Kushandiswa kweMemTrax uye Muchina Kudzidza Modelling muKupatsanura kweMild Cognitive Impairment

Chinyorwa Chekutsvagisa

Vanyori: Bergeron, Michael F. | Landset, Sara | Zhou, Xianbo | Ding, Tao | Khoshgoftaar, Taghi M. | Zhao, Feng | Du, Bo | Chen, Xinjie | Wang, Xuan | Zhong, Lianmei | Liu, Xiaolei| Ashford, J. Wesson

DOI: 10.3233/JAD-191340

Journal: Journal of Dambudziko reAlzheimer, vol. 77, no. 4, pp. 1545-1558, 2020

ngovepo

Background:

Chiitiko chakapararira uye kupararira kwe Chirwere cheAlzheimer uye hunyoro hwekuziva kukanganisa (MCI) yakakurudzira kufona nekukurumidza kwetsvagurudzo kuti isimbise kukurumidza kuona kwekuziva kuongorora uye kuongorora.

Chinangwa:

Chinangwa chedu chekutanga chekutsvagisa chaive chekuona kana MemTrax mashandiro ekuita metrics uye akakodzera huwandu hwevanhu uye hutano chimiro hunhu hunogona kushandiswa nemazvo mumamodheru ekufembera akagadzirwa nemuchina kudzidza kurongedza hutano hwekuziva (zvakajairwa neMCI), sezvaizoratidzwa ne Kuongorora kweMontreal Kuziva (MoCA).

Nzira:

Takaita chidzidzo chemuchinjika pa 259 neurology, kiriniki yekurangarira, uye yemukati mishonga yevarwere vakuru vakatorwa kubva kune vaviri. zvipatara muChina. Murwere wega wega akapihwa chiChinese-mutauro weMoCA uye akazvipa-yega iyo inoenderera yekuzivikanwa MemTrax online episodic. ndangariro bvunzo online nezuva rimwe chetero. Predictive classification modhi akavakwa pachishandiswa muchina kudzidza negumi-fold muchinjikwa kusimbiswa, uye kuita modhi kwakayerwa pachishandiswa Area Under the Receiver Operating Characteristic Curve (AUC). Mamodheru akavakwa pachishandiswa maviri MemTrax performance metrics (muzana chaiyo, nguva yekupindura), pamwe chete nemasere anozivikanwa ehuwandu hwevanhu uye nhoroondo yemunhu.

Results:

Tichienzanisa vadzidzi pamisanganiswa yakasarudzwa yezvibodzwa zveMoCA uye zvikumbaridzo, Naïve Bayes ndiye aiwanzova mudzidzi anoita zvepamusoro nekuita kwemhando ye0.9093. Kupfuurirazve, pakati pevadzidzi vatatu vepamusoro, MemTrax-yakavakirwa classification performance yakazara yaive yepamusoro ichishandisa iyo yepamusoro-yepamusoro maficha (0.9119) kana ichienzaniswa nekushandisa ese gumi akajairika maficha (10).

mhedziso:

Kuita kweMemTrax kunogona kushandiswa zvinobudirira mumuchina wekudzidza classification yekufungidzira modhi yekuongorora application yekuona yekutanga nhanho yekuziva kukanganisa.

NHANGANYAYA

Izvo zvinozivikanwa (zvisinei zvisina kuongororwa) zvakapararira-zvakapararira chiitiko uye kuwanda uye zvakafanana kuwedzera kwekurapa, magariro, uye veruzhinji. utano mari uye mutoro wechirwere cheAlzheimer (AD) uye kunyorova kwekuziva (MCI) zviri kuramba zvichinetsa kune vese vanobatana [1, 2]. Iyi inoshungurudza uye inoshungurudza mamiriro akakurudzira kushevedzwa kwekukurumidzira kuti tsvagiridzo isimbiswe kutanga kuona kucherechedzwa kwekuongorora uye zvigadziriso zvekushandisa zvenguva dzose zvinoshanda mune zvega uye zvekiriniki marongero evarwere vakura munzvimbo dzakasiyana siyana uye vanhu [3]. Zviridzwa izvi zvinofanirwa kupawo shandurudzo isina musono yeruzivo rwezvabuda mumarekodhi ehutano emagetsi. Mabhenefiti anozoonekwa nekuzivisa varwere uye kubatsira vanachiremba mukuziva shanduko dzakakosha kare uye nokudaro kugonesa kukurumidza uye nenguva stratification, kuita, uye kurondwa kweakakodzera ega uye anodhura-inoshanda kurapwa uye kutarisirwa kwevarwere kune avo vanotanga kusangana. kunzwisisa kunonoka [3, 4].

Iyo yekombuta yeMemTrax chishandiso (https://memtrax.com) iri nyore uye pfupi inoenderera mberi yekucherechedzwa yekuongorora iyo inogona kuzvidzora yega pamhepo kuyera inonetsa yenguva episodic memory performance apo mushandisi anopindura kune inodzokororwa mifananidzo uye kwete kune yekutanga mharidzo [5, 6]. Tsvagiridzo ichangoburwa uye zvinokonzeresa zvinoshanda zviri kutanga zvishoma nezvishoma uye pamwe chete kuratidza kushanda kwekiriniki kweMemTrax mukutanga AD uye MCI kuongorora [5-7]. Nekudaro, kuenzanisa kwakananga kwekiriniki utility kune iripo kuziva hutano ongororo uye zvakajairwa zviyero zvinotenderwa kuzivisa maonero ehunyanzvi uye kusimbisa MemTrax utility mukuonekwa kwekutanga uye tsigiro yekuongorora. van der Hoek et al. [8] akaenzanisa akasarudzwa MemTrax performance metrics (yekuita nekukurumidza uye muzana chaiyo) kune yekuziva mamiriro sezvakatemwa neMontreal. Cognitive Assessment (MoCA). Nekudaro, chidzidzo ichi chaive chakaganhurirwa kubatanidza ma metrics ekuita aya nekumisikidzwa kwechimiro chekuziva (sezvinotsanangurwa neMoCA) uye kutsanangura hukama hwepakati uye cutoff kukosha. Saizvozvo, kuwedzera pakuferefeta uku nekuvandudza mashandiro emhando uye kugona, mubvunzo wedu wekutanga wekutsvagisa waive:

  • Inogona here kusarudzwa nemunhu MemTrax performance metrics uye yakakodzera huwandu hwevanhu uye hutano profile hunhu huchashandiswa zvinobudirira mumuenzaniso wekufungidzira wakagadzirwa nemuchina kudzidza kurongedza hutano hwekuziva dichotomously (zvakajairwa kupesana neMCI), sezvazvingazoratidzwa nemunhu weMoCA mamakisi?

Chechipiri kune izvi, taida kuziva:

  • Kusanganisira iwo maficha akafanana, iyo MemTrax inoshanda-yakavakirwa muchina yekudzidza modhi inogona kunyatso shandiswa kumurwere kufanotaura hukasha (hunyoro kana hwakanyanya) mukati mezvikamu zvakasarudzwa zvekukanganisika kwenjere sezvazvingazotemerwa nekuzvimiririra kuchipatara kuongororwa?

Kuuya uye kushanduka kwekushandisa kwehungwaru hwekugadzira uye kudzidza muchina mukuongorora / kuona zvakatoratidza zvakanakira zvinoshanda, nekufungidzira modhi inotungamira zvinobudirira kutungamira varapi mukuongorora kwakaoma kwekuziva / hutano hwehuropi uye manejimendi evarwere. Muchidzidzo chedu, takasarudza nzira yakafanana muMCI yemhando yekuenzanisira uye kusarura kwekusarura kwekusarura sekusimbiswa nekuongororwa kwekiriniki kubva kumaseti matatu anomiririra vakasarudzwa vanozvipira varwere uye varwere vekunze kubva kuzvipatara zviviri muChina. Tichishandisa muchina kudzidza kufanofembera modhi, takaona vadzidzi vanoita zvepamusoro-soro kubva kune akasiyana dhataseti/musanganiswa wevadzidzi uye takaisa zvinhu kuti zvititungamirire mukutsanangura zvakanyanya kliniki inoshanda modhi mashandisirwo.

Mafungiro edu aive ekuti MemTrax-yakavakirwa modhi yakagadziriswa inogona kushandiswa kuronga hutano hwekuziva dichotomously (yakajairika kana MCI) zvichibva pane yeMoCA aggregate mamakisi danho, uye kuti yakafanana MemTrax fungidziro modhi inogona kushandiswa zvinobudirira mukusarura kuomarara muzvikamu zvakasarudzwa zve kurwara nechipatara kukanganiswa kwepfungwa. Kuratidza zvinotarisirwa mhedzisiro zvingave zvakakosha mukutsigira kushanda kweMemTrax seyekutanga yekuona skrini yekudzikira kwekuziva uye yekusagadzikana kwekusaziva. Kuenzanisa kwakanaka kune indasitiri inonzi chiyero chinowedzerwa nekureruka kukuru uye nekukasira kwekushandisa kwaizove nesimba mukubatsira varapi kutora ichi chiri nyore, chakavimbika, uye chinosvikika chishandiso seyekutanga skrini mukuona kwekutanga (kusanganisira prodromal) nhanho yekuziva kushomeka. Maitiro akadaro uye kushandiswa zvinogona kukurudzira zvakanyanya panguva uye zvirinani stratified murwere kutarisirwa uye kupindira. Aya maonero ekufunga kwemberi uye akagadziridzwa metrics uye modhi anogona zvakare kubatsira mukudzikamisa kana kumisa kuenderera mberi kwedementia, kusanganisira AD uye AD-inoenderana nedementia (ADRD).

ZVINHU NEMITARIDZO

Kudzidza kuwanda

Pakati paNdira 2018 naNyamavhuvhu 2019, kutsvagisa-chikamu kwakapedzwa pavarwere vakatorwa kubva kuzvipatara zviviri muChina. Kutungamirwa kweMemTrax [5] kune vanhu vane makore makumi maviri nerimwe zvichikwira uye kuunganidzwa nekuongororwa kweiyo data kwakaongororwa uye kubvumidzwa uye kutungamirwa maererano nehunhu hwetsika. munhu Chidzidzo Chekudzivirira Komiti yeStanford University. MemTrax uye kumwe kuyedzwa kwese kwechidzidzo ichi chakaitwa maererano neHelsinki declaration ye1975 uye yakatenderwa neInstitutional Review Board yeFirst Affiliated Hospital yeKunming Medical University muKunming, Yunnan, China. Wese mushandisi akapihwa chibvumirano chakanaka fomu rekuverenga/kuongorora wobva wabvuma nokuzvidira kutora chikamu.

