In recent years， artificial intelligence （AI） technologies have developed rapidly in the medical field and have gradually been applied to the diagnosis， treatment decision-making， and long-term management of neurological disorders.Neurological disorders are generally characterized by complex etiologies， marked heterogeneity in clinical manifestations， diverse sources of diagnostic and therapeutic information， and a high demand for full-course disease management. Conventional diagnostic and treatment models still have certain limitations in early identification， risk prediction， prognostic evaluation， and individualized management. AI technologies， including machine learning， deep learning， multimodal fusion models， and large language models， can integrate medical imaging， electroencephalographic signals， clinical data， laboratory indicators， pathological and molecular testing results， omics data， and information from wearable devices. These technologies have shown potential applications in emergency stroke imaging recognition and prognostic prediction， electroencephalographic analysis and seizure monitoring for epilepsy， early screening and progression prediction for cognitive impairment， quantification of motor symptoms and long-term management of Parkinson disease， lesion segmentation and molecular marker prediction for brain tumors， as well as auxiliary differentiation and risk assessment of neuroimmune disorders. However， the clinical translation of AI in neurological disorders still faces several challenges， including insufficient data quality and standardization， limited external validation， inadequate model generalizability and interpretability， and unclear boundaries of ethical regulation and responsibility.This review focuses on the research progress and application value of AI in major neurological disorders， and further analyzes current challenges and future development directions， aiming to provide a reference for intelligent diagnosis and treatment as well as human-AI collaborative management of neurological disorders.

Neurodegenerative diseases， particularly Alzheimer disease （AD） and Parkinson disease （PD）， are prominent and urgent public health problems in the context of global population aging. With the rapid development of artificial intelligence （AI） technologies in recent years， the integration of multimodal imaging analysis and digital biomarkers has led to important breakthroughs in the early screening， accurate diagnosis， and progression prediction of these two diseases. Therefore， this paper systematically reviews the latest research advances in the application of AI to the diagnosis of neurodegenerative diseases， including multimodal image fusion technologies， digital biomarker extraction methods， and the latest innovations in deep learning algorithms， as well as various challenges encountered in clinical translation. This paper provides a detailed analysis of cutting-edge technologies such as machine learning models based on blood biomarkers， deep learning architectures， smartphone application-based assessments， and monitoring with wearable devices， confirming their potential value in the early detection of AD and PD. On this basis， the paper demonstrates the superiority of multimodal AI models in the diagnosis of AD and PD， discusses the prediction models for the conversion from mild cognitive impairment to AD， and outlines future research directions. This paper provides certain technical guidance and future development directions for AI-aided diagnosis of neurodegenerative diseases.

The systematic application of artificial intelligence （AI） in the medical field is profoundly reshaping the paradigms for disease diagnosis， prognostic assessment， and treatment decision-making. Neurological diseases are characterized by high phenotypic heterogeneity， complex multimodal data structures， and uneven distribution of professional resources， thereby becoming one of the most strategically valuable and challenging fields for AI-enabled medical applications. This article systematically reviews the evolution of AI in neuroscience and neurology， from early rule-based expert systems to modern deep learning frameworks. It also summarizes the clinical application of AI in major neurological diseases including stroke， Alzheimer disease， Parkinson disease， brain tumors， and multiple sclerosis and elaborates on the latest targeted advances of AI in rare neurological diseases such as rare epilepsy syndromes， rare neuromuscular disorders， leukoencephalopathy， Huntington disease， and mitochondrial encephalopathy. Finally， it looks ahead to the future development trends of explainable AI， federated learning， and digital biomarker systems. This article aims to provide a systematic reference framework for neurologists， researchers， and clinical policymakers regarding AI-powered precise diagnosis and treatment of neurological diseases.

Contrast-enhanced computed tomography （CECT） uses iodinated contrast agents to improve the contrast between the lesion and adjacent tissue， and it plays a critical role in tumor staging， evaluation of vascular diseases， emergency screening， radiation therapy planning， and preoperative planning； however， its clinical application is limited by allergy-like reactions， contrast-associated acute kidney injury， complicated workflows， increased medical costs， and prolonged emergency management time. Non-contrast computed tomography （NCCT） is safe and convenient， and deep learning models have shown the promising performance of NCCT in the screening of pancreatic cancer and colorectal cancer and the assessment of acute ischemic stroke； nevertheless， insufficient soft tissue contrast still limits the detection of hypovascular lesions and small lesions. Generation of virtual CECT images from NCCT based on artificial intelligence aims to retrieve enhancement information of blood vessels， parenchymal organs， lymph nodes， and tumors without administering additional contrast agents， thereby reducing the risk of patients and optimizing the allocation of medical resources. This article summarizes the mainstream technical approaches including generative adversarial network （GAN）， conditional GAN， CycleGAN， Wasserstein GAN， multi-stage cascaded networks， CNN-Transformer hybrid architectures， diffusion probabilistic models， and conditional diffusion models， as well as the application of anatomical priors， segmentation labels， organ/vessel masks， multi-task learning， clinical parameters， and multimodal information. Related studies have been extended to multiple scenarios such as CTA generation， multi-phase hepatic contrast-enhanced imaging， abdominal tumor screening， identification of large-vessel occlusion in acute ischemic stroke， and segmentation of organs at risk for radiotherapy. Overall， although this technique holds great potential for reducing the use of contrast agents， assisting clinical diagnosis， and optimizing clinical workflows， it still faces several challenges including a lack of paired datasets， limited cross-center generalization capability， imaging artifacts and model hallucinations， as well as undefined safety thresholds and regulatory frameworks in clinical practice.

Artificial intelligence （AI）-assisted electroencephalography （EEG） analysis is profoundly transforming the diagnosis and treatment paradigm of disorders of consciousness（DOC）. In recent years， through innovative approaches such as machine learning， deep learning， multimodal fusion， and brain-computer interfaces， this technique has achieved remarkable progress in several key areas including the diagnosis and classification of DOC， detection of covert consciousness， prognosis prediction， and therapeutic intervention. This article systematically reviews the latest research advances in AI-assisted EEG analysis for the diagnosis and treatment of DOC， analyzes the core challenges， and proposes future development directions， in order to provide a systematic reference for the translational application of this technique in the clinical diagnosis and treatment of DOC.

Drug research and development for central nervous system （CNS） diseases has long been hindered by challenges including difficulties in target identification， restricted blood-brain barrier （BBB） penetration， and low efficiency of clinical translation. In recent years， artificial intelligence （AI）， combined with single-cell and spatial multi-omics， protein structure prediction， and generative models， has provided new technical strategies for drug research and development in the field of CNS diseases. This article reviews the advances in the application of AI in drug research and development for CNS diseases， including disease-related cell state identification， the analysis of cell-cell communication and spatial microenvironment， protein structure prediction and molecular design， as well as the evaluation of BBB permeability， brain exposure， and drug safety. In addition， this article summarizes the role of perturbomics， brain organoids， and BBB-on-a-chip platforms in validating AI predictions and discusses the closed-loop research framework of "prediction-validation-iteration". Finally， this article analyzes the issues such as insufficient data standardization， limited model interpretability， and limited physiological relevance of in vitro models and proposes the future directions of AI-driven drug research and development for CNS diseases.

Objective To develop an artificial intelligence-based multimodal model for stroke risk screening， and to improve the accuracy and efficiency of stroke screening. Methods With reference to the mainstream models like the “8+2” model of the National Stroke Prevention and Control Committee， the Prediction for ASCVD Risk in China （China-PAR） model， and the China Kadoorie Biobank （CKB） model， a multimodal model for stroke risk screening and prevention was established by integrating multi-dimensional data including blood biochemistry， clinical diagnosis， individual medical history， signs， and living habits. This model was applied in the clinical verification of 6 511 individuals who participated in stroke risk screening in Shaoxing People’s Hospital from January to December 2023， and the above models were compared in terms of the consistency of prediction results and the acceptance rate among physicians. Results The consistency rate of risk classification was 71.31% between the self-developed model and the China-PAR model and 70.80% between the self-developed model and the CKB model， which were significantly higher than the consistency rates between the CKB model and the China-PAR model and between the “8+2” model and the other models. The consistency rate of risk classification was 89.79% between the self-developed simplified model （with 10 biomarkers） and the complete model （with 17 biomarkers）， suggesting that the self-developed simplified model is suitable for primary outpatient screening or large-scale physical examination. In addition， the self-developed model further improved the accuracy of risk assessment through percentage conversion and refinement of risk levels. Conclusion The intelligent screening model based on multimodal fusion can effectively improve the accuracy and timeliness of stroke screening， thereby providing a feasible solution for large-scale stroke screening.

This article reports a case of delayed oculomotor nerve palsy after basal ganglia hemorrhage breaking into the ventricles and discusses its clinical features and possible pathogenic mechanisms，so as to provide a reference for clinical diagnosis and treatment. A retrospective analysis was performed for the clinical data of a male patient with intracerebral hemorrhage aged 57 years，including medical history，physical examination，imaging examination，diagnosis，and treatment，and the possible pathogenic mechanisms were analyzed with reference to related articles in China and globally. Head computed tomography angiography showed no signs of aneurysm. The patient achieved absorption of intracerebral hemorrhage after treatment and complete resolution of blepharoptosis and recovery of eye movement at 2 months after discharge. Delayed oculomotor nerve palsy may occur after intracerebral hemorrhage breaking into the ventricles，and its pathogenesis may be associated with factors such as nerve compression caused by organization and contraction of intraventricular blood clots and vasospasm. Conservative treatment may help to achieve a good prognosis. This case can provide a reference for clinical practice.

Migraine is a neurovascular disorder and ranks as the second leading cause of disability worldwide. The Global Burden of Disease Study 2021 estimated that approximately 1.2 billion people are affected by migraine，making it one of the major contributors to years lived with disability years attributable to neurological disorders. The pathogenesis of migraine is complex and involves multiple processes，including activation of the trigeminovascular system，neuroinflammation，abnormal meningeal vasomotion，and central sensitization. Among these，potassium channels play a vital role in maintaining the resting membrane potential of cells，regulating neuronal excitability，modulating neurotransmitter release，and controlling vascular tone. In recent years，the inwardly rectifying potassium channel （Kir） family has become a hotspot in headache research. This family includes Kir2.x，G protein-coupled inwardly rectifying potassium channels （GIRK/Kir3.x），and ATP-sensitive potassium channels （K_ATP，Kir6.x/SUR），which are involved in membrane potential stabilization，receptor-coupled inhibition，spatial potassium buffering in glial cells，and metabolic-electrical activity coupling. Notably，the K_ATP channel opener levcromakalim can reliably induce migraine attacks in humans，making K_ATP channels one of the most translationally promising ion channel targets at present. This article focuses on the role of inwardly rectifying potassium channels in migraine，summarizes the distribution，mechanisms，and physiological functions of various Kir channels，and discusses the evidence supporting the involvement of the Kir family in migraine. Furthermore，it reviews the interactions among different Kir channels，as well as their crosstalk with canonical migraine pathways including CGRP，PACAP，and NO. Finally，it analyzes relevant therapeutic targets，current research limitations，and future directions. Overall，inwardly rectifying potassium channels are situated at critical nodes of functional coupling among neurons，glial cells，and vascular smooth muscle，representing an important entry point for understanding migraine heterogeneity and developing precision therapeutics.

Vestibular migraine （VM） is not only a common cause of episodic vertigo but also a systemic disorder involving extensive brain network dysfunction. In addition to typical headache and severe vestibular symptoms，VM is often accompanied by significant cognitive alterations that seriously impair the daily functioning and social capacity of patients. This article systematically reviews the latest research advances in VM-associated cognitive impairment in terms of clinical phenotypes，potential pathophysiological mechanisms，biomarkers，and assessment scales. Clinical assessments show that VM-associated cognitive impairment mainly manifests as retardation of thinking，impaired attention and working memory，increased cognitive load，and impaired executive function，as well as a potential increase in the risk of falls. Neuroimaging and neuroelectrophysiological evidence has revealed the presence of microstructural damage along with macroscopic functional network reorganization and structural alterations in the brains of VM patients. Besides，5-hydroxytryptamine，calcitonin gene-related peptide，macrophage migration inhibitory factor，and 25-OH vitamin D may participate in the development and progression of VM-associated cognitive impairment. In conclusion，through a comprehensive assessment of scales，neuroimaging，electrophysiological examination，and serological biomarkers，cognitive screening in VM patients can achieve early detection of cognitive dysfunction，and effective interventions should be performed to control the symptoms of VM while protecting cognitive function，so as to improve long-term prognosis and quality of life for these patients.

Migraine is a common disabling trigeminovascular disorder，and its comorbidity with ocular diseases has attracted increasing attention in clinical practice. This review focuses on the comorbid relationship between migraine and dry eye disease，constructs a whole-process clinical management pathway，emphasizes the core value of multidisciplinary team collaboration between ophthalmology and neurology in optimizing patient prognosis，and points out that prospective multicenter studies are needed in the future to further reveal the mechanism of such comorbidity and explore biomarkers and novel targeted therapies，so as to improve the long-term quality of life of patients with this comorbidity.

Cluster headache（CH） is a severe primary headache disorder characterized by excruciating pain，with a complex pathophysiology involving intricate interactions among the trigeminovascular system，the hypothalamic circadian clock，and the autonomic nervous system. In recent years，calcitonin gene-related peptide （CGRP），recognized as a core mediator in migraine pathophysiology，has also attracted substantial attention regarding its role in CH. This review aims to thoroughly dissect the central role of CGRP in the pathogenesis of CH，systematically elaborate on the relationship between the circadian rhythmicity of CH attacks and the regulation of CGRP pathways，and，through differential diagnostic analysis of CH versus other trigeminal autonomic cephalalgias，reveal the distinct roles of CGRP pathways across different headache subtypes. By integrating the latest preclinical studies，clinical trial data，and biomarker explorations，this article seeks to provide a comprehensive reference for understanding the complex pathophysiology of CH and for developing novel targeted therapeutic strategies.

Migraine is a common primary headache disorder with a high disability burden，and temporomandibular disorders （TMD） are among the most prevalent chronic pain conditions in the orofacial region. Studies have shown a significant bidirectional comorbidity between migraine and TMD，with pain-related TMD being particularly closely associated with migraine. Patients with this comorbidity often experience greater pain burden，more pronounced psychological distress，and poorer quality of life. The underlying mechanisms may involve multiple factors，including the convergence of pain signals within the trigeminal system，peripheral and central sensitization，abnormal neuropeptide signaling involving calcitonin gene-related peptide，oral parafunctional behaviors，sleep disturbances，and psychological stress. The third edition of the International Classification of Headache Disorders （ICHD-3） classifies “headache attributed to temporomandibular disorder” as an entry for secondary headache diagnosis. Clinically，attention should be paid to differentiating migraine comorbid with TMD from TMD-related secondary headache. This review focuses on the epidemiological association，potential comorbid mechanisms，diagnostic differentiation，and oral clinical implications of migraine and TMD，aiming to provide a reference for the standardized identification and comprehensive management of affected patients.

Objective To investigate the association of the polymorphisms of vitamin D receptor ApaI and BsmI genes with the risk of vestibular migraine（VM）. Methods A retrospective case-control study was conducted. A total of 200 patients with VM who were admitted to The Affiliated Xuzhou Municipal Hospital of Xuzhou Medical University from April 2024 to June 2025 were enrolled as VM group，and 100 healthy individuals who underwent physical examination in the same hospital during the same period of time were enrolled as control group. Peripheral blood DNA samples were collected from all subjects to investigate the polymorphisms of the ApaI gene at the rs7975232 locus and the BsmI gene at the rs1544410 locus，and the genotype and allele frequencies at these loci were analyzed for both groups，as well as their association with the onset of VM. Results There were significant differences between the VM group and the control group in the genotype distribution of the ApaI gene at the rs7975232 locus and the BsmI gene at the rs1544410 locus（P<0.05），and compared with the control group，the VM group had a significantly higher frequency of A allele at the rs7975232 locus of the ApaI gene and a significantly higher frequency of G allele at the rs1544410 locus of the BsmI gene（P<0.05）. Compared with the control group，the VM group had a significantly higher proportion of patients with a family history of migraine and significant increases in the levels of total cholesterol and serum uric acid（P<0.05）. The Logistic regression analysis showed that a family history of migraine，a high level of serum uric acid，the AA genotype at the rs7975232 locus of the ApaI gene，and the GG genotype at the rs1544410 locus of the BsmI gene were risk factors for the onset of VM（odds ratio>1，P<0.05）. Conclusion The AA genotype at the rs7975232 locus of the ApaI gene and the GG genotype at the rs1544410 locus of the BsmI gene may be associated with the susceptibility to VM. Detection of the genotypes of these two genes can help to assess the risk of VM.

Trigeminal autonomic cephalalgias（TACs）are a group of primary headache syndromes characterized by unilateral headache and ipsilateral craniofacial autonomic symptoms. TACs include cluster headache，paroxysmal hemicrania，short-lasting unilateral neuralgiform headache attacks（including short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing and short-lasting unilateral neuralgiform headache attacks with craniofacial autonomic symptoms），and hemicrania continua. The pathophysiology involves the trigeminovascular system，the autonomic nervous system，the hypothalamus，and the recently identified role of the vagus nerve. Diagnosis mainly relies on headache attack frequency，duration，and accompanying symptoms. Each TAC has its unique clinical manifestations，pathophysiology，and treatment methods，which are discussed in depth in this review.

Migraine，as a globally high-burden neurological disorder，imposes a dual burden on individual quality of life and the public health system due to its high prevalence rate，high misdiagnosis rate，and significant disability during attacks. Traditional diagnosis and treatment modes face problems such as a lack of individualized treatment regimens，the difficulty in predicting attacks，insufficient objective basis for diagnosis，and the risk of excessive drug use. In recent years，machine learning-based research methods have shown strong potential in integrating clinical data，thereby providing new research ideas for the precise prediction，diagnosis and classification，and individualized treatment decision-making of migraine. This article reviews the application of machine learning in migraine in recent years.

Primary headaches，particularly migraine and tension-type headache，are among the leading causes of disability worldwide. This article systematically reviews recent international and domestic advances in the field of nursing care for primary headaches and elaborates on the critical role of nurses in the comprehensive management of headaches. Evidence indicates that nurses serve as a vital bridge connecting patients with the healthcare system，coordinating multidisciplinary teams，and facilitating patient self-management. In terms of nursing assessment，headache nurses perform systematic history taking，physical and psychological evaluation，and psychosocial assessment to lay the foundation for individualized care plans. Regarding treatment implementation，headache nurses are responsible for monitoring adverse reactions to acute-phase and preventive medications as well as for preventing and monitoring medication-overuse headache. In terms of non-pharmacological interventions，headache nurses lead the delivery of multifaceted measures including lifestyle modification，cognitive behavioral therapy，biofeedback，and health education. Of particular importance，headache nurse-led health education models，through headache diary management，digital tool applications，and multidisciplinary collaboration，significantly enhance the disease awareness，treatment adherence，and self-efficacy of patients. In the future，establishing headache nurse-led headache centers and community-based nursing systems will shift headache management from passive symptom control toward active functional rehabilitation and improvement of quality of life.

Epilepsy is a neurological disorder with marked genetic heterogeneity， and genetic factors play an important role in the development and progression of epilepsy. High-throughput sequencing， genome-wide association studies， and large-scale genomic databases have greatly accelerated the discovery of epilepsy-associated genes and precision medicine； however， the pathogenic interpretation of low-frequency variants， complex inheritance patterns， and variants with mild functional impairment remains challenging. The gene dependency principle addresses how life activities depend on gene function and proposes the key concepts including functional dependency， genetic dependent nature， genetic dependent quantity， and genetic dependent stage， which provides a new medical genetics framework for understanding gene-disease relationship and the pathogenicity of variants.Based on this principle， the 3-Ⅰ approach emphasizes individualized phenotype analysis， individualized case analysis， and individualized gene analysis， integrating whole-exome sequencing， gene dependency assessment， and the China Epilepsy Gene 1.0 Project cohort， and this framework has facilitated the discovery or identification of more than 40 novel pathogenic genes or candidate pathogenic genes for epilepsy， thereby providing a theoretical basis for precise diagnosis， individualized therapy， genetic counseling，and disease prevention.Recently， the new pathogenic gene phenotypes including USP25， ZFHX3， and ATP6V0C have been recognized by OMIM， which fully reflects the application value of the gene dependency principle in the discovery of novel epilepsy-associated genes.

This article summarizes the research advances in the clinical manifestations， related factors， and underlying mechanisms of cognitive impairment in elderly adults with epilepsy. The prevalence rate of cognitive impairment ranges from 26% to 87% in elderly adults with epilepsy， with the main manifestations of impairment in short- and long-term visual memory， verbal memory， executive function， attention， and psychomotor speed. Cognitive function in elderly adults with epilepsy shows a progressive decline over time， which can be explained by the parallel decline model， the two-hit model， and the accelerated aging model. Epilepsy comorbid with cognitive impairment in elderly adults may result from the combined effect of epilepsy itself， brain structural and functional changes， comorbidity burden， anti-seizure medications， and pathological changes associated with Alzheimer disease. Frequent seizures and subclinical epileptiform discharges may accelerate cognitive decline by disrupting neural network activity and memory consolidation processes. The bidirectional relationship between epilepsy and dementia suggests the existence of shared pathological mechanisms leading to cognitive impairment.

Electroencephalography（EEG） is a fundamental tool in the diagnosis and management of epilepsy， and in recent years， its role in epilepsy associated with malformations of cortical development（MCD） has expanded from a conventional tool for adjunctive diagnosis to one used to reveal structural etiologies and assess prognosis. This article systematically reviews the scalp EEG features of patients with MCD， summarizes the electrophysiological manifestations of common MCD subtypes， and highlights the value of EEG in the diagnosis of MCD and the potential significance of its discharge patterns in prognostic evaluation， in order to provide a reference for clinical diagnosis and treatment.