SEARCH · Search PubMed

Results for “Neural oscillations”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

67 records · Page 4Linked to original sources

PubMed2026

Brain network alterations underlying cue reactivity and craving in abstinent methamphetamine users: a systematic review of functional MRI findings.

BACKGROUND: Methamphetamine use disorder (MUD) is marked by intense craving and high relapse risk, often triggered by drug-related cues. Functional magnetic resonance imaging (fMRI) provides key insight into the neural basis of this cue reactivity, implicating large-scale brain networks for reward, motivation, and control. Yet, findings remain inconsistent across studies due to differences in task design, abstinence duration, and participant characteristics. OBJECTIVE: This systematic review synthesises evidence on how abstinence influences brain network alterations underlying cue reactivity and craving in methamphetamine users, integrating task-based and resting-state fMRI findings within leading neurobiological models of addiction. METHODS: A systematic search of PubMed, Scopus, Web of Science, and Ovid was conducted up to August 10, 2025, following PRISMA 2020 guidelines. Eligible fMRI studies examined cue reactivity or craving in abstinent methamphetamine users. Data were extracted on activation, connectivity, and brain-behaviour associations, and synthesised narratively. RESULTS: Task-based studies revealed heightened activation across reward, salience, and control networks during cue exposure, which diminished as parietal and executive control systems re-engaged with longer abstinence. Resting-state findings showed disrupted intrinsic connectivity among default mode, salience, and frontoparietal networks, reflecting persistent imbalances linked to craving and use severity. CONCLUSION: fMRI evidence shows that MUD is marked by network-level disruption linking reward, salience, and control systems. Task-based findings reveal strong cue reactivity in reward circuits, while resting-state data show persistent imbalance among default mode and control networks. With abstinence, partial restoration of network integrity emerges, highlighting both vulnerability and opportunities for targeted, recovery-based interventions.

Humans↗

PubMed2026

Age-related differences in motor unit behaviours and maximal strength: A systematic review and meta-analysis.

Ageing is associated with a decline in strength; however, the neural mechanisms underpinning these changes remain poorly understood. Motor unit discharge rate (MUDR) and recruitment threshold (MURT) regulate the magnitude of motoneuron output through rate coding and orderly recruitment, while discharge rate variability (MUDRV) reflects the steadiness of motoneuron output. Yet, age-related differences in these properties remain inconsistent across the literature. Therefore, this systematic review and meta-analysis quantified age-related differences in motor unit behaviours and their contribution to maximal isometric strength. Electronic databases (Medline, Embase, Scopus, PsycINFO, Ovid Emcare, CENTRAL, and Web of Science) were searched up to May 2025, yielding 1493 records; of these, 48 studies met the inclusion criteria. Standardised mean differences (SMDs) were calculated using random-effects models to compare older and younger adults, and methodological quality was assessed using the AXIS tool. Older adults exhibited markedly lower maximal strength than younger adults (SMD = -1.01; 95% CI -1.22, -0.79). MUDR was lower in older adults across all contraction intensities, with greater reductions at high forces (> 60% maximal voluntary contraction (MVC): SMD = -0.65; 95% CI -0.96, -0.34) compared to low forces (< 30% MVC: SMD = -0.34; 95% CI -0.50, -0.18). Discharge rate variability was greater (SMD = 0.44; 95% CI 0.15, 0.72), whereas recruitment thresholds relative to MVC were lower (SMD = -0.42; 95% CI -0.80, -0.03) in older adults. Collectively, these findings suggest that age-related alterations in motor unit discharge behaviour may contribute, at least in part, to reduced maximal strength in older adults.

Aging↗

PubMed2026

The Role of Artificial Intelligence Combined With Digital Cholangioscopy for Indeterminant and Malignant Biliary Strictures: A Systematic Review and Meta-analysis.

