Modeling the hemodynamic response in fMRI using smooth FIR filters

Abstract

Abstract-Modeling the haemodynamic response in functional magnetic resonance (fMRI) experiments is an important aspect of the analysis of functional neuroimages. This has been done in the past using parametric response function, from a limited family. In this contribution, we adopt a semi-parametric approach based on finite impulse response (FIR) filters. In order to cope with the increase in the number of degrees of freedom, we introduce a Gaussian process prior on the filter parameters. We show how to carry on the analysis by incorporating prior knowledge on the filters, optimizing hyper-parameters using the evidence framework, or sampling using a Markov Chain Monte Carlo (MCMC) approach. We present a comparison of our model with standard haemodynamic response kernels on simulated data, and perform a full analysis of data acquired during an experiment involving visual stimulation.

Rapidly acquired multisensory association in the olfactory cortex

“…This is because delays in BOLD response to stimuli can vary across regions and subjects (Goutte et al. 2000; Ollinger et al. 2001).…”

“…We also investigated the hemodynamic response function for odor + visual and visual‐only conditions by convolving respective condition vectors of onset times with a finite impulse response function (IRF) with eight 2‐sec time bins (Goutte et al. 2000; Ollinger et al. 2001; Lindquist et al.…”

Bayesian spatiotemporal model of fMRI data using transfer functions

“…Since then, they have proven to be a powerful and flexible tool to infer on the HRF shape. Goutte et al (2000) adopt a transfer function model to estimate the HRF in fMRI, introducing a Gaussian process prior on the coe cients of the TF. The authors prove not just the ability of the FIR filter to recover the shape of traditional linear filters (Poisson, Gamma filter and Gaussian filters) but also that the smooth FIR filter is able to model additional features in real data, when the three traditional filter shapes fail.…”

“…The authors prove not just the ability of the FIR filter to recover the shape of traditional linear filters (Poisson, Gamma filter and Gaussian filters) but also that the smooth FIR filter is able to model additional features in real data, when the three traditional filter shapes fail. Ciuciu et al (2003) generalise Goutte et al (2000) in order to account for estimation of the HRF for any fMRI experiment. Makni et al (2008) propose a parcel-based spatiotemporal approach under a Bayesian formalism.…”

“…Regions are first identified as functionally homogeneous parcels in the mask of the grey matter and a region-based transfer function model of the HRF is adopted (Goutte et al (2000); Ciuciu et al (2003)). Detection is achieved by modelling activating, deactivating and nonactivating voxels through a mixture model (gamma-Gaussian) within each parcel.…”