Vatori vechikamu vakatorwa kubva mudziva revarwere vari kunze mukiriniki yeurology paYanhua Hospital (YH sub-dataset) uye ndangariro kiriniki paFirst Affiliated Hospital yeKunming Medical Yunivhesiti (XL sub-dataset) muBeijing, China. Vatori vechikamu vakatorwawo kubva kuNeurology (XL sub-dataset) uye mushonga wemukati (KM sub-dataset) varwere paFirst Affiliated Hospital yeKunming Medical University. Maitiro ekusanganisirwa aisanganisira 1) varume nevakadzi vangangoita makore makumi maviri nerimwe ekuzvarwa, 21) kugona kutaura chiChinese (Mandarin), uye 2) kugona kunzwisisa zvirevo zvemashoko uye zvinyorwa. Maitiro ekusarudzika aive ekuona uye kukanganiswa kwemota kutadzisa vatori vechikamu kupedzisa MemTrax bvunzo, pamwe nekusakwanisa kunzwisisa mirairo chaiyo yebvunzo.

Shanduro yeChinese yeMemTrax

Inthanethi MemTrax test platform yakashandurwa kupinda muchiChinese (URL: https://www.memtrax.com.cn) uyezve yakagadziridzwa kuti ishandiswe kuburikidza neWeChat (Shenzhen Tencent Computer Systems Co. LTD., Shenzhen, Guangdong, China) kuti uzvitungamirire. Dhata rakachengetwa pasevha yegore (Ali Cloud) iri kuChina uye ine rezinesi kubva kuAlibaba (Alibaba Technology Co. Ltd., Hangzhou, Zhejiang, China) neSJN Biomed LTD (Kunming, Yunnan, China). Tsanangudzo dzakananga paMemTrax uye bvunzo dzechokwadi nzira dzakashandiswa pano dzakatsanangurwa kare [6]. Ongororo yakaitwa pasina muripo kuvarwere.

Maitiro ekudzidza

Kune vanorwara nevarwere vekunze, pepa remibvunzo yakajairika yekuunganidza huwandu hwevanhu uye ruzivo rwemunhu sezera, zvepabonde, makore edzidzo, basa, kurarama oga kana nemhuri, uye nhoroondo yezvokurapa yaitungamirirwa nenhengo yeboka rezvidzidzo. Mushure mekupedzwa kwebvunzo, iyo MoCA [12] uye MemTrax bvunzo dzakapihwa (MoCA kutanga) pasina anopfuura maminetsi makumi maviri pakati pebvunzo. MemTrax muzana yakarurama (MTx-% C), kureva nguva yekupindura (MTx-RT), uye zuva uye nguva yekuedzwa zvakanyorwa papepa nenhengo yeboka rekudzidza kune mumwe nomumwe anotora chikamu akaedzwa. Gwaro remibvunzo rakapedzwa uye zvabuda zveMoCA zvakaiswa muExcel spreadsheet nemuongorori aiongorora uye akasimbiswa nemumwe waaishanda naye mafaera eExcel asati achengetwa kuti aongororwe.

MemTrax bvunzo

Muedzo wepamhepo weMemTrax waisanganisira mifananidzo makumi mashanu (50 yakasarudzika uye makumi maviri neshanu inodzokororwa; 25 seti yemifananidzo mishanu yezvinoitika kana zvinhu) inoratidzwa mune chaiyo pseudo-random order. Mutori wechikamu aibata (pamurairo) kubata bhatani reKutanga pachiratidziro kuti atange bvunzo uye otanga kuona iyo mifananidzo yakatevedzana uye obatazve mufananidzo uri pachiratidziro nekukasira pese painodzokororwa mufananidzo. Mufananidzo wega wega wakaonekwa kwe25s kana kusvika mufananidzo uri pachiratidziri wabatwa, izvo zvakakurudzira kuratidzwa kweiyo inotevera mufananidzo. Uchishandisa wachi yemukati yemudziyo wenzvimbo, MTx-RT yemufananidzo wega wega yakatemwa nenguva yakapfuura kubva pakuratidzwa kwechifananidzo kusvika payakabatwa chidzitiro nemutori wechikamu mukupindura kuratidza kucherechedzwa kwechifananidzo sechaive chatoratidzwa. panguva yekuedzwa. MTx-RT yakarekodhwa kumufananidzo wega wega, iine yakazara 5 s yakarekodhwa ichiratidza kuti hapana mhinduro. MTx-% C yakaverengerwa kuratidza chikamu chekudzokorora uye mifananidzo yekutanga iyo mushandisi akapindura nenzira kwayo (yechokwadi yakanaka + yechokwadi yakaipa yakakamurwa ne5). Zvimwe zvekuwedzera zveMemTrax manejimendi uye kuita, kuderedzwa kwedata, kusashanda kana "hapana mhinduro" data, uye yekutanga data kuongororwa inotsanangurwa kumwe kunhu [3].

Muedzo weMemTrax wakatsanangurwa zvakadzama uye bvunzo yekudzidzira (ine mifananidzo yakasarudzika kunze kweiyo yakashandiswa muyedzo yekurekodha mhinduro) yakapihwa vatori vechikamu muchipatara. Vatori vechikamu muYH neKM sub-datasets vakatora bvunzo yeMemTrax pane smartphone yaive yakatakurwa nechikumbiro paWeChat; nepo nhamba shoma yevarwere veXL sub-dataset vakashandisa iPad uye vamwe vose vakashandisa smartphone. Vese vatori vechikamu vakatora bvunzo yeMemTrax nemuongorori wekudzidza asingatarise.

Montreal cognitive ongororo

Iyo Beijing vhezheni yeChinese MoCA (MoCA-BC) [13] yaitungamirwa uye yakapihwa nevanoongorora vakadzidziswa zvinoenderana nemirairo yepamutemo yebvunzo. Zvakakodzera, iyo MoCA-BC yakaratidzwa kuva yakavimbika bvunzo yekuziva kuongororwa pamatanho ese edzidzo muvakuru vechiChinese vakura [14]. Muedzo wega wega wakatora maminetsi gumi kusvika makumi matatu kuita bvunzo zvichienderana nekugona kuziva kwemutori wechikamu.

MoCA classification modelling

Paive nezvikamu makumi maviri nemapfumbamwe zvinoshandisika, kusanganisira maviri MemTrax test performance metrics uye 27 maficha ane hukama nehuwandu hwevanhu uye hutano ruzivo rwemunhu wese anotora chikamu. Murwere wega wega MoCA aggregate test mamakisi akashandiswa seye kuongorora kuongorora "benchmark" yekudzidzisa mamodheru edu ekufungidzira. Saizvozvo, nekuti MoCA yakashandiswa kugadzira iyo kirasi label, isu hatina kukwanisa kushandisa aggregate mamakisi (kana chero eMoCA subset zvibodzwa) sechinhu chakazvimirira. Takaita zviyedzo zvekutanga umo takatevedzera (kuronga hutano hwekuziva hunotsanangurwa neMoCA) zvipatara zvitatu zvepakutanga / kiriniki (s) sub-datasets yega uye ndokusanganiswa uchishandisa ese maficha. Zvisinei, zvose zvakafanana zvinyorwa zve data hazvina kuunganidzwa mune imwe neimwe yemakiriniki mana anomiririra matatu-sub-datasets; saka, zvakawanda zvezvinhu zvedu mudheta rakasanganiswa (pakushandisa zvinhu zvose) zvaiva nehuwandu hwehuwandu hwekushayikwa. Isu takazovaka mamodheru ane dataset yakasanganiswa tichishandisa zvakajairika maficha izvo zvakakonzera kuvandudzwa kwemaitiro. Izvi zvingangove zvakatsanangurwa nemusanganiswa wekuva nemamwe mamiriro ekushanda nawo nekubatanidza matatu evarwere sub-datasets uye pasina maficha ane kushaikwa kwehuwandu husina kukosha (chinhu chimwe chete mu dataset yakasanganiswa, rudzi rwebasa, raive nechero hunhu husipo, hunokanganisa. zviitiko zvitatu chete zvevarwere), nekuti chete maficha akajairwa akanyorwa panzvimbo dzese nhatu akasanganisirwa. Zvakanyanya, isu takanga tisina chaiyo yekurambwa kwechinhu chimwe nechimwe icho chakazopedzisira chisina kubatanidzwa mune yakasanganiswa dataset. Nekudaro, mune yedu yekutanga yakasanganiswa dataset modelling, isu takatanga kushandisa ese maficha kubva kune yega yega matatu akapatsanurwa evarwere sub-dataset. Izvi zvakakonzera zvakanyanya mukuita kwemodhi iyo yaive yakadzikira zvakanyanya pane yekutanga yekutanga modhi pane yega yega sub-dataset. Uyezve, nepo kuita kwemhando yemhando dzakavakwa pachishandiswa ese maficha kwaikurudzira, kune vese vadzidzi uye zvirongwa zvekuisa, kuita kwakagadziridzwa kwekaviri mamodheru kana uchingoshandisa zvakajairika. Muchokwadi, pakati pezvakazopedzisira zvave vadzidzi vedu vepamusoro, zvese kunze kweimwe modhi yakagadziridzwa pakubvisa zvisiri zvakajairika.

Yekupedzisira aggregate dataset (YH, XL, uye KM yakasanganiswa) yaisanganisira mazana maviri nemakumi mashanu nemapfumbamwe ezviitiko, imwe neimwe ichimiririra akasarudzika akatora bvunzo dzeMemTrax neMoCA. Paive ne259 yakagovaniswa yakazvimirira maficha: MemTrax performance metrics: MTx-% C uye zvinoreva MTx-RT; ruzivo rwehuwandu hwevanhu uye nhoroondo yezvokurapa: zera, zvepabonde, makore edzidzo, rudzi rwebasa (blue collar/white collar), rubatsiro rwekugarisana (kunyangwe muongorori achigara ega kana nemhuri), uye hongu/kwete mhinduro dzekuti mushandisi aive ne nhoroondo yechirwere cheshuga, hyperlipidemia, kana kukuvara kwehuropi. Mamwe ma metrics maviri, MoCA aggregate mamakisi uye MoCA aggregate mamakisi akagadziridzwa kwemakore edzidzo [10], akashandiswa zvakasiyana kugadzira anotsamira emhando yezvinyorwa, nekudaro kugadzira maviri akasiyana emuenzaniso zvirongwa kuti zvishandiswe kune yedu yakasanganiswa dataset. Kune imwe neimwe shanduro (yakagadziridzwa uye isina kugadziriswa) yeMoCA mamakisi, iyo data zvakare yakapatsanurwa yakapatsanurwa yebhinari classification uchishandisa maviri akasiyana criterion mazinga-iyo yekutanga yakakurudzirwa imwe [12] uye imwezve kukosha yakashandiswa uye inosimudzirwa nevamwe [12, 8]. Mune imwe nzira yekuisa pachikumbaridzo chirongwa, murwere aionekwa seane hutano hwekuziva kana s / akawana ≥15 pabvunzo yeMoCA uye aine MCI kana mamakisi aive 23 kana pasi; nepo, mune yekutanga yakakurudzirwa kupatsanura fomati, murwere aifanira kuwana 22 kana zvirinani paMoCA kuti inyore seane hutano hwekuziva.

