Limits...
MEG and fMRI Fusion for Non-Linear Estimation of Neural and BOLD Signal Changes.

Plis SM, Calhoun VD, Weisend MP, Eichele T, Lane T - Front Neuroinform (2010)

Bottom Line: Neural activity estimates were derived using a dynamic Bayesian network with continuous real valued parameters by means of a sequential Monte Carlo technique.The highly non-linear model of the BOLD response poses a difficult inference problem for neural activity estimation; computational requirements are also high due to the time and space complexity.We show that joint analysis of the data improves the system's behavior by stabilizing the differential equations system and by requiring fewer computational resources.

View Article: PubMed Central - PubMed

Affiliation: The Mind Research Network Albuquerque, NM, USA.

ABSTRACT
The combined analysis of magnetoencephalography (MEG)/electroencephalography and functional magnetic resonance imaging (fMRI) measurements can lead to improvement in the description of the dynamical and spatial properties of brain activity. In this paper we empirically demonstrate this improvement using simulated and recorded task related MEG and fMRI activity. Neural activity estimates were derived using a dynamic Bayesian network with continuous real valued parameters by means of a sequential Monte Carlo technique. In synthetic data, we show that MEG and fMRI fusion improves estimation of the indirectly observed neural activity and smooths tracking of the blood oxygenation level dependent (BOLD) response. In recordings of task related neural activity the combination of MEG and fMRI produces a result with greater signal-to-noise ratio, that confirms the expectation arising from the nature of the experiment. The highly non-linear model of the BOLD response poses a difficult inference problem for neural activity estimation; computational requirements are also high due to the time and space complexity. We show that joint analysis of the data improves the system's behavior by stabilizing the differential equations system and by requiring fewer computational resources.

No MeSH data available.


Activity localized by the MNE software inverse solution and overlayed on the inflated cortical surface. Peak stimulus driven activity is shown by red and yellow colors. Green areas indicate the region selected from the FreeSurfer atlas and corresponding to cuneus, precuneus, and pericalcarine ROIs for the left (A) and the right (B) hemispheres.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC2991230&req=5

Figure 12: Activity localized by the MNE software inverse solution and overlayed on the inflated cortical surface. Peak stimulus driven activity is shown by red and yellow colors. Green areas indicate the region selected from the FreeSurfer atlas and corresponding to cuneus, precuneus, and pericalcarine ROIs for the left (A) and the right (B) hemispheres.

Mentions: MNE software (Hämäläinen, 2005) was used to localize cortical areas in the source space that are active while the task was performed. The areas that exhibited the most stimulus driven activity were identified from the expert knowledge of similar experiments, since visual stimulation mostly activates the primary visual cortex, (threshold levels set in MNE were 4.0, 8.5, and 9.5 for fthresh, fmid, and fmax parameters respectively) and corresponding ROIs from FreeSurfer (Fischl et al., 1999) atlas were chosen to represent the forward model (see Figure 12). A similar data collection scheme was used for EEG/fMRI fusion in (Ostwald et al., 2010). All dipoles orthogonal to the cortical surface and belonging to selected ROIs were used to construct the MEG forward model corresponding to the single hidden neural activity node in the DBN structure of Figure 2. The same ROIs were combined to build the HFM. This means that the model had single parameter for dipole amplitude (the same for all dipoles) in the MEG forward model and the absolute value parameter (with appropriate scaling) was also used as neural activity in the fMRI forward model. The activity source in the real data experiment is localized and active voxels do not differ substantially in their dynamics to cause problems when averaging.


MEG and fMRI Fusion for Non-Linear Estimation of Neural and BOLD Signal Changes.

Plis SM, Calhoun VD, Weisend MP, Eichele T, Lane T - Front Neuroinform (2010)

Activity localized by the MNE software inverse solution and overlayed on the inflated cortical surface. Peak stimulus driven activity is shown by red and yellow colors. Green areas indicate the region selected from the FreeSurfer atlas and corresponding to cuneus, precuneus, and pericalcarine ROIs for the left (A) and the right (B) hemispheres.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC2991230&req=5

Figure 12: Activity localized by the MNE software inverse solution and overlayed on the inflated cortical surface. Peak stimulus driven activity is shown by red and yellow colors. Green areas indicate the region selected from the FreeSurfer atlas and corresponding to cuneus, precuneus, and pericalcarine ROIs for the left (A) and the right (B) hemispheres.
Mentions: MNE software (Hämäläinen, 2005) was used to localize cortical areas in the source space that are active while the task was performed. The areas that exhibited the most stimulus driven activity were identified from the expert knowledge of similar experiments, since visual stimulation mostly activates the primary visual cortex, (threshold levels set in MNE were 4.0, 8.5, and 9.5 for fthresh, fmid, and fmax parameters respectively) and corresponding ROIs from FreeSurfer (Fischl et al., 1999) atlas were chosen to represent the forward model (see Figure 12). A similar data collection scheme was used for EEG/fMRI fusion in (Ostwald et al., 2010). All dipoles orthogonal to the cortical surface and belonging to selected ROIs were used to construct the MEG forward model corresponding to the single hidden neural activity node in the DBN structure of Figure 2. The same ROIs were combined to build the HFM. This means that the model had single parameter for dipole amplitude (the same for all dipoles) in the MEG forward model and the absolute value parameter (with appropriate scaling) was also used as neural activity in the fMRI forward model. The activity source in the real data experiment is localized and active voxels do not differ substantially in their dynamics to cause problems when averaging.

Bottom Line: Neural activity estimates were derived using a dynamic Bayesian network with continuous real valued parameters by means of a sequential Monte Carlo technique.The highly non-linear model of the BOLD response poses a difficult inference problem for neural activity estimation; computational requirements are also high due to the time and space complexity.We show that joint analysis of the data improves the system's behavior by stabilizing the differential equations system and by requiring fewer computational resources.

View Article: PubMed Central - PubMed

Affiliation: The Mind Research Network Albuquerque, NM, USA.

ABSTRACT
The combined analysis of magnetoencephalography (MEG)/electroencephalography and functional magnetic resonance imaging (fMRI) measurements can lead to improvement in the description of the dynamical and spatial properties of brain activity. In this paper we empirically demonstrate this improvement using simulated and recorded task related MEG and fMRI activity. Neural activity estimates were derived using a dynamic Bayesian network with continuous real valued parameters by means of a sequential Monte Carlo technique. In synthetic data, we show that MEG and fMRI fusion improves estimation of the indirectly observed neural activity and smooths tracking of the blood oxygenation level dependent (BOLD) response. In recordings of task related neural activity the combination of MEG and fMRI produces a result with greater signal-to-noise ratio, that confirms the expectation arising from the nature of the experiment. The highly non-linear model of the BOLD response poses a difficult inference problem for neural activity estimation; computational requirements are also high due to the time and space complexity. We show that joint analysis of the data improves the system's behavior by stabilizing the differential equations system and by requiring fewer computational resources.

No MeSH data available.