BACKGROUND: Current endoscopic retrograde cholangiopancreatography (ERCP) and cholangioscopic-based diagnostic sampling for indeterminant biliary strictures remain suboptimal. Artificial intelligence (AI)-based algorithms by means of computer vision in machine learning have been applied to cholangioscopy in an effort to improve diagnostic yield. The aim of this study was to perform a systematic review and meta-analysis to evaluate the diagnostic performance of AI-based diagnostic performance of AI-associated cholangioscopic diagnosis of indeterminant or malignant biliary strictures. METHODS: Individualized searches were developed in accordance with PRISMA and MOOSE guidelines, and meta-analysis according to Cochrane Diagnostic Test Accuracy working group methodology. A bivariate model was used to compute pooled sensitivity and specificity, likelihood ratio, diagnostic odds ratio, and summary receiver operating characteristics curve (SROC). RESULTS: Five studies (n=675 lesions; a total of 2,685,674 cholangioscopic images) were included. All but one study analyzed a deep learning AI-based system using a convoluted neural network (CNN) with an average image processing speed of 30 to 60 frames per second. The pooled sensitivity and specificity were 95% (95% CI: 85-98) and 88% (95% CI: 76-94), with a diagnostic accuracy (SROC) of 97% (95% CI: 95-98). Sensitivity analysis of CNN studies (4 studies, 538 patients) demonstrated a pooled sensitivity, specificity, and accuracy (SROC) of 95% (95% CI: 82-99), 88% (95% CI: 72-95), and 97% (95% CI: 95-98), respectively. CONCLUSIONS: Artificial intelligence-based machine learning of cholangioscopy images appears to be a promising modality for the diagnosis of indeterminant and malignant biliary strictures.

Humans↗

PubMed2026

Six weeks of isometric resistance training led to evidence of corticospinal but not reticulospinal adaptation in previously untrained adult males.

The latest hypothesis regarding the source of enhanced neural activation from resistance training is the reticulospinal rather than the corticospinal tract, based on invasive animal and emerging human data. The present study employed a six-week isometric resistance training intervention in a randomized controlled design to address this knowledge gap. Thirty-nine healthy, untrained males (age ~23 y, sustained contraction group n = 13, explosive contraction group n = 9, control group n = 17) underwent neuromuscular and electrophysiological testing and completed all study requirements. Maximal isometric torque (MVC) and rate of torque development (RTD) were measured during a familiarization session as well as before and after the six-week period. Transcranial magnetic stimulation was used to assess motor-evoked potential (MEP) area and silent period duration while subjects contracted to 10% of MVC. Loud sound (120 dB) was used to modulate MEP area and reaction time to visual stimuli during the StartReact test. Only the intervention groups demonstrated significant improvements in MVC (27%) and RTD (60%) (both P < 0.01), along with reduced MEP area (- 21%) and silent period duration (- 23%) (both P < 0.01). The sustained contraction group showed reduced modulation of reaction time and increased MEP suppression due to loud sound. Short-term resistance training seemed to reduce cortical inhibition and corticospinal excitability in both training groups. The study showed conflicting changes in measures purported to evaluate reticulospinal functioning. It is recommended to examine different forms of resistance training and longer training exposure in future.

Humans↗

PubMed2026

Effects of acute resistance exercise on prefrontal oxygenation and task-switching performance: Considerations of loading strategies and blood flow restriction.

Although acute resistance exercise (RE) has been proposed to influence cognitive flexibility and underlying neural mechanisms, it remains unclear whether these effects vary across loading strategies and whether exercise-induced prefrontal hemodynamic responses translate into cognitive outcomes. The present study examined (1) prefrontal cortex (PFC) oxygenated hemoglobin (O2Hb) responses across exercise sets and conditions, (2) the effects of low-load (LL), LL with blood flow restriction (BFR), and high-load (HL) RE on task-switching performance, and (3) whether exercise-related PFC O2Hb responses were associated with pre- to post-exercise changes in task-switching performance. Thirty physically active adults completed three randomized, counterbalanced RE conditions consisting of four sets of barbell squats. LL was performed at 30% one-repetition maximum (1RM) with and without BFR, whereas HL was performed at 70% 1RM. Cognitive flexibility was assessed pre- and post-exercise using a modified Stroop task, indexed by switch-cost reaction time (RT) and accuracy. PFC O2Hb was assessed using functional near-infrared spectroscopy during exercise and expressed as changes from the resting baseline for each set (Sets 1-4). PFC O2Hb increased across sets, rising from Set 1 to Set 3 before plateauing, with no differences observed across conditions. Switch cost RT and accuracy did not improve from pre- to post-exercise, and no differences across conditions were detected. PFC O2Hb during the final set was not associated with changes in switch cost. These findings suggest that although acute RE elicits robust increases in prefrontal hemodynamic activity, such responses may not translate into acute improvements in cognitive flexibility.

