ObjectiveThe essence of syndrome manifestation recognition in traditional Chinese medicine (TCM) is to infer the body’s latent pathogenesis state from clinical observational information, rather than to perform simple label matching. However, previous studies have largely modeled this task as syndrome pattern classification within a fixed label space, which does not adequately reflect the cognition process of TCM syndrome differentiation centered on pathogenesis reasoning, and is also insufficient to capture the openness, semantic variability, and cross-disease reusability of syndrome manifestation expression. This study aimed to investigate whether introducing pathogenesis reasoning chain-of-thought (PR-CoT) supervision into large language models (LLMs) could improve the quality and cognitive consistency of syndrome manifestation recognition and support cross-disease transfer. MethodsSyndrome manifestation recognition was formulated as a conditional generation task under the framework of clinical observational information (X)→pathogenesis structure (Z)→syndrome pattern output (Y), where Z serves as an explicit intermediate structural variable linking the clinical evidence and syndrome judgment. Within this framework, a PR-CoT-supervised dataset for syndrome manifestation recognition was constructed based on medical case records of spleen-stomach disorders. After preprocessing, information extraction, manual proofreading, and data cleaning, the dataset comprised 4 800 training cases, 400 development cases, and 400 test cases. Each sample was annotated with a structured PR-CoT consisting of three progressive levels: clinical information summarization, comprehensive pathogenesis analysis, and syndrome pattern output. Supervised fine-tuning was conducted on open-source LLMs, with an end-to-end model serving as the baseline. Qwen3-32B was used as the primary experimental model, and Qwen3-14B as the scale comparison model. A progressive multidimensional evaluation framework was further established, comprising a structural parsing level, a semantic similarity level, and an expert blind review level. At the structural parsing level, syndrome pattern expressions were decomposed into structural elements and evaluated using Precision, Recall, F1 score, and Jaccard similarity. At the semantic similarity level, independent LLMs scored the theoretical proximity between predicted and reference syndrome patterns. At the expert blind review level, three TCM experts independently evaluated model outputs on two dimensions: syndrome differentiation consistency and terminology standardization of syndrome patterns. In addition, zero-shot cross-disease transfer evaluation was conducted on gynecological and heart-system disorder test sets. ResultsAt the structural parsing level, PR-CoT supervision did not lead to a stable improvement in the element-wise overlap of syndrome pattern structural components. Compared with the corresponding baselines, neither Qwen3-32B nor Qwen3-14B showed consistent advantages in structural matching metrics after the introduction of PR-CoT supervision. In contrast, at the semantic similarity level, PR-CoT supervision produced stable positive gains across different model scales and evaluation systems. The average semantic score of Qwen3-32B increased from 6.425 8 in the baseline model to 6.585 0 after PR-CoT supervision, and that of Qwen3-14B increased from 5.870 0 to 5.964 2. At the expert blind review level, the overall score of Qwen3-32B (PR-CoT) was 7.026 0±0.107 7, higher than 6.416 3±0.288 9 for its baseline. In zero-shot cross-disease testing, the PR-CoT model still showed advantages in semantic evaluation and expert evaluation on both gynecological and heart-system disorder test sets, indicating a certain degree of transferability. ConclusionThe benefits of PR-CoT supervision are mainly reflected in TCM semantic consistency and clinical plausibility, rather than in improved hard matching of structural elements. These findings support understanding syndrome manifestation recognition as a process of generating and expressing latent pathogenesis structures, rather than as a classification task within a traditional fixed label space. By introducing pathogenesis reasoning as an explicit intermediate structure into the modeling process and combining it with a progressive multidimensional evaluation framework, this study provides a methodological pathway for intelligent TCM syndrome differentiation that integrates theoretical alignment, interpretability, and multi-level evaluation.