Yakasefa data yeMoCA classification modelling

Isu takaongororazve MoCA kupatsanura tichishandisa mana anowanzo shandiswa maitiro ekuisa maitiro: Chi-Squared, Gain Ratio, Ruzivo Ruzivo, uye Symmetrical Kusavimbika. Nekuona kwechinguvana, takaisa marenji kune yese yakasanganiswa dhataset tichishandisa yega yega yedu mana emuenzaniso zvirongwa. Vese vamiriri vakabvumirana pazvinhu zvakafanana zvepamusoro, kureva, zera, nhamba yemakore edzidzo, uye ese MemTrax performance metrics (MTx-% C, zvinoreva MTx-RT). Isu tinozovaka patsva mamodheru tichishandisa imwe neimwe nzira yekusarudza maficha kudzidzisa mamodheru pazvinhu zvina chete zvepamusoro (ona Feature sarudzo pasi).

Mhedzisiro misere yekupedzisira misiyano yeMoCA mamakisi emhando yekuenzanisira zvirongwa zvinounzwa muTafura 1.

Tafura 1

Pfupiso yekumisikidza hurongwa hwekusiyanisa inoshandiswa kuMoCA classification (Kazhinji Cognitive Health maringe neMCI)

Modelling SchemeNormal Cognitive Health (Negative Kirasi)MCI (Positive Kirasi)
Yakagadziridzwa-23 Isina kusefa/Isina kusefa101 (39.0%)158 (61.0%)
Yakagadziridzwa-26 Isina kusefa/Isina kusefa49 (18.9%)210 (81.1%)
Kunadjusted-23 Unfiltered/Sefa92 (35.5%)167 (64.5%)
Kunadjusted-26 Unfiltered/Sefa42 (16.2%)217 (83.8%)

Nhamba inoenderana uye muzana yevarwere vese mukirasi yega yega inosiyaniswa nekugadzirisa zvibodzwa zvedzidzo (Kugadziriswa kana Kusagadziriswa) uye kupatsanurwa chikumbaridzo (23 kana 26), sekushandiswa kune ese maficha seti (Asina kusefa uye Akasefa).

MemTrax-yakavakirwa kukiriniki yekuongorora modhi

Pamatatu edu ekutanga madiki-datasets (YH, XL, KM), chete varwere veXL sub-dataset ndivo vakazvimiririra vakaongororwa nekiriniki nekuda kwekukanganisa kwekuziva (kureva, yavo yeMoCA mamakisi haina kushandiswa mukutanga kupatsanurwa kweyakajairwa kupesana nekukanganisa). Kunyanya, varwere veXL vakaonekwa vaine chero Kuongororwa kwechirwere cheAlzheimer (AD) kana vascular dementia (VaD). Mukati meimwe neimwe yeaya ekutanga ekuongororwa zvikwata, pakanga paine kumwe kudomwa kweMCI. Kuongororwa kweMCI, dementia, vascular neurocognitive disorder, uye neurocognitive disorder nekuda kweAD yakabva pane zvakananga uye zvakasiyana-siyana zvekuongororwa zvinotsanangurwa muDiagnostic uye Statistical Manual yeMental Disorders: DSM-5 [16]. Tichifunga nezvekuongororwa kwakanatswa uku, zvirongwa zviviri zvekuenzanisira zvakaiswa zvakasiyana kune iyo XL sub-dataset kusiyanisa nhanho yekuomarara (dhigirii rekukanganisika) kune yega yega yekutanga kuongororwa. Dhiyabhorosi inoshandiswa mune imwe neimwe yeiyi diagnostic modeling schemes (AD neVaD) yaisanganisira ruzivo rwevanhu uye murwere nhoroondo, pamwe chete neMemTrax performance (MTx-% C, zvinoreva MTx-RT). Kuongororwa kwega kwega kwakanyorwa kunyoro kana kwakasarudzwa MCI; zvikasadaro, zvairangarirwa zvakakomba. Isu takatanga kufunga kusanganisira iyo MoCA mamakisi mumhando dzekuongorora (zvinyoro zvichienzaniswa neyakaomarara); asi isu takasarudza izvo zvaizokunda chinangwa chechipiri chekufungidzira modelling chirongwa. Pano vadzidzi vaizodzidziswa vachishandisa humwe hunhu hwemurwere hunowanikwa zviri nyore kumupi uye metrics ekuita kweyedzo yakapfava yeMemTrax (panzvimbo yeMoCA) vachipesana nereferensi "yegoridhe mwero", iyo yakazvimirira yekuongororwa kwekiriniki. Paive nezviitiko makumi matanhatu nepfumbamwe muAD diagnostic dataset uye makumi manomwe nenomwe zviitiko zveVaD (Table 2) Mune ese madatasets, pakanga paine gumi nemaviri akazvimirira maficha. Pamusoro pezvikamu zve12 zvinosanganisirwa muchikamu cheMoCA, nhoroondo yevarwere yaisanganisirawo ruzivo rwenhoroondo yehypertension uye sitiroko.

Tafura 2

Pfupiso yekumisikidza hurongwa kusiyanisa inoshandiswa pakuongorora kuomarara kwemhando (Mild versus Yakakomba)

Modelling SchemeNyoro (Negative Kirasi)Zvakaoma (Positive Kirasi)
MCI-AD inopesana neAD12 (17.4%)57 (82.6%)
MCI-VaD inopesana neVaD38 (50.0%)38 (50.0%)

Nhamba inoenderana uye muzana yevarwere vese mukirasi yega yega vanosiyaniswa nekutanga kuongororwa chikamu (AD kana VaD).

Statistics

Kuenzaniswa kwemaitiro evatori vechikamu uye mamwe maficha enhamba pakati pe-sub-datasets yega yega modhi yekuisa nzira (kufanotaura MoCA yekuziva hutano uye kuongororwa kuomarara) kwakaitwa uchishandisa Python programming mutauro (version 2.7.1) [17]. Kusiyana kwekuita kwemuenzaniso kwakatanga kushandiswa kushandisa imwe- kana mbiri-chinhu (sezvakakodzera) ANOVA ine 95% nguva yekuvimba uye Tukey yakatendeseka musiyano wakakosha (HSD) kuyedza kuenzanisa nzira dzekuita. Kuongorora uku kwekusiyana pakati pemaitiro emuenzaniso wakaitwa uchishandisa musanganiswa wePython neR (version 3.5.1) [18]. Isu takashandisa izvi (zvisinei, zvine nharo zvishoma pane yakakwana) nzira chete serubatsiro rwerubatsiro pane izvi. nhanho yekutanga kwekutanga muenzaniso wekuita kuenzanisa mukutarisira zvingangoita kliniki application. Isu takazoshandisa bvunzo yeBayesian yakasaina-renji tichishandisa kugovera kwemashure kuti tione mukana wekusiyana kwekuita kwemuenzaniso [19]. Pakuongorora uku, takashandisa nguva -0.01, 0.01, zvichireva kuti kana mapoka maviri ane mutsauko wekuita wepasi pe 0.01, aionekwa seamwechete (mukati medunhu rekuenzana kunoshanda), kana kuti zvakasiyana (rimwe riri nani pane chimwe chacho). Kuti tiite kuenzanisa kweBayesian kwevadzidzisi uye kuverenga zvingangoitika izvi, takashandisa raibhurari yebaycomp (vhezheni 1.0.2) yePython 3.6.4.

Predictive modelling

Isu takavaka mamodheru ekufungidzira tichishandisa misiyano gumi yakazara yezvirongwa zvedu zvekuenzanisira kufanotaura (kuronga) mhedzisiro yemurwere wega wega MoCA bvunzo kana kuomarara kwekuongororwa kwekiriniki. Vese vadzidzi vakaiswa uye mamodheru akavakwa pachishandiswa yakavhurika sosi software chikuva Weka [20]. Pakuongorora kwedu kwekutanga, takashandisa gumi anowanzo shandiswa kudzidza algorithms: 10-Nearest Nevakidzani, maviri mavhezheni eC5 sarudzo muti, Logistic Regression, Multilayer Perceptron, Naïve Bayes, maviri maBhaibheri eRandom Forest, Radial Basis Function Network, uye Tsigira Vector. Machine. Hunhu hwakakosha uye misiyano yealgorithms iyi yakatsanangurwa kumwe kunhu [4.5] (ona Appendikisi yakatarisana). Aya akasarudzwa nekuti anomiririra akasiyana emhando dzakasiyana dzevadzidzi uye nekuti isu takaratidza kubudirira tichiashandisa muongororo yapfuura pane yakafanana data. Hyper-parameter marongero akasarudzwa kubva kutsvagurudzo yedu yapfuura achiratidza kuti ane simba pane akasiyana siyana data [21]. Zvichienderana nemhedzisiro yekuongorora kwedu kwekutanga tichishandisa iyo yakafanana yakasanganiswa dataset ine zvinhu zvakajairika zvakashandiswa muongororo izere, takaona vadzidzi vatatu avo vaipa kuita kwakasimba kwakasimba pazvikamu zvese: Logistic Regression, Naïve Bayes, uye Support Vector Machine.

Muchinjikwa-kusimbisa uye modhi yekuita metric

Kune ese akafanofungidzira modhi (kusanganisira yekutanga ongororo), yega yega modhi yakavakwa uchishandisa 10-kupeta muchinjikwa kusimbiswa, uye kuita kwemuenzaniso kwakayerwa uchishandisa Area Under the Receiver Operating Characteristic Curve (AUC). Kugadziriswa kwemuchinjikwa kwakatanga nekuparadzanisa imwe neimwe yezvikamu gumi zvekuenzanisa zvirongwa muzvikamu gumi zvakaenzana (kupeta), uchishandisa zvipfumbamwe zvezvikamu izvi zvakasiyana kudzidzisa modhi uye chikamu chakasara chekuongorora. Iyi nzira yakadzokororwa ka10, ichishandisa chikamu chakasiyana seyedzo yakaiswa mune imwe neimwe iteration. Mibairo yacho yakabva yasanganiswa kuti iverenge mhedzisiro yemuenzaniso wekupedzisira / kuita. Pamudzidzi wega wega / dhatasethi musanganiswa, iyi yese maitiro akadzokororwa ka10 nedata richipatsanurwa zvakasiyana nguva imwe neimwe. Iyi nhanho yekupedzisira yakaderedza kusarura, yakavimbisa kudzokororwa, uye yakabatsira mukuona iyo yese modhi maitiro. Pakazara (yeMoCA mamakisi uye diagnostic severity classification schemes zvakasanganiswa), 10 modhi dzakavakwa. Izvi zvinosanganisira zviuru zana nemazana masere emhando dzisina kusefa (zvirongwa zvitanhatu zvekuenzanisira zvakashandiswa kudhatabheti×10 vadzidzi×6,600 inomhanya×1,800 kupeta = 6 modhi) uye 3 modhi yakasefa (zvirongwa zvina zvekuenzanisira zvakashandiswa kudhatabheti×10 vadzidzi×10 maitiro ekusarudza maitiro×1,800 anomhanya× 4,800 kupeta = 4 mhando).

