Comparing within-subject classification and regularization methods in fMRI for large and small sample sizes

• Published in: Human brain mapping Vol. 35; no. 9; pp. 4499 - 4517
• Main Authors: Churchill, Nathan W., Yourganov, Grigori, Strother, Stephen C.
• Format: Journal Article
• Language: English
• Published: New York, NY Blackwell Publishing Ltd 01.09.2014; Wiley-Liss; John Wiley & Sons, Inc; John Wiley and Sons Inc
• Subjects: Adult; Biological and medical sciences; BOLD fMRI; Brain - blood supply; Brain - physiology; Brain Mapping - methods; Cerebrovascular Circulation - physiology; data-driven metrics; Female; head motion; Humans; Investigative techniques, diagnostic techniques (general aspects); Logistic Models; Magnetic Resonance Imaging - methods; Male; Medical sciences; Miscellaneous; model optimization; multivariate analysis; Nervous system; Neuropharmacology; Oxygen - blood; Pharmacology. Drug treatments; physiological noise; preprocessing; Principal Component Analysis; Radiodiagnosis. Nmr imagery. Nmr spectrometry; Reproducibility of Results; Sample Size; Signal Processing, Computer-Assisted; Signal-To-Noise Ratio; Support Vector Machine; Young Adult; Human; Motion; Nervous system diseases; Head; Radiodiagnosis; Sample size; Noise; preprocessing; head motion; physiological noise; Multivariate analysis; data-driven metrics; Nuclear magnetic resonance imaging; Optimization; BOLD fMRI; Regularization method; Classification; Models; model optimization; multivariate analysis
• ISSN: 1065-9471; 1097-0193; 1097-0193
• DOI: 10.1002/hbm.22490

In recent years, a variety of multivariate classifier models have been applied to fMRI, with different modeling assumptions. When classifying high‐dimensional fMRI data, we must also regularize to improve model stability, and the interactions between classifier and regularization techniques are still being investigated. Classifiers are usually compared on large, multisubject fMRI datasets. However, it is unclear how classifier/regularizer models perform for within‐subject analyses, as a function of signal strength and sample size. We compare four standard classifiers: Linear and Quadratic Discriminants, Logistic Regression and Support Vector Machines. Classification was performed on data in the linear kernel (covariance) feature space, and classifiers are tuned with four commonly‐used regularizers: Principal Component and Independent Component Analysis, and penalization of kernel features using L1 and L2 norms. We evaluated prediction accuracy (P) and spatial reproducibility (R) of all classifier/regularizer combinations on single‐subject analyses, over a range of three different block task contrasts and sample sizes for a BOLD fMRI experiment. We show that the classifier model has a small impact on signal detection, compared to the choice of regularizer. PCA maximizes reproducibility and global SNR, whereas Lp‐norms tend to maximize prediction. ICA produces low reproducibility, and prediction accuracy is classifier‐dependent. However, trade‐offs in (P,R) depend partly on the optimization criterion, and PCA‐based models are able to explore the widest range of (P,R) values. These trends are consistent across task contrasts and data sizes (training samples range from 6 to 96 scans). In addition, the trends in classifier performance are consistent for ROI‐based classifier analyses. Hum Brain Mapp 35:4499–4517, 2014. © 2014 Wiley Periodicals, Inc.