ObjectiveThe essence of syndrome manifestation recognition in traditional Chinese medicine (TCM) is to infer the body’s latent pathogenesis state from clinical observational information, rather than to perform simple label matching. However, previous studies have largely modeled this task as syndrome pattern classification within a fixed label space, which does not adequately reflect the cognition process of TCM syndrome differentiation centered on pathogenesis reasoning, and is also insufficient to capture the openness, semantic variability, and cross-disease reusability of syndrome manifestation expression. This study aimed to investigate whether introducing pathogenesis reasoning chain-of-thought (PR-CoT) supervision into large language models (LLMs) could improve the quality and cognitive consistency of syndrome manifestation recognition and support cross-disease transfer. MethodsSyndrome manifestation recognition was formulated as a conditional generation task under the framework of clinical observational information (X)→pathogenesis structure (Z)→syndrome pattern output (Y), where Z serves as an explicit intermediate structural variable linking the clinical evidence and syndrome judgment. Within this framework, a PR-CoT-supervised dataset for syndrome manifestation recognition was constructed based on medical case records of spleen-stomach disorders. After preprocessing, information extraction, manual proofreading, and data cleaning, the dataset comprised 4 800 training cases, 400 development cases, and 400 test cases. Each sample was annotated with a structured PR-CoT consisting of three progressive levels: clinical information summarization, comprehensive pathogenesis analysis, and syndrome pattern output. Supervised fine-tuning was conducted on open-source LLMs, with an end-to-end model serving as the baseline. Qwen3-32B was used as the primary experimental model, and Qwen3-14B as the scale comparison model. A progressive multidimensional evaluation framework was further established, comprising a structural parsing level, a semantic similarity level, and an expert blind review level. At the structural parsing level, syndrome pattern expressions were decomposed into structural elements and evaluated using Precision, Recall, F1 score, and Jaccard similarity. At the semantic similarity level, independent LLMs scored the theoretical proximity between predicted and reference syndrome patterns. At the expert blind review level, three TCM experts independently evaluated model outputs on two dimensions: syndrome differentiation consistency and terminology standardization of syndrome patterns. In addition, zero-shot cross-disease transfer evaluation was conducted on gynecological and heart-system disorder test sets. ResultsAt the structural parsing level, PR-CoT supervision did not lead to a stable improvement in the element-wise overlap of syndrome pattern structural components. Compared with the corresponding baselines, neither Qwen3-32B nor Qwen3-14B showed consistent advantages in structural matching metrics after the introduction of PR-CoT supervision. In contrast, at the semantic similarity level, PR-CoT supervision produced stable positive gains across different model scales and evaluation systems. The average semantic score of Qwen3-32B increased from 6.425 8 in the baseline model to 6.585 0 after PR-CoT supervision, and that of Qwen3-14B increased from 5.870 0 to 5.964 2. At the expert blind review level, the overall score of Qwen3-32B (PR-CoT) was 7.026 0±0.107 7, higher than 6.416 3±0.288 9 for its baseline. In zero-shot cross-disease testing, the PR-CoT model still showed advantages in semantic evaluation and expert evaluation on both gynecological and heart-system disorder test sets, indicating a certain degree of transferability. ConclusionThe benefits of PR-CoT supervision are mainly reflected in TCM semantic consistency and clinical plausibility, rather than in improved hard matching of structural elements. These findings support understanding syndrome manifestation recognition as a process of generating and expressing latent pathogenesis structures, rather than as a classification task within a traditional fixed label space. By introducing pathogenesis reasoning as an explicit intermediate structure into the modeling process and combining it with a progressive multidimensional evaluation framework, this study provides a methodological pathway for intelligent TCM syndrome differentiation that integrates theoretical alignment, interpretability, and multi-level evaluation.