Feature sarudzo

Kumamodheru akasefa, kusarudzwa kwechimiro (uchishandisa nzira ina dzenzvimbo) yakaitwa mukati memuchinjiko-kusimbisa. Pane imwe neimwe yegumi, sechikamu chakasiyana che10% chedataset yaive data rekuyedza, chete mana epamusoro akasarudzwa ega ega edzidzo dataset (kureva, mamwe mapfumbamwe, kana 10% yasara yedataset rese) yakashandiswa. kugadzira mamodheru. Hatina kukwanisa kusimbisa kuti ndezvipi zvinhu zvina zvakashandiswa mumuenzaniso wega wega, sezvo ruzivo irworwo rwusina kuchengetwa kana kuitwa kuti rwuwanikwe mukati mepuratifomu yekuenzanisira yatakashandisa (Weka). Nekudaro, tichipihwa kuenderana mukusarudzwa kwedu kwekutanga kwezvimiro zvepamusoro apo marenji akaiswa kune rese rakasanganiswa dhata uye kufanana kwakatevera mukuita modhi, zvinhu izvi zvakafanana (zera, makore edzidzo, MTx-% C, uye zvinoreva MTx-RT. ) ndiwo angangove akanyanya kushandiswa mana epamusoro anoshandiswa pamwe chete nesarudzo yemhando mukati memuchinjiko-yekusimbisa maitiro.

RESULTS

Unhu hwenhamba dzevatori vechikamu (kusanganisira zvibodzwa zveMoCA uye MemTrax performance metrics) yemaseti akateedzana ega ega ega ega ega ega ega ekufanotaura neMoCA-inotaridzirwa hutano hwekuziva (yakajairika maringe neMCI) uye kuomarara kwekuongorora (zvinyoro kana zvakakomba) zvinoratidzwa muTebhu 3.

Tafura 3

Mutori wechikamu maitiro, zvibodzwa zveMoCA, uye kuita kweMemTrax kune yega yega modhi yekuronga zano

Classification StrategyzeradzidzoMoCA YakagadziriswaMoCA isina KugadziriswaMTx-% CMTx-RT
MoCA Category61.9 y (13.1)9.6 y (4.6)19.2 (6.5)18.4 (6.7)74.8% (15.0)1.4 s (0.3)
Diagnosis Severity65.6 y (12.1)8.6 y (4.4)16.7 (6.2)15.8 (6.3)68.3% (13.8)1.5 s (0.3)

Values ​​inoratidzwa (kureva, SD) inosiyaniswa nekuenzanisira kupatsanura nzira inomiririra dataset rakasanganiswa rinoshandiswa kufanotaura MoCA-yakaratidza cognitive hutano (MCI maringe neyakajairwa) uye XL diki-dhataset rinoshandiswa kufanotaura kuomarara kwechirwere (chinyoro kana chakaomarara).

Pamusanganiswa wega wega weMoCA mamakisi (yakagadziriswa / isina kugadziriswa) uye chikumbaridzo (26/23), pakanga paine mutsauko wehuwandu (p = 0.000) mukuenzanisa kwega kwega (hutano hwekuziva hunoenderana neMCI) yezera, dzidzo, uye kuita kweMemTrax (MTx-% C uye MTx-RT). Murwere wega wega sub-dataset mukirasi yakatarisana yeMCI yemusanganiswa wega wega waive paavhareji anenge 9 kusvika 15 makore ekure, akashuma makore mashanu mashoma edzidzo, uye aive nekushomeka kweMemTrax kuita kwema metrics ese ari maviri.

Predictive modelling performance results for the MoCA score classifications using the top three learners, Logistic Regression, Naïve Bayes, and Support Vector Machine, zvinoratidzwa muTable 4. Izvi zvitatu zvakasarudzwa zvichibva pakuita kwemudzidzi kusingachinjiki kwepamusoro pamhando dzose dzakasiyana-siyana. inoshandiswa kune datasets kune ese mamodeli zvirongwa. Kune iyo isina kusefa dataset uye modhi, imwe neimwe ye data data iri muTable 4 inoratidza mashandiro emodhi yakavakirwa paAUC inorehwa inotorwa kubva kumazana zana (100 anomhanya×10 kupeta) yakavakirwa ega ega mudzidzi/yemuenzaniso chirongwa musanganiswa, neyakasiyana-siyana. mudzidzi anotamba anoratidzwa nemavara matema. Nepo pakusefa dataset modelling, mibairo yakataurwa muTebhurari 10 inoratidza avhareji yemhando yemhando dzemamodheru kubva kumazana mana emhando dzemudzidzi wega wega achishandisa imwe neimwe yemaitiro ekuisa maitiro (4 maitiro ekuisa maitiro×400 anomhanya×4 kupeta).

Tafura 4

Dichotomous MoCA mamakisi echikamu chekuita (AUC; 0.0–1.0) mhedzisiro yevatatu vevatatu vanoita-pamusoro-soro kune ese akateedzana zvirongwa zvekuenzanisira.

Feature Set YakashandiswaMoCA ScoreCutoff ThresholdKudzvinyirira KunogadzirisaNaïve BayesTsigira Vector Machine
Haina kusefa (10 maitiro)Zvakashandurwa230.88620.89130.8695
260.89710.92210.9161
Kusagadziriswa230.91030.90850.8995
260.88340.91530.8994
Sefa (4 zvinhu)Zvakashandurwa230.89290.89540.8948
260.91880.92470.9201
Kusagadziriswa230.91350.91340.9122
260.91590.92360.9177

Uchishandisa mutsauko weiyo seti seti, MoCA mamaki, uye MoCA mamakisi cutoff chikumbaridzo, kuita kwepamusoro-soro kwechirongwa chega chega chemuenzaniso chinoratidzwa mu. oneka (hazvina hazvo kuti zvakasiyana pane vamwe vese vasiri mukati oneka kune iyo yakatarisana modhi).

Tichienzanisa vadzidzi pamisanganiswa yese yeMoCA mamakisi eshanduro uye zvikumbaridzo (yakagadziridzwa/isina gadziriso uye 23/26, zvichiteerana) mune yakasanganiswa isina kusefa dataset (kureva, kushandisa gumi akajairika maficha), Naïve Bayes ndiye aiwanzove mudzidzi anoita zvepamusoro ane yakazara. classification performance ye10. Tichifunga nezvevadzidzi vatatu vepamusoro, iyo Bayesian-yakabatana-yakasaina-yenzvimbo bvunzo yakaratidza kuti mukana (Pr) yeNaïve Bayes yakapfuura Logistic Regression yaive 99.9%. Zvakare, pakati peNaïve Bayes neSupport Vector Machine, mukana we21.0% wekuenzana kunoshanda mukuita kwevadzidzi (saka, mukana we 79.0% weNaïve Bayes inokunda Support Vector Machine), pamwe chete ne 0.0% mukana weSupport Vector Machine kuita zvirinani, zvinoyerwa. inosimbisa mukana wekuita weNaïve Bayes. Kumwe kuenzaniswa kweMoCA mamakisi vhezheni kune vese vadzidzi / zvikumbaridzo zvakaratidza mukana wekuita zvishoma uchishandisa zvisina kurongeka MoCA zvibodzwa zvichienzaniswa nekugadziriswa (0.9027 maringe ne0.8971, zvichiteerana; Pr (isina kugadziriswa> yakagadziriswa) = 0.988). Saizvozvo, kuenzanisa kweiyo cutoff chikumbaridzo kune vese vadzidzi uye MoCA mamakisi mavhezheni airatidza diki danho rekuita mukana uchishandisa makumi maviri nenhanhatu sechikamu chechikamu maringe ne26 (23 maringe ne0.9056, zvichiteerana; Pr (26 > 23) = 0.999). Chekupedzisira, tichiongorora maitirwo emhando yemamodheru tichishandisa zvakasefa chete (kureva,, ari pamusoro-soro maficha mana chete), Naïve Bayes (0.9143) ndiye aive mudzidzi aita zvepamusoro pamhando dzese dzeMoCA zvibodzwa. Nekudaro, pamativi ese ezvimiro maitiro akasanganiswa, vese vadzidzi vepamusoro vakaita zvakafanana. Bayesian vakasaina-renji bvunzo dzakaratidza 100% mukana wekuenzana kunoshanda pakati pevaviri vevadzidzi vakasefa. Sezvakaita nedata risina kucheneswa (uchishandisa ese gumi akajairika maficha), pakanga paine zvakare mukana wekuita weiyo isina kurongeka vhezheni yeMoCA mamakisi (Pr (isina kugadziriswa> yakagadziridzwa) = 1.000), pamwe chete nemukana wakasiyana wakasiyana wechikamu chechikamu che26 (Pr (26> 23) = 1.000). Zvikuru, avhareji yekushanda kwemumwe nemumwe wevatatu vepamusoro vadzidzi pane ese MoCA mamakisi mavhezheni / zvikumbaridzo vachishandisa chete-yepamusoro-zvimiro zvina zvakapfuura avhareji yekuita kwechero mudzidzi pane iyo data isina kusefa. Hazvishamise kuti kuita kwemhando yemhando dzakacheneswa (uchishandisa iyo yepamusoro-nzvimbo ina maficha) zvakazara zvaive zvepamusoro (0.9119) kune asina kusefa mamodheru (0.8999), zvisinei nemhando yemhando yemamodheru akafananidzwa neaya mamodheru achishandisa ese gumi akajairika. features. Kune yega yega nzira yekusarudza maitiro, pakanga paine 10% mukana wemubairo wekuita pane iwo asina kusefa.

Nevarwere vanotariswa kuti AD yekuongororwa kuomarara kupatsanurwa, pakati-boka (MCI-AD neAD) misiyano yezera (p = 0.004), dzidzo (p = 0.028), chibodzwa cheMoCA chakagadziridzwa / chisina kugadziriswa (p = 0.000), uye MTx-% C (p = 0.008) zvaive zvakakosha; nepo MTx-RT yanga isiri (p = 0.097). Nevaya varwere vanotariswa kuVaD kuongororwa kuomarara kwechikamu, pakati-boka (MCI-VaD versus VaD) misiyano yeMoCA mamakisi akagadziridzwa / asina kugadziriswa (p = 0.007) uye MTx-% C (p = 0.026) uye MTx-RT (p = 0.001) zvaive zvakakosha; nepo zera (p = 0.511) uye dzidzo (p = 0.157) pakanga pasina kukosha pakati-kusiyana kweboka.

