LYLi Yi

Li Yi

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Li Yi has a total of 101 co-authors. Here are the top co-authors by number of shared publications:

  • Ying Song

    Department of Neurology, Peking University Shenzhen Hospital, Shenzhen, China

    8 shared publications
  • Qizhi Xie

    Rehabilitation Department, Shenzhen Longhua District Central Hospital, Shenzhen, Guangdong, China

    6 shared publications
  • Jingjing Jia

    Department of Neurology, Peking University First Hospital, No.8 Xishiku Street, Xicheng District, Beijing 100034, China

    5 shared publications
  • Xiaoyun Zhang

    Neurology Department, Peking University Shenzhen Hospital, Shenzhen, China

    5 shared publications
  • Zhiqiang Wei

    Neurology Department, Peking University Shenzhen Hospital, Shenzhen, China

    5 shared publications
  • Jingjing Sun

    Department of Neurology, Peking University Shenzhen Hospital, Shenzhen, China

    4 shared publications

Latest Publications

  • Independent and Joint Associations of Circadian Rhythm Patterns and Physical Activity Levels With Venous Thromboembolism: A Large-Scale Cohort Study Using Accelerometry Measurements

    by Qian Zhang, Yao Yao, Yanling Wei, Yansun Sun, Siqi Zhu, Li Yi, Ying Shan

    Background: Venous thromboembolism (VTE) constitutes a significant global health burden, yet the contribution of circadian rhythm disruptions to VTE risk remains insufficiently characterized. This study investigated the independent and joint associations of rest-activity circadian rhythm and physical activity (PA) with VTE incidence in a large cohort. Methods: A total of 89,473 UK Biobank participants were analyzed (median follow-up: 12.7 y). Rest-activity rhythm was assessed using accelerometer-derived relative amplitude (RA), and PA was measured by average daily acceleration. VTE events were ascertained via International Classification of Diseases, 10th Revision codes. Cox regression models were used to evaluate linear, nonlinear, and joint associations, adjusting for sociodemographic and clinical covariates. Sensitivity analyses, subgroup analyses, and E -value calculations further supported the robustness of the findings. Results: Each SD increase in RA was associated with a 12% lower VTE risk (hazard ratio, 0.88; 95% CI, 0.85–0.91). Nonlinear analysis of PA identified an inflection point at 30.4m g (milligravity); PA below this level was associated with a 37% higher risk (hazard ratio, 1.37; 95% CI, 1.21–1.54). Participants with both low RA and low PA had a 54% higher VTE risk compared with those with high RA and high PA (hazard ratio, 1.54; 95% CI, 1.34–1.76). Chronotype showed no significant association with VTE. Conclusions: This large-scale cohort study demonstrates that disrupted circadian rhythms and low PA are independently and jointly associated with elevated VTE risk. These findings suggest that circadian and activity monitoring, for example, via wearable devices, may have potential utility in future VTE risk assessment particularly among high-risk populations.


  • Open Access

    Development and validation of a prognostic model for neurological deterioration in acute posterior circulation cerebral infarction

    by Zedan Guo, Qian Zhang, Jiani Wu, Li Yi

    Background: Neurological deterioration occurs in 8%-40% of patients with acute posterior circulation cerebral infarction (APCCI), leading to higher disability and mortality. We developed and validated a prognostic model integrating dynamic physiological indicators and quantitative neuroimaging indicators for early risk stratification. Methods: We retrospectively enrolled patients with APCCI from Peking University Shenzhen Hospital into training set ( n = 447) and testing set ( n = 94). The model incorporates quantitative neuroimaging (mean apparent diffusion coefficient value (MAV) of infarct lesions) and dynamic physiological monitoring (systolic blood pressure variability), combining qualitative and quantitative approaches to better characterize infarction and hemodynamic instability. A nomogram was constructed and evaluated using discrimination, calibration, and clinical decision curve analysis (DCA). Results: The model demonstrated moderate discrimination, with an AUC of 0.74 (95% CI: 0.62–0.86) in the testing set, and ideal calibration (Brier score 0.15). DCA confirmed superior clinical benefit within an 8%–56% threshold range. Both MAV and blood pressure variability were significant predictors, highlighting their added prognostic value beyond conventional factors. Conclusions: The prognostic model effectively identifies APCCI patients at risk of neurological deterioration by incorporating innovative neuroimaging and dynamic monitoring indicators. Temporal validation shows promising performance in the validation cohort, and further multi-center validation is warranted for clinical adoption.


  • Open Access

    Constructing a risk screen for attention difficulty in U.S. adults using six machine learning methods

    by Ying Song, Yansun Sun, Zedan Guo, Li Yi

    Background: Concentration difficulty is recognized as a hallmark of various neurologic and neuropsychiatric disorders. However, an accurate estimation of epidemiological risk factors for concentration difficulty remains severely limited. Aims: The study aimed to develop an interpretable machine-learning (ML) model to predict risk factors of concentration difficulty among adults in the United States. Methods: A total of 9,971 participants were included from the 2015–2016 cycle of the National Health and Nutrition Examination Survey (NHANES). Six ML algorithms, including Logistic Regression, ExtraTrees classifier, Bagging, Gradient Boosting, Extreme Gradient Boosting (XGBoost), and Random Forest (RF), were applied in this study. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), accuracy, precision, specificity, decision curve analysis (DCA), and calibration plots. Finally, a nomogram was constructed based on the best performing model. Results: Of these, 2,146 participants aged 20 years and older were analyzed. Logistic regression exhibited the best clinical predictive value in both internal and external validation sets, with AUCs of 0.881 and 0.818, respectively. The DCA curve revealed that logistic regression exhibited the greatest net benefits in the internal cohort, whereas the RF model provided the largest net benefits in the external cohort (threshold: 0.2–0.3). Conclusion: Logistic regression exhibited the highest clinical value in predicting concentration difficulty. These findings provide valuable insights for the recognition, management, and effective interference strategies for concentration difficulty.


  • Open Access

    Using machine learning model for predicting risk of memory decline: A cross sectional study

    by Ying Song, Yansun Sun, Qi Weng, Li Yi


  • Open Access

    Effects of multiple fluid intake on Urolithiasis by mendelian randomization study

    by Qian Zhang, Zedan Guo, JinYuan Zhang, Hang Liu, Li Yi

    Urolithiasis, a common and recurrent condition, imposes a significant global health burden. This study investigates the causal relationship between various types of fluid intake and urolithiasis using univariable and multivariable Mendelian randomization (MR) approaches. Genetic data for urolithiasis were obtained from the FinnGen Consortium (n = 400,681), while genetic measures for nine fluid types were sourced from meta-analyses of genome-wide association studies (GWAS). The analysis revealed that genetically predicted coffee intake (OR: 0.571, 95% confidence interval (CI): 0.341–0.958, p = 0.034) and ground coffee intake (OR: 0.219, 95% CI: 0.005–0.923, p = 0.038) were protective factors for upper urinary calculi. Additionally, a standard deviation (SD) increase in tea intake was associated with a 53% reduction in risk (OR: 0.473, 95% CI: 0.320–0.700, p < 0.001). Conversely, milk intake (OR: 1.072, 95% CI: 1.018–1.130) and alcohol intakefrequency (OR: 1.304, 95% CI: 1.083–1.570, p = 0.005) increased the risk of upper urinary calculi. These findings highlight the association of tea with reduced risk of urolithiasis, instead for milk and alcohol intake, in formulating prevention strategies for urolithiasis.


  • Open Access

    Serum neurofilament light chain levels are associated with depression among US adults: a cross-sectional analysis among US adults, 2013–2014

    by Ying Song, Huili Jiao, Qi Weng, Hang Liu, Li Yi

    Background: Serum neurofilament light chain (sNfL) has been identified as a biomarker for neurologic diseases. However, sNfL remains unknown to be responsible for depression. Aims: The aim of this research was to explore the relationship between sNfL levels and depression in US adults. Methods: In this cross-sectional survey of the general population, we investigated representative data involving 10,175 participants from the 2013–2014 cycle of the National Health and Nutrition Examination Survey (NHANES). Depression was diagnosed using the Patient Health Questionnaire-9 (PHQ-9). The effect of related factors on depression was analyzed by conducting a univariate analysis. Stratified analysis was utilized to detect the stability and sensitivity of the relationship. After adjusting for race, education, marital status, smoking status, body mass index (BMI), sleep duration, income, and a history of hypertension, sedentary behavior and stroke, multivariable linear regression was performed to demonstrate the correlation between sNfL and depression. Results: A total of 1301 individuals between the ages of 20 and 75 were involved in this investigation, of which 108 (8.3%) were diagnosed with depression. A significant positive correlation between sNfL and depression among adults in the US was observed by conducting univariable analyses. After adjusting for confounding factors, the multivariate analyses indicated that elevated sNfL levels might play a pivotal role in the development of depression (odds ratio (OR) = 3.0; 95% confidence interval (CI): (1.5, 6.1), P = 0.002). Conclusion: These results indicated that sNfL is closely linked to depression in a nationally representative individual. However, further studies are needed to confirm the biological mechanism as well as the clinical implications of sNfL and depression.


  • Open Access

    Long noncoding RNA UNC5B-AS1 suppresses cell proliferation by sponging miR-24-3p in glioblastoma multiforme

    by Ying Song, Baodong Chen, Huili Jiao, Li Yi

    Background: Glioblastoma multiforme (GBM) is the most common primary CNS tumor, characterized by high mortality and heterogeneity. However, the related lncRNA signatures and their target microRNA (miRNA) for GBM are still mostly unknown. Therefore, it is critical that we discover lncRNA markers in GBM and their biological activities. Materials and methods: GBM-related RNA-seq data were obtained from the Cancer Genome Atlas (TCGA) database. The “edger” R package was used for differently expressed lncRNAs (DELs) identification. Then, we forecasted prospective miRNAs that might bind to lncRNAs by Cytoscape software. Survival analysis of those miRNAs was examined by the starBase database, and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of the miRNAs’ target genes was conducted by the Gene Set Enrichment Analysis (GSEA) database and R software. Moreover, the proliferative ability of unc-5 netrin receptor B antisense RNA 1 (UNC5B-AS1) cells was evaluated by Cell Counting Kit-8 (CCK-8) analysis. Mechanistically, the regulatory interaction between UNC5B-AS1 and miRNA in GBM biological processes was studied using CCK-8 analysis. Results: Our results indicated that overexpression of UNC5B-AS1 has been shown to suppress GBM cell growth. Mechanistically, miR-24-3p in GBM was able to alleviate the anti-oncogenic effects of UNC5B-AS1 on cell proliferation. Conclusion: The discovery of the novel UNC5B-AS1-miR-24-3p network suggests possible lncRNA and miRNA roles in the development of GBM, which may have significant ramifications for the analysis of clinical prognosis and the development of GBM medications.


  • Open Access

    The prevalence and risk factors of anxiety in multiple sclerosis: A systematic review and meta-analysis

    by Xiaoyun Zhang, Ying Song, Zhiqiang Wei, Xiao Chen, Xiaojia Zhuang, Li Yi

    Background: Patients with multiple sclerosis (MS) suffer from repetitive neurological deterioration, while anxiety may play a significant role in the disease’s progression.Objective: To explore the prevalence of anxiety in MS and to investigate the risk factors related to anxiety in MS patients.Methods: An analysis of four databases, PubMed, Web of Science, EMBASE, and Cochrane Library, has been conducted to determine the prevalence or risk factors for anxiety in MS published before May 2021.Results: In total, 32 studies were found to be eligible. Anxiety prevalence was estimated to be 36% based on the pooled estimates [the 95% confidence interval (CI) = [0.30–0.42],I2= 98.4%]. Significant risk factors for developing of anxiety were as follows: age at survey [the weighted mean difference (WMD) = 0.96, 95% CI = [0.86–1.06],I2= 43.8%], female [the odd ratio (OR) = 1.78, 95% CI = [1.38–2.30],I2= 0%], living together (OR 2.83, 95% CI = [1.74–4.59],I2= 0%), past psychiatric history (OR 2.42, 95% CI = [1.56–3.75],I2= 0%), depression (OR 7.89, 95% CI = [3.71–16.81],I2= 0%), not taking MS medication (OR 2.33, 95% CI = [1.29–4.21],I2= 77.8%), relapsing-remitting MS (RRMS) (OR 1.50, 95% CI = [0.94–2.37],I2= 53.5%), and baseline Expanded Disability Status Scale (EDSS) (OR 0.84, 95% CI = [0.48–1.21],I2= 62.2%).Conclusion: An estimated 36% of people with MS suffer from anxiety. And anxiety rates in MS patients are significantly associated with age, gender, living together, prior psychiatric history, depression, drug compliance, RRMS, and baseline EDSS.Systematic review registration: https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=287069, identifier CRD42021287069.


  • Open Access

    Management of Rheumatoid Arthritis With a Digital Health Application

    by Chun Li, Jianlin Huang, Huaxiang Wu, Fen Li, Yi Zhao, Zhenchun Zhang, Shengguang Li, Hua Wei, Miaojia Zhang, Hongsheng Sun, et al.

    IMPORTANCE: Digital health applications have been shown to be effective in the management of chronic diseases with simple treatment targets. The potential clinical value of digital health applications in rheumatoid arthritis (RA) has not been well studied.OBJECTIVE: To investigate whether assessing patient-reported outcomes using digital health applications could result in disease control for patients with RA.DESIGN, SETTING, AND PARTICIPANTS: This is a multicenter, open-label randomized clinical trial in 22 tertiary hospitals across China. Eligible participants were adult patients with RA. Participants were enrolled from November 1, 2018, to May 28, 2019, with a 12-month follow-up. The statisticians and rheumatologists who assessed disease activity were blinded. Investigators and participants were not blind to group assignment. Analysis was conducted from October 2020 to May 2022.INTERVENTIONS: Participants were randomly assigned at a 1:1 ratio (block size of 4) to a smart system of disease management group (SSDM) or a conventional care control group. Upon the completion of the 6-month parallel comparison, patients in the conventional care control group were instructed to use the SSDM application for an extension of 6 months.MAIN OUTCOMES AND MEASURES: The primary outcome was the rate of patients with disease activity score in 28 joints using the C-reactive protein (DAS28-CRP) of 3.2 or less at month 6.RESULTS: Of 3374 participants screened, 2204 were randomized, and 2197 patients with RA (mean [SD] age, 50.5 [12.4] years; 1812 [82.5%] female) were enrolled. The study included 1099 participants in the SSDM group and 1098 participants in the control group. At month 6, the rate of patients with DAS28-CRP of 3.2 or less was 71.0% (780 of 1099 patients) in the SSDM group vs 64.5% (708 of 1098 patients) in the control group (difference between groups, 6.6%; 95% CI, 2.7% to 10.4%; P = .001). At month 12, the rate of patients with DAS28-CRP of 3.2 or less in the control group increased to a level (77.7%) that was comparable with that (78,2%) in the SSDM group (difference between groups, −0.2%; 95% CI, −3.9% to 3.4%; P = .90).CONCLUSIONS AND RELEVANCE: In this randomized clinical trial of RA, the use of a digital health application with patient-reported outcomes was associated with an increase in disease control rate.TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT03715595


  • Open Access

    Non-invasive brain stimulation therapy on neurological symptoms in patients with multiple sclerosis: A network meta analysis

    by Xiaoyun Zhang, Yaping Huai, Zhiqiang Wei, Weiwei Yang, Qizhi Xie, Li Yi

    Objective: The aim of the study was to evaluate non-invasive brain stimulation (NIBS) [including transcranial magnetic stimulation (TMS) and transcranial electrical stimulation (tES)] on neurological symptoms in patients with multiple sclerosis (PwMS).Method: We searched PubMed, Embase, Cochrane Library, Web of Science and Ovid MEDLINE until February 2022. And we evaluated the included studies for methodological quality by the Cochrane bias risk assessment tool and assessed the studies' certainty of evidence using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework. We performed network meta analysis (NMA) by using Stata 15 and ranked the results of the NMA by using the surface under the cumulative ranking curve (SUCRA) ranking chart.Result: Twenty seven clinical trials were finally included (N = 596, 66.4% women). For the immediate effects, rTMS over M1 yielded the most optimal scheme for fatigue reduction among all the interventions compared to the sham stimulation groups [MD = −0.85, 95% CI (−1.57, −0.14)] (SUCRA = 82.6%). iTBS over M1 yielded the most signifcant reduced pain level than the sham groups did [MD = −1.26, 95% CI (−2.40, −0.11)] (SUCRA = 98.4%). tDCS over F3 was the best protocol of NIBS to improve quality of life (QOL) [MD = 1.41, 95% CI = (0.45,2.36)] (SUCRA = 76.7%), and iTBS over M1 may significantly reduce spasticity compared to sham stimulation [MD = −1.20, 95% CI = (−1.99, −0.41)] (SUCRA = 90.3%). Furthermore, rTMS, tRNS, and tDCS on certain areas may improve PwMS accuracy, response time, manual dexterity, pain relief and QOL, but does not show statistically significant differences. The evidence assessed using GRADE is very low.Conclusion: Based on the NMA and SUCRA ranking, we can conclude that symptoms including fatigue, pain, spasticity, and QOL can be improved by following NIBS protocol after treatment. Nonetheless, most of the included studies lack a good methodology, and more high-quality randomized clinical trials are needed.