Bootstrap generation and evaluation of an fMRI simulation database

Abstract

Computer simulations have played a critical role in functional magnetic resonance imaging (fMRI) research, notably in the validation of new data analysis methods. Many approaches have been used to generate fMRI simulations, but there is currently no generic framework to assess how realistic each one of these approaches may be. In this paper, a statistical technique called parametric bootstrap was used to generate a simulation database that mimicked the parameters found in a real database, which comprised 40 subjects and 5 tasks. The simulations were evaluated by comparing the distributions of a battery of stastical measures between the real and simulated databases. Two popular simulation models were evaluated for the first time by applying the bootstrap framework. The first model was an additive mixture of multiple components and the second one implemented a non-linear motion process. In both models, the simulated components included the following brain dynamics : a baseline, physiological noise, neural activation and random noise. These models were found to successfully reproduce the relative variance of the components and the temporal autocorrelation of the fMRI time series. By contrast, the level of spatial autocorrelation was found to be drastically low using the additive model. Interestingly, the motion process in the second model intrisically generated some slow time drifts and increased the level of spatial autocorrelations. These experiments demonstrated that the bootstrap framework is a powerful new tool that can pinpoint the respective strengths and limitations of simulation models.

How ignoring physiological noise can bias the conclusions from fMRI simulation results

“…However, these data contain a lot of unintended and often unknown factors, such as activity from the default mode network (Raichle and Snyder, 2007). Some authors tried to avoid this unwanted activity by constructing the noise from summary statistics based on the real data (Bellec et al, 2009) or by modelling the noise as the resampled residuals from a GLM-analysis of the real data (Havlicek et al, 2010).…”

“…All data in both simulation studies were generated according to an additive simulation model (Bellec et al, 2009). More specifically, we considered data consisting of two separate layers, namely (1) an activation layer and (2) a noise layer.…”

“…Some studies look at test-reliability and reproducibility of results (see for example Schuyler et al, 2010), while others use resampling techniques (e.g. bootstrapping) and use the distributional properties of the resampled data as validation measures (see Bellec et al, 2009 for an example of using parametric bootstrapping). However, validation studies often need to know the ground truth of their data, and this is almost impossible for human data without very invasive procedures (David et al, 2008) or technically challenging constructions (Cheng et al, 2006;Brosch et al, 2002).…”

Iterative approach of dual regression with a sparse prior enhances the performance of independent component analysis for group functional magnetic resonance imaging (fMRI) data

“…ion.ucl.ac.uk/spm) with the following steps: slice timing correction, head motion correction, normalization to Montreal Neurological Institute (MNI) coordinates (3-mm isotropic voxels), spatial smoothing with an 8-mm isotropic full-width-at-half-maximum Gaussian kernel, and linear detrending using a high-pass filter with a 0.0033-Hz cut-off frequency. The non-neuronal components were subsequently decomposed using a parametric bootstrapping model with additional shrinkage correction (Bellec et al, 2009). A portion of the volumes of each fMRI scan (i.e., 70 and 30 volumes for block-based and Fig.…”

Functional parcellation of memory related brain networks by spectral clustering of EEG data

“…In order to achieve statistically significant electrodes from group-wise clusters, bootstrap technique is employed [19]. A simulation database is generated by mimicking the subjectwise clusters to test the significance of group-wise clusters.…”