Predictive modelling performance results for the diagnosis severity classifications tichishandisa vadzidzi vatatu vakasarudzwa kare, Logistic Regression, Naïve Bayes, uye Support Vector Machine, inoratidzwa muTable 5. Nepo vamwe vadzidzi vakaongororwa vakaratidza maitiro akasimba zvishoma mumwe nomumwe neimwe yezvikamu zviviri zvekuongororwa kwekliniki. , vadzidzi vatatu vatakanga taona kuti ndivo vainyanya kufarirwa mumuenzaniso wedu wekare vakapa kushanda kwakanyatsoenderana nehurongwa hutsva hwekuenzanisira. Tichienzanisa vadzidzi pane imwe neimwe yemhando dzekutanga dzekuongorora (AD neVaD), pakanga pasina mutsauko unowirirana wekuita musiyano pakati pevadzidzi veMCI-VaD vachipesana neVaD, kunyangwe Support Vector Machine yaiwanzoita zvine mukurumbira. Saizvozvo, pakanga pasina misiyano yakakosha pakati pevadzidzi veMCI-AD vachipesana neAD, kunyangwe Naïve Bayes (NB) yaive nekashoma kuita mukana pamusoro peLogistic Regression (LR) uye kungowanda kusingaite pamusoro peSupport Vector Machine, ine mukana we61.4% uye 41.7% maererano. Pakati pemaseti ese ari maviri, pakanga paine mukana wekuita basa reSupport Vector Machine (SVM), ine Pr (SVM> LR) = 0.819 uye Pr (SVM > NB) = 0.934. Kuita kwedu kwese kwechikamu chevadzidzi vese mukufanotaura kuomarara kwekuongororwa muXL sub-dataset kwaive nani muchikamu cheVaD chekuongorora chinopesana neAD (Pr (VAD> AD) = 0.998).

Tafura 5

Dichotomous kiriniki yekuongorora severity classification performance (AUC; 0.0-1.0) mhedzisiro yemumwe wevatatu vepamusoro-vatatu vadzidzi kune ese ari maviri akateedzana mamodeli zvirongwa.

Modelling SchemeKudzvinyirira KunogadzirisaNaïve BayesTsigira Vector Machine
MCI-AD inopesana neAD0.74650.78100.7443
MCI-VaD inopesana neVaD0.80330.80440.8338

Kushanda kwepamusoro kwechimwe nechimwe chirongwa chemuenzaniso chinoratidzwa mukati oneka (hazvina hazvo kuti zvakasiyana pane vamwe vasiri mukati oneka).

KUKURUKURA

Kuonekwa kwekutanga kwekuchinja kwehutano hwekuziva kwakakosha zvinoshanda mukutonga kwehutano hwemunhu uye hutano hweveruzhinji zvakafanana. Hongu, zvakare zvakanyanya kukoshesa zvakanyanya mumakiriniki marongero evarwere pasi rese. Chinangwa chakagovaniswa ndechekuzivisa varwere, vatarisiri, uye vapeji uye nekukurumidza kurapwa kwakakodzera uye kusingadhure uye kutarisira kwenguva refu kune avo vanotanga kuona kuderera kwekuziva. Kubatanidza zvipatara zvedu zvitatu / makiriniki (s) data subsets, takaona vadzidzi vatatu vanosarudzika (vane imwe yakakurumbira standout -Naïve Bayes) kuvaka mhando dzekufungidzira vachishandisa. MemTrax performance metrics inogona kurongedza nekuvimbika hutano hwekuziva mamiriro dichotomously (yakajairika cognitive hutano kana MCI) sezvaizoratidzwa neMoCA aggregate mamakisi. Zvikuru, kuita kwese kwese kwevadzidzi vatatu kwakavandudzika apo modhi yedu yakangoshandisa iyo yepamusoro-yepamusoro maficha ayo ainyanya kubatanidza aya MemTrax performance metrics. Zvakare, isu takaratidza mukana wakasimbiswa wekushandisa vadzidzi vakafanana uye MemTrax performance metrics mune yekuongorora tsigiro yemhando yekuenzanisira chirongwa chekusiyanisa kuomarara kwemapoka maviri echirwere chedementia: AD neVaD.

Memory test ndiyo yakakosha pakuonekwa kwekutanga kweAD [23, 24]. Saka, mukana wekuti MemTrax inogamuchirwa, inobata, uye iri nyore kuita online. kuongorora bvunzo ye episodic memory muhuwandu hwevanhu [6]. Kuzivikanwa kwechokwadi uye nguva dzekupindura kubva kune inoenderera mberi kuita basa iri kunyanya kuburitsa mukuziva kwekutanga uye kubuda kuderera uye kukonzeresa kushomeka kwemaitiro neuroplastic ane chekuita nekudzidza, ndangariro, uye cognition. Ndiko kuti, mamodheru ari pano akavakirwa zvakanyanya paMemTrax performance metrics anonzwa uye anowanzo kuve nyore uye nemutengo wakaderera anoratidza biological neuropathologic deficits panguva yekuchinja asymptomatic nhanho pamberi pekurasikirwa kwakanyanya kwekushanda [25]. Ashford et al. akanyatsoongorora maitiro uye maitiro ekuzivikanwa ndangariro kurongeka uye nguva yekupindura muvashandisi vepamhepo vakapinda vega neMemTrax [6]. Nekuremekedza kuti kugoverwa uku kwakakosha mukuenzanisira kwakaringana uye kugadzira zvikumbiro zvinogoneka uye zvinoshanda zvevarwere, kutsanangura kucherechedzwa kunoshanda kwekiriniki uye nguva yekupindura maprofile kwakakosha pakumisa hwaro hwakakosha hwekushandisa kiriniki uye yekutsvagisa. Iko kukosha kunoshanda kweMemTrax muAD yekuongorora kwekutanga nhanho yekukanganisika uye kusiyanisa tsigiro yekuongorora inoda kuongororwa zvakanyanya mumamiriro ezvinhu ekiriniki apo comorbidities uye yekuziva, manzwiro, uye kugona kwemota kunokanganisa kuita kwebvunzo kunogona kutariswa. Uye kuzivisa maonero eunyanzvi uye kukurudzira zvinoshanda kukiriniki utility, zvakakosha kutanga kuratidza kuenzanisa neyakagadzika yekuziva yehutano yekuongorora bvunzo, kunyangwe iyo yekupedzisira ingave ichioneswa kudzorwa nekuomesera kwekuedza zvinhu, dzidzo uye zvinodzivirira mutauro, uye tsika nemagariro [26] . Panyaya iyi, kuenzanisa kwakanaka kweMemTrax mukushanda kwekiriniki kuMoCA iyo inowanzonzi chiyero cheindasitiri kwakakosha, kunyanya kana tichiyera kureruka kwekushandisa uye kugamuchirwa kwemurwere kweMemTrax.

Kuongorora kwakapfuura kuenzanisa MemTrax neMoCA kunosimbisa humbowo uye humbowo hwekutanga hunobvumidza kuongorora kwedu kwemuenzaniso [8]. Nekudaro, kuenzanisa kwekutanga uku kwakangobatanidza maviri makiyi eMemTrax maitiro ekuita metrics isu takaongorora nemamiriro ekuziva sekutsanangurwa kwazvinoitwa neMoCA uye yakatsanangudza anoenderana siyana uye cutoff kukosha. Isu takadzamisa kiriniki utility ongororo yeMemTrax nekuongorora fungidziro yekuenzanisira-yakavakirwa nzira iyo yaizopa kutarisisa kwakasiyana-siyana kwezvimwe zvingangove zvakakosha zvakatarwa nemurwere. Kusiyana nevamwe, isu hatina kuwana mukana mukuita modhi tichishandisa kururamisa dzidzo (kugadziridzwa) kune chibodzwa cheMoCA kana mukusiyanisa hutano hwekuziva kusarura MoCA aggregate mamakisi zvibodzwa kubva pakutanga yakakurudzirwa 26 kusvika 23 [12, 15]. Muchokwadi, iyo yemhando yekuita mukana inofarirwa uchishandisa iyo isina kurongeka MoCA mamakisi uye yepamusoro chikumbaridzo.

Pfungwa huru mukuita kwekiriniki

Kudzidza kwemuchina kunowanzo shandiswa zvakanyanya uye kunoshanda zvakanyanya mukufungidzira modhi kana iyo data yakakura uye yakawanda-dimensional, ndiko kuti, kana paine akawanda ekutarisa uye inopindirana yakafara dhizaini yemhando yepamusoro (inopa) hunhu. Asi, neaya data aripo, mamodhi akasefa ane mana chete akasarudzika maficha akaita zvirinani pane ayo anoshandisa ese gumi akajairika maficha. Izvi zvinoratidza kuti dhatabheti yedu yechipatara yakaunganidzwa yakanga isina yakanyanya kukodzera kliniki (yakakwirira kukosha) maficha ekunyatsoronga varwere nenzira iyi. Zvakangodaro, iyo ficha inosimbisa pakiyi MemTrax performance metrics-MTx-% C uye MTx-RT-inotsigira zvakasimba kuvaka nhanho yekutanga yekuziva kushomeka yekuongorora mhando dzakatenderedza bvunzo iyi iri nyore, iri nyore kubata, yakaderera-mutengo, uye kunyatso ratidza maererano. Memory performance, ingangoita izvozvi seyekutanga chidzitiro chebinary classification yehutano hwekuziva. Tichifunga nezvekugara-kukwira kune vanopa uye hutano masisitimu, maitiro ekuongorora varwere uye makiriniki mashandisirwo anofanirwa kuvandudzwa nekusimbisa kuunganidza, kutevera, uye kuenzanisira iwo hunhu hwevarwere uye bvunzo metrics anonyanya kubatsira, anobatsira, uye anoratidzwa anoshanda mukuongorora. uye rubatsiro rwekutarisira murwere.

Iine maviri makiyi eMemTrax metrics ari pakati peMCI class, wedu-anoita-pamusoro mudzidzi (Naïve Bayes) aive nepamusoro-soro ekufungidzira kuita mumamodheru mazhinji (AUC pamusoro pe0.90) ine yechokwadi-yakanaka kune yenhema-positive ratio iri pedyo kana kuti inodarika mana. 4 kushayikwa iyo kupatsanurwa mune avo vane kusaziva kwekuziva (nhema dzisina kunaka). Chero chimwe chezviitiko izvi zvekusarudzika zvinogona kuisa mutoro usina kufanira wepfungwa-wemagariro kune murwere nevachengeti.