Objective:To report the nationwide surveillance results of pathogenic profiles and antimicrobial resistance patterns of Gram-positive bloodstream infections in China in 2023.Methods:The clinical isolates of Gram-posttive bacteria from blood cultures were collected in member hospitals of National Bloodstream Infection Bacterial Resistant Investigation Collaborative System（BRICS）during January to December 2023. Antimicrobial susceptibility testing was performed using the dilution method recommended by the Clinical and Laboratory Standards Institute（CLSI）. Statistical analyses were conducted using WHONET 5.6 and SPSS 25.0 software.Results:A total of 4 385 Gram-positive bacterial isolates were obtained from 60 participating center. The top five pathogens were Staphylococcus aureus（ n=1 544，35.2%），coagulase-negative Staphylococci（ n=1 441，32.9%）， Enterococcus faecium（ n=574，13.1%）， Enterococcus faecalis（ n=385，8.8%），and α-hemolytic Streptococci（ n=187，4.3%）. The prevalence of methicillin-resistant Staphylococcus aureus（MRSA）and methicillin-resistant coagulase-negative Staphylococci（MRCNS）was 26.2%（405/1 544）and 69.8%（1 006/1 441），respectively. Notably，all Staphylococci remained susceptible to glycopeptide or daptomycin. Staphylococcus aureus demonstrated excellent susceptibility（>97.0%）to cephalobiol，rifampicin，trimethoprim-sulfamethoxazole，linezolid，minocycline，tigecycline，and eravacycline. No Enterococcus exhibiting resistance to linezolid were detected. Glycopeptide resistance was uncommon but more frequent in Enterococcus faecium（resistance to vancomycin and teicoplanin：both 1.7%）compared to Enterococcus faecalis（both 0.3%）. The detection rates of MRSA and MRCNS exhibited significant regional variations across the country（ χ2=17.674 and 148.650，respectively，both P<0.001）. No vancomycin-resistant Enterococci were detected in central China. Institutional comparison demonstrated higher prevalence of MRSA（ χ2=14.111， P<0.001）and MRCNS（ χ2=4.828， P=0.028）in provincial hospitals than that in municipal hospitals. Socioeconomic analysis identified elevated detection rates of both MRSA（ χ2=18.986， P<0.001）and MRCNS（ χ2=4.477， P=0.034）in less developed regions（per capita GDP

Objective:To report the results of bacterial resistant investigation collaborative system（BRICS）on the distribution and antimicrobial resistance profile of clinical Gram-negative bacteria isolates from bloodstream infections in China in 2023，and provide reference for clinical tretment of bloodstream infections and prevention and control of bacterial resistance.Methods:The clinical isolates of Gram-negative bacteria from blood cultures in member hospitals of BRICS were collected during January 2023 to December 2023. Antibiotic susceptibility tests were conducted by agar dilution or broth dilution methods recommended by Clinical and Laboratory Standards Institute（CLSI）. WHONET 5.6 and SPSS 25.0 were used to analyze the data.Results:During the study period，11 492 strains of Gram-negative bacteria were collected from 60 hospitals，of which 10 098（87.9%）were Enterobacterales and 1 394（12.1%）were non-fermentative bacteria. The top 5 bacterial species were Escherichia coli（50.0%）， Klebsiella pneumoniae（26.1%）， Pseudomonas aeruginosa（5.1%）， Acinetobacter baumannii complex（5.0%）and Enterobacter cloacae complex（4.1%）. The ESBL-producing rates in Escherichia coli， Klebsiella pneumoniae and Proteus mirablilis were 46.8%（2 685/5 741），18.3%（549/2 999）and 44.0%（77/175），respectively. The prevalence of carbapenem-resistant Escherichia coli（CREC）and carbapenem-resistant Klebsiella pneumoniae（CRKP）were 1.3%（76/5 741）and 15.0%（450/2 999）；32.9%（25/76）and 78.0%（351/450）of CREC and CRKP were sensitive to ceftazidime/avibactam combination，respectively. 94.7%（72/76）and 90.2%（406/450）of CREC and CRKP were sensitive to aztreonam/avibactam combination. Furthermore，57.9%（44/76）and 79.1%（356/450）were sensitive to imipenem/relebactam combination. The prevalence of carbapenem-resistant Acinetobacter baumannii（CRAB）complex was 64.6%（370/573），while more than 80.0% of CRAB complex was sensitive to tigecycline，eravacycline and polymyxin B. The prevalence of carbapenem-resistant Pseudomonas aeruginosa（CRPA）was 17.0%（99/581）. There were differences in the composition ratio of Gram-negative bacteria in bloodstream infections and the prevalence of important Gram-negative bacteria resistance among different regions in China，with statistically significant differences in the prevalence of CREC，CRKP，CRPA and CRAB complex（ χ2=10.6，28.6，10.8 and 19.3， P<0.05）. The prevalence of ESBL-producing Escherichia coli， CREC，CRAB complex and CRKP were higher in provincial hospitals than those in municipal hospitals（ χ2=12.5，9.8，12.7 and 57.8，all P<0.01）. Conclusions:Gram-negative bacteria are the main pathogens causing bloodstream infections in China，and Escherichia coli is ranked in the top，while the trend of Klebsiella pneumoniae increases continuously with time. CRKP infection shows a slow upward trend，CREC infecton maintains a low prevalence level，and CRAB complex infection continues to exhibit a high prevalence rate. The composition and resistance patterns of pathogens causing bloodstream infections vary to some extent across different regions and levels of hospitals in China.

