuroimaging markers of white matter integrity. These regional cortical alterations may be relevant to post-stroke outcomes and require validation in larger longitudinal studies.

Repetitive transcranial magnetic stimulation (rTMS) to the primary motor cortex (M1) provides significant pain relief in &#x223c;45% of chronic pain patients. Identifying biomarkers that predict treatment response before starting rTMS is essential for guiding clinical decision-making. Here, we used TMS combined with electroencephalography (TMS-EEG) to assess pre-treatment cortical function in 43 patients with chronic pain before receiving 12 sessions of therapeutic 10 Hz rTMS to M1 over eight weeks as a secondary analysis from a trial comparing effects of rTMS in different cortical targets. Responders were defined as individuals reporting a &#x2265; 30% reduction in pain intensity on a visual analogue scale at week 8. Pre-therapy TMS-evoked cortical reactivity was quantified using global mean field power (GMFP) and local mean field power (LMFP) within an early post-stimulus window (20-120 ms). Oscillatory dynamics were measured by event-related spectral perturbation (ERSP) and intertrial coherence (ITC) in alpha (8-12 Hz), low-beta (13-20 Hz), and high-beta (21-30 Hz) bands. Compared with non-responders, responders (n = 20; 47%) showed lower GMFP, LMFP, alpha-band ERSP, and ITC at the stimulation site (p < 0.05). These low measures correlated with greater reductions in pain intensity (p < 0.05). Exploratory supervised machine-learning analysis using three TMS-EEG features (GMFP, alpha-band ERSP, alpha-band ITC) predicted responder status with acceptable performance (ROC-AUC=0.70, PR-AUC=0.76). These findings suggest that lower pre-treatment TMS-evoked cortical reactivity and alpha-band oscillatory dynamics may identify patients more likely to benefit from rTMS. Prospective clinical trials should test pre-therapy reactivity and connectivity metrics to select patients more likely to benefit from therapy.

Autism spectrum disorder (ASD) is an increasingly diagnosed neurodevelopmental condition characterized by persistent difficulties in social communication and restricted, repetitive patterns of behavior and sensory processing that leads to functional impairment. The diagnosis of ASD relies on behavioral and clinical assessment as there are no currently available biomarkers. Recent brain imaging studies have suggested abnormalities in the brain fluid flow in individuals with ASD. Cardiorespiratory and vasomotion-induced very low frequency (VLF &#x2264; 0.1 Hz) brain pulsations are now considered to facilitate the cerebrospinal- and interstitial fluid exchange in the brain, thus contributing to maintaining cerebral homeostasis and fluid clearance. In this study, we utilized ultrafast resting-state functional magnetic resonance imaging (fMRI) to capture and compare the powers of each physiological pulsation in groups of 18 young adults diagnosed with ASD and 19 neurotypical controls (NTC). We further probed the clinical significance of findings by undertaking regression analyses examining the associations of both Autism Spectrum Quotient (AQ) and Autism Diagnostic Observation Schedule (ADOS) scores with pulsation powers, and by receiver operating characteristics (ROC) analysis. Compared to the NTC group, the ASD group showed significantly higher VLF pulsation power, which was located predominantly in subcortical grey matter nuclei and the white matter, indicating increased vasomotor power in ASD. In addition, the individual VLF power enabled good accuracy (ROC area under the curve = 75%-93%) for discriminating ASD subjects from NTCs. In conclusion, present findings of increased VLF power are postulated as possible indication of altered driving force of cerebral neurofluid dynamics and could potentially serve as a useful clinical classifier.

Friedreich's ataxia (FRDA) is an inherited neurodegenerative disorder characterized by progressive ataxia and multisystem manifestations resulting from involvement of the peripheral and central nervous systems. While regional atrophy is known to be associated with symptoms, functional network alterations may represent a critical pathological mechanism; however, their specific contribution to motor and cognitive impairment remains unclear. We combined T1-weighted anatomical MRI and resting-state functional MRI (rs-fMRI) in 37 individuals with FRDA and 41 age- and sex-matched healthy controls and explored how functional connectivity differences are related to atrophy, clinical severity and cognitive performance. Regional volumes were quantified using morphometry analyses, spontaneous rs-fMRI activity was assessed via amplitudes of low-frequency fluctuations, and functional co-activation was evaluated among regions showing structural and neuronal activity alterations. Volume reductions were most pronounced in the brainstem, cerebellar white matter, hemisphere of lobules VI, X, and thalamus. Functionally, individuals with FRDA showed decreased fronto-cerebellar connectivity alongside increased intracerebellar, thalamo-striatal, and hippocampal-cerebellar coupling. Infratentorial and thalamic volume loss correlated strongly with clinical disease severity, whereas reduced frontal co-activation with cerebellar lobules VI, Crus I and II was moderately associated with poorer motor and cognitive performance. In contrast, increased intracerebellar and hippocampal-cerebellar coupling was observed particularly in individuals with more advanced disease and was partly associated with better cognitive outcomes. These findings indicate widespread disruptions of long-range cerebro-cerebellar connectivity together with increased intraregional coupling and potential network reorganization, underscoring the importance of network-level mechanisms for understanding clinical heterogeneity in FRDA and guiding future prognostic and therapeutic studies.