Nepo muongororo yekutanga uye yakazara takashandisa vadzidzi gumi muchirongwa chega chega chekuenzanisira, takaisa mhedzisiro yedu pavatatu vekirasi vanoratidza kuita kwakasimba kwakasimba. Izvi zvakare zvaive zvekusimbisa, zvichibva pane idzi data, vadzidzi vangatarisira kuita zvakavimbika pamwero wepamusoro mune inoshanda yekiriniki application mukusarudza yekuziva mamiriro ekuisa. Pamusoro pezvo, nekuti chidzidzo ichi chakaitirwa seyekutanga kuferefeta mashandisiro emuchina kudzidza pakuongorora kwenjere uye aya matambudziko akakodzera ekiriniki aya, takaita danho rekuchengeta nzira dzekudzidza dziri nyore uye dzakajairwa, paine kushomeka kweparamita. Isu tinoonga kuti nzira iyi inogona kunge yakaganhura mukana wezvakanyanya kutsanangurwa nemurwere-chaiwo kufungidzira kugona. Saizvozvowo, nepo kudzidzisa mamodheru tichishandisa chete epamusoro maficha (yakasefa nzira) inotizivisa zvimwe nezve data iri (chaiyo kune zvikanganiso mudhata rakaunganidzwa uye kuratidza kukosha mukugadzirisa yakakosha nguva yekiriniki uye zviwanikwa), isu tinoziva kuti nguva isati yakwana kupfupika. hupamhi hwemamodheru uye, naizvozvo, ese (nezvimwe zvinhu) zvinofanirwa kutariswa netsvagurudzo yeramangwana kudzamara tave neruzivo rwakanyanya rwezvinhu zvakakosha izvo zvinozoshanda kune ruzhinji rwevanhu. Nokudaro, isu tinonzwisisawo zvizere kuti dhata rinosanganisira uye rakapamhama rinomiririra uye optimization yeizvi uye mamwe mamodheru zvingave zvakakosha tisati tazvibatanidza muchirongwa chekiriniki chinoshanda, kunyanya kugadzirisa comorbidities inokanganisa mashandiro ekuziva izvo zvingada kutariswa mune imwezve kliniki kuongororwa.

Kushandiswa kweMemTrax kwakagadziridzwa zvakare nekuenzanisira kwekuoma kwechirwere zvichienderana nekuongororwa kwekiriniki kwakasiyana. Kuita kurinani kwese kwese kwekufanotaura kuomarara kweVaD (kuenzaniswa neAD) kwakanga kusiri zvinoshamisa zvakapihwa chimiro chemurwere mune mamodheru akananga kuhutano hwevascular uye njodzi yekurohwa, kureva, hypertension, hyperlipidemia, chirwere cheshuga, uye (zvechokwadi) nhoroondo yesitiroko. Kunyangwe zvingave zvakanyanya kudiwa uye zvakafanira kuve neiyo ongororo yekiriniki inoitwa pavarwere vakafanana vane hutano hwekuziva kudzidzisa vadzidzi neaya anosanganisirwa data. Izvi zvinonyanya kutenderwa, sezvo MemTrax inotarisirwa kushandiswa zvakanyanya padanho rekutanga kuona kushomeka kwekuziva uye kunotevera kunotevera shanduko yemunhu. Zvinogoneka zvakare kuti kugovera kunonyanya kudiwa kwedata muVaD dataset kwakabatsira muchikamu mukuenzaniswa kurinani kwekuita kwemuenzaniso. Iyo dataset yeVaD yaive yakanyatso kurongeka pakati pemakirasi maviri, nepo AD dhata nevarwere vashoma veMCI yanga isina. Kunyanya mumaseti madiki, kunyangwe mashoma ekuwedzera zviitiko anogona kuita mutsauko unoyerwa. Maonero ese ari maviri nharo dzine musoro dziri pasi pekusiyana kwechirwere chekuomarara modhi maitiro. Nekudaro, kuverengera kuverengera kuita kwakagadziridzwa kune dataset nhamba hunhu kana zvemukati maficha akananga kune yekiriniki mharidzo iri kutariswa ndeye nguva isati yakwana. Zvakangodaro, riini iri rakaratidza kushanda kweMemTrax yekufungidzira kupatsanura modhi muchikamu chekiriniki yekuongorora tsigiro inopa yakakosha maonero uye inosimbisa kutsvaga kwekuwedzera kuongororwa nevarwere mukati mekuenderera mberi kweMCI.

Kuitwa uye kwakaratidza kushandiswa kweMemTrax uye aya mamodheru muChina, uko mutauro netsika zvakasiyana zvakanyanya kubva kune mamwe matunhu ezvakagadziriswa zvekushandisa (semuenzaniso, France, Netherlands, uye United States) [7, 8, 27], inosimbisa izvo zvinogoneka. kugamuchirwa kwepasirese uye kukosha kwekiriniki kweMemTrax-based platform. Uyu muenzaniso unoratidzwa mukuvavarira kuwiriranisa data uye kugadzira tsika dzepasirese uye zviwanikwa zvekuenzanisira zvekuongorora zvine hungwaru zvakamisikidzwa uye zvinogadziriswa zviri nyore kuti zvishandiswe pasi rese.

Matanho anotevera mukuziva kuderera modhi uye kushandisa

Kusagadzikana kwekuziva muAD zvechokwadi kunoitika pane inoenderera, kwete mumatanho akasiyana kana matanho [28, 29]. Nekudaro, panguva ino yekutanga, chinangwa chedu chaive chekutanga kumisa kugona kwedu kuvaka modhi inobatanidza MemTrax iyo inogona kusiyanisa "zvakajairika" kubva "zvisiri zvakajairika". Yakawanda inosanganisirwa empirical data (semuenzaniso, kufungidzira kwehuropi, genetic maitiro, biomarker, comorbidities, uye anoshanda mamaki ezvakaoma. mabasa anoda njere control) [30] munzvimbo dzakasiyana-siyana dzepasirese, vanhu, uye mapoka ezera kudzidzisa nekusimudzira zvakanyanya (kusanganisira zvinorema zvakaringana ensemble) mhando dzekudzidza dzemuchina dzichatsigira huwandu hukuru hwekunatsiridza kupatsanura, ndiko kuti, kugona kugovera mapoka evarwere vane MCI mune zvidiki uye zvakanyanya kujekesa subset pamwe nekuziva kuderera kuenderera. Zvakare, kuwirirana kwekiriniki yekuongororwa kwevanhu mumatunhu akasiyana-siyana evarwere kwakakosha kuti zvinobudirira dzidzisa aya mamodheru anosanganisirwa uye anofungidzira akasimba. Izvi zvinogonesa mamwe stratified kesi manejimendi kune avo vane mamiriro akafanana, pesvedzero, uye zvakanyanya kutsanangurwa zvishoma hunhu hwekuziva maprofiles uye nekudaro kukwidziridza tsigiro yekiriniki yesarudzo uye kutarisirwa kwevarwere.

Zvakawanda zvekutsvagisa kwekiriniki kwakakodzera kusvika-zvino zvakataura nevarwere vane kanenge nyoro dementia; uye, mukuita, kazhinji kupindira kwevarwere kunongoedzwa pamatanho epamusoro. Nekudaro, nekuti kuderera kwenjere kunotanga pasati pasangana nzira dzekiriniki yedementia, chidzitiro cheMemTrax-based early screen chinogona kukurudzira dzidzo yakakodzera yevanhu nezvechirwere uye mafambiro acho uye nekukurumidza uye kupindira kwakawanda panguva. Nokudaro, kuonekwa kwepakutanga kunogona kutsigira kubatanidzwa kwakakodzera kubva pakurovedza muviri, kudya, kutsigirwa mupfungwa, nekuvandudza kushamwaridzana kune kupindira kwemishonga uye kusimbisa shanduko dzine chokuita nemurwere mumafambiro uye maonero ayo ega kana kuti pamwe chete anogona kuderedza kana kumisa kufambira mberi kwedementia [31, 32] . Uyezve, nekubudirira kuongororwa kwekutanga, vanhu nemhuri dzavo vanogona kukumbirwa kuti vafunge nezvezviedzo zvekiriniki kana kuwana mazano uye mamwe rubatsiro rwemagariro evanhu kuti vabatsire kujekesa zvinotarisirwa uye zvinangwa uye kutarisira mabasa ezuva nezuva. Kumwe kusimbiswa uye kupararira kwekushandisa munzira idzi kunogona kubatsira mukuderedza kana kumisa kufambira mberi kweMCI, AD, uye ADRD kune vanhu vazhinji.

Zvechokwadi, kupera kwakaderera kwezera remurwere mukudzidza kwedu hakumiriri huwandu hwemagariro evanhu vane AD. Zvakangodaro, avhareji yezera reboka rega rega rinoshandiswa muzvikamu zvekuenzanisira zvirongwa zvinoenderana neMoCA mamakisi / chikumbaridzo uye kuomarara kwekuongorora (Tafura 3) inosimbisa ruzhinji rwakajeka (kupfuura 80%) kuve angangoita makore makumi mashanu. Uku kugovera nekudaro kwakanyatso kukonzeresa, kutsigira kushandiswa kwemamodheru aya muhuwandu hunoratidza avo vanowanzobatwa nazvo kutanga kwekutanga uye kuputika kwechirwere cheurocognitive nekuda kweAD neVaD. Zvakare, humbowo huchangobva kuitika uye maonero anosimbisa izvo zvinozivikanwa zvinhu (semuenzaniso, hypertension, kufutisa, chirwere cheshuga, uye kuputa) zvinogona kukonzera kukurumidza kukurumidza. mukuru uye midlife vascular risk scores uye zvinozotevera zvisinganzwisisiki vascular brain kukuvara izvo zvinokura zvisingaite nemhedzisiro inooneka kunyangwe mudiki. vakuru [33–35]. Saizvozvo, yakanyanya kunaka yekutanga yekuongorora mukana wekuona nekukurumidza danho rekuziva deficits uye kutanga kunoshanda kudzivirira uye nzira dzekupindira mukubudirira kugadzirisa dementia ichabuda kubva pakuongorora zvinopa zvinhu uye zviratidziro zvakapfuura muzera rose, kusanganisira kukura uye zvingangove zvehudiki (tichicherekedza kukosha kweiyo genetic zvinhu zvakaita se apolipoprotein E kubva pakubata pamuviri).

Mukuita, kuongororwa kwekiriniki kwakakodzera uye maitiro anodhura ekufungidzira kwepamberi, genetic profiling, uye kuyera anovimbisa biomarkers hazviwanzo kuwanikwa zviri nyore kana kunyange zvinogoneka kune vazhinji vanopa. Nekudaro, muzviitiko zvakawanda, kwekutanga kwehutano hwekuziva mamiriro ehutano kurongwa kunogona kutorwa kubva kumamodheru uchishandisa mamwe mametric akareruka anopihwa nemurwere (semuenzaniso, kuzvitaura. zvinetso zvemusoro, mishonga yemazuva ano, uye zvigadziro zvekuita nguva dzose) uye zvinowanzoonekwa zvevanhu [7]. Registries yakadai seYunivhesiti yeCalifornia Hutano Hutano Registry (https://www.brainhealthregistry.org/) [27] nevamwe vane hupamhi hukuru hwekuzvizivisa zviratidzo, maitiro ehunhu (semuenzaniso, kurara uye kuziva zuva rega rega), mishonga, mamiriro ehutano, uye nhoroondo, uye huwandu hwakadzama hwehuwandu huchabatsira mukugadzira uye kusimbisa mashandisirwo anoshanda eaya emhando dzechinyakare mukiriniki. Kupfuurirazve, bvunzo yakaita seMemTrax, iyo yakaratidza kushanda mukuongorora ndangariro basa, inogona kupa fungidziro iri nani yeAD pathology pane biological mamaki. Tichifunga kuti chinonyanya kukosha cheAD pathology kukanganisa neuroplasticity uye kurasikirwa kwakanyanya kwakaoma kwemasynapses, izvo zvinoratidzwa seepisodic. memory dysfunction, chiyero chinoongorora episodic memory chingangoitika kupa fungidziro iri nani yeAD pathological mutoro pane biological mamaki mumurwere mupenyu [36].

