A popular method for investigating whether stimulus information is present in fMRI response patterns is to attempt to "decode" the stimuli from the response patterns with a multivariate classifier. The sensitivity for detecting the information depends on the particular classifier used. However, little is known about the relative performance of different classifiers on fMRI data. Here we compared six multivariate classifiers and investigated how the response-amplitude estimate used (beta or t-value) and different pattern normalizations affect classification performance. The compared classifiers were a pattern-correlation classifier, a k-nearest-neighbors classifier, Fisher's linear discriminant, Gaussian naïve Bayes, and linear and nonlinear (radial-basis-function-kernel) support vector machines. We compared these classifiers' accuracy at decoding the category of visual objects from response patterns in human early visual and inferior temporal cortex acquired in an event-related design with BOLD fMRI at 3T using SENSE and isotropic voxels of about 2-mm width. Overall, Fisher's linear discriminant (with an optimal-shrinkage covariance estimator) and the linear support vector machine performed best. The pattern-correlation classifier often performed similarly as those two classifiers. The nonlinear classifiers never performed better and sometimes significantly worse than the linear classifiers, suggesting overfitting. Defining response patterns by t-values (or in error-standard-deviation units) rather than by beta estimates (in % signal change) to define the patterns appeared advantageous. Cross-validation by a leave-one-stimuluspair-out method gave higher accuracies than a leave-one-run-out method, suggesting that generalization to independent runs (which more safely ensures independence of the test set) is more challenging than generalization to novel stimuli within the same category. Independent selection of fewer more visually responsive voxels tended to yield better decoding performance for all classifiers. Normalizing mean and standard deviation of the response patterns either across stimuli or across voxels had no significant effect on decoding performance. Overall our results suggest that linear decoders based on t-value patterns may perform best in the present scenario of visual object representations measured for about 60-minutes per subject with 3T fMRI.

Objective Patients with depression show blunted amygdala hemodynamic activity to positive stimuli, including autobiographical memories. The authors examined the therapeutic efficacy of real-time functional MRI neurofeedback (rtfMRI-nf) training aimed at increasing the amygdala’s he-modynamic response to positive memories in patients with depression. Method In a double-blind, placebo-controlled, randomized clinical trial, unmedicated adults with depression (N=36) were randomly assigned to receive two sessions of rtfMRI-nf either from the amygdala (N=19) or from a parietal control region not involved in emotional processing (N=17). Clinical scores and autobiographical memory performance were assessed at baseline and 1 week after the final rtfMRI-nf session. The primary outcome measure was change in score on the Montgomery-Åsberg Depression Rating Scale (MADRS), and the main analytic approach consisted of a linear mixed-model analysis. Results In participants in the experimental group, the hemo-dynamic response in the amygdala increased relative to their own baseline and to the control group. Twelve participants in the amygdala rtfMRI-nf group, compared with only two in the control group, had a >50% decrease in MADRS score. Six participants in the experimental group, compared with one in the control group, met conventional criteria for remission at study end, resulting in a number needed to treat of 4. In participants receiving amygdala rtfMRI-nf, the percent of positive specific memories recalled increased relative to baseline and to the control group. Conclusions rtfMRI-nf training to increase the amygdala hemo-dynamic response to positive memories significantly decreased depressive symptoms and increased the percent of specific memories recalled on an autobiographical memory test. These data support a role of the amygdala in recovery from depression.