Objective To discuss the effects of the"twelve needles for hands and feet"acupuncture on the brain response mechanism and acupoint specificity in hemiplegic patients after ischemic stroke using fMRI.Methods Totally 43 inpatients from Dongzhimen Hospital of Beijing University of Chinese Medicine from December 2018 to March 2023 were included.The included patients were randomly divided into a true acupoint group of 23 cases and a fake acupoint group of 20 cases using the random number table method for clinical scale evaluation and magnetic resonance scanning.The true acupoint group received"twelve needles for hands and feet"acupuncture;the fake acupoint group was treated with acupuncture by opening 1 inch next to the meridian point,once a day,for consecutive 10 d.Intervention completed,excluding subjects who fell out during the intervention process,had unsatisfactory MRI data quality and had incomplete clinical scales,and conducting data processing and analysis.Bilateral anterior central gyrus(PreCG),amygdala(AMYG),caudate nucleus(CAU),putamen lenticularis(PUT),pallidus lenticularis(PAL)and thalamus(THA)were selected as regions of interest(ROI),and the changes in Granger causality(GC)and differences in scale scores between the two groups before and after treatment were compared.Results After intervention in the true acupoint group,the GC of CAU.R → PreCG.R,AMYG.R → PAL.R was stronger than before intervention,and the GC of THA.L→PAL.L,AMYG.R→THA.R in the true acupoint group was weaker than before intervention(P<0.05);after intervention with fake acupoints,the GC of AMYG.R→PUT.L in the false acupoint group was stronger than before intervention,and the GC of THA.L→PreCG.R,AMYG.L→CAU.L was weaker than before intervention(P<0.05);the GC of THA.R→PAL.L before intervention was stronger than false acupoint group,the GC of CAU.L→PreCG.R,PUT.L→PreCG.R before intervention was weaker than fake acupoint group(P<0.05);the GC of CAU.R→PreCG.R,THA.L→PreCG.R,AMYG.R→CAU.L,THA.L→PAL.R,AMYG.L→THA.L after intervention in the true acupoint group was stronger than the false acupoint group,and the GC of AMYG.R→AMYG.L after intervention in the true acupoint group was weaker than the fake acupoint group(P<0.05).Conclusion"Twelve needles for hands and feet"acupuncture can stimulate preferential compensation of the right motor circuit,bilateral basal ganglia remodeling,and balance regulation of excitatory inhibitory effects on the thalamus in patients with radiation-induced coronary stroke,thereby improving the degree of motor dysfunction and neurological deficits;acupuncture at acupoints stimulates brain activity more than at non-acupoints,indicating acupoint specificity.

Objective To investigate the effect and mechanism of nodakenin(Nod)in neuropathic pain(NP).Methods Differential expression genes in the primary somatsensory cortex(S1)of NP data and overlapping genes between the dataset and mitochondrial data were screened and analyzed.Overlapping gene interaction networks were overlapped and core genes were screened.A total of 27 mice were randomly divided into the sham operation group,the model group and the drug administration group(9 mice/group).The chronic compression injury model of sciatic nerve was constructed in the model group and the drug administration group.Nod 10 mg/kg was intraperitoneally injected into the drug administration group for 1 week.Changes of pain behavior and motor ability in mice were detected.HE staining and Nissl staining were used to detect effects of nerve injury and inflammation on brain tissue of S1 region of mice.The expression levels of interleukin-1β,early gene(c-Fos),panthenol-cytochrome c reductase complex III subunit(Uqcrq)and ubiquinone oxidoreductase subunit(Nduf)b5 in S1 brain region were analyzed by Western blot assay.Molecular docking was used to study the target of Nod.PC12 cells were divided into the control group,the IL-1β group(1 μmol/L IL-1β treatment)and the IL-1β+Nod group(1 μmol/L IL-1β+1 μmol/L Nod treatment),and mitochondrial membrane potential was detected in each group.Results In the NP dataset GSE180627,S1 brain region contained 293 differentially expressed genes,and the mitochondrial data contained 1 082 genes.There were 34 overlapping genes,and genes related to oxidative phosphorylation and electron transport chain were enriched.The protein interaction network showed that core genes included electron transport chain related proteins Ndufb5,Uqcrq,Ndufs8,Ndufa7,Ndufa3,Cox6b1 and Mrps33.Compared with the model group,the mechanical foot shrinkage threshold,thermal foot shrinkage reflex latency and rod rotation residence time of mice were increased in the drug administration group,the number of inflammatory infiltrating cells in S1 tissue and the number of Nislet bodies in neurons,expression levels of c-Fos and IL-1β in neurons were decreased,and expression levels of Uqcrq and Ndufb5 were increased(P＜0.05).Molecular docking showed that Nod could bind Uqcrq and Ndufb5.Compared with the IL-1β group,the fluorescence signal of mitochondrial membrane potential was enhanced in the IL-1β+Nod group(P＜0.05).Conclusion Nodakenin can improve pain behavior in mice,and its mechanism involves ameliorating mitochondrial damage in S1.