Nemamodheru ese ekufungidzira-angave akatsigirwa nedata rakaoma uye rinosanganisirwa kubva kune-ye-ye-iyo-hunyanzvi tekinoroji uye yakanatswa yekiriniki yekuziva munzvimbo dzakawanda kana idzo dzinoganhurirwa kune mamwe ekutanga uye ari nyore ruzivo ruzivo rwemaitiro evarwere aripo-mubairo unozivikanwa wehungwaru hwekugadzira. uye kudzidza kwemuchina ndekwekuti mamodheru anozobuda anogona kubatanidza uye zvine mutsindo "kudzidza" kubva kune yakakosha data uye maonero anopihwa nekuenderera mberi kwekushandisa kwekushandisa. Kutevera kushandiswa kwehunyanzvi hwekuita tekinoroji, sezvo mamodheru ari pano (uye achakudziridzwa) ari kushandiswa uye kupfumiswa neakawanda nyaya uye data yakakodzera (kusanganisira varwere vane comorbidities iyo inogona kuunza nekutevera kuderera kwekuziva), kufanotaura kwekuita uye kurongeka kwehutano hwekuziva kuchave kwakasimba, zvichiita kuti pave nekuwedzera kushanda kwekiriniki sarudzo yekutsigira utility. Kushanduka uku kuchave kwakazara uye nekunyatso zivikanwa nekuisa MemTrax mutsika (yakanangwa kune iripo kugona) mapuratifomu anogona kushandiswa nevarapi vehutano munguva chaiyo mukiriniki.

Izvo zvakakosha pakusimbisa uye kushandiswa kweiyo MemTrax modhi yerutsigiro yekuongorora uye kuchengetwa kwevarwere inotsvakwa zvakanyanya-kutevera ine revo rengitudinal data. Nokucherechedza nekunyora shanduko dzinoenderana (kana dziripo) muhutano hwemakiriniki mukati mehutano hwakakwana hwemazuva ose kuburikidza nekutanga-danho reMCI, mienzaniso yekuongorora kwakakodzera kunoenderera mberi uye kurongedza kunogona kudzidziswa uye kuchinjwa sezera revarwere uye vanorapwa. Izvi zvinoreva kuti, kushandiswa kwakadzokororwa kunogona kubatsira nekutevera kwenguva refu kwekuchinja kwehunyoro hwekuziva, kushanda kwekupindira, uye kuchengetedza ruzivo rwakarongwa. Iyi nzira inoenderana zvakanyanya nekuita kwekiriniki uye murwere uye kutonga kwenyaya.

Nokuremara

Tinoonga dambudziko uye kukosha kwekuunganidza data rakachena rekiriniki mukiriniki inodzorwa/muchipatara. Zvakangodaro, zvingadai zvakasimbisa kuenzanisa kwedu kana dhatabhesi redu raisanganisira varwere vakawanda vane maitiro akafanana. Uyezve, zvakanangana nemuenzaniso wedu wekuongorora, zvingave zvakanyanya kudiwa uye zvakakodzera kuti tive nekuongorora kwekiriniki kwakafanana kunoitwa kune vanofananidzwa nevarwere vane hutano hwekuziva kudzidzisa vadzidzi. Uye sekusimbiswa nepamusoro-soro dhizaini mashandiro uchishandisa iyo yakasefa dataset (chete yepamusoro-chinzvimbo mana maficha), zvakajairika uye yekuziva hutano matanho / zviratidzo zvingangove zvakavandudza kuenzanisira kuita nehuwandu hukuru hwezvinhu zvakajairika kune vese varwere.

Vamwe vatori vechikamu vanogona kunge vakasangana nezvimwe zvirwere zvinogona kunge zvakakonzera kushaya njere kana kusingaperi. Zvimwe kunze kweiyo XL sub-dataset apo varwere vakarongedzerwa seane AD kana VaD, comorbidity data haina kuunganidzwa / kutaurwa muYH murwere dziva, uye iyo inonyanya kutaurwa comorbidity kusvika kure muKM sub-dataset yaive chirwere cheshuga. Zvine nharo, zvisinei, kuti kusanganisira varwere muzvirongwa zvedu zvekuenzanisira zvine comorbidities zvinogona kukurudzira kana kuwedzera chiyero chekushaya ruzivo uye mhedzisiro yakaderera yeMemTrax kuita kungave kunomiririra yechokwadi-yepasirese yakanangwa huwandu hwevarwere kune iyi yakajairika yepakutanga yekuongorora yekuongorora. uye modelling nzira. Kufambira mberi, kuongororwa kwakaringana kwema comorbidities angangokanganisa kugona kwekuziva kunobatsira zvakanyanya pakugadzirisa mamodheru uye mhedzisiro yekuchengeta varwere.

Pakupedzisira, YH neKM sub-dataset varwere vakashandisa smartphone kuti vatore MemTrax bvunzo, nepo nhamba shoma yeXL sub-dataset varwere vakashandisa iPad uye vamwe vose vakashandisa smartphone. Izvi zvinogona kunge zvakaunza mutsauko mudiki-une hukama muMemTrax mashandiro eMoCA classification modelling. Nekudaro, misiyano (kana iripo) muMTx-RT, semuenzaniso, pakati pemidziyo ingangove isingaite, kunyanya nemutori wechikamu wega wega achipihwa bvunzo "yekudzidzira" nguva isati yarekodhwa kuita kwebvunzo. Zvakangodaro, kushandiswa kwezvishandiso zviviri izvi zvakabatwa nemaoko zvinogona kukanganisa kuenzanisa kwakananga uye / kana kusanganisa nemamwe maMemTrax mhinduro apo vashandisi vakapindura kudzokorora mapikicha nekubata spacebar pakhibhodi yekombuta.

Mapoinzi akakosha paMemTrax yekufungidzira modhi yekushandisa

  • • Mhando dzedu dzepamusoro-soro dzekufungidzira dzinosanganisira dzakasarudzwa MemTrax performance metrics dzinogona kurongedza nekuvimbika mamiriro ehutano hwekuziva (yakajairika cognitive hutano kana MCI) sezvaizoratidzwa neyakazivikanwa bvunzo yeMoCA.
  • • Mhedzisiro iyi inotsigira kubatanidzwa kweMemTrax performance metrics yakasarudzwa kuita classification predictive model screening application yepakutanga kukanganiswa kwekuziva.
  • • Kuenzanisira kwedu kwechikamu kwakaratidzawo mukana wekushandisa MemTrax kuita mumashandisirwo ekusiyanisa kuoma kwechirwere chedementia.

Izvi zvitsva zvakawanikwa zvinosimbisa humbowo hwakakwana hunotsigira kushanda kwemuchina kudzidza mukuvaka yakagadziridzwa yakasimba MemTrax-yakavakirwa mutsara modhi yerutsigiro rwekuongorora mune inoshanda kiriniki kesi manejimendi uye kuchengetwa kwevarwere kune vanhu vane dambudziko rekuziva.

ZVINONYANYA

Tinoziva basa raJ. Wesson Ashford, Curtis B. Ashford, uye vatinoshanda navo pakugadzira uye kusimbisa iyo online inoenderera yekuzivikanwa basa uye chishandiso (MemTrax) chinoshandiswa pano uye tinotenda kune vazhinji varwere vane dementia vakabatsira mukukosha kwekutanga tsvakurudzo. . Isuwo tinotenda Xianbo Zhou nevamwe vake paSJN Biomed LTD, vaanoshanda navo uye vashandi panzvimbo dzezvipatara / makiriniki, kunyanya Dr. M. Luo naM. Zhong, vakabatsira nekutora vatori vechikamu, kuronga bvunzo, uye kuunganidza, kurekodha, uye kumberi-kugadzirisa data, uye vatori vechikamu vanozvipira vakapa nguva yavo yakakosha uye vakaita kuzvipira kutora bvunzo nekupa. iyo data yakakosha yekuti isu tiongorore muchidzidzo ichi. Izvi kudzidza kwakatsigirwa muchikamu neMD Scientific Research Chirongwa cheKunming Medical University (Grant no. 2017BS028 kusvika XL) uye Research Program yeYunnan Science and Technology Department (Grant no. 2019FE001 (-222) kusvika XL).

J. Wesson Ashford akaisa chikumbiro chemvumo yekushandiswa kweiyo paradigm inoenderera yekuzivikanwa inotsanangurwa mubepa rino kune general kuedza kwendangariro.

MemTrax, LLC ikambani yaCurtis Ashford, uye kambani iyi iri kutarisira ndangariro kuongorora system inotsanangurwa mubepa rino.