Early identification of Alzheimer's disease (AD) and its prodromal stage, mild cognitive impairment (MCI), is important for timely clinical assessment and disease management. Structural T1-weighted magnetic resonance imaging (MRI) captures macroscopic neurodegenerative changes associated with disease progression; however, developing deep learning models that are both methodologically rigorous and clinically interpretable remains challenging. This study presents an explainable deep learning framework for longitudinal classification of cognitively normal (CN), MCI, and AD subjects using MRI data from the Alzheimer's Disease Neuroimaging Initiative (ADNI). A two-and-a-half-dimensional (2.5D) convolutional neural network based on a modified ResNet-18 architecture was used to incorporate contextual information from adjacent sagittal slices while remaining computationally efficient. To preserve longitudinal validity and reduce potential information leakage, multiple follow-up visits per subject were included and a strict subject-level data splitting strategy was adopted. Under this evaluation protocol, the proposed model achieved moderate classification performance, highlighting the difficulty of three-class longitudinal neurodegenerative disease classification under a stringent split design. To further examine the effect of split level, we additionally evaluated the same framework using a scan-level splitting strategy. In our implementation, scan-level splitting did not produce a substantial performance improvement, supporting the use of subject-level splitting primarily as a methodological choice for longitudinal validity. To enhance transparency, Gradient-weighted Class Activation Mapping (Grad-CAM) was used to interpret model predictions. The resulting explanations showed plausible temporo-frontal attention patterns in correctly classified AD cases and more heterogeneous responses in MCI cases. Overall, this work demonstrates that combining longitudinal MRI analysis with explainable deep learning under a rigorous evaluation framework can provide an interpretable and reproducible baseline for Alzheimer's disease research. By jointly emphasizing interpretability, temporal consistency, and evaluation validity, the proposed framework contributes toward the development of more trustworthy AI methods in neuroimaging, while also highlighting the remaining challenges in robust AD classification.

Magnetic resonance imaging (MRI) is a valuable clinical and research tool for patients managed using deep brain stimulation (DBS). Unfortunately, MRI under these conditions is associated with substantial risks, necessitating stringent regulations that limit its clinical and research utility. In addition, magnetic susceptibility differences between DBS hardware and surrounding tissues significantly compromise spatial encoding mechanisms of conventional MRI sequences, resulting in image signal loss and geometric distortions. The impact on gradient-recalled echo echo-planar imaging sequences commonly utilised for functional MRI in DBS settings is particularly severe. This review presents a range of mitigation strategies aimed at both safety and enhanced image quality, spanning innovations in DBS hardware design to advanced MRI sequences capable of addressing issues inherent to the presence of DBS hardware, such as electrode heating and susceptibility artefacts. Additionally, we highlight approaches incorporating the discussed novel postprocessing techniques and functional MRI acquisition protocols, along with their limitations and associated challenges to enable their wider dissemination, with the overarching objective of improving the quality of life of DBS patients.