Objective To identify novel biomarkers for predicting the risk of latent tuberculosis infection(LTBI)activation using bio-informatics and machine-learning algorithms and to establish a risk model.Methods The GSE112104 and GSE193777 datasets were obtained from the Gene Expression Omnibus.Differential gene expression and weighted gene co-expression network analyses were per-formed to identify ferroptosis-related differentially expressed genes(FRG-DEGs)associated with LTBI activation.Three machine-learning algorithms,least absolute shrinkage and selection operator,support vector machine-recursive feature elimination,and random forest,were used to identify ferroptosis-related hub genes(FRG-hubs).The reliability of these genes was validated using independent validation datasets and reverse transcription polymerase chain reaction(PCR).A risk model was established using R software.Results In the GSE 112104 dataset,296 genes were upregulated and 1 569 genes were downregulated in active tuberculosis compared to those in LTBI.Among the LTBI progressors,506 genes were upregulated and 1 132 genes were downregulated.Weighted correlation network analysis identified five gene modules,with the blue module showing the strongest correlation with LTBI activation(cor=0.62,P=0.000 04),con-taining 1 340 genes.Intersections with 728 ferroptosis-related genes resulted in eight FRG-DEGs.The machine-learning algorithms iden-tified four FRG-hubs:PLA2G6,GLS2,JUN,and AMN,whose expression decreased with LTBI activation.Reverse transcription PCR con-firmed this trend.A risk model based on these genes yielded an area under the curve of 0.98 to 1.00.Conclusion This study successfully identified novel biomarkers for predicting the risk of LTBI activation and developed an accurate predictive risk model.

Objective To investigate the effect and mechanism of nodakenin(Nod)in neuropathic pain(NP).Methods Differential expression genes in the primary somatsensory cortex(S1)of NP data and overlapping genes between the dataset and mitochondrial data were screened and analyzed.Overlapping gene interaction networks were overlapped and core genes were screened.A total of 27 mice were randomly divided into the sham operation group,the model group and the drug administration group(9 mice/group).The chronic compression injury model of sciatic nerve was constructed in the model group and the drug administration group.Nod 10 mg/kg was intraperitoneally injected into the drug administration group for 1 week.Changes of pain behavior and motor ability in mice were detected.HE staining and Nissl staining were used to detect effects of nerve injury and inflammation on brain tissue of S1 region of mice.The expression levels of interleukin-1β,early gene(c-Fos),panthenol-cytochrome c reductase complex III subunit(Uqcrq)and ubiquinone oxidoreductase subunit(Nduf)b5 in S1 brain region were analyzed by Western blot assay.Molecular docking was used to study the target of Nod.PC12 cells were divided into the control group,the IL-1β group(1 μmol/L IL-1β treatment)and the IL-1β+Nod group(1 μmol/L IL-1β+1 μmol/L Nod treatment),and mitochondrial membrane potential was detected in each group.Results In the NP dataset GSE180627,S1 brain region contained 293 differentially expressed genes,and the mitochondrial data contained 1 082 genes.There were 34 overlapping genes,and genes related to oxidative phosphorylation and electron transport chain were enriched.The protein interaction network showed that core genes included electron transport chain related proteins Ndufb5,Uqcrq,Ndufs8,Ndufa7,Ndufa3,Cox6b1 and Mrps33.Compared with the model group,the mechanical foot shrinkage threshold,thermal foot shrinkage reflex latency and rod rotation residence time of mice were increased in the drug administration group,the number of inflammatory infiltrating cells in S1 tissue and the number of Nislet bodies in neurons,expression levels of c-Fos and IL-1β in neurons were decreased,and expression levels of Uqcrq and Ndufb5 were increased(P＜0.05).Molecular docking showed that Nod could bind Uqcrq and Ndufb5.Compared with the IL-1β group,the fluorescence signal of mitochondrial membrane potential was enhanced in the IL-1β+Nod group(P＜0.05).Conclusion Nodakenin can improve pain behavior in mice,and its mechanism involves ameliorating mitochondrial damage in S1.

Objective To observe the value of artificial intelligence iterative reconstruction(AIIR)for low-dose chest CT images of infants with congenital heart disease.Methods Totally 262 infants with congenital heart disease who would undergo chest CT scanning were prospectively enrolled and randomly divided into low-dose group(n=142)and conventional dose group(n=120).Chest CT scanning with tube voltage of 80 kVp and tube current of 10 mAs was performed in low-dose group,and hybrid iterative reconstruction(HIR,group A)and AIIR(group B)were used to reconstruct images,respectively.In conventional dose group,chest CT scanning with tube voltage of 80 kVp and tube current of 100 mAs was performed,and HIR was used to reconstruct images(group C).Then subjective and objective evaluation on image quality were performed,the results were compared among 3 groups,and the value of AIIR was analyzed.Results Significant differences of image quality and clarity of displaying structures were found among 3 groups(all P＜0.001).Among them,except for the clarity of interlobar fissure,no significant difference of subjective scores was found between low-dose AIIR images and conventional dose HIR images(all corrected P＞0.05),while subjective scores of low-dose HIR images were all lower than those of low-dose AIIR images and conventional dose HIR images(all corrected P＜0.05).Significant differences of standard deviation(SD),signal-to-noise ratio(SNR)and contrast-to-noise ratio(CNR)were found among 3 groups(all P＜0.001)and between each 2 groups(all corrected P＜0.05).The effective dose of low-dose group and conventional dose group was 0.09(0.08,0.10)and 0.85(0.75,1.03)mSv,respectively,and the former was lower than the latter(Z=-13.942,P＜0.001).Conclusion Using AIIR could obtain low-dose chest CT images of infants with quality comparable to conventional chest CT images.

Objective To discuss the effects of the"twelve needles for hands and feet"acupuncture on the brain response mechanism and acupoint specificity in hemiplegic patients after ischemic stroke using fMRI.Methods Totally 43 inpatients from Dongzhimen Hospital of Beijing University of Chinese Medicine from December 2018 to March 2023 were included.The included patients were randomly divided into a true acupoint group of 23 cases and a fake acupoint group of 20 cases using the random number table method for clinical scale evaluation and magnetic resonance scanning.The true acupoint group received"twelve needles for hands and feet"acupuncture;the fake acupoint group was treated with acupuncture by opening 1 inch next to the meridian point,once a day,for consecutive 10 d.Intervention completed,excluding subjects who fell out during the intervention process,had unsatisfactory MRI data quality and had incomplete clinical scales,and conducting data processing and analysis.Bilateral anterior central gyrus(PreCG),amygdala(AMYG),caudate nucleus(CAU),putamen lenticularis(PUT),pallidus lenticularis(PAL)and thalamus(THA)were selected as regions of interest(ROI),and the changes in Granger causality(GC)and differences in scale scores between the two groups before and after treatment were compared.Results After intervention in the true acupoint group,the GC of CAU.R → PreCG.R,AMYG.R → PAL.R was stronger than before intervention,and the GC of THA.L→PAL.L,AMYG.R→THA.R in the true acupoint group was weaker than before intervention(P<0.05);after intervention with fake acupoints,the GC of AMYG.R→PUT.L in the false acupoint group was stronger than before intervention,and the GC of THA.L→PreCG.R,AMYG.L→CAU.L was weaker than before intervention(P<0.05);the GC of THA.R→PAL.L before intervention was stronger than false acupoint group,the GC of CAU.L→PreCG.R,PUT.L→PreCG.R before intervention was weaker than fake acupoint group(P<0.05);the GC of CAU.R→PreCG.R,THA.L→PreCG.R,AMYG.R→CAU.L,THA.L→PAL.R,AMYG.L→THA.L after intervention in the true acupoint group was stronger than the false acupoint group,and the GC of AMYG.R→AMYG.L after intervention in the true acupoint group was weaker than the fake acupoint group(P<0.05).Conclusion"Twelve needles for hands and feet"acupuncture can stimulate preferential compensation of the right motor circuit,bilateral basal ganglia remodeling,and balance regulation of excitatory inhibitory effects on the thalamus in patients with radiation-induced coronary stroke,thereby improving the degree of motor dysfunction and neurological deficits;acupuncture at acupoints stimulates brain activity more than at non-acupoints,indicating acupoint specificity.