Zvakaburitswa nevanyori zviripo online (https://www.j-alz.com/manuscript-disclosures/19-1340r2).

ndangariro test dementia test memory kurasikirwa bvunzo yenguva pfupi ndangariro kurasikirwa bvunzo gondohwe kuyedza pfungwa kudya kwakasiyana mabhuku cognitive bvunzo online.
Curtis Ashford - Cognitive Research Coordinator

References

[1] Alzheimer's Association (2016) 2016 Alzheimer's chirwere chokwadi uye manhamba. Alzheimers Dement 12, 459-509.
[2] Gresenz CR , Mitchell JM , Marrone J , Federoff HJ (2019) Mhedzisiro yekutanga-nhanho Chirwere cheAlzheimer pamigumisiro yemari yemhuri. Hutano Econ 29, 18–29.
[3] Foster NL , Bondi MW , Das R , Foss M , Hershey LA , Koh S , Logan R , Poole C , Shega JW , Sood A , Thothala N , Wicklund M , Yu M , Bennett A , Wang D (2019) Quality improvement in neurology: Mild cognitive impairment quality measurement set. Neurology 93, 705-713.
[4] Tong T , Thokala P , McMillan B , Ghosh R , Brazier J (2017) Mutengo wekushandisa cognitive screening bvunzo dzekuona dementia uye nyoro yekuziva kukanganisa mukutarisira kwekutanga. Int J Geriatr Psychiatry 32, 1392–1400.
[5] Ashford JW , Gere E , Bayley PJ (2011) Kuyera ndangariro mumapoka makuru eboka uchishandisa bvunzo inoenderera yekuzivikanwa. J Alzheimers Dis 27, 885-895.
[6] Ashford JW , Tarpin-Bernard F , Ashford CB , Ashford MT (2019) Basa rekombuta rinoramba richizivikanwa rekuyera episodic memory. J Alzheimers Dis 69, 385-399.
[7] Bergeron MF , Landset S , Tarpin-Bernard F , Ashford CB , Khoshgoftaar TM , Ashford JW (2019) Episodic-memory performance mumuchina kudzidza modhi yekufembera yekuziva yekuziva hutano mamiriro. J Alzheimers Dis 70, 277-286.
[8] van der Hoek MD , Nieuwenhuizen A , Keijer J , Ashford JW (2019) The MemTrax bvunzo zvichienzaniswa neMontreal cognitive assessment fungidziro yehunyoro hwekuziva kukanganisa. J Alzheimers Dis 67, 1045-1054.
[9] Falcone M , Yadav N , Poellabauer C , Flynn P (2013) Kushandisa manzwi evhawero ega ega ekuronga kwekukuvara kwehuropi hunyoro. Muna 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, Vancouver, BC, pp. 7577–7581.
[10] Dabek F, Caban JJ (2015) Kushandisa data hombe kuenzanisira mukana wekuvandudza mamiriro epfungwa mushure mekuputika. Procedia Comput Sci 53, 265–273.
[11] Climent MT , Pardo J , Munoz-Almaraz FJ , Guerrero MD , Moreno L (2018) Muti wechisarudzo chekuonekwa kwekutanga kwekukanganiswa kwepfungwa nevashambadziri vemunharaunda. Pamberi Pharmacol 9, 1232.
[12] Nasreddine ZS , Phillips NA , Bedirian V , Charbonneau S , Whitehead V , Collin I , Cummings JL , Chertkow H (2005) The Montreal Cognitive Assessment, MoCA: A short screening tool for mild cognitive impairment. J Am Geriatr Soc 53, 695–699.
[13] Yu J , Li J , Huang X (2012) The Beijing version yeMontreal cognitive assessment sechidimbu chekuongorora chekugadzirisa kukanganisa kwekuziva: Chidzidzo chemunharaunda. BMC Psychiatry 12, 156.
[14] Chen KL , Xu Y , Chu AQ , Ding D , Liang XN , Nasreddine ZS , Dong Q , Hong Z , Zhao QH , Guo QH (2016) Kusimbiswa kweChinese version yeMontreal cognitive assessment basic for screening mild cognitive impairment. J Am Geriatr Soc 64, e285–e290.
[15] Carson N , Leach L , Murphy KJ (2018) Ongororo zvakare yeMontreal Cognitive Assessment (MoCA) cutoff scores. Int J Geriatr Psychiatry 33, 379–388.
[16] American Psychiatric Association (2013) Task Force Diagnostic and statistical manual yezvirwere zvepfungwa: DSM-5™, American Psychiatric Publishing, Inc., Washington, DC.
[17] Python. Python Software Foundation, http://www.python.org, Yakapinda munaNovember 15, 2019.
[18] R Core Boka, R: Mutauro uye nharaunda yenhamba komputa R Nheyo yeStatistical Computing, Vienna, Austria. https://www.R-project.org/, 2018, Yakasvika munaNovember 15, 2019.
[19] Benavoli A , Corani G , Demšar J , Zaffalon M (2017) Nguva yekuchinja: Chidzidzo chekuenzanisa vagadziri vakawanda kuburikidza nekuongorora kweBayesian. J Mach Dzidza Res 18, 1–36.
[20] Frank E , Hall MA , Witten IH (2016) The WEKA Workbench. In Data Mining: Inoshanda Machine Kudzidza Zvishandiso uye Matekiniki, Frank E, Hall MA, Witten IH, Pal CJ, ed. Morgan Kaufmann https://www.cs.waikato.ac.nz/ml/weka/Witten_et_al_2016_appendix.pdf
[21] Bergeron MF , Landset S , Maugans TA , Williams VB , Collins CL , Wasserman EB , Khoshgoftaar TM (2019) Kudzidza nemuchina mukuenzanisira chesekondari sport concussion chiratidzo kugadzirisa. Med Sci Sports Exerc 51, 1362-1371.
[22] Van Hulse J , Khoshgoftaar TM , Napolitano A (2007) Maonero ekuedza pakudzidza kubva kune data isina kuenzana. In Zvakaitika zve24th International Conference pamusoro peKudzidza kweMichina, Corvalis, Oregon, USA, mapeji 935-942.
[23] Ashford JW , Kolm P , Colliver JA , Bekian C , Hsu LN (1989) Kuongororwa kwevarwere veAlzheimer uye mini-mental state: Item characteristic curve analysis.P. J Gerontol 44, 139–146.
[24] Ashford JW , Jarvik L (1985) Chirwere cheAlzheimer: Ko neuron plasticity inofanotungamira kune axonal neurofibrillary degeneration? N Engl J Med 313, 388–389.
[25] Jack CR Jr , Therneau TM , Weigand SD , ​​Wiste HJ , Knopman DS , Vemuri P , Lowe VJ , Mielke MM , Roberts RO , Machulda MM , Graff-Radford J , Jones DT , Schwarz CG , Gunter JL , Rocca WA , Petersen RC (2019) Kuwandisa kwebiologically vs kuchipatara kunotsanangurwa neAlzheimer spectrum entities uchishandisa National Institute on Aging-Alzheimer's Association Research framework. JAMA Neurol 76, 1174-1183.
[26] Zhou X, Ashford JW (2019) Kufambira mberi mukuongorora zviridzwa zve. Chirwere cheAlzheimer. Kuchembera Med 2, 88–93.
[27] Weiner MW , Nosheny R , Camacho M , Truran-Sacrey D , Mackin RS , Flenniken D , Ulbricht A , Insel P , Finley S , Fockler J , Veitch D (2018) The Hutano Hutano Registry: Indaneti-yakavakirwa papuratifomu yekutsvaga, kuongorora, uye kutarisa kwenguva refu kwevatori vechikamu kune neuroscience zvidzidzo. Alzheimers Dement 14, 1063-1076.
[28] Ashford JW , Schmitt FA (2001) Kutevedzera nguva-kosi ye Alzheimer dementia. Curr Psychiatry Rep 3, 20–28.
[29] Li X , Wang X , Su L , Hu X , Han Y (2019) Sino Longitudinal Chidzidzo pamusoro peCognitive Decline (SILCODE): Protocol yeChinese longitudinal Observational study yekuvandudza njodzi yekufungidzira mamodheru ekushandurukira kune yakapfava yekuziva kukanganisa muvanhu vane subjective cognitive. kuderedza. BMJ Vhura 9, e028188.
[30] Tarnanas I , Tsolaki A , Wiederhold M , Wiederhold B , Tsolaki M (2015) Five-year biomarker progression variability for Chirwere cheAlzheimer's dementia kufanotaura: Zviitiko zvakaoma zvechishandiso zvemazuva ese ekurarama mucherechedzo zvinogona kuzadza maburi? Alzheimers Dement (Amst) 1, 521-532.
[31] McGurran H , Glenn JM , Madero EN , Bott NT (2019) Kudzivirira uye kurapwa kwechirwere cheAlzheimer: Biological nzira dzekurovedza muviri. J Alzheimers Dis 69, 311-338.
[32] Mendiola-Precoma J , Berumen LC , Padilla K , Garcia-Alcocer G (2016) Kurapa kwe kudzivirira uye kurapwa kwechirwere cheAlzheimer. Biomed Res Int 2016, 2589276.
[33] Lane CA , Barnes J , Nicholas JM , Sudre CH , Cash DM , Malone IB , Parker TD , Keshavan A , Buchanan SM , Keuss SE , James SN , Lu K , Murray-Smith H , Wong A , Gordon E , Coath W , Modat M , Thomas D , Richards M , Fox NC , Schott JM (2020) Kudyidzana pakati pengozi yevascular mukukura uye kurwara kwehuropi munguva yekupedzisira: Humbowo kubva kuboka rekuzvarwa reBritish. JAMA Neurol 77, 175-183.
[34] Seshadri S (2020) Kudzivirirwa kwedementia-kufunga kupfuura zera uye amyloid mabhokisi. JAMA Neurol 77, 160-161.
[35] Maillard P , Seshadri S , Beiser A , Himali JJ , Au R , Fletcher E , Carmichael O , Wolf PA , DeCarli C (2012) Migumisiro ye systolic blood pressure pa white-matter integrity in young people in the Framingham Heart Study: A cross -chikamu kudzidza. Lancet Neurol 11, 1039–1047.
[36] Fink HA , Linskens EJ , Silverman PC , McCarten JR , Hemmy LS , Ouellette JM , Greer NL , Wilt TJ , Butler M (2020) Kururama kwekuongororwa kwebiomarker kune neuropathologically defined Chirwere cheAlzheimer muvakuru vakuru vane dementia. Ann Intern Med 172, 669-677.

Affiliations: [a] SIVOTEC Analytics, Boca Raton, FL, USA | [b] Dhipatimendi reComputer uye Electrical Engineering uye Computer Science, Florida Atlantic University, Boca Raton, FL, USA | [c] SJN Biomed LTD, Kunming, Yunnan, China | [d] Center for Tsvakurudzo yeAlzheimer, Washington Institute of Clinical Research, Washington, DC, USA | [e] Dhipatimendi reRehabilitation Medicine, The First Affiliated Hospital yeKunming Medical University, Kunming, Yunnan, China | [f] Dhipatimendi reNeurology, Dehong People's Hospital, Dehong, Yunnan, China | [g] Dhipatimendi reNeurology, Chipatara Chekutanga Chakabatana cheKunming Medical University, Wuhua District, Kunming, Yunnan Province, China | [h] Chirwere Chinoenderana neHondo uye Kukuvara Chidzidzo, VA Palo Alto Health Care System, Palo Alto, CA, USA | [I] Dhipatimendi rePsychiatry & Behavioral Sayenzi, Stanford University Chikoro cheMishonga, Palo Alto, CA, USA.

Kunyorerana: [*] Kunyorerana ku: Michael F. Bergeron, PhD, FACSM, SIVOTEC Analytics, Boca Raton Innovation Campus, 4800 T-Rex Avenue, Suite 315, Boca Raton, FL 33431, USA. E-mail: mirgeron@sivotecanalytics.com.; Xiaolei Liu, MD, Dhipatimendi reNeurology, First Affiliated Hospital yeKunming Medical University, 295 Xichang Road, Wuhua District, Kunming, Yunnan Province 650032, China. E-mail: ring@vip.163.com.

Keywords: Kuchembera, Chirwere cheAlzheimer, dementia, kuongororwa kwevanhu vakawanda