Pattern classification of fMRI data: Applications for analysis of spatially distributed cortical networks

详细信息    查看全文

• 作者：Grigori Yourganov^a ; ^{yourgano@mailbox.sc.edu" class="auth_mail} ; Tanya Schmah^b ; Nathan W. Churchill^b ; Marc G. Berman^a ; Cheryl L. Grady^b ; ^c ; Stephen C. Strother^b ; ^d ; ^e
• 关键词：Multivariate pattern analysis ; Classification ; Regularization ; Spatial maps ; Principal component analysis (PCA)
• 刊名：NeuroImage
• 出版年：1 August 2014
• 年：2014
• 卷：96
• 期：Complete
• 页码：117-132
• 全文大小：2211 K

文摘

<dl class="listitem" id="list_l0005"><dt class="label">•dt><dd>

d="p0660">A set of pattern classifiers was evaluated on fMRI data (experimental and simulated).

dd><dt class="label">•dt><dd>

d="p0665">Spatial maps were created for each classification task.

dd><dt class="label">•dt><dd>

d="p0670">Linear and quadratic discriminants with PCA regularization produce the best maps.

dd><dt class="label">•dt><dd>

d="p0675">The worst maps are produced by univariate classifiers.

dd><dt class="label">•dt><dd>

d="p0680">The maps for SVM and ridge-regression linear discriminant are very similar.

dd>dl>