最新[精品]以头夹肌肉肌电图分析实现重郁症之电脑辅助诊断(PartPPT课件.ppt

《最新[精品]以头夹肌肉肌电图分析实现重郁症之电脑辅助诊断(PartPPT课件.ppt》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、[精品]以头夹肌肉肌电图分析实现重郁症之电脑辅助诊断(PartAnartificialneuralnetwork(ANN)ArtificialintelligenceMathematicalmodelLearningsystemComputer-aiddiagnosisMethodsandMaterialsAnartificialneuralnetwork(ANN):類神經網路2MethodsandMaterialsBackpropagationneuralnetwork(BPNN)SupervisedneuralnetworkBackpropagationneuralnetwork(BPNN

2、):倒傳遞類神經網路f1234kLf123451234abcdekLInputlayerHiddenlayerOutputlayer3MethodsandMaterialsBackpropagationneuralnetwork(BPNN)Adjustweights7MethodsandMaterialsSupportvectormachine(SVM)ClassificationStatisticsHyperplaneoptimalseparatinghyperplane(OSH)supporthyperplaneMarginOSH:最佳分割超平面Supporthyperplane:支持超平

3、面8ResultsT.O.V.AMDDHigheromissionratesHighermeanresponsetimesMorevariabilityEMGfeaturesMDDLowerEAandRMSvaluesHigherMFandMPFvalues9ResultsT.O.V.ACompareEMGComparisonbygroupsduringrestGroupsMeanSDPEA(uv)MDD22.2034.840.001Control56.5953.51RMS(uv)MDD0.120.160.000Control0.280.25MDF(PSD)MDD80.407.310.000C

4、ontrol73.117.77MPF(PSD)MDD84.984.190.000Control80.514.14*p<0.0510ResultsT.O.V.ACompareEMGComparisonbygroupduringTOVAGroupsMeanSDPEA(uv)MDD28.4638.270.013Control65.0188.40RMS(uv)MDD0.140.180.010Control0.320.40MDF(PSD)MDD80.836.590.000Control72.796.45MPF(PSD)MDD85.223.970.000Control80.433.31*p<0.0511R

5、esultsAccuracyAccuracyRestTOVATrainTestTrainTestBPNN92.06%76.67%87.67%76.67%SVM100%83.33%98.33%83.56%12ConclusionandDiscussionMDDbecomesmoredistinguishablewhenrestingHealthcontrolshavewiderrangeofEAandMFMDDpatientshavethelowercapabilityonphysiologicalregulationHopefully,thesystemcanbeusedtodetectand

6、tocontroltheMDDdisorderinthefuture.13References[1]J.M.DonohueandH.A.Pincus,Reducingthesocietalburdenofdepression:areviewofeconomiccosts,qualityofcareandeffectsoftreatment,Pharmacoeconomics25(2007)7.[2]P.Sobocki,B.Jonsson,J.Angst,C.Rehnberg.CostofdepressioninEurope.JMentHealthPolicyEcon9(2006)87.[3]A

7、mericanPsychiatricAssociation,Diagnosticandstatisticalmanualofmentaldisorders(AmericanPsychiatricAssociation,Washington,DC,2000).[4]R.C.Kessler,P.Berglund,O.Demler,R.Jin,D.Koretz,K.R.Merikangas,A.J.Ru