Migraine is a prevalent and disabling neurological disorder, characterized by impaired regulation of migraine burden, sensory processing, and cognitive-emotional states. Brain entropy quantifies the complexity of neural dynamics, where reduced entropy may reflect diminished neural adaptability, but its assessment with fMRI in migraine remains limited. Here, we examined alterations in brain entropy and their associations with clinical burden, migraine phase, and symptomatology. Resting-state fMRI data were acquired from adults with episodic migraine, chronic migraine, and healthy controls. Following standard preprocessing, voxel-wise sample entropy was computed, and group differences were assessed using ANCOVA with age and sex as covariates. Associations with clinical burden and symptom measures were examined within affected regions. In chronic migraine, attack timing-related changes in entropy were further explored, and the Largest Lyapunov Exponent (LLE) was estimated to characterize chaotic dynamics underlying attack-related complexity changes. Migraine patients showed reduced entropy in visual, dorsal attention, and default mode network regions compared to controls, most pronounced in chronic migraine. Lower entropy correlated with greater headache frequency and longer illness duration. In chronic migraine, entropy relatively increased during attacks in multisensory integration regions and was associated with positive and elevated LLEs, indicating partially restored complexity with weakly chaotic dynamics. Patients experiencing phonophobia and nausea also exhibited increased entropy in multisensory integration and default mode network regions. Our findings demonstrate widespread reductions in brain entropy in migraine, reflecting impaired neural adaptability, whereas attacks may transiently restore complexity partially through weakly chaotic dynamics. These results advance understanding of migraine pathophysiology and highlight potential targets for therapeutic intervention.

Foundation models (FMs), large neural networks pretrained on extensive and diverse datasets, have revolutionized artificial intelligence and demonstrated significant promise in medical imaging by enabling robust performance with limited labeled data. Although numerous surveys have reviewed the application of FMs in healthcare, brain imaging remains underrepresented, despite its critical role in the diagnosis and treatment of neurological diseases using modalities such as magnetic resonance imaging (MRI), computed tomography (CT), and positron emission tomography (PET). To address this gap, we present the first comprehensive and curated review of FMs for brain imaging. We systematically analyze 161 brain imaging datasets and 143 FMs up to Jan, 2026, providing insights into key design choices, training paradigms, and optimizations driving recent advances. Our review highlights that the race for larger models has stabilized in 2026 towards more efficient models. FMs for brain imaging heavily rely on MRI (92%) and CT (57%) inputs, while PET imaging remains vastly underexplored (supported by only 15% of models). Our study also demonstrates architectural vulnerabilities caused by homogenization and lack of diversity, with Vision Transformers utilized in 48% of visual encoders, and models predominantly built by patching pre-existing natural image backbones like SAM (19%), and CLIP (12%) rather than utilizing native domain-specific 3D medical imaging innovations. For each of the eight tasks of the study the systematic review identifies the best models and discusses their innovations. Our study also uncovers critical gaps in the tasks, pathologies and clinical validation. We demonstrate that the literature is disproportionately skewed toward brain cancer research (37% of models) and neurodegenerative diseases (24%), and discuss the potential causes and remedies. Similarly, tasks are heavily weighted toward anomaly classification (44%) and segmentation (32%), leaving areas like mental health and image synthesis underrepresented. Besides, most models rely exclusively on traditional machine learning metrics (e.g., DICE or SSIM) rather than medically relevant measures, and only seven out of the 143 models incorporated human expert evaluations to verify real-world utility. Our systematic review concludes by outlining future research directions to advance FMs in brain imaging and actionable recommendations to build better FMs and to evaluate and deploy them in clinical and research settings.

Multiple system atrophy (MSA) is a rapidly progressive neurodegenerative disorder for which sensitive and objective imaging biomarkers for early diagnosis and disease monitoring remain limited. Iron dysregulation is implicated in MSA pathophysiology and represents a promising therapeutic target. We used quantitative susceptibility mapping (QSM) MRI to characterize regional patterns of iron accumulation in early-stage MSA and to explore its potential diagnostic and longitudinal utility. 3T MRI data were acquired from 38 individuals with MSA, including 10 early-stage patients with 12-month follow-up, along with 43 patients with Parkinson's disease and 23 age-matched healthy controls. Magnetic susceptibility was quantified in the substantia nigra, globus pallidus, putamen, and dentate nucleus using median and 75th-percentile metrics, and patient-specific voxel-wise z-score maps were generated relative to normative distributions. Compared with Parkinson's disease and healthy controls, individuals with MSA exhibited significantly elevated susceptibility in the globus pallidus and substantia nigra, with smaller effects in the putamen. Seventy-fifth percentile metrics demonstrated greater sensitivity to group differences than median measures, capturing focal iron accumulation. Globus pallidus susceptibility was strongly associated with clinical severity. Longitudinal analyses revealed significant increases in susceptibility over 12 months in early-stage MSA in the substantia nigra and globus pallidus. These findings demonstrate that QSM detects progressive, clinically relevant iron accumulation in MSA, and that higher-order susceptibility metrics and patient-level normative mapping enhance sensitivity to focal pathology, supporting QSM-derived measures as imaging biomarkers for diagnosis, disease monitoring, and assessment of iron-modulating therapies.