bangodi: Ling, Zou; Yue, Chen; Wenjie, Li; Fan, Jing
Source: Journal ea Litšoantšo tsa Bongaka le Bophelo bo Botle ba Bophelo, Buka ea 5, Nomoro ea 7, Pulungoana 2015, maq. 1499-1503 (5)
mohoeletsi: Bahatisi ba Saense ba Amerika
Sekhetho:
Leqepheng lena, lipontšo tsa electroencephalogram (EEG) li tlalehiloe ho tsoa ho baithuti ba univesithi ba leshome ba nang le ts'ebeliso e mpe ea inthanete (IA) nakong ea pono ea oddball. Taba ea mantlha, lipontšo tsa mantlha li ile tsa ntlafatsoa ho tlosa lisebelisoa tse ling tse sebelisang algorithm ea Independent Component Analysis (ICA). Joale, Principal Component Analysis (PCA) e ile ea hiroa ho khetha seteishene sa liteishene tse bolokang tlhaiso-leseling e ngata ha e bapisoa le sete e felletseng ea liteishene tse 64. Kamora nako, likarolo tsa maqhubu a P300 li ntšitsoe ho menyetla e amanang le ketsahalo (ERPs) mme ea bapisoa ho pholletsa le sepheo sa li-ERP le li-ERP tse sa reretsoeng, hammoho le sehlopha ka seng sa IA le sehlopha sa taolo. Likarolo tse ntšitsoeng li ile tsa sebelisoa hape ho koetlisa li-classifiers tse 'ne: Fisher Linear Discriminate Analysis (FLDA), Back Propagation (BP) Neural Network, Bayesian Classifier (BC) le Bayesian Regularization Back Propagation (BRBP) Neural Network. Likanale tse sebetsang li ne li le libakeng tse ka pele, tsa parietal, tsa occipital le tsa parietal-occipital bakeng sa baithuti ba univesithi ba phetseng hantle le ba nang le IA. Latency ea liteko tse 42 tsa 'li-ERP' tse lekantsoeng tlasa ts'usumetso ea sepheo e ne e le telele ho feta tsa li-ERP tse 558 tse lekantsoeng ka tlasa ts'usumetso e seng ea sepheo (leq 0.05), 'me boholo ba li-ERP tse 42 tsa liteko tse tlasa sepheo sa sepheo li ne li le kholo ho feta tsa li-ERP tse 558 tse neng li lekantsoe tlasa sepheo se seng (leq 0.05). E bonts'itse phapang e kholo litabeng tsa P300 lipakeng tsa lithuto tse phetseng hantle le lithuto tsa Keketso ea inthanete. Bokaholimo ba tlatsetso ea inthanete bo ne bo le tlase (leq 0.05). Ho nepahala ha tlhahlobo ho ka fihla kaholimo ho 93% ho sebelisa mokhoa o thehiloeng ho Bayesian libakeng tse sebetsang, ha o ntse o le tlase ho 90% libakeng tse bohareng. Liphetho li bonts'a hore ho na le litšusumetso tse mpe karabong ea boko le bokhoni ba ho hopola ba liithuti tsa univesithi tse nang le bothata ba IA. Pampiri e sebetsana le ts'ebetso ea ts'ebetso ea filimi ea dijithale ho hatella lerata la matla a 50 Hz u sebelisa li-filters tsa cogerfficients. Tharollo e potlakileng haholo ebile e bonolo e nolofalletsa ho hatella likarolo tsa mantlha le tse lumellanang tsa lerata la matla ka litšitiso tse sa bonahaleng. Matšoao a 'nete a ECG a ne a sebelisetsoa ho leka katleho ea khatello ea lerata ea matla. Ho nepahala ho hlahlojoa hore na molumo oa "sinusoidal" le "leqhubu" oa molumo ke eng?
Keywords: KHOTSO CHANNEL; LIPOTSO TSE KHOLO; TLHALOSO E NANG KOTSI; P300; TLHOKOMELISO PATTERN
Mofuta oa Tokiso: Sengoloa sa Patlisiso
DOI: https://doi.org/10.1166/jmihi.2015.1570
Letsatsi la phatlalatso: November 1, 2015
;