Humans↗

PubMed2026

Triggering modalities to synchronise non-invasive respiratory support in preterm infants: a systematic review and meta-analysis.

BACKGROUND: Increasing evidence suggests that synchronised nasal intermittent positive pressure ventilation (sNIPPV) may be the optimal mode of non-invasive respiratory support. However, no comprehensive review of sNIPPV modes is available. This review aims to describe different synchronisation methods for sNIPPV and compare their effectiveness in clinical, physiological and technical outcomes with other modes of non-invasive respiratory support. METHODS: This review identified all clinical studies in preterm infants that compared sNIPPV to other non-invasive respiratory support modes or compared different trigger modalities between 1990 and 2026. A search was carried out in MEDLINE, Embase and the Cochrane Library. Main outcomes were categorised as clinical (eg, extubation failure (EF)), physiological (eg, breathing effort) and technical (eg, synchronisation rate). RESULTS: 49 studies (2864 infants) were included, with a low to moderate risk of bias. Meta-analysis showed a reduction in EF (risk ratio=0.38; 95% CI 0.17 to 0.85; p=0.03) in favour of sNIPPV compared with nasal continuous positive airway pressure (nCPAP). Physiological outcomes were significantly improved during sNIPPV compared with nCPAP and nasal intermittent positive pressure ventilation (NIPPV), especially breathing effort. When reviewing technical outcomes, non-invasive neurally adjusted ventilatory assist showed lower patient-ventilator asynchrony (PVA) (index ranging from 7% to 50%), a higher synchronisation rate (80%-99%) and a shorter trigger delay (35 ms) compared with other sNIPPV modes. CONCLUSIONS: This review shows that synchronising NIPPV results in consistent physiological (reduced patient effort) and technical benefits (reduced PVA). However, evidence on positive effects on (long-term) major clinical outcomes remains limited and requires further studies. PROSPERO REGISTRATION NUMBER: CRD420251022479.

Humans↗

PubMed2026

Effects of anodal transcranial direct current stimulation over the right primary motor cortex on a sequential motor finger tapping task in developmental stuttering.

INTRODUCTION: This study investigates the impact of anodal transcranial direct current stimulation (tDCS) on non-speech sequential motor practice in adults who stutter (AWS), compared to non-stuttering controls (ANS). Recent research has explored the effects of tDCS on speech fluency in stuttering. However, its effect on non-speech motor tasks has not yet been studied. METHODS: 20 AWS and 30 ANS right-handed participants were randomly assigned to anodal or sham tDCS conditions, performing a sequential finger tapping task. We targeted over the right primary motor cortex, stimulating at 2 mA for 20 min. Sequence duration and reaction time were analyzed. RESULTS: AWS analysis revealed that the anodal condition had significantly slower reaction times in the second half of the task compared to sham. For sequence durations, AWS in the anodal condition had slower overall sequence durations than the sham condition. However, there were no block-by-block differences in sequence duration. When comparing AWS and ANS, no significant differences were observed for sequence duration. However, there were significant differences in reaction time between AWS and ANS, specifically in earlier blocks. Additionally, there was no significant Group × Condition interaction. DISCUSSION: The findings suggest that anodal stimulation impeded finger sequencing in AWS, showing overall slower sequence durations and a diminishing effect on reaction times in the second half of the experiment, suggesting anodal tDCS may interact uniquely with the neural mechanisms in stuttering. Future studies should explore the effects of anodal tDCS on non-speech motor tasks to gain a broader understanding of its impact on motor control and motor learning.

Humans↗

PubMed2026

Acupuncture improves depressive symptoms and prefrontal cortical function in mild to moderate depressive disorder: A randomized sham-controlled trial and fNIRS study.

BACKGROUND: Depressive disorder is a common mental illness associated with substantial functional impairment. Although pharmacotherapy is widely used, its effectiveness is often limited by adverse effects and poor adherence. Acupuncture has been increasingly applied as a complementary treatment for depression, and its neurobiological characteristics remain unclear. OBJECTIVE: This randomized, sham-controlled trial aimed to evaluate the clinical efficacy of acupuncture for mild to moderate depressive disorder and to investigate its effects on prefrontal cortical function using functional near-infrared spectroscopy (fNIRS). METHODS: Patients with mild to moderate depressive disorder were randomly assigned to a real acupuncture (RA) group or a sham acupuncture (SA) group and received standardized treatment for 8 weeks. Clinical outcomes were assessed using the Self-Rating Depression Scale (SDS), Self-Rating Anxiety Scale (SAS), Short Form-36 Health Survey (SF-36), and a traditional Chinese medicine syndrome score. A subset of participants underwent fNIRS assessment during resting-state and task-based conditions to evaluate prefrontal cortical activation and functional connectivity. RESULTS: Compared with baseline, the RA group showed significant reductions in SDS and SAS scores and significant improvements in SF-36 emotional domains, with effects emerging at Week 4 and persisting up to 12 weeks after treatment. Improvements were greater and more stable in the RA group than in the SA group. fNIRS analyses revealed enhanced activation in dorsolateral and medial prefrontal regions and strengthened prefrontal functional connectivity following acupuncture, whereas neural changes in the SA group were limited. CONCLUSION: Acupuncture is effective for improving depressive and anxiety symptoms and quality of life in patients with mild to moderate depressive disorder. Modulation of prefrontal cortical activation and connectivity may underlie its antidepressant effects.

Humans↗

PubMed2026

A contextual activity score (CAS) for inferring ADAR-associated transcriptional activity across RNA-seq, single-cell, and spatial transcriptomics.

BACKGROUND AND OBJECTIVE: Adenosine-to-inosine RNA editing, catalyzed by Adenosine Deaminases Acting on RNA (ADARs), is a widespread modification involved in neural function, immune regulation, and cancer. The Alu Editing Index (AEI) is the standard metric to estimate ADAR activity but requires raw sequencing reads and is poorly suited for single-cell and spatial transcriptomic data. This study aimed to develop an alternative framework for inferring ADAR-associated transcriptional activity from gene expression data across diverse transcriptomic technologies. METHODS: We developed the Contextual Activity Score (CAS), a framework based on transcriptional signatures from ADAR perturbation experiments. Context-specific signatures were generated for human neurons, mouse neurons, and cancer models to infer ADAR1 and ADAR2 activity. CAS was computed from normalized gene expression matrices using regulon-based enrichment analysis. Performance was evaluated by comparing with the Alu Editing Index across bulk RNA sequencing datasets, simulated sequencing depths, and library preparation protocols. RESULTS: CAS showed strong concordance with the Alu Editing Index across multiple datasets, while remaining robust to reduced sequencing depth and different library protocols. Unlike the Alu Editing Index, CAS can be applied to single-cell and spatial transcriptomic data and enables the independent assessment of ADAR2 activity. In cancer and neuronal contexts, CAS captured biologically meaningful variations in ADAR-associated transcriptional activity at sample, cell-type, and spatial levels. CONCLUSION: CAS provides a scalable approach applicable across multiple RNA-seq protocols for estimating ADAR-associated transcriptional activity using gene expression data. This method, implemented in an open-source R package for broad adoption, expands the ability to study ADAR-associated transcriptional activity across transcriptomic modalities where direct editing quantification is challenging, such as single-cell and spatial transcriptomics.

Adenosine Deaminase↗

PubMed2026

Micro- and nanoplastics-induced neurotoxicity: a CNS-centered, evidence-graded adverse outcome pathway framework based on systematic weight-of-evidence assessment.

Micro- and nanoplastics (MPs/NPs) are ubiquitous anthropogenic particulate pollutants posing emerging threats to human neurological health. Severe heterogeneity in particle physicochemical properties, environmental aging status, exposure paradigms and experimental platforms has created persistent mechanistic uncertainties in MP/NP neurotoxicology, hindering reliable hazard characterization and risk translation. Here, we systematically consolidate empirical toxicological evidence and construct a dedicated central nervous system (CNS)-targeted adverse outcome pathway (AOP) network integrated with rigorous weight-of-evidence (WoE) grading to elucidate the hierarchical, particle-specific toxic cascades underlying MP/NP-induced neural injury. Our synthesis overturns the conventional linear toxicity paradigm, demonstrating that MPs/NPs trigger neurotoxicity via a complex multi-input mechanistic network. We definitively establish oxidative stress as a robust early convergent key event-rather than a universal molecular initiating event-orchestrating ROS overproduction, lipid peroxidation, mitochondrial dysfunction, and neuroinflammation to propagate neuronal damage. This core module is driven by five distinct particulate upstream triggers: particle-biomolecule interfacial perturbation, corona-facilitated cellular internalization, plastic-associated chemical leaching, aging-derived free radical reactivity, and gut-borne systemic neurotoxic signaling. Downstream pathogenic outcomes encompass glial overactivation, neurotransmitter dyshomeostasis, autophagy-lysosome dysfunction, metabolic reprogramming, regulated neuronal cell death, and behavioral impairments. Tiered WoE analysis confirms strong validation for early oxidative/inflammatory cascades, moderate support for gut-brain axis crosstalk and intracellular trafficking disruption, and nascent evidence for synaptic dysfunction and neurodegeneration-linked proteostatic defects. Extrapolation to human health risk remains constrained by the frequent use of high-dose exposure paradigms, limited validated data on internal dosimetry in the human brain, discrepancies between effective concentrations in experimental models and environmentally relevant human tissue burdens, and insufficient causal validation of distal adverse outcomes. We highlight key research priorities including aged mixed-particle exposure systems, leachate-controlled assays, quantitative internal dose evaluation, and mechanistic intervention verification. This evidence-stratified AOP framework resolves longstanding mechanistic ambiguities in particulate neurotoxicity, providing a standardized, causality-based foundation for future mechanistic exploration and health risk assessment of global plastic pollution.

Adverse outcome pathway↗

PubMed2026

Predicting ACL injury risk in athletes: A systematic review of machine learning-based models.

BACKGROUND: Early ACL injury risk identification in athletes is essential. This systematic review examines machine learning (ML) models for predicting ACL injuries, evaluating their methodological quality, performance, and reliability. METHOD: A comprehensive electronic search was conducted across PubMed, Scopus, Web of Science, and IEEE Xplore databases, supplemented by Google Scholar for grey literature, covering articles published between January 1, 2015, and August 30, 2025. Eligible studies were appraised using the Prediction Model Study Risk of Bias Assessment Tool (PROBAST) for methodological quality and risk of bias, and the Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) guidelines for quality of evidence. RESULTS: Ten studies were included. PROBAST showed eight studies had moderate risk of bias and two low risk. TRIPOD found only two studies met quality criteria. ML models included logistic regression (n = 5), support vector machines (n = 4), k-nearest neighbor (n = 3), decision trees (n = 3), random forests (n = 5), neural networks (n = 2), linear discriminant analysis (n = 1), and pre-trained CNNs (n = 1). AUC ranged from 0.63 to 0.98. Accuracy (reported in six studies) ranged from 26% to 95%; however, these values should be interpreted with caution due to the absence of confidence intervals, lack of class imbalance handling, and limited external validation across studies. Tree-based ensemble methods such as random forest achieved competitive accuracy (74-86%), while SVM, a non-ensemble classifier, reported accuracy ranging from 71% to 95%; however, the highest values were obtained in studies with notably small sample sizes (n = 12 to n = 39), raising concerns about overfitting and generalizability. CONCLUSION: Current ML algorithms show promise for identifying athletes at high ACL injury risk and detecting relevant risk factors. Although study quality was generally satisfactory, future research should prioritize external validation and model interpretability to support clinical translation.

Humans↗

PubMed2026

The effect of tDCS on emotion-related risk-taking behavior and delay discounting in adults with ADHD.

INTRODUCTION: Adults with Attention Deficit Hyperactivity Disorder (ADHD) often engage in risky behaviors due to impaired decision-making processes. This study aims to investigate the effects of transcranial direct current stimulation (tDCS) over the dorsolateral prefrontal cortex (dlPFC) and ventromedial prefrontal cortex (vmPFC) on emotion-related risk-taking behavior and delay discounting in adults with ADHD. METHODS: Thirty adults with ADHD underwent three tDCS conditions, administered in a randomized order with at least one week between sessions: (1) left dlPFC anode/right vmPFC cathode, (2) left dlPFC cathode/right vmPFC anode, and (3) sham stimulation. In each session, participants completed the Delay Discounting Task (DDT) and the Modified Balloon Analogue Risk Task (mBART) under three emotional conditions (neutral, positive, and negative) which were induced using emotionally congruent photographs and sounds. Galvanic skin responses (GSR) were also recorded. In the DDT, both area under the curve (AUC) values and log-transformed discounting rates (log k) were calculated for small, medium, and large reward magnitudes (RM). Exploratory electric field modeling was also performed to characterize current distribution. RESULTS: The findings demonstrated task-specific effects of tDCS on decision-making. Although no overall tDCS effect was observed on DDT performance, significant tDCS × RM interactions emerged, particularly for smaller rewards. In contrast, exploratory analyses suggested that tDCS affected all mBART scores. Emotional condition did not influence consistently behavioral performance in either task, whereas both emotional stimulation and tDCS significantly affected GSR responses. However, exploratory electric field modeling indicated a broad prefrontal current distribution extending beyond the intended cortical targets. CONCLUSIONS: These findings suggest preliminary evidence that prefrontal tDCS can influence risk-related decision-making and autonomic responses in adults with ADHD. However, its effects on delay discounting appear to be context-dependent and limited to specific RMs. Future studies combining neuroimaging with individualized electric field modeling are needed to clarify the neural mechanisms underlying the observed effects of tDCS and to optimize stimulation protocols in adults with ADHD.

Humans↗

PubMed2026

Artificial intelligence (AI) uses in stereotactic radiosurgery (SRS): diagnosis with brain metastasis (BM) - A systematic review.

BACKGROUND: Brain metastases (BM) are the most common intracranial tumors in adults, and stereotactic radiosurgery (SRS) has become a mainstay of management. However, several diagnostic challenges persist in the SRS pathway, particularly the differentiation of radiation necrosis (RN) from true tumor progression, which conventional MRI and even advanced imaging techniques often cannot reliably resolve. Recent advances in artificial intelligence (AI) offer the potential to address these diagnostic limitations. This systematic review synthesizes current literature on AI applications for MRI-based diagnostic decision support in BM patients undergoing SRS, with a focus on radiomics and deep learning tools for distinguishing RN from progression, classifying molecular and histologic subtypes, and predicting treatment response. METHODS: A systematic review was performed in accordance with PRISMA guidelines. PubMed, Web of Science, and Scopus were searched using a targeted query combining terms related to AI, brain metastasis, diagnosis or imaging, and SRS. After screening 483 records and applying strict inclusion and exclusion criteria, 18 studies published between 2015 and 2025 were included. Data were extracted on study design, cohort characteristics, imaging modality, AI methodology, validation strategy, and reported diagnostic performance. RESULTS: Among the 18 included studies, AI models demonstrated strong performance across diagnostic tasks in the BM-SRS pathway. The differentiation of RN from true tumor progression was the most extensively studied application, addressed by 14 of 18 studies, with reported AUCs ranging from 0.71 to 0.94. Support vector machines, random-forest ensembles, convolutional neural networks, and transformer-based multimodal architectures were widely used. The literature evolved from single-sequence radiomic classifiers in 2018 to multimodal deep learning frameworks fusing imaging with clinical and genomic data in 2025. Contrast-enhanced T1-weighted MRI was the dominant imaging input, and texture-based radiomic features (GLCM, GLSZM, GLDM, and wavelet-derived features) were the most consistently predictive. The highest-performing models reached AUCs of 0.85-0.91 through multimodal integration of imaging with clinical and genomic features, and consistently outperformed expert neuroradiologist read on matched cases. Remaining studies addressed longitudinal segmentation-based detection of local failure and adverse radiation effects, BRAF mutation status in melanoma BM, early Gamma Knife treatment response, and primary tumor histology classification, with more variable performance. CONCLUSION: AI models, particularly those integrating MRI-derived radiomic features with clinical and genomic data, show high accuracy in supporting diagnostic decisions for BM patients treated with SRS. The post-SRS differentiation of radiation necrosis from true tumor progression has reached the greatest level of maturity and is closest to clinical translation, with potential to reduce unnecessary biopsies, personalize surveillance intervals, and rationalize treatment-pathway decisions. Other diagnostic applications, including molecular subtyping and primary tumor histology classification, remain exploratory and require further multicenter validation. Integration of AI tools into multidisciplinary tumor-board workflows, combined with prospective validation and standardized reporting, will be essential to realize the full clinical benefits of AI in SRS for brain metastases